Been outside playing with the car this evening, mainly giving the LPG system a tweak. Decided to take it for a quick blast down the road with the laptop on the passenger seat so I could see what the LPG system was doing. It had been idling for a while so was up to temperature and I dropped it into reverse still at idle. Slowly reversed about 2 feet and the engine cut out. Tried to restart it and nothing. Then noticed the check engine light wasn't coming on and the LPG switch wasn't lighting up. Tried the EKA code, no different. Tried the Nanocom and it said everything was fine. Started using a bit of logic and figured that the LPG switch would be getting an ignition switched supply so no check engine light and no starter either could be down to a lack of ignition switched electric somewhere. Fuse 26 was blown. Put in a new 20A fuse, check engine light came on, LPG switch lit up, starter turned and the engine fired up. Into reverse, managed about 10 feet this time and it cut out again. Fuse 26 blown. Another 20A fuse put in, reversed out of the driveway, into drive and set off for test drive (with a box of spare fuses on the passenger seat). Drove about a mile and a half to my local test road (a bit that used to be the southbound carriageway of the A1 but now doesn't go anywhere). Tried it in normal, floored it in Sport and confirmed that the LPG system was working spot on and keeping the mixture correct no matter how much I tried to make it work (foot to the floor in Sport and it didn't change up into top until I hit 85mph!). Got to the end of the road, went to turn round and managed no more than 10 feet in reverse and it cut out again. Fuse 26 had blown.

Now Fuse 26 supplies ignition switched volts to the engine ECU (which would explain the lack of a check engine light), the ignition coils and the lambda sensor heaters. As far as I can see from the diagram, putting it into reverse would have no affect on the current draw on fuse 26. So why the hell does it blow when I put it into reverse and not when It's idling or going forwards? When reversing out of the driveway, I don't even give it any throttle, just let it creep back at idle so it isn't likely to be a torque reaction when reversing causing something to short out, so what the hell is it?

Does it do it when reverse selected but not moving Richard?
Does it do it when rolling backwards down a suitable slope in neutral rather than reverse?

Could it be one of the lambda sensors wiring looms is sat rubbing against one of the driveshafts, and has rubbed through in a particularly odd way to short out when its spinning in reverse? Seems unlikely to be that repeatable but that's all I could think of.

Yeah- I was thinking same thing. I suppose theoretically, if enough cable slack to wrap round shaft, drive shaft rotating one way would tighten and short. Other way would loosen and not short. It'd be a one in a million scenario though.

No slopes in Cambrideshire to see if it does it when rolling in neutral, but I'll try putting it in reverse and holding it on the brake later. I don't think there's enough slack in any of the cables to be touching the propshaft but I'll have a quick look at that too. I've got a standard 0-1V Zirconia lambda sensor in the RH downpipe solely to drive the single point LPG controller. The reason why I was playing was that the system was giving me a lambda sensor failed error. When I drove it I found that it was switching between 0.00 and 0.7 volts rather than 0.1 - 0.9V and then when running at constant speed it would stop working so the system was reporting a constant 0.48V bias from the controller.

So I changed the sensor a couple of weeks back. At that time I was under it looking at the lambda sensor wiring but on the other side to the propshaft. The new one was still showing a low output so went through the earths to get rid of any voltage drop (I was seeing 0.2V on the earth at the sensor with respect to the battery negative). Although I got that voltage drop down, I was still never seeing the full 0.9V when rich that I should see so tried swapping the lambda input to the LPG system over to one of the GEMS 5-0V sensors. Changed it in the LPG software and it all seemed fine. Lambda switching between 1.2 and 4.3V and this matched what the Nanocom said it was doing. Car running fine except for fuse 26 blowing when reversing.

The heater for the lambda sensor takes it's power from the LPG system output so is live when switched to running on gas but the LPG system is taking the ignition switched supply from the petrol injector supply and not the fuse 26 supply so it isn't going to be connected with that. Or at least I don't think it is. But even if it was, why only when reversing? I've got an appointment at the dentist later so a fuse that blows when going backwards is going to make parallel parking an interesting experience.......

Maybe pull the lpg lambda sensor wiring from the petrol sensor and lpg fuses and try it reversing on petrol? That'd eliminate any weirdness that might be associated with the lpg system

Had a quick look at the lambda sensor wiring and it's nowhere near the propshaft all neatly clipped away as it should be. Disconnected the lambda sensor wire to the LPG system. Locked the LPG stepper at the default so it would run on LPG but in open loop. So all I've changed is disconnected the one wire. Reversed out of the driveway and fuse didn't blow. Reversed up and down my lane (about 100m) a few times and the fuse didn't blow. Reconnected the lambda sensor wire but this time put a 1k resistor inline (working on the principle that it's only sniffing off the voltage so it wont alter the reading) and removed the default lock so I was running closed loop on LPG again. Tried more reversing up and down the lane, fuse still didn't blow. Went for my dental appointment, parallel parked and the fuse didn't blow.

