Indicator/turn signals
and hazard flashers
Indicator/turn signal schematics
Hazard warning schematics
Indicators
Hazards
The indicator/turn switch Added August 2008
Where is the flasher unit?
Dash tell-tales
Fault diagnosis
Indicator Flasher Replacement May 2016
LED flashers
Adding hazards to earlier cars October 2010
A louder audible warning
the innards of each flasher unit
  June 2018: Warning! It seems that new Lucas indicator flashers SFB115 (for Mk2 cars and later) from some suppliers at least contain hazard flasher internals, three from different suppliers so far, with two different batch numbers. The differences are subtle but have safety implications. The most obvious difference is immediately you operate the indicator switch. With the correct indicator flasher the lights should come on straight away to give an immediate indication to other road users, then after a pause start flashing off-on-off-on. Hazard flashers by contrast do nothing when first operating the switch, then after a short delay start flashing on-off-on-off, which delays how soon other road uses will see your intentions. The other more important difference is that the correct indicator flasher unit will only flash if both main bulbs are working, if one has failed the other (and the tell-tale) is permanently lit so alerting the driver. Hazard flasher units are designed to flash anything from one to four bulbs, some of which may be out of action following a collision, so when used as an indicator flasher the driver has no idea that a corner may have failed, and other road uses may not be aware that you are about to slow down or change lane or direction. I've written to Lucas/Elta and they were not aware of the problem, but were interested enough to ask for pictures of the markings on the item and packaging, and the Technical Manager said he would get some out of stock for testing. No feedback was ever received. There is an issue even with nominally correct indicator flasher units in that some are designed to flash two corners and a tell-tale as with the MGB, and others to flash those plus a side repeater and if these are used in an MGB they give a slow flash rate. For information on a modern flasher unit with a good flash rate as well as lighting up immediately and indicating if a corner has failed see here.

Indicator and hazard flasher units are not 'relays' as often stated, if anything they are 'switches', but that would cause confusion with the manual switches, hence 'units'. A relay receives an input voltage from one source which powers an electro-magnet to close a contact which completes a separate electrical circuit. It isolates the two circuits so that the electrical characteristics of one can't affect or damage the other. By contrast an indicator or hazard flasher unit takes the input voltage and extends it through to the output i.e. it is simply switching the input voltage on and off.

Indicators: For Mk2 and later go here, but for Mk1 read on - it's complicated, as complicated as Mk2 but in different ways.

Mk1 cars had a cylindrical 3-terminal indicator flasher unit originally GFU103 FL5 but it's completely different to 1980s 3-pin electronic flasher units. Terminals can be screwed or Lucar/spade, one is for the 12v supply in (B or X, green), one for output to the indicator switch and the lights at the corners of the car (L, light-green/brown), the third terminal (P, light-green/purple) is used for flashing the dash tell-tales. This supplies 12v to both tell-tales at the same time, the other side of each tell-tale goes to additional contacts on the indicator switch which provide an earth to the appropriate tell-tale when the switch is operated to one side or the other. The tell-tale flashes in anti-phase to the corners of the car, and comes on as soon as you operate the switch and only goes off after a short delay, which means the corners of the car don't light up until after a short delay, then start flashing on-off-on-off. From brief exposure to one of these units with one bulb not working the other bulb only makes a very brief flash, barely visible to other road users, the tell-tale may be on for much longer with only a brief period off. Note that to use a two-terminal flasher unit you would need to connect the light-green/brown and light-green/purple wires to one of the terminals with green connected to the other. But note the warnings on using modern SFB115 units at the top of this page, units intended for hazards, and any number of so-called 'universal' and LED types, elsewhere in these pages.

If that wasn't enough there are at least two types of this flasher unit - one for MGBs and one for Triumph TR4 and earlier, and the difference is in how the tell-tale is operated. Helping a pal with the indicators on his TR4 he had bought what he thought was the correct flasher unit ... but the tell-tale was on all the time with the ignition, albeit flashing when the indicators were being used. From what I knew about the Mk1 MGB system I realised that what he had was an MGB unit, which has the additional contacts on the indicator switch so the tell-tale is off when they are not being used. The TR4 only has one tell-tale for both sides, and no extra contacts on the indicator switch, so it needs a flasher unit where the tell-tale contact is not live until the unit starts flashing. Paradoxically it wouldn't matter if the TR4 and earlier unit was used on an MGB, so the question has to be why the MGB had a different one (the MGA apparently had two different systems, one the same as the Triumph and the 1500 which is completely different as it involves flashing the brake lights).

Current replacements can be stocked as GFU101, GFU2101, GFU103, GFU2103, SFB105, FL5, FL54, C16729, 35002, 35003, 35010, 35028, 35028A, RTC3560, UD1511 and probably others! Some of them are for Triumph but will work equally well in the MGB (unlike the other way round). Some Triumph sources report similar problems to the Lucas SFB115 two-pin flasher unit for Mk2 including being a hazard flasher and not an indicator flasher.

Mk2: A simpler system with only Lucar/spade two terminals on an 8FL flasher unit - 12v to B (green) and L (light-green/brown) to the indicator switch, and no extra contacts on that switch. The two dash-tell-tales are simply wired in parallel with the bulbs at the corners of the car - one per side, so come on and go off with the corners. These units have the safety feature that as soon as you operate the indicator switch both the corners that side (and the dash tell-tale) light up, and after a pause start flashing off-on-off-on. However there are a number of 8FL units and some are not correct for the MGB. There is a 'type' number on the original Lucas can which can be 36, 41, 50 and maybe others and this relates to the current they are designed to work with which is dependant on the number of bulbs and their wattage.

A second safety feature is that if one corner fails the other corner and the tell-tale will glow steadily instead of flashing, which is a warning to the driver that they may need to start using hand-signals.

A drawback of this system is that the flasher unit is very sensitive to voltage and current, and with low voltage or increasing resistance from ageing connections and bulbs the flash rate gets slower and slower as well as reducing the brightness of the bulbs, and this is where some have changed to using three-pin electronic flasher units. Many of these (but not all ...) have the same safety features of lighting up straight away and indicating when a corner has failed, except that in the latter case with one corner not working the other corner and the dash tell-tale will flash at double speed. Being electronic they don't suffer from slowing down with falling voltage or increasing resistance issue and give a much more consistent flash rate, BUT ... bulbs that are dim from poor connections will remain dim. It's a significant exercise to investigate that, and there can be very many small extra resistances throughout the circuit to find and fix, although fixing the greatest resistances first will have the greatest effect.

Current stock replacements are the aforementioned suspect SFB115 as well as SFB114, GFU125, GFU2125, 8FL, 35049. Note some of these may be cylindrical, the can won't fit the original rectangular clip but may come with its own clip. Also that rectangular ones may be double-height but they fit the original clip. Confirm that the packaging or can states '48W' or '50W', and/or '2 x 21W' (may also include '+6w +2w') and NOT '12A', '96W' or '4 x 21W', however even then some types will flash slower than they should. When fitted immediately test and confirm that as soon as you operate the indicator switch the lamps light up, then after a short pause start to flash - may well need the engine to be running to flash at an acceptable rate, also that with one bulb disconnected the flash rate changes markedly.

That leaves LED bulbs which are whole different world of pain in the indicator flasher unit department.

Hazard flashers:

When can they be used?
Supplementary hazard light

Flasher unit originally ATJ8880 9FL, now SFB130 (many equivalents) labelled '21w x 4 + 5w', 94W, or similar. These operate differently to indicator flashers and are a different construction internally. For a start when you first turn on the hazard switch nothing happens, then they start flashing on-off-on-off. Secondly they are capable of flashing anything from one to four main bulbs (plus side indicators and tell-tales) if for example the vehicle has been in an accident and one or more corners are damaged. Secondly they are relatively insensitive to falling voltage as they may be left operating for some time without the engine running, and flash at a relatively consistent rate. Two-pin devices (like the Mk2 indicator flasher) - 12v to B (brown) and L (light-green/brown) to the hazard switch which links the two sides and connects the flasher unit to them. The flasher switch has two more contacts that are normally closed with the hazards off and these feed fused ignition power (green) to the indicator flasher unit. These contacts disconnect that when the hazards are operated, otherwise if the indicator switch was left operated power from the hazard flasher to the corners of the car could feed back through the indicator switch and the indicator flasher onto the green and white circuits, and hence power the fuel pump and the ignition even with the ignition key in your hand - an obvious safety hazard in the event of damaged fuel lines.

Unfortunately having two spade terminals like the Mk2 indicator flasher unit they can be substituted for each other in error. With a hazard flasher in place of an indicator flasher things will appear to work, but there will be a delay in the corners lighting up when the indicator switch is operated, and there will be no indication of bulb failure. The other way round is more obvious as with an indicator flasher trying to light all four corners there is only the briefest of flash lighting them before they go off again.

With LED bulbs at the corners the standard hazard flasher may work if there is are incandescent bulbs in the tell-tale, but probably not if they are LEDs as well.

Hazard lights were standard on North American spec cars from the start of Mk1 production in 1967, on all V8s, and non-North American 4-cylinder cars from the start of the 1974 model year i.e. the last chrome bumper cars according to the list of detailed changes in Clausager ... or is it? He says in the text that they were also available as an optional extra on home market before that, but that the information available is 'conflicting', and it is. He says the 1973/74 GT brochure for home market cars quotes them as an optional extra, but no mention at all in the 1974 roadster brochure. '1973/74' could be taken to mean both 1973 and 1974 models, but could also be the model that was built during the second part of 1973 and the first part of 1974 which would be the 1974 model built from August 73 at chassis number 328101 (roadster) and 328801 (GT) to September 74 at chassis number 360301 (roadster and 361101 (GT). Elsewhere he says they were 'probably' fitted to all home market cars from the start of the 1974 model year. The Leyland schematics show it for 1974 models but not for 1973 models, in neither case as an option. The Parts Catalogue shows hazard switch BHA5267 being used for RHD chassis number HD5-328801 which is specifically GTs from the start of the 1974 model year, but it doesn't indicate it is an option unlike for fog and spot light switches. Flasher unit ATJ8880 is specified for chassis number 328101 (roadster) and 328801 (GT) i.e. all 1974 model cars and again no indication of being optional. The only clear thing is that all RB cars had them. It may be that when the 1974 model brochures were being produced it got into the GT one but not the roadster one, and by the time it came to build the cars they had decided to fit them to both roadster and GT anyway (why would they fit it to one but not the other?). Who knows, other than looking at actual cars of the period, but even that is fraught.

