Another update...
Despite a lot of searching, I couldn't find PEH169 capacitors for sale anywhere...and to be honest, they'd be such old stock, I wouldn't trust the dielectric not to have dried out by now, anyway. On the basis that a capacitor is a capacitor is a capacitor (except when they aren't...but mostly they are), I ordered two replacements based on the capacity and operating Voltages of the old RIFA ones:
As you can see, they don't have mounting studs, nor do they have screw-terminals, so some modifications were required:
This seemed like a good idea at the time - the eyelets would use the same bolts as the old capacitors; I just had to find appropriately sized nuts:
As you can see, the new capacitors are much smaller - they're still polarised, though, so I had to mount them the right way round (double checked that about a hundred times while I fitted them!). Reading on the internet suggests dielectric materials have improved a lot in the last 20-30 years, meaning smaller capacitors are possible, hence the reduce size. The wires I'd soldered onto the capacitors were too long and the eye connectors were also too long to fit above the PCB, so I had to bend them and the legs of the capacitors, which wasn't ideal. Fortunately, nobody will ever know, as they'll be hidden inside the ECU... To make up for the lack of mounting stud, I used the tried and tested engineering solution of hot-glue to stop the capacitors vibrating against the PCB and causing damage.
It was -8C here this morning with 5cm of snow on the ground, so I wasn't keen to go out and test the new, improved ECU but, by 16:00, it was a balmy -1.5C, so there was no excuse not to try it. I put the lid on the ECU loosely, bolted the ECU itself into the Defender, wired it up and turned the key...
[imagine a picture of a Defender with smoke coming out of its exhaust]
I took a picture as proof it at least started, but I seem to have lost it. Anyway, after some coughing and reluctance, the engine started and ran nicely. I revved the engine and it behaved - no stalling or engine light coming on. I left it running for half an hour and, when I came back, it was still running and still revved - if there'd been a driver's seat, I'd have tried moving it!
So, is it fixed? Who knows - it's yet to exhibit the previous fault is all I can say for sure. Can you crack open an ECU, attack it with a soldering iron and still have it work at the end of the day? Seemingly, yes, but time will tell... I'm not going to say "Ha! Saved myself three hundred quid there with just a fivers worth of capacitors off ebay!" just yet, because that would obviously be tempting fate, but I'll post an update on continued success or abject failure in due course...
Next jobs are:
Clear the ECU codes, run the engine and check them again for anything nasty.
Re-seal the ECU and bolt it back in under the seat.
Strip out the CKS sensor cable bypass I went to all the effort of fitting and put the CKS sensor back on the loom connector.
Refit the driver's seat.
Get in, drive about and either feel smug or sit in a layby waiting for the AA again, rolling my eyes and thinking "hmm, wasn't that, then..."
Watch this space, as they say