17 replies to this topic

[tiger99]

Yes, it is extremely easy to make a load-agnostic flasher. The difficulty is that, unlike the old style metal can with bimetallic mechanism, it doesn't easily provide a way of detecting that a lamp has failed. That is a lot more difficult, and I will be looking out for something simple and readily available, which the average Mini owner can install. I do know how to  make one that will work reliably, but it is beyond the skill and knowledge of most people here. In return, there are many people here who can knock out dents and spray paint far better than I can, so I am not having a go at people who can't do anything electronic, just that not everyone is going to have every skill. If they did, we would not need garages, or hospitals, or builders, or politicians, because we would do it all ourselves. Oh wait, that last one we really don't need.....

 

Remember that it should flash 90 times a minute (approximately) regardless of load, which is fairly easy to achieve, should drive about 90 watts of lamps, including side repeaters, in hazard mode, with maybe 25 amps peak into cold lamps, and in indicator mode should drive half that, about 45 watts, giving a "normal" indication if the load is somewhere about 40 watts, and an "abnormal" indication, or "lamp failed" caption, or whatever you want, if the load is only about 25 watts. It would be nice to get it to detect a failed side repeater, but you would need rather close tolerance components, and compensation for battery voltage changes, to get sufficiently accurate current detection for that. I know how to do all that, but rather would not, because it is all extra work, complexity, and potential unreliability. But just performing the flashing function, with supreme reliability, is far from being rocket science, as you know.

 

My next project(s), time permitting, will have modern electronics, including lamp failure detection, and that will be accomplished by using a protected solid state switch for each lamp, with status flag that will show that no current is being drawn when the lamp is supposed to be on. Probably Infineon PfoFets, because I have used them very successfully on aircraft in the past. It is a great way of improving seperation between left and right headlight circuits to avoid simultaneous failure of both, while incorporating correct switching of fogs and spots, without needing huge numbers of relays and fuses.

 

http://www.infineon....af2ab3486&ack=t

http://www.infineon....132f7151b4a7955

http://www.infineon....132b527d9173083

 

I may even use a 24V electrical system, which has many advantages, provided that I can get a suitable alternator and starter motor. Actually a 12V starter motor may survive, not everyone who converted a 6V car to 12V found the need to change it, so that "may" be ok, or now that very low on-resistance MOSFETs are available, a chopper to limit the average starter current to a safe value is feasible, and a 12V alternator can be made to generate 24V by upgrading the diodes, modifying the regulator, and if you need charge at low revs, and have a safety margin at the top end, modifying the drive pulley ratio.

 

I will not even think about the half-baked 36/12V (aka 24/14V) systems, designed by incompetents, on behalf of marketeers, whose only purpose is to make spare batteries and other vital components much more expensive than usual, and tied to a specific manufacturer. The several modern cars which use these systems utterly disgust me, as they have numerous additional failure modes never seen in any normal car. I say "incompetents" because on every single published article that I have seen, the alternator load dump voltage in a 42V system has been scaled up by adding (42V-14V)=28V, or (36V-12V)=24V to the known load dump voltage of 45V to 60V in a standard system, to get values around 70V to 80V, which is just plain wrong, as the load dump voltage actually multiplies by the voltage ratio, and is in fact 135V to 180V in a 36/42V system, which may well explain why certain newish models are frequently seen broken down at the roadside.

 

Oh, and I am not exaggerating the problem, or prone to over-engineering, as I was able to use a BTS412, intended for a 12V car system, in a nominally 28V aircraft system, while respecting all derating margins, and they proved to be supremely reliable. The design was fully visible to the certification authorities, including Transport Canada, who unlike the US FAA, do their job properly. The spikes in the 28V aircraft power systems, including allowance for lightning strikes, really did not exceed the safe ratings of the BTS412 in the configuration in which I was using it, with several additional inexpensive, but vital components. But needing to understand how to get it right, by correctly interpreting the data sheet, is why I never encourage anyone except a professional electronics design engineer (who will not need my advice anyway) to try it at home. There are pitfalls, and it can go horribly wrong.

 

Feel free to use any of these thoughs in any way that you wish, and if you do devise something useful, please do show it here.

[Icey]

I have the bones of fault managing digital flasher in an SVN repo somewhere (not used it for a while since moving to Git) based on a PIC 12f675. It has some fairly basic ADCs which I had started using to watch the voltage across the output. This was in an all digital system, making it work with filament bulbs would be a little more difficult but not impossible.

 

I dropped that project because it just wasn't worth the effort in my application, maybe one day I'll pick it up again....one day!

[tiger99]

Yes, that is the sort of project that has the potential to fulfil the need. I would use an AVR rather than a PIC (purely personal preference and experience, some would use an MSP430 or an ARM, all adequate).

 

It just occurred to me that high side current sensor chips are very readily available nowadays and solve the problem of shifting down from a small voltage drop in the nominally 12V line, to sense current, to refer it to a grounded ADC. Useful for everything from a remote ammeter to a light bulb monitor. I used one successfully at work, from memory a Linear Tech one.

 

Quite a few ways of skinning this cat! I wish I had the time right now. The other thing I would like to make is a tiny rectangular instrument panel, like one of Sir Alec's early sketches, which goes just under the top dash rail in front of the driver, leaving a completely clear full-width parcel shelf. A Digidash is not the right shape, and too big, but with a rectangular LCD a speedo and full set of warnings is possible. Actually, a LED bargraph speedo could fit in a narrow slot in the top dash rail, and individual warning LEDs can be dotted about anywhere, which gives an even more intriguing option.

 

Apart from engine management, which has to be worth it on absolutely any car for economy, engine life, performance, and reliability, and satnav, I don't think that the Mini needs any other modern frippery. No power steering, for instance, although respect where it is due for those who have successfully fitted it, but I just like the nice, moderately heavy feel as it is.

 

As I think we both know, there are once again lots of people dabbling in electronics at home, some of these are also car enthusiasts, and computing has become ridiculously cheap, so a convergence of technologies is inevitable, but not, I hope, on any rare models that should be kept as original as possible. I will hopefully use a boring model such as a City E for my next project, hopefully at the best value end of the Mini spectrum, and mildly butchering one might not offend too many people. I really would not modify any form of Cooper, or very limited edition special, and vans are scarce and too valuable to mess about. But if someone gave me a derelict but saveable van, it would be severely tempting.....