Brian Wall has said his November 72-built GT has them i.e. a 1973 model built between August 72 and August 73, which raises the possibility that they were an option in 1973. But that could be a mod by a previous owner, possibly with a rewire, it looks like that was the only electrical change between 1973 and 1974 models. If it had been an option the later harness could have been installed on the line as required, or maybe it was on all of them and the switch and flasher unit not provided. In which case there would need to be a dummy 'switch' on the connector to link the two green wires together or the indicators wouldn't work, in which case to add hazards would just be a case of adding the switch and flasher unit to the connections. The option could also have been a sub-harness added to the earlier standard harness, which is what I did.

Adding hazard flashers is not simply a case of tapping into a 12v supply and the wiring to the corners of the car which is easily done. The hazard switch when on has to interrupt the power to the indicator flasher unit for safety reasons. Feasible with a sub-harness as it would be a case of simply removing the 12v feed from the indicator flasher, connecting that to the sub-harness, then connecting another wire in the sub-harness back on the indicator flasher. But if the main harness had provision for optional hazards there would need to be an link (insulated from coming into contact with anything else as it would be part of a bunch of wires not attached to anything i.e. flapping around) between the two green wires going to the switch, removed when the switch was provided.

May 2019: A puzzling problem for Steve Henson-Webb on the MGOC forum. When using his indicators the dash tell-tales worked, but when using the hazards they didn't. The appropriate corners of the car flashed when they should. Unfortunately he didn't mention until later that when using the indicators both tell-tales were flashing, but dimmer than usual, which made things much easier to understand. The problem was the earth wire to both tell-tales had become detached. That meant that when using the indicators the tell-tales were effectively in series to earth via the indicator bulbs on the side of the car that wasn't flashing, so both flash, albeit dimly as each has only 6v instead of 12v. When the hazards are on both tell-tales have 12v on their 'live' sides, but no earths on the other, so neither flashed. The puzzling thing is that it was a 1974 CB, but on Mk2 tin-dash cars (e.g. my 73) there is no earth wire for the indicator tell-tales that I'm aware of, the bulb holder picks that up from the bracket they are pushed in to, which is attached to the back of the dash. He says it happened after fitting a radio, maybe that interrupted the dash earthing somehow, but seems unlikely. Mk1 cars (which didn't have hazards from the factory of course) did use 2-wire indicator bulb-holders as the 12v supply was switched from the flasher unit and the earth supply came from the indicator switch. North American Mk2 with the padded dash, all V8s and all RB cars also have 2-wire indicator bulb-holders as they are mounted in plastic panels so need a wired earth.

The indicators are powered from one of two green (fused ignition) circuits - originally the one in the fusebox. But sometime in 1978 when the ignition relay circuit on RHD cars was modified to add a second in-line fuse between brown/white and green wires under the fusebox, the indicators (and tach, heater fan and GT GRW) were powered from one of these - the one with the thinner wires, the other with thick wires being for the cooling fan. For more information see the ignition schematics. On cars equipped with hazard flashers the green circuit goes via the hazard flasher switch to be connected when the hazards are off, and disconnected when they are on.

By all accounts indicators are the bane of an LBC-ers life. But like all things, they worked when it came out of the factory. If it doesn't work now then there must be a reason (or two or three), so it can be found and fixed.

But first - know the difference between indicator/turn flasher units (they are not relays, strictly speaking they are thermo-switches) and hazard flasher units:

  • With indicator flasher units as soon as you operate the switch the lamps should light up, and after a short pause they should start flashing off-on-off-on.
  • With hazard flasher units as soon as you operate the switch nothing happens for a short period, then they should start flashing on-off-on-off.
However that only applies to the 2-pin indicator flasher units used on Mk2 and later cars. Mk1 cars used a 3-pin cylindrical flasher unit that operates differently. It appears to be the same - i.e. when you operate the column stalk and look at the dash tell-tales they come on straight away, then after a pause they start flashing off-on-off-on. But if you look at the corners of the car you will see they are operating in anti-phase, i.e. when you first operate the stalk nothing happens, then after a pause they start flashing on-off-on-off, i.e. the same as hazards! This is because the tell-tale contact on the flasher unit shows 12v when switched off, and is why the column stalk needs two additional contacts to connect the appropriate tell-tale when the stalk is moved. This caused some confusion on a pal's TR3, which only has one dash tell-tale for both sides, which is wired direct to the flasher unit. This was glowing all the time the ignition was on, and operated in anti-phase to the corners of the car. He had been supplied an MGB flasher unit, whereas the correct Triumph flasher unit has the tell-tale contact normally off, so the tell-tale operates in synch with the corners of the car, and the indicator switch doesn't need any additional contacts. Incorrect supply for Triumph owners must be very common, as many suppliers quote the MGB item as being suitable for the TR2/3/4, but I did find the correct Triumph Lucas FL number (since mislaid) for my pal and he was able to get the correct item from the same supplier.

If you substitute an indicator flasher for a hazard flasher it will appear to click at the correct rate but will only send power to the bulbs very briefly, not long enough to fully light them, and may burn-out quickly as they are only designed to flash two 21w bulbs (plus a 5w wing repeater on the 2-pin types) not four. But if you substitute a hazard flasher for a 2-pin indicator flasher it will at first seem to work correctly, unless you notice the sequence (as above) is incorrect. This is a safety hazard, as it delays the lights coming on and hence the warning to road users. So many people these days seem to operate the indicators at the same time as they turn the wheel that the rest of us need all the warning we can get!

There are also 'universal' or 'heavy duty' flasher units that although they may have the correct sequence for indicators i.e. they come on as soon as you operate the switch, don't have the built-in 'bulb failure warning' of the originals. Really when fitting an alternative flasher unit you need to disconnect one corner and confirm that you still get this warning. On original MGB types this warning is that flashing stops altogether, and only one external bulb will be lit. On modern electronic units the remaining bulb should flash at double-speed.

To complicate matter even further there is another type of after-market indicator flasher unit intended for use with after-market LED bulbs, more on those here.

Originally the MGB used a cylindrical 3-pin flasher unit (GFU103, Lucas FL5), but this is not the same as modern electronic 3-pin flashers. On the originals the third pin is used to flash the dash repeaters via additional contacts in the indicator switch, whereas on electronic units the third pin is connected to earth. Mk2 MGBs used a rectangular 2-pin indicator flasher unit (SFB115 (was GFU107), Lucas 8FL) as the dash repeaters are now connected to the wiring going out to the corners of the car. Both the original MGB types are 'thermal' type flashers and the following information is from Steve Blakeway who was an employee of Lucas working on the 2-pin thermal flasher units and their electronic replacements for nearly 30 years:

The moving contact is on a metal plate pressed from thin spring steel. The blanking and pressing of the spring steel gives it a 'set'. The thin strip was then welded diagonally such that the plate was deformed against the set. When the indicator switch is operated and passes bulb current through the thin strip it heats up and expands, which allows the plate to ping back to it's original set, opening the contact and extinguishing the bulbs. The current ceases and allows the thin strip to cool and contract. This pings the spring-steel plate back to it's previous position, closing the contact, illuminating the bulbs, and heating up the thin strip again, so repeating the process. It works rather like the Pop-o-Matic Dice Shaker in the game 'Frustration'.

As well as straight-forward disconnections causing non-working bulbs, the 2-pin MGB indicator/turn flasher is very particular about the amount of current it needs to work - it doesn't have to drop very far due to low battery voltage or bad connections before you start to get slow, or non-flashing where the lamps stay lit all the time. Incorrectly rated lamps will cause problems, as will 'tired' bulbs that were originally the correct rating but have become high-resistance internally with age. This is another difference to the Mk1 3-pin flasher units, which even with one bulb disconnected will try to operate and can give a faint click and a very brief flash of the other bulb. It's unfortunate that the change in design that meant they light the lamps immediately the switch is operated, also made them so sensitive to current. But on new cars out of the factory that wouldn't have been a problem, only to us 40 years down the road!

Hazard flashers, on the other hand, are designed to work irrespective of how many lamps are working. The car may have been in an accident and a corner may be smashed, but you want as many lamps as you have left to flash a warning to other road users, even if only one lamp is left working. Hazard flashers will also continue to flash even as the battery discharges and the voltage and hence current drops. Again, you want to warn other road users for as long as possible. Hazard flashers can be useful in de-bugging indicators. North American Mk2 cars had hazard flashers fitted from the factory. The circuit diagrams shows these as being cylindrical 3-pin with the third pin flashing an additional hazard tell-tale lamp even though the indicator/turn signal repeaters are also flashing. However the ATJ8880, Lucas 9FL flasher unit listed in the Parts Catalogue only has two pins, and from 1972 the circuit diagrams show the additional tell-tale being connected to the hazard switch instead, and a 2-pin hazard flasher unit is shown. V8s have the same part number listed which is shown as 2-pin in the diagram, and the same applies to UK cars when they got hazard flashers in 1974. UK cars never had the additional tell-tale. However there was a Lucas cylindrical 3-pin hazard flasher available at the time, similar to the original MGB indicator flasher, so perhaps that is where the confusion on the diagram stems from.

Some suppliers are showing SFB115, GFU2124 or GFU2125 as replacements for ATJ8880, but these are indicator flashers only capable of operating two 21w bulbs, they are not hazard units which need to flash four 21w bulbs. The correct item is SFB130, aka 35053.

When can they be used? Passing a car travelling slowly on the inside lane of a dual carriage way with its hazard warning lights on the other day I said to The Navigator "I'm not sure that's legal". From years ago the only scenario I was aware of was to warn of a stationary hazard, such as a breakdown or obstruction in the road. More recently on motorways and dual-carriageways people use them to warn of sudden braking ahead - and that is what The Highway Code rule 116 basically says, although it's not entirely clear. It says "You MUST NOT use hazard warning lights while driving or being towed unless you are on a motorway or unrestricted dual carriageway and you need to warn drivers behind you of a hazard or obstruction ahead" with no punctuation as if the whole sentence applies to 'driving' OR 'being towed'. I suspect the actual case is that you must not use them while being towed, but can use them to warn of a hazard or obstruction ahead. Of course if you are being towed AND see an obstruction ahead ...

