The target vehicle turned out to have serious problems which I am still battling with, but I belatedy started building the board anyway with the thought of using it as a test for edis on a rover v8.

The board arrived from Andy with smd's already installed. A work of art before I started my slash and burn.

I decided to get the components locally which led to some frustration but had the unexpected benefit of finding the map sensors nos at a bargain price.

First step was to drill some random holes on the traces.

Then build cpu power supply. I could not get the right size tants at this stage. I might eventually replace them.

Tests OK at exactly 5.00V on my fluke.

Then I built the sensor power supply. By this time I had found the rights tants. The only MOV I could find was a bit of a monster but it fitted. Tests OK. Pic later.

then it was some patches for the Q12 mods. Untested at this stage. If it looks wrong tell me now!

Hooray for this thread! :-)

Highly recommend replacing the regulators with bent legged regulators. Sad, as your current one seems very accurate, and it's a shame to waste them. Thing is, if/when you mount this up, those leads are on borrowed time without strain relief. You can wing it and repair as necessary, but it will fail eventually, likely sooner than later, without a strain relief.

I have no idea if the clamp fix is OK or not, but you can carefully test it by attempting to raise the voltage gradually and ensure it stops at 5.15 or whatever it's supposed to be, and stop at ~5.6 or so if you don't get that result, to avoid damage. Use resistors to a higher voltage to achieve this, start with higher ohms, lower step by step.

Fred.

The regulators are at risk even from handling the board (I'm clumsy). I haven't found any enclosure even vaguely the right size so I don't know what I'm going to put this in.

Here's the analog PS. If you look carefully you can see some traces cut in preparation for the smd map sensors.

I'm now stuck because the analog regulator is getting too hot and I can't see why. I can see it will always flow a little current even when no sensors are being powered but not enough to make it hot. I am using 2940's instead of 2937's.

From cold, if you power it up, is it putting out a steady 5V? How long does it take to get hot? Could there be a short on the output somewhere? If it heats up slowly then that's likely just normal loading generating heat? What is your bench supply Voltage? 15? 7? 10? This makes a HUGE difference to regulator heat.

So, either:

1) You have a short on the output
2) It's oscillating badly at high frequency
3) It just needs a heat sink
4) Something else?

I was having a hot regulator problem a while back. I had mixed up the leads on a tantalum capacitor. It took a surprisingly long time to release its magical hydrogen cyanide smoke. It might be worth double checking.

Solid suggestion! On the 12V side they release it with a bang, but on the 5V side, that's a gooooood point. I'd have "liked" or "+1ed" your post instead of posting, but it's not possible, so here is some worthless commentary instead.

Sadly it wasn't that. It was the very last thing I could possibly unsolder, the regulator itself. Seems to have internal low resistance. almost a short. Replaced it and it still gets hot, but I can still hold it after 60 secs whereas before I had 3rd degree burns after 20 secs. Nothing a heat sink won't solve.

In the process of unsoldering I broke a leg off the smd regulator over voltage thingy, the TL431. Might replace it with a through hole version since all the tracks around it are cut anyway.

Theres something I really really don't understand and that is the smd resistors R106 R107 and R109. 109 measures correctly at 600 or so but the others measure at 6.5K when they are meant to be 10K. When I measure across the 3 in series it measures 6.2K. WTF?

Other connected components interacting, I guess. Measuring parts in-circuit is never easy.

You don't need the clamp, you could just forget about it, and be careful with your wiring.

That's the logical explanation but I can't see what. Everything that I can see on that circuit is either unsoldered or the traces cut.
Yes I will proceed without the clamp for the meantime and maybe add it back later. The smd tl431 seems to be unobtanium here and not sure about the specs for the through hole ones that I can get.

Think of it as semi-idiot-proof technology development, and then realise you don't need it, and stop caring :-)