One motor insurance company has this to say "You're driving slowly in poor weather conditions, such as fog. In this case, your hazard warning lights will make you more visible to other road users." Maybe so, but that's what rear fog lights are for, and they have their own rules. It goes on to say "Some specialist vehicles might also make use of hazard warning lights in certain situations. For example, an ice cream van may flash its hazard warning lights when it's stopped to sell ice cream."!!

It's becoming more common for vehicles to turn on the hazard warning lights automatically if involved in a collision, and some turn them on automatically in very heavy braking, as well as flashing the brake lights automatically.

Indicator/Turn Switch Updated December 2009

Cancelling
Cowl Positioning

No less than 13 variations over the years, although several were to cater for LHD and RHD of course, plus other territorial variations.

  1. Cars to chassis number 161086 had a switch with contacts for selecting which dash tell-tale would be lit as well as contacts for controlling which side of the car would be lit. There were three variations on that - the second being to angle the stalk closer to the steering wheel, the third had what Clausager describes as "longer peg for more positive location". Only one replacement switch seems to be available - BHA4628
  2. From chassis number 161087 non-North American cars up to chassis number 187169 (roadster) and 187840 (GT) had a switch with a headlamp flasher - BHA4898, except Japan which had a switch without the headlamp flasher - BHA4897.
  3. North American Mk2 cars and non-North American cars from chassis number 187170 (1969 model year) up to chassis number 219000 (1970 model year) had a switch with dip/main, headlamp flasher and horn functions - BHA4948
  4. From chassis number 219001 (1971 model year) until 410001 (1977 model year) all markets had the horn button back on the steering wheel, 37H8050 for chrome bumper, 37H8101 for RHD rubber bumper and all V8 with legends for function, 37H8523 for LHD with words for functions.
  5. For the 1977 model year for the remainder of production the horn button returned to the indicator stalk, AAU4991 for RHD, AAU4995 for USA, AAU4993 for Canada. These act as the mounting plate for wiper switch (AAU4992, AAU4996, AAU4994).

August 2023: Replacing the indicator/lighting/horn stalk on a pal's 78 (indicator contacts not functioning) with a new Lucas-boxed unit the mounting collar on the switch has a tab that is supposed to locate into a cut-out on the column outer. This tab can clearly be seen on the earlier switch (above the legend 'Contact fingers') and prevents the switch assembly from rotating about the column outer as the stalks are operated. Or it should but the tab on this new switch will not go into the cut-out, I think I can feel where it is trying, but is maybe too wide. On a subsequent visit I was determined to resolve this so took the switch off and did a trial fit of the old switch which went straight into the location notch. Took that off and put the new one back on ... and that also went straight into the location notch! Weird.

Cancelling:
For cancelling, cars up to chassis number 187211 (basically 1970 models that began at chassis number 187170 in September/October 1969) had a peg screwed into the column. After that until 1977 the steering column has a clip which is a tight fit but can be slid round to the correct position. For 1977 on a different arrangement was used where the steering wheel itself interfaces directly with the indicator switch.

The inconvenience with the early peg is that the whole column has to be turned in the UJ to get the correct alignment, and then the wheel turned on the column, whereas the clip can just be slid round to the correct position. Both types slide under fingers on the switch and lift them out of the way as you make the turn. With the early type as the wheel is returned the peg catches the metal finger, which lifts up the spring that is holding the stalk to one side, and the stalk should return. June 2015: Note that this type of column inner slides freely in the tube and if removing and refitting or replacing the column as a whole you may have to adjust the position of the outer in its clamp brackets, i.e. slide it up or down relative to the inner, to get the indicator switch in the correct position relative to the cancelling peg, even though the switch position on the tube can be adjusted to some extent. The position of the inner is determined by the U-joint and rack.

I've always had my clips facing the indicator switch with the wheel in the straight-ahead position and that way the wheel only needs a 1/4-turn in order to cancel when the wheel is straightened again. However three people on the MGOC MGB forum stated that their pegs face away from the switch, and the wheel needs to be turned 3/4 before it will cancel the indicators. The 77 and later indicator switch has a collar that is continuously rotated by the wheel and has an identification rib on one side of the collar. If that is facing the switch then again only a 1/4-turn of the wheel is needed before it will cancel as the wheel is straightened. If it faces away from the switch then it needs a 3/4 turn before it will cancel. Obviously off the car it can be positioned anywhere, but having looked at half a dozen suppliers photos they all show the rib facing the switch, so I'm pretty sure that is how they should be on the car. Also checking on my Mercedes A-class that only needs the wheel to be turned 45 degrees to hear the click of the cancelling mechanism (how agricultural ...) which will then cancel the indicators as the wheel is straightened.

On the later type of indicator switch with plastic fingers the cancelling cam engages with the end of the finger and physically pushes the switch back to the central position. The fingers can wear such that the cancelling cam just lifts the fingers up again rather than bearing on them to cancel the switch, as well as the fingers having broken off or the cam being in the wrong place or missing.

Note that the clip-type cancelling cam or striker changed twice - once in June 73 on 4-cylinder cars from BHH254 to BHH1301, and again in September 74 for rubber bumpers to BHH402. This later change was for the full energy-absorbing column and column-stalk mounted OD switch that V8s had always had, but neither column nor switch seem to have changed on 4-cylinder cars in June 73. Roadwarrior says one was taller than the other, but he also says that when that is fitted to the wrong car the problem is that it causes the indicators to cancel as you start making the turn as well as when you straighten up again. But that is a different problem to the one that led up to him making that comment on the MG Enthusiasts Forum - non-cancelling - and may be the same cause but in the other direction i.e. the lower cam fitted where there should be the taller one. I've had to build-up the one on Bee, possibly after I changed the switch but I can't be sure. The V8 with the original switch (not changed by me at any rate) and striker has never been a problem.

February 2020:
I get both cowls off and compare the cams and columns. The upshot is that the V8/RB cam is 'taller', but as well as that the column shaft is wider. So whilst in error the CB cam could perhaps be forced onto the V8/RB shaft it may well not be tall enough to push the indicator switch fingers back. Also whilst the V8/RB cam being taller may operate the CB indicator switch better, it will be a looser fit on the smaller column so may not stay in place. Possibly 'pinch it up' enough to grip, but the curvatures would still be different.

Update September 2007:
1977 (and later) model-year cars have a special wheel boss with two projections that engage with a cancellation collar on the indicator/turn switch itself. In some ways this 77-on arrangement is best because all that needs to be done is to correctly align the steering wheel for the straight-ahead position. But if an after-market wheel is fitted, or if the later dual-stalk column switch is fitted to an earlier column, the wheel won't have the necessary protrusions to engage with the slots in the cancelling collar. The later clip could possibly be fitted to the column shaft, but is too wide to fit in one of the slots in the cancelling collar. A peg screwed into the column shaft would work, but I would draw the line at drilling a hole for it. On a friends car with a non-standard wheel I made a part out of a bit of scrap metal which joined together two handy holes in the back of the wheel boss, to the two slots in the switch cancelling collar.

Shortly before getting my hands on this 1980 UK model Barrie Robinson was seeking advice on cancelling indicator/turn switches on his car, which is a bit of a mish-mash of years, and he wasn't sure which column he had. He had bought a new 77 and later switch as the old one broke, but having a Moto-Lita wheel was left with this problem and didn't really want to splash-out for a new switch. I sent photos of what I had done to him, which gave him the ideas as to what to do with his wheel, making a neater job of it than I did.

June 2018:
Another possibility where there is an existing through-hole in the boss, is to use a rod or bar in the hole to engage with one of the slots in the indicator switch. Pre-1977 and after-market wheels may have a suitable hole, originally used for the centre horn-push pencil or connection wire, once the slip-ring for the horn connection is removed.

Where is the flasher unit? April 2020

Indicator flasher units seem to be in much the same place for various years and markets, to the left of the wiper motor. Originally the hazard flasher and its fuse were behind the centre console. Eventually - probably 1977-on - the fuse moved to a more logical (and accessible) position below the fusebox, and the flasher unit to beside the indicator flasher.

Dash tell-tales: October 2019

Mk1 cars have two wires to the tell-tales as they have connections to both the flasher unit and the indicator switch and need insulated bulb holders 37H5181 similar to the ignition warning light. The bulb is an MES E10 screw-fitting (GLB987).

Mk2 CB RHD (and LHD without the padded dash) cars have a holder with only one wire to the tell-tale (13H1924), the holder picking up an earth from the bracket on the back of the dash panel. The same MES E10 bulb as above. In front of this bracket is a tube which concentrates the light onto a green plastic lens, which is positioned behind an arrow-shaped cut-out in the dash panel.

  Early padded dash have a unique dash fitting and a two-wire (green/white or green/red and black) claw-type bulb holder similar to Mk1 cars, and takes the same MES E10 bulb.

Later padded-dash, V8s and rubber bumper tin dash have a tubular lens pushed into the front of the dash panel, secured from the back with a spire clip retainer. The bulb holder has two wires, the second providing an earth as the mounting panel is insulated, and pushes into the lens from the back. The bulb is a BA7S (GLB281) i.e. bayonet-type.

From 1977-on the Parts Catalogue shows the same bulbholders and retainers securing the four warning lights round the fuel gauge. With the smaller squarer rocker switch the rear fog-light switch has an internal bulb (LES GLB921 screw fitting) for the tell-tale function showing orange as well as the night-time illumination showing green.

Fault diagnosis:

Indicators/turn signals
Hazards
The curious case of the fuel pump that didn't click when turning on the ignition February 2022

The first and easiest step for either flasher unit - once you have located - is to bridge the two wires at the unit - green and light-green/green for indicator and brown and light-green/green for hazards. If the lights now come on with the appropriate switches operated (ignition on and indicator switch operated to either side for indicators, hazard switch for hazards) but don't flash then the flasher unit is faulty and these are GFU103 or FL5 for Mk1 indicators, SFB118 or 8FL for Mk2 and later indicators, and GFU2204 or SFB130 for hazards. But if still no lights, read on.

 Indicators/turn signal faults: Updated July 2015:

From 1962-67 a cylindrical 3-terminal flasher unit was used, not be confused with later electronic 3-terminal units. After that a smaller 2-terminal rectangular unit was used, and the two differ as far as fault diagnosis is concerned. On the later type if a bulb fails or there is a disconnection in the wiring to it, then the indicator won't tick, the bulbs won't flash, but the working bulb and the dash tell-tale will glow continuously, making it easy to see which end you need to investigate. But on the early type this won't happen, except that you may get a very brief and faint flicker on the working bulb.

If neither side lights (including dash tell-tale) or flashes it's probably the flasher unit, the voltage supply to it, the indicator switch, or connections between them. If you have hazards and they don't light or flash either then it's either a fundamental problem at the hazard switch, or two or more separate faults.

  Apart from that where both corners light but don't flash one side it is probably weak connections that side and this printable schematic and chart should help you to plot the voltages through your circuits and locate those connections. You don't need the engine to be running but it is more realistic if it is from a voltage point of view, and won't flatten the battery (dynamo-equipped car may need a bit of fast idle to extinguish the ignition warning light). If you don't run the engine (saving fuel) then disconnect the coil to prevent it overheating which it almost certainly will do with the ignition left on for a long time (with the exception of certain electronic ignition systems). Non-flashing could be due to a failed flasher unit of course as well as bad connections. An ammeter in place of the indicator flasher unit should ideally show 3.5 amps, if it shows 3.2 amps or higher but doesn't flash then the flasher unit is probably faulty. With bad connections a new flasher unit may well start them flashing again, but this can simply be due to its being new and more sensitive, as it 'burns in' they will probably slow and stop again. If you are investigating slow flashers, and they continue to flash slowly, it makes measurements easier if you bridge the two connections on the flasher unit to stop it flashing. See here for the results of the tests on Vee.

  This sensitivity to current was deliberate to give warning to the driver if a corner should have failed, otherwise traffic around you may not realise you are preparing to turn. Modern flashers use electronics (instead of a heated bi-metallic strip) and flash rapidly if one of the main lamps fails and are nowhere near as sensitive to slightly bad connections as the original 2-pin units. Some people fit a modern electronic flasher to their classic car when they get slow flashing from poor connections, not realising that they will still be there and causing the lights to be dimmer than they should be. A pal fitted one of these a while ago, but on doing voltage tests for another issue more recently found he was getting less than 8v at each rear light. Whilst fitting an electronic flasher unit to get round 'slow flashing' problems because of bad connections may seem to have done the trick - temporarily, eventually you may have to find and fix the root cause(s), you might as well do it now and get brighter lights. Be aware that some aftermarket types flash at the same rate regardless of current and therefore give no warning of lamp failure, so if you wimp-out and fit an alternative flasher unit disconnect one of the bulbs with it fitted and make sure that the flashing speed changes. If not, you run the risk of being rammed up the back because the person behind had no idea you were going to turn so didn't expect you to slow down. Electronic flasher units have their own problems - or at least people have problems with electronic flashers, as I have known of at least two occasions where the driver was blissfully unaware that because one side flashed at twice the rate of the other it indicated bulb failure, they hadn't even noticed! As Einstein reputedly said "Only two things are infinite - the universe and human stupidity, and I'm not sure about the former." Likewise LED Bulbs have their own issues.

  As mentioned before the indicators circuit is: Battery - starter cable - brown circuit - ignition switch - white circuit - fuse (note 1) - green circuit - hazard flasher switch (note 2) - green circuit (note 3) - indicator flasher - light-green/brown circuit - indicator flasher switch - green/white (RH side) and green/red circuits (LH side) - indicator bulb holder - indicator bulb - indicator bulb holder - body earth (note 4) - battery earth cable. CB cars with twin-6v batteries also have the battery link cable.

Note 1: Later cars have an ignition relay and white/brown circuit between the white circuit and the No. 2 fuse.

Note 2: The indicator flasher is wired via the hazard switch so that it is disconnected when the hazard flashers are turned on, and only works when the hazards are turned off. This prevents the outputs from the hazard and indicator flashers from conflicting with each other, but more importantly prevents the hazards feeding power back through the indicator switch, indicator flasher, green circuit, fusebox and onto the white circuit and so energising the fuel pump and ignition (my thanks to Mark Childers for pointing this out). So don't be tempted to bypass the hazards switch if it is that which is causing your indicators problems.

Note 3: The 'green' circuit from the hazard flasher switch to the indicator flasher should really have its own tracer colour as it is no longer part of the 'real' green circuit.

Note 4: Rear light clusters on all cars and front indicator/parking light clusters on CB cars have the cluster picking up an earth from their physical fixings to the body. RB front indicators have a wired earth shared with the headlights/front parking lights.

Typical indicators faults can be "They don't work at all" or "They don't work on one side" or "They light but don't flash or flash too slowly" or "They flash but so do other lamps" or "They don't cancel".

  "They don't work at all"

Do you have hazards?

Yes - do they work?

Yes, but only some of them work - follow through the continuity of the lamps that don't work. Could be bad connectors, corroded lamp holders, blown lamp or bad earths. If only one side flashes with the hazards switched on it could be a dirty contact inside the switch or a bad connector in the green/white or green/red circuit as applicable. Then follow the following paragraph.

Yes, they all work - if all lamps are flashing then that indicates that there is continuity at the lamp ends of the green/white and green/red circuits, although they could still have connections bad enough to affect the rate of flashing of the indicators. Now check the green circuit for 12v through the hazard switch (which needs to be off. Note that dirty contacts in the hazard switch are a frequent cause of indicator problems that affect both sides) and the indicator flasher to the indicator switch (if you suspect the flasher unit itself just bridge its two contacts. The lights should light, but not flash). Then through either the green/white or green/red circuits out toward the lamps. Pay particular attention to any volt drops anywhere except across the indicator flasher itself, which typically drops about 0.25 volts when the lamps are lit (and 12v when they are in the 'off' part of the cycle).

No - see below.

No - check the green circuit for 12v through the indicator flasher to the indicator switch. Then through either the green/white or green/red circuits out toward the lamps. Pay particular attention to any volt drops anywhere except across the indicator flasher itself, which typically drops about 0.25 volts when the lamps are lit (and 12v when they are in the 'off' part of the cycle). If you suspect the flasher unit itself just bridge its two contacts. The lights should light, but not flash.

  "They don't work on one side"

If neither lamps on one side flash or light you could have one fault in the common circuitry e.g. the switch or the connectors by the steering column (in which case the dash tell-tale won't light either), or two (or more) independant faults in the wiring out towards the lamps. Track the 12v through the indicators switch and the green/white(RH side) or green/red (LH side) wiring out towards the lamps.

  "They light but don't flash or flash too slowly" Updated April 2013

This is an indication of either a failed flasher (which affects both sides equally) or bad connections out towards one or more lamps. Pay particular attention to the front indicator, although chrome bumper and rubber bumper are obviously different, both are subject to water and salt being thrown forwards by the wheels and hence corrosion. Chrome bumper cars earth through their physical fixings, and whilst rubber bumper units have a wired earth shared with the headlights it's connection to the light unit is external and unprotected, using a type of bullet. The bullet is large and hollow i.e. not crimped to the wire as elsewhere. The wire is stripped, pushed up the middle, and the conductors folded down the outside, then this is pushed into a flimsy clip. The result is poor - worse than the body earth arrangement of CB fronts and all rear light units. One of these on mine was losing nearly a volt, and the other nearly half a volt. Removed and cleaned up got then down to 0.2v and 0.1v respectively - a big improvement but twice the rears. If the shared earth i.e. from the bullets by the headlights to the body earth is high resistance the indicators may not flash with the headlights on. If you replace the flasher unit with an ammeter ideally you will see about 3.5 amps drawn per side. The more this drops, the slower the flash rate will be, particularly with the engine stopped. But if you see 3.2 amps or above (i.e. satisfactory connections through the whole of the circuit) and it doesn't flash then almost certainly the flasher unit has failed. Note that replacing a slow flasher does have a good chance of speeding things up, but they usually have a 'burn in' period then slow a bit to their 'normal' flash rate, meaning you end up no better off. If you see 3 amps or less it might be an idea to go to the corners of the car and do the last few tests first, as it could be that you have incorrectly rated or very tired bulbs. As bulbs age the filament thins, which reduces the current, and that will slow the flasher. The symptom of this would be a good voltage at the light unit but still a low current at the flasher.

Investigating slow or non-flashing where the cause is low current is probably the most difficult electrical job on the car, and can be very frustrating, the only way to deal with it is in a logical and methodical manner. As well as being the most sensitive circuit on the car to bad connections, there are more connections in this circuit than any other - around 30 just to flash two bulbs on one side! Any electrical circuit will 'lose' some voltage in wiring and connections when carrying current (and ours have up to 50 years of oxidisation to contend with), so my recommended methodology involves taking voltage measurements at certain points along the circuit, all of which can be done with minimal disturbance to wiring and connections. By working along the circuit you can spot a sudden drop in voltage, which means there must be a problem between this point and the previous one. However rather than testing every single one in strict order, it's more efficient and will save time if you test certain key points first, then use that to decide whether the intermediate points need investigation or not. For example if you only see a 0.1v drop between two points that have three other connections between them, there is no point testing those three other connections. The first half of the tests are all on circuitry that is common to both sides, but after the indicator switch you needs to take one set of measurements for each side, and shortly after that one set for each corner. Whilst slow or non flashing both sides will lead you to think it must be a common problem, it's just as likely for there to be problems on both sides.

The ignition will be on with the engine stopped for these tests, so the coil should be disconnected to prevent that overheating. It also reduces the load on the battery. With 3 amps or more you will still be discharging the battery noticeably, and you will need to know when to stop or whether to connect a battery charger during the tests to avoid discharging it too much. The other thing is that while testing, and discharging the battery, its voltage will be dropping anyway, so you need to take this into account when you are testing along the circuit by periodically remeasuring the first test point, or you could be led to think there is more voltage being lost the further you go along the circuit than there actually is. If you only operate the indicator switch long enough to take each measurement, and turn it off while moving the meter from point to point, you will minimise the drain on the battery. Finally it's not going to be easy measuring voltage on either an analogue or most digital meters if the flasher is going, even slowly, so it makes sense to bridge the green and green/brown wires at the flasher unit so the lamps are glowing continuously while taking the measurements.

This schematic and list should help you plot the voltages through the circuit. The list works along the circuit connection by connection, but some are conditional i.e. only performed if a test earlier in time, but later in the circuit, shows a bad connection. You will probably end up with a progressively dropping voltage as you go from point to point. Writing these down you will see where the biggest drops are, and tackling those first will give you the biggest improvement. That way, when you get fed up, the worst ones should have been done! Note that the last few are earth tests so in an ideal world these will all be zero, so switch your meter to a lower range if appropriate. Any voltage seen in these tests indicates a bad earth. Note that as well as the centre contact of the bulb being a possible cause of a bad connection, which is inaccessible without removing the bulb, there is also a connection between the bulb base and the holder, the holder and the light unit, and the light unit and either the body (chrome bumper front lights and all rear lights) or the wired earth (rubber bumper front flashers), all of which can cause problems.

The attached shows the measurements on my V8, not because I had a problem but as a practical indication of the sort of figures you might get. The first thing to say is that I have a quality AVO analogue meter, and a cheap digital, and I got some weird and inconsistent results between the two. The first problem was that at the battery connections the analogue read 12.2v but the digital only 11.2v, both with everything switched off. I've seen this before with a digital dash voltmeter - which rather goes against the point of its existence. Subsequently I compared those two with a third, analogue Gunson's instrument, and with all three connected at the same time I got 0.5v difference between the original two, and the additional analogue instrument was lower again! It would be tempting to say the digital must be the most accurate, and the AVO reading high and the Gunson's low. But when doing earth tests at the first light unit (right front) with the digital it didn't matter whether the probes were connected or disconnected, the display kept hunting around the 200-300mV area. If I connected the probes together, or even put my thumbs on them, the reading stepped down to zero. So I tried my AVO and that immediately showed 0.9v on the bulb base, holder and light unit. So that, together with previously having found it increasingly more inaccurate as the resistance value got higher, means I don't have full confidence in it. Nevertheless, it is comparative values along the circuit that we are going to be looking for rather than absolutes, and as the digital is much smaller than the analogue I used the digital to move around the car and left the analogue connected to the battery so I could monitor it's reducing voltage through the test process. In the event I got half way through without seeing any drop (from 11.9v under indicator load) so stopped recording it for a while. I checked again near the end and it had dropped to 11.3v, after maybe 3 hours of switching the indicators on and off and moving from point to point.

  "They flash but so do other lamps"

This usually affects rear lamp clusters and front lamp holders of CB cars and is usually caused by a bad earth. Most noticed when another circuit in the cluster is powered at the same time as the indicators as other lamps flashing in time with the indicators, it is caused by current flowing backwards through any other lamps that share the same faulty earth to whatever other earth it can find. All rear lamp clusters, and CB front parking/indicator light units, earth via their physical fixings to the wings.

  "They don't cancel" Updated September 2007

A mechanical problem, this, rather than electrical. Up to the 77 model year cancelling is performed by a cam or peg at the top of the steering shaft engaging on one of two fingers projecting out from the switch. Early columns have a peg screwed into the shaft in a fixed position, later columns have a cam which is a tight sliding-fit on the shaft. With the indicator switch in the 'off' position the cam or peg clears the switch fingers as the wheel is turned in either direction. Operating the switch moves one or other of the fingers into the path of the cam or peg as the wheel is turned. When turning in to the corner the peg or cam passes under the finger, lifting it out of the way. Then when you straighten up the peg or cam pushes against the end of the finger to cancel the switch. From 1977 on the switch (both stalks on a single plate) fitted over the steering column instead of bolting to the side and includes a rotating 'cancellation' collar with two notches. The steering wheel itself engages with these notches to turn the collar and cancel the indicator stalk if it is operated. After-market wheels probably won't have the ability to engage with this collar, and so won't cancel, see here for suggestions on how to interface an after-market wheel to this type of switch. Cancellation is the same as before, i.e. on straightening up from a corner, but the mechanics of the operation are concealed inside the switch.

For both peg and cam types, with the wheel straight-ahead the peg or cam should be pointing at the middle of the indicator/turn switch. For the 77 and later type there is a rib on one side of the cancellation collar, and again this should be pointing at the middle of the switch. If the peg is in the wrong position on the shaft to cancel the switch correctly the column shaft, UJ and rack shaft have been incorrectly assembled. The UJ is clamped onto each shaft with a bolt, and this bolt passes through a cut-out in the shaft so that even if the bolt becomes loose the shaft cannot pull out of the UJ (the bolt has to be completely removed to withdraw either shaft from the UJ). However, although the column shaft only has a notch for the bolt, meaning that it can only be inserted into the joint in one position, the rack shaft has a groove machined all the way round so that it can be assembled in any position. Use this feature to get the peg in the correct position. You will probably then have to alter the position of the wheel on the column shaft (click here for how to remove the steering wheel) to get the correct 'straight-ahead' orientation of the wheel.

Problems can be caused by worn or broken fingers on the switch. Building up the height of the cam or judicious bending of the fingers with heat (don't break them!) can compensate for this. Broken fingers may be able to be jury-rigged - you will have to judge.

The sliding cam can become loose on the column and slip round instead of cancelling the switch. You could try removing the cam and closing it up a bit making it a tighter fit on the shaft, or degreasing and roughing-up both surfaces, gluing, or as a last resort drilling a hole and fitting a small screw through cam and shaft (but drilling holes in things like steering shafts isn't really recommended).

 Hazard faults:

The hazard warning circuit is: Battery - heavy current cable - brown circuit - in-line fuse - another brown circuit - hazard flasher - light-green/brown - hazard flasher switch - then out on the green/white (RH side) and green/red (LH side) circuits to the lamps at the corners of the car as with the indicators. The hazard flasher fuse was originally behind the centre console - inconvenient! - moving to under the fusebox, possibly around 1977.

Note: The indicator flasher is wired via the hazard switch so that it is disconnected when the hazard flashers are turned on, and only works when the hazards are turned off. This prevents the outputs from the hazard and indicator flashers from conflicting with each other, but more importantly prevents the hazards feeding power back through the indicator switch, indicator flasher, green circuit, fusebox and onto the white circuit and so energising the fuel pump and ignition (my thanks to Mark Childers for pointing this out).

These tests should be done with the ignition off, and all wiring connected, except where specified otherwise.

First check for 12v on the brown and light-green/brown terminals of the hazard flasher. Note: North American cars prior to 1972 have a third terminal on the hazard flasher with a light-green/purple wire. From 1972 on this wire was on the hazard switch. This is for the hazards tell-tale and should be ignored in all tests.

No 12v on either - check the in-line fuse.

12v on the brown but not the light-green/brown - hazard flasher faulty

12v on both - move on to the hazard switch

Hazard switch: Check for 12v on the light-green/brown. No 12v - break in the light-green/brown back towards the hazard flasher

12v present - operate the switch and check for 12v on the light-green/brown again

12v drops to zero - check the light-green/brown back at the hazard flasher again No voltage - hazard flasher faulty

12v still present at the light-green/brown at the flasher but not at the hazard switch - bad connection between these two points. Note late model cars have a multi-way plug and socket concealed behind the dash.

Light-green/brown at hazard switch still at 12v - turn off the hazard switch, turn on the ignition and operate the indicators to either side. Check for 12v on the green/white (RH side) or green/red (LH side) at the hazard switch No 12v on either green/white or green/red with the indicators flashing - break in the green/white and/or green/red between the flasher switch and the wiring between the indicator switch and the corners of the car. Note that on North American spec cars the green/red joins at a six-way bullet connector in the mass where the main and rear harnesses join together at the firewall by the fusebox, whereas the green/white joins on the back of the multi-plug for the indicator switch. After that, and on UK cars, both green/red and green/white join at the multi-plug.

12v flashing on and off with the indicators - hazard switch faulty. This can be confirmed by cancelling the indicators and turning the ignition off again, then bridging the light-green/brown to either the green/red or the green/white (or both together) wires that go to the hazard switch, removing the plug from the switch if required. If the remainder of the hazard circuit is good it will start to flash the lights.

Indicator Flasher Replacement Updated June 2024

Fraught with pitfalls even as far as the original as far as the two-pin Mk2 MGB type is concerned. The 'Type' number on original Lucas 8FL indicator units tells you what bulbs they are designed to flash, and there are Type 36, Type 41, Type 43 and Type 51 where the type number specifies the current they are designed to work with i.e. 3.6 amps, 4.1 amps, 4.3 amps and 5.1 amps, each with a different part number 35048, 35049, 35057 and 35052 respectively. The colour of the printing also varies - Type 36 are usually black, Type 41 yellow, Type 43 blue, and Type 51 green but seemingly not exclusively so. The reason for the subtle variation between units is that this design of flasher unit is very sensitive to current flow, which is dependant on the number and wattage of bulbs, and will give an incorrect flash rate when used on the wrong vehicle. The MGB has two 21w bulbs at the corners and one 2W tell-tale on the dash. Together this comes to 44W and at a nominal 12v this results in a current of 3.6 amps and the type 36 flasher unit is the correct one to use. Type 41 are expecting 4.1A i.e. 49 watts and that needs an additional 5W repeater flasher bulb front wing to flash at the correct rate (the other two need even more bulb current) and on the MGB they will either flash slowly or not at all. Current-stock generic no-name flasher units labelled '21WX2+5W' are effectively Type 41 and even with the engine running i.e. dynamo/alternator charging they can be flashing at below the legal limit which is 60 fpm, and those labelled '(2X21W)+(6W)+(2W)' as Lucas SFB115 will be the same unless you connect up additional bulbs somewhere! However three SFB115 I tried had hazard flasher internals instead of indicator internals which raises two completely different issues, and so-called 'universal' flasher units are the same.

The upshot is that the MGB needs 8FL Type 36 30548 only available as NOS and rare, or (supposedly!) Lucas 35048. Googling that brings up loads of motorbike sources, one ordered for testing. However on arrival it is effectively a hazard flasher as it only comes on after a pause instead of immediately, and it doesn't change the flash rate on bulb failure. So as three SFB115 I purchased did exactly the same which I put down to a manufacturing error, I'm wondering now if they are all deliberately like that to get round complaints of slow flashing! Maybe the only alternative to NOS 8FL Type 36 (rare!) is the next generation electronic flasher unit for filament bulbs used in the 80s and 90s which is nowhere near as sensitive to voltage and current and has the bulb failure warning which in this case flashing the remaining bulb at double-speed when one corner has failed. But remember, if the slow flashing of the standard flasher unit is down to some high-resistance connections that will be making the bulbs dimmer than they should be, even though they will flashing at a good rate. As an aside my Mercedes A-Class with LED lighting has the indicators lighting immediately for both indicators and hazards, good flash rate engine not running, don't know about 'corner failure' though.

July 2024:
I was beginning to think I had dreamt the 'flashing at double speed if a corner has failed' as flasher units for MG, Rover and land Rover models from 1983 such as YWT10003 state "21Wx2+(5W) 21Wx4+(2X5W) Max.98W" which implies they will flash either four or two at the correct rate, so how can they flash at a different rate with only one indicator bulb working? But Googling throws up loads of references that flashing at double-speed does occur to show that a corner has failed when using the indicators. I purchased one and it is indeed clever enough to flash either two (indicators) or four (hazards) at a better rate than the standard 2-pin flashers, ignoring changes in voltage, but if only one corner is working it will flash at double-speed. Unlike the original 2-pin which will work connected either way round these have to be connected correctly - green (12v supply) to terminal 49, green/brown (indicator switch) to 49a, and earth to 31. The first click lights the bulbs at the corners of the car and is almost instantaneous. But as I say above if the slow flashing of the standard flasher unit is down to high-resistance connections they will be making the bulbs dimmer than they should be, and that will still be the case even with this electronic unit flashing at a good rate.

In the following video with the engine not running see how the flash rate of a good standard flasher unit (as indicated by the voltmeter tick rate) is slow with no additional electrical loads, and slows even further with lowered voltage when the twin cooling fans and HRW on the V8 are turned on. With the YWT10003 unit the flash rate is better to begin with, and maintains that rate with the increased load/lowered voltage. The final clip shows this unit flashing one bulb at double-speed when the other bulb has been disconnected:

However having tested it in Bee with a buzzer in circuit to give a louder audible warning it is buzzing faintly all the time the indicators are off, so the output terminal must have a partial path to earth - probably part of the electronics and bulb failure warning, whereas with the original unit the only path to earth is via the indicator switch to the bulbs, so more research required. It might be possible to use a diode to block the earth from the buzzer, but because that results in a volt-drop that would reduce the brightness of the bulbs slightly so not a good idea. A pair of diodes on the output of the indicator switch (to prevent cross-feeding from one side to the other) would work, but then the buzzer would sound with the hazards. Further down the diode rabbit hole using a full-wave bridge rectifier with the output wires connected to the polarity-sensitive buzzer and the input wires connected one to the left side and the other to the right will only sound the buzzer when only one side is powered i.e. the indicators. When both sides are powered i.e. hazards there is no potential difference between the input wires to sound the buzzer. This does mean that the buzzer is earthing through the unlit side i.e. passing a current through the bulbs, but in a test there is no sign of the corners glowing that side which means the current is too low to have an effect, so is my preferred solution. That may be different if using LED indicators at the corners of the car, but I don't use them, and they may make prevent the buzzer from working.

August 2024: Duly installed it works as expected, so I get another flasher unit of the same type for Vee. No buzzer on Vee and the new one is definitely quieter, so I'll see how it goes. Doesn't take many miles to determine it is too quiet - virtually inaudible. So I go to resurrect the plans to use a solenoid as a sounder from Feb 2020 ... but of course I can't find the sounder I bought then so have to order another.

May 2016:
I'd noticed Bee's flash rate was getting quite sedate, even driving along i.e. full voltage, and they wouldn't flash at all with the engine stopped apart from a brief click as you switched from one side to the other. I did my voltage tests and with one exception there was very slightly less voltage lost end-to-end than Vee, the one exception being in the 'new' indicator switch assembly which was slightly higher, but that only brought it back to the same as Vee overall. Two 21w bulbs directly on the output of the flasher unit were the same, the only thing that got them to flash - albeit slowly - (with the engine off remember) was when I added a 2.2W bulb to the 21W bulbs. So I reckoned the 1978-vintage flasher unit was probably getting tired, and ordered a new one from the MGOC. Generic and no-name I was surprised to find that was twice the height of the old one labelled '2 x 21W + 5W'. Installed it does flash with the engine stopped albeit quite slowly of course, although I'm sure it's quieter. However I know these units are slightly more sensitive when new, which reduces after a short period of use to a 'long term' flash rate. After a weekend away it definitely is quieter, but more importantly noticeably slower then the old unit when the engine is running. The old unit gives 80 flashes per minute, the new one only 64, which is only just above the MOT minimum of 60 flashes per minute. Roger Parker said that in these days of frantic traffic one needs them flashing faster rather than slower, and recommended that the club shop send me another. They did, but that is even slower at 56 fpm and so below the legal minimum. I did some more tests with my three bulbs, and also powering the flasher unit off the purple circuit, which when combined eliminates almost all the cars wiring, on both the roadster and the V8. The upshot was that the fastest I could get a new flasher to run at was 76 fpm which is just about OK. The original unit connected the same way was flashing at 116 fpm, which is almost too fast, the legal maximum being 120 fpm. Whereas the original unit showed a 60% increase in flashing rate between the two extremes of connection, the new units showed only a 15% and 20% increase. Neither do the new units exhibit the slight slowing when applying the brakes that the original units do. Whether this is just because they are new, or whether the bigger can means they are different inside I don't yet know. The new units do seem to be able to ignore connection and wiring resistances much better than the original units, much as the 3-pin electronic units do. If only they flashed at a better rate when connected normally, they would be an improvement but as they are I think they are incorrect in this application. I've sent the results of my tests to Roger, and await developments. In the meantime I opened up one of the new flasher units to find - not surprisingly - it is the same as another I have. I tried tweaking the contacts, which did make it flash at an acceptable rate at 12v, however with the engine running i.e. at 14v it's initially very erratic either galloping or not flashing at all before settling down to an acceptable rate, so on balance I have put the original unit back in. It seems to me this unit would only be acceptable if you added a 5W bulb or 30 ohm resistance to the circuit to draw the current that this unit is expecting. You could add one to each 'leg' of the indicator circuit i.e. one per side, however a single one connected on the output side of the flasher unit would do the same job - at the 'expense' of the bulb glowing all the time or the resistor taking current all the time the ignition was on even when the indicator switch hasn't been operated. This is because the indicator flasher is 'normally' closed, to give the immediate lighting of the lamps at the corner of the car as soon as the switch is operated.

July 2018:
Happened to come across an NOS 8FL indicator flasher in a red bubble pack for £5.50 on eBay so snapped it up. Ticks faster than either of the units that came with Bee and Vee, and on Bee flashes with the engine stopped, so far so good. However it's noticeably quieter, so a couple of times I've found I've left it on after a turn. September 2020: while testing other aspects of the lighting the indicators have gone back to not flashing with the engine off, just one tick changing from one side to the other as before, proving that there is a 'burn-in' period and they deteriorate slightly for the bulk of their life. Still plenty fast enough with the engine running. Subsequently I purchased two more supposedly NOS but without packaging and whilst they lit they didn't flash even at charging voltage. Took it up with the supplier who pointed out mine was type 36 and his were type 41 which needed an additional 5W repeater on the front wing to work correctly. Mea Culpa, but he did refund me on return.

July 2023: Just lately the indicators have been noticeably slower, so I was thinking that perhaps the NOS 8FL above had aged a bit now and was as bad as the others. I still have the after-market ones that were slower than an old FL, tried them and they were no better, in fact one of them kept blowing the fuse because one of the terminals had pushed through the plastic base when I was fitting the spades and was shorting-out internally. That was because when I pulled the spades off the 8FL they didn't seem very tight so I had pinched them up a bit. Went back to the 8FL and with the tighter connections it speeded up significantly, and has remained so.

Note that while modern 3-pin electronic flasher units don't suffer from the slow flashing problem, neither do they warn you of worsening electrical connections that will be resulting in dimming bulbs.

Adjustable flasher units: June 2024 The question of indicator flasher units cropped up on the MGOC forum again and I recommended looking for NOS 8FL as bought a few years ago. I found an eBay seller with two, so bought them, but on receipt whilst they lit the bulbs they didn't flash, even with the battery charger connected. I complained sending a photo of my existing unit which worked (of course), and he said that's because mine is a Type 36 whereas his are Type 41 and only suitable with additional bulbs, which led to the paragraph at the top of this section, so returned them for a refund. Whilst looking to see if I could find any more I discovered loads of 'adjustable' types advertised for motorcycles, saying they were suitable for 2X21W filament bulbs or LED bulbs which take a very much lower current, just turn the adjuster on the base for anywhere between 50 and 200 flashes per minute. The first one I came across was £15, but referenced in the eBay ad were loads at £5 and less, so one of those was ordered for testing. But they're exactly the same as the Lucas 35038 above with delayed bulb lighting and no bulb failure i.e. they are no better than hazard flasher units except that you can increase the flashing speed to whatever you want.

LED indicators Updated December 2017

Many new cars these days are being fitted with external LED lighting including indicators, so it is inevitable these are being offered by aftermarket suppliers as replacements for filament bulbs. The first thing to be aware of is that in the UK at least, and at the time of writing, external LED lighting is not legal for use on public roads, only off-road or at shows, and vendors for the UK market have to state 'off-road use only' somewhere in their advert . They are not specifically banned for cars before 1986, but there is enough conflict between two sets of regulations currently in force in the UK to make it a 'grey area'. Whilst the Police almost certainly won't stop you, and the MOT only checks for function (and adjustment of headlights), you should check with your insurance company first. I did, and was told they would not be acceptable, and they could affect any claim, even if I had informed them of the change (which you have to do). UK law changed on 6th February 2023 to allow them but you may still need to check with your insurance company.

Secondly they will not work correctly with either the original thermal or 'standard' electronic flashers - in the former case both bulbs will light but not flash, in the second case they will flash but very quickly indicating bulb failure. Some vendors supply a load resistor with LED lamps so that the original flasher units (thermal and electronic) flash at the correct rate, but then the 'bulb failure warning' feature in the original flashers will only detect resistor failure, not LED or wiring failure!

You can get indicator flasher units specifically for LED lamps which should flash them at the correct rate - 2-pin and 3-pin where the third pin goes to earth. But again, these do not tell you when a corner has failed and so are equally as unsafe as load resistors with standard flashers, and with some types if you connect more than one filament bulb to them you will burn them out. Also some contain a relay and make an audible click and some do not. April 2018: Out of interest I purchased a 3-pin type advertised as being for 0.02A to 20A so apparently suitable for anything from all-LED to all-filament indicators and hazards. But on all-LED they just flashed once, even though there was a current of 0.9A. It needed an filament tell-tale bulb added to boost the current to about 1.3A before they would flash repeatedly. Ends

A supplier of a 2-pin type claims that they are the units fitted to BMWs (but at only £14 I find that unlikely). Also I'd be surprised if modern cars were allowed to get away with there being no indication of failure as IMO it is a significant safety hazard. On questioning the supplier they defended themselves by saying LED bulbs were much less prone to failure than filament which is correct, but there can still be wiring or connector failure as before which will have the same effect. They defended that by saying failure warning types were being developed but had no date for availability, something I find difficult to believe when they are supposed to be OE units. I've not been able to establish yet whether OE use does or doesn't have failure warning, but recently I was behind a Range Rover with these bulbs which stopped in the middle of the road prior to turning across the traffic without indicating. Cussing the driver under my breath I then noticed as he turned that the side marker was flashing at the correct rate! Which indicates (ho ho) that Land Rover at least may well be fitting these with no failure warning.

Note that if you fit LED lamps and the LED indicator flasher unit you will probably need to change the hazard flasher as well. Whilst hazard flashers are designed to flash anything from one to four 21w filament bulbs, they do still need a certain amount of current flowing through them, and the current from even four LEDs can be less than one filament (21 watts i.e. 1.75 amps). My 'standard' hazard flasher will - just about - flash a single 6w filament but it won't flash two 21W-equivalent LEDs and I don't have four to try.

About the only excuse for fitting LED indicators is they can be left operating for a lot longer in the event of a breakdown or accident without flattening the battery - unless you have installed load resistors at the same time and retained the original flasher unit!

Direct replacements? October 2024:

I have become aware of a type of replacement LED bulb with an internal load resistor which would seem to be a direct replacement for filament bulbs and work correctly with the original flasher units, variously described as 'CANbus error free' or 'no hyper-flash' for use in 1993 and later cars that originally had filament bulbs such as MG, Rover, and Land Rover. 'CANbus' includes a system that senses the current flowing in external bulbs and raise a warning on the dashboard if that drops below a certain level i.e. a bulb has failed. 'Hyperflash' is what you get with 80s/90s cars with filament bulbs - if a corner fails the tick, the tell-tale and the remaining bulb all go at about double speed to warn the driver. With the original MGB thermal flasher units bulb failure stops ticking and the remaining bulb and dash-tell-tale glow permanently while the switch is operated which still warns the driver and by looking at the corners makes it easy to see where the problem is likely to lie.

So far so good, but to be directly equivalent LED replacements would have to take the same current as filament bulbs, which means they would get at least as hot (I have seen one type described as having an internal fan!), and still be subject to slow flashing if there are weak connections anywhere in the circuit, which may also affect brightness. Also the correct flash rate only comes from the load resistor, if some other component inside the bulb fails and it stops lighting up then you won't know about it from the tick rate or dash tell-tales. However CANbus systems accept currents within a range - which can vary from vehicle manufacturer to manufacturer. Anything in that range will flash correctly, but step outside that and it instantly goes into hyperflash mode and/or raises a dashboard warning. For that reason suppliers like Classic Car LEDs say that despite their BA15s (the MGB fitting) being 'no hyperflash' they are not suitable for the original thermal flasher units, and even on more modern cars with CANbus may still beed a ballast resistor to extinguish 'bulb failure' warnings. They also say "We recommend that you only use your Hazard lights in an emergency. Unnecessary prolonged use of your hazard lights can lead to early failure of the LEDs." This may be a reference to the heat generated by the internal load resistors - semi-conductors do not like heat.

Some LED replacements are described as 'ballast included' but look carefully, they may consist of a 'standard' low current LED with an external ballast component you will have to wire in, where the tick rate is simply controlled by the ballast resistor (typically one 6 ohms 50W per side) and the dash tell-tale will flash correctly even if both corners fail to light up.

I've already replaced the flasher units on both MGBs with 1980s electronic units to get a good flash rate but would still like brighter bulbs. I tried a couple described as 'Built-in load Resistor & IC Driver: Canbus error free, plug-n-play, Wide voltage & constant current' (but no mention of 'hyperflash') and with both replaced one side they flashed at about double-speed i.e. hyperflash. Even with only one bulb replaced the flash rate was slightly higher and the reason for this is that each LED bulb only takes half the current of a filament bulb (probably to reduce the heat generated which is bad for electronic components) so two of these bulbs on the same side take the same current as filament bulbs where one has failed - hence the double-speed flashing. The ones I had were significantly brighter (described as 1200lM each, which Classic Car LEDs says is headlight power so almost certainly incorrect, their indicator bulbs being 700lM and their headlights 700 dipped 1200 main) but are obviously unusable without an additional load to slow them down when used as indicators. Used as hazards here (LED on the right obviously) that flasher unit is independant of current so they flash at the 'normal' rate:

Far more 'urgent' as well as bright, the question is are they too bright! If the suppliers claim of 1200 lumens per bulb is correct this equates to 95 candelas (1 candela = 12.57 lumens). I can't see any intensity information for indicators in The Road Vehicles Lighting Regulations 1989 but for stop lights which have the same wattage on the MGB as indicators it should be between 20 and 60 candelas. 60 candelas converts to 754 lumens, with Classic Car LEDs versions being 700 lumens i.e. just within the maximum. Out of interest the Regulations documents don't have an intensity figure for headlamps but they do for 'running lamps' e.g. DRLs and that is 200 to 800 candelas or 2500 to 10,000 lumens, so 1200 lumens for main beam and 700 for dipped beam as quoted by Classic Car LEDs seems pitiful, but it is in the right area quoted by other sources.

Powered from the indicator flasher I calculated I would need an additional 8.6 ohms in parallel to flash at the correct rate, and a resistor for that would need to be rated at 20 watts to cope with the current and heat. However in tests with two LEDs on one side it needed closer to 6 ohms (which is the value of the ballast resistors sold for use with non-CANbus LEDs) to flash correctly. Is it worth it? Especially bearing in mind the warning about extended use? Another thought is that if I do fit ballast resistors will the flasher unit still detect loss of one bulb? If not then it is no better than cheap non-CANbus LEDs with ballast, but in more tests it did detect the unplugging of either bulb, clicking and flashing the dash tell-tale and the remaining bulb at double-speed as it should. But at £11 per corner with VAT on top (the bright pair I tried were £11 the pair delivered) they come out at £55 (at least shipping is 'free' for four or more ...) for one car or in my case £110! Plus ballast resistors!

Remember that cheap non-CANbus LEDs need either ballast resistors or LED-specific flasher units and will lose the failure warning which means the flasher unit will tick and the tell-tale will flash at the correct speed even when one and maybe both of the corners has failed.

Adding hazards to earlier cars

Adding to Bee

Hazard flashers were standard in the North American market from the start of Mk2 production in late 1967, but not added to other models until the either the 1974 model year or rubber bumpers, it's not entirely clear. Before adding hazard flashers to earlier cars it is vital to understand that the hazard switch must disconnect the standard indicators in some way, otherwise it is possible to have the fuel pump and ignition powered even with the ignition off and the key in your hand, which is obviously a serious safety hazard (pun not intended) especially in the event of a collision. If a hazard circuit is simply added to the indicator wiring then with the hazards turned on and flashing the lights, if the indicator switch should happen to be operated, power will feed backwards through the indicator switch, indicator flasher, onto the green circuit and through the fusebox onto the white circuit. Factory flashers power the indicator flasher unit through the hazard switch in the 'off' position, disconnected when the hazards are turned on, for retro-fitting this reverse current path can be disconnected either with a suitable switch, a relay or blocking diode.

Options are:

  • Obtain a factory-style rocker switch and wire it as per factory wiring. This will require diverting the green circuit that feeds the indicator flasher unit as well as connecting to the green/white and green/red lamp circuits, and connecting the new hazard flasher via an in-line fuse to a brown circuit. A good option for UK cars with the earlier style of rocker switches as there is an empty space in the centre console, and could also be used on toggle-switch cars in a supplementary panel.
  • Use a generic double-pole double-throw switch with the green/white and green/red wires from the corners of the car connected to the two common terminals, the normally closed contacts connected to the green/white and green/red wires from the indicator switch, and the normally open terminals connected together and to the hazard flasher. Requires diverting both the green/white and green/red circuits which can be done at the bullet connectors between the main and rear harnesses in the engine compartment if you don't want to cut wires.
  • Use a generic on/off switch and two relays, one relay tapping into the red/green and red/white wires for the hazards, and the other disconnecting the indicator circuit when the hazards are in. This disconnection can be done where the light-green/brown wire connects to the indicator flasher unit.
  • Use a generic on/off switch and diodes to connect the hazard flasher to the green/white and green/red circuits via two diodes, with a third blocking diode in the green circuit feeding the indicator flasher unit.
    Note: Inserting diodes into circuits results in a small volt-drop which means that slightly less voltage will be reaching the lamps. With incandescent bulbs there there will be a small reduction in brightness, it's up to you whether you consider this to be acceptable, MGB indicators are not the brightest to begin with. Note also that diodes can fail, disabling either hazards or indicators, or could cause both sides to flash with the indicators and damaging the indicator flasher (but then so can relays and after-market wiring).
  • Hazard flasher conversions are available as after-market add-ons, make sure these do disconnect or block the existing indicator circuit to prevent this reverse current flow. The vendors may have no idea what you are talking about, so always test after fitting by looking for 12v on the white or green wires at the fusebox with the ignition off, hazards on, and indicator switch operated to one side or the other. If the circuit is not blocking this reverse path you will see 12v switching on and off as the hazard flasher unit clicks. It's up to you whether you are happy to use it or send it back. Note that a kit advertised in the MGOC 'Enjoying MG' magazine showed the detail of the connections in the October 2013 issue and did not interrupt the indicator circuit.

There is also the hazard tell-tale to consider. According to the schematics UK cars don't seem to have had these (why would you need one when both the dashboard indicator tell-tales will be flashing anyway?), but North American spec always did. From 1972 this was fed by a light-green/purple wire off one of the contacts of the hazard switch and a 2-wire hazard flasher unit was used. In this case only the first option above can be used, as the tell-tale needs to be isolated from the hazard flasher unit and the indicator wiring when not in use, and only the factory switch (or similar hazard-specific switch) does this. The schematics show that from 1968 to 1971 a 3-wire hazard flasher unit was used, with the third wire feeding the tell-tale, and this type of flasher and tell-tale wiring can be used with any of the options above. 3-wire hazard flashers seem to have been used on a number of British cars of the era, still seemingly available from the likes of Rimmer, Canley Classics and others. Check they are capable of driving at least 4 21w bulbs, they may also be marked 'heavy duty'. Alternatively it may be possible to wire a tell-tale bulb in parallel with the 2-wire hazard flasher i.e. directly to its two terminals. This will flash the tell-tale in anti-phase to the corners of the car instead of in phase as with the factory and 3-terminal options, but should be legally acceptable for inspections.

Finally power to the hazard flasher must come from an always on, fused source. 'Always on' because the hazards need to be available with the ignition off and the key out, fused in case one of the corners of the car is damaged and the lamp holder or wiring is shorting out. Without a fuse this could cause a fire, adding to your woes. Factory cars were wired from the brown circuit via an in-line fuse solely for the hazard flasher, originally in the very inconvenient location of behind the centre console! Whilst it is technically feasible to power it from the purple circuit which is also always on and already fused, as this feeds the horns and other circuits accident damage may have shorted out that wiring elsewhere on the car and blown that fuse. This means that if you are going to the trouble of adding hazard flashers, a separate fuse off a brown wire is really the only sensible option.

Adding to Bee: August 2013
After a pal had his TR6 written off by being rear-ended, just a couple of weeks after completing a two-year restoration to make things even worse, I decide I really need to fit hazards to Bee (Vee has them as standard). The TR6 didn't have them, although in that case I don't think it would have made any difference. The car had broken down on a dual-carriageway, was only half on the carriageway and half on a grass verge next to a crash-barrier, in clear visibility on a straight road just after a roundabout, with my pal back up the road warning people to keep over. Nevertheless this ... chap seemed totally oblivious of both my pal's warning as well as the car, almost hit him, then smacked right into the TR6's off-side rear corner which caused the perpetrator to spin and roll, coming to rest on its side. But anything that might improve visibility of these, by today's standards, small cars has to be of benefit and I decide to fit hazards before going any further with DRLs which a pal and I have been pondering for some time.

August 2023: Checking Bee over prior to an MOT - lights, wipers, washers, horn. Main lights - fine. Hazards - fine. Brake lights - fine. Indicators ... no indicators. I'd only been out in her a couple of days previously and they were fine then. The hazard switch disconnects power from the indicators (as explained above) so flipped the hazard switch on and off again and the indicators worked again. Only installed 10 years using the standard switch as there was a spare position in the centre-console, they are probably only used twice a year - pre-MOT and MOT, so hardly 'worn out'. I tried the switch a few times and most of the time it worked as it should but every now and again if I switched it off slowly the indicators didn't work. Quite co-incidentally a pal I'm helping with his electrics gave me a new hazard switch of the correct type just the other day which he had bought for his car not realising it used a different switch. Fitted that, working again.

I'd had the same thing on Vee's hazards not long after buying her 30 years ago, opened up the switch and it was full of hardened grease. Dug that out and replaced with fresh, worked again and has done ever since. I'm almost certain that is the original switch. Opened up this switch, only a little grease and certainly not hard. Cleaned that out anyway and scraped the contacts (as I had to do with said pal's brand-new main lighting switch straight out of the box) and with a test meter that seems OK now, so has gone in the spares box.

When I restored Bee just over 30 years I replaced all the dash switches as the logos were a bit faded, then some years later the new main lighting switch started getting intermittent. As the logo on that was getting a bit faded I put what was almost certainly the original switch back, and that has worked ever since.

New stuff is rubbish even though it looks exactly the same as the originals.

A louder audible warning

Never the loudest ticking, particularly at higher speeds in either roadster or GT, and some without the hearing sensitivity of a bat might find themselves inadvertently leaving the indicators on when they shouldn't be, you can add a buzzer to give more of an audible warning. Get a 12v dc buzzer and simply connect its two wires to the two terminals on the indicator flasher. Some electronic 'buzzers' are polarity conscious and will have red and black wires in this case, and for negative earth cars connect the red wire from the buzzer to the green on the indicator flasher and the black to the light-green/brown wire. For the earlier positive earth cars connect the buzzer the other way round. This may work on 'modern' cars with three-terminal flasher units but see below.

When you first operate the indicators you won't hear anything - don't panic! It is only when the lights go out on the first click of the flasher that the buzzer will sound, i.e. it operates in anti-phase to the lights. If you find the buzzer too loud you can always wrap a couple of turns of insulating tape or similar round it. However the cheap electronic buzzers make a pretty horrible sound, and even the piezo type can get annoying, so I'm experimenting with something to make a louder clicking noise with my NOS original which although flashing quicker is quieter than the old one.

August 2024: I've been wanting to change the flasher unit for a more modern OEM electronic unit from the 80s/90s for a while now to counteract the slowing of the flash rate when sitting waiting to turn with the brakes on, which I've never been happy about, and they are slow at the best of times. It took a while but I did eventually find one, which worked well, but in the roadster with my GPO sounder (which has a more pleasant sound) connected it emitted a slight sound all the time which wasn't acceptable, so the output terminal must have a partial connection to earth internally - probably part of the electronics and bulb failure warning, whereas with the original unit the only path to earth is via the indicator switch to the bulbs. Reluctant to do without it, it would be possible to use a diode to block the earth from the buzzer, but because that results in a volt-drop that would reduce the brightness of the bulbs slightly so not a good idea. A pair of diodes on the output of the indicator switch (to prevent cross-feeding from one side to the other) would work, but then the buzzer would sound with the hazards. Further down the diode rabbit hole using a full-wave bridge rectifier with the output wires connected to the polarity-sensitive buzzer and the input wires connected one to the left side and the other to the right will only sound the buzzer when only one side is powered i.e. the indicators supplying 12v and the other side has a virtual earth through the bulbs. When both sides are powered i.e. hazards with 12v on both sides there is no potential difference between them to sound the buzzer. This does mean that the buzzer is earthing through the unlit side i.e. passing a current through the bulbs, but in a test there is no sign of the corners glowing that side which means the current is too low to have an effect, so is my preferred solution. That may be different if using LED indicators at the corners of the car, but I don't use them, they may make prevent the buzzer from working, and you cant use one of these flasher units with LEDs anyway. Wired as described it works well although at the first flash instead of sounding straight away at full volume as it did with the thermal flasher unit it fades in over a fraction of a second, although after that it is as before, and I will ignore it.

September 2024:

Fitted the same flasher unit to Vee with the same benefits, but even in the GT it is very quiet when driving. I only have one GPO buzzer so go back to the solenoid idea, needless to say I can't find the one I bought four years ago so bought another. It's a bit quiet by itself particularly as one end has a rubber O-ring to cushion the return stroke so it needs to hit something. Both ends are 'open' so with the right size 'box' one end can hit on the 'go' stroke and the other on the return, and a short section of square drain down-pipe (again!) is the perfect size. A simple bracket clamps it to one wall of the 'box', cable-tied on top of the bracing tube behind the dash, wires ScotchLok'ed to the green/white and green/red wires in the main harness by the connector to the indicator switch. One wire from the solenoid to each, no need for diodes as it's not polarity sensitive, and as before power to the indicators one side supplies 12v and the earth comes from the bulbs on the other side to power the solenoid well. Also as before being wired like that means it doesn't sound with the hazards on as then the solenoid has 12v both sides, and no visible 'leakage' to the bulbs on the unused side. Cable-tied to the top of the bracing bar behind the dash it works well - too well as it is quite loud with a 'clack' sound so will experiment with damping it down a bit - inserting a suitable resistor probably the easiest.

Then it struck me that a standard 12v Bosch-type relay clicks as it operates and releases. Dangling from wires it's not really any louder the the flasher unit (which is screwed to the bulkhead), held on top of the bracing bar behind the dash (where the solenoid was) it's slightly louder but probably still not loud enough. Then I thought if I could attach it to the back of the dash by the fuel gauge it would use the dash as an 'amplifier' - and it did. But how to secure it? Double-sided sticky pads is a possibility, but damps the sound back down again. Then trying various positions round by the fuel gauge I discover it wedges into the space between the bottom of the fuel gauge and the top of the lower edge of the dash that turns back under the gauges. Sounds pretty good and is held firmly, so I'll see how that goes. Still too 'clacky', so move it to the bracing bar with a sticky-pad. Still a bit quiet so cable-tied a non-sticky side to the bracing bar - better when driving, but 'clacky' stationary! At this rate I should have reverse-engineered the GPO sounder! But doing some Googling just on the offchance not only did I find a circuit diagram for the GPO buzzer, but also two sources of the buzzer itself, and for the princely sum of £5 Vee now has the same buzz as Bee.

February 2020: An enquiry on the MGOC forum reminded me that I hadn't updated this. My first thought was to wire a small loudspeaker in series with a capacitor. The theory being that although the impedance of a typical speaker is far too low to wire across the flasher unit and would affect both it and the lamps, adding a capacitor would prevent that but allow to capacitor to 'charge up' and discharge via the speaker each time the flasher contact opened and closed. I used to have a couple of speakers from small transistor radios I messed with decades ago, but couldn't find them. Got one off eBay but it was much smaller than expected and was way too quiet.

Next thought was a solenoid where the ends of the plunger were exposed, and positioned in a box or by a bulkhead the plunger would hit the sides of the box or the bulkhead so making a noise each time it operated and released (in a box) or just on the one stroke (bulkhead). Ordered one of those, but in the meantime I discovered an old GPO electronic sounder with three volume settings. Tried that and like Goldilocks and the three bears one setting was too loud, another wasn't loud enough, but the third setting was just right. It's also quite a melodic sound very different to the harsh 12v buzzer, so that's as far as I went.

The MGOC forum reminded me that I had also seen buzzers for motorcycles, including one type with a delayed response of 20 flashes or so before it started buzzing. Some of these have an additional feature that if you sit there with your brake lights on the buzzer is muted until you release them. As Dave Birkby said - "bizarre". Whilst I can understand car drivers sitting at traffic lights with a foot on the brake (instead of selecting neutral and applying the handbrake ...), do bikers habitually do that? Even though modern bikes light the brake lights from both front and rear brakes? And one motorbike forum was discussing a system where the buzzer only sounded once every ten or so flashes. Both relatively easy using a chain of bistables or JK flip-flops, should you be so inclined - four to give you a count of 8, five for 16. One of the products included the following in the description: "Built by my strusted and quilfied staff.". Fills one with confidence ...