Tonight I'm going to try measuring the current drawn on the fuse 26 circuit, although all of my 3 meters only read up to 10A, so I'll use the cheapo Maplin one just in case and put a 20A fuse in line. I'll also try tapping off the other lambda sensor as the one I'm currently connected to (RH bank) takes at least a minutes running before it starts to give an output. It'll probably be the same as they were both new when I put the rebuilt engine in a year and a half (and 30,000 miles) ago but it's worth a try.

Didn't try it on petrol as it runs like a dog and throws the fuel trims out by miles. At least one, if not two, of the petrol injectors are clogged so the air that doesn't get mixed with any petrol tells the lambda sensors that it's running very lean. That makes the trims go fully rich and it really does run rough as it's nearly drowning on petrol. I know exactly how they got clogged and have been meaning to sort it out for ages but as all it ever does is start on petrol and immediately change over to LPG, it's one of those jobs that is about as far down the list of things to do as it can get (it comes below renewing the headlining even) I'll get around to it one day. Probably around August when I'm about to set off on a 3,000 mile road trip around Europe.

Used one of these on my lpg lambda. Dirt cheap, turned up next day and works fine.

I used to use those, or the 4 wire version anyway, but found they only last about 9 months if you get a good one. The problem with a standalone is that if there is a bit of resistance anywhere, the voltage will be slightly low (as mine was) so the controller will be setting the mixture to aim for 0.5V average. If you're losing 0.1V, ideally it wants to aim for 0.45V so will be setting the mixture permanently slightly rich.

Anyway, got to go to the other side of town now, so I'll see how it behaves itself.......

Five hours and he still isn't back - time for a search party?

No, got back but had to go and do some work. Drove to the other side of town, reversed into a parking space, drove back again. Reversed into the drive and it's behaving itself now. After blowing 3 20A fuses in the space of driving it 20 feet backwards, I've now driven almost as far backwards as forwards and everything is fine. Which is a worry as I've no idea what caused it, haven't found anything amiss that could have caused it so have no way of knowing if it is suddenly going to start doing it again. Good job I bought a big box of fuses last time I ordered some stuff from Auto Electric Supplies......

Now it's getting even sillier. Took the 1k resistor out so the wiring was exactly as it was before. Reversed out of the driveway, went to Flogas to fill up with LPG, reversed out of their yard, went to Halfords to pick up the DAB adapter I'd ordered to put in the company car and reversed into a parking space, drove home again and reversed into the driveway. At no time did the bloody fuse blow!!!!

May be getting somewhere here. Decided to sort out the running on petrol at the weekend so pulled the injectors out. Squirted carb cleaner into each one, blew through it and applied volts to it and got a nice stream coming out of each one except for the one from number 5 which did nothing. It clicked with volts on it but I couldn't blow through it (got a gobful of ether though). Tried various different solvents and an air compressor and still no flow so fitted the other 7 injectors to the fuel rail and attached it to the fuel rail upside down. Ignition on to pressurise the rail, applied volts and it squirted fuel out, a lot of fuel, all over the engine. Pulled it out of the rail and even more fuel squirted all over the engine. Tried blowing through it again and this time it worked. So put everything back together.

Reset the adaptive values and started the engine on petrol and for the first time in over a year and a half, it ran on all 8. Reversed out of the driveway and bloody fuse 26 blew! So, it's still blowing on petrol and the LPG controller was disconnected from the lambda sensor at that time too. Ran it down the road on petrol with the Nanocom connected and bank 2 lambda sensor was showing a constant 5V whereas bank 1 was switching as you'd expect. Connected the LPG system to the additional lambda sensor in the bank 2 downpipe, changed over to LPG and it was switching. So was the bank 1 sensor but bank 2 sensor stayed at 5V all the time. So it looks like a problem with the bank 2 sensor, maybe a short between the heater and signal (I suspect that as the Nanocom and the LPG controller are expecting to see a 5-0V signal they will both display 5v if they see anything greater that that). An internal short (or a problem on the wiring) would explain the blowing fuse but not why it only does it when going backwards and not why it doesn't just blow. Further investigation required still but I think I'm getting somewhere.......

Current state of things is that the LPG controller is using the additional sensor so everything back to as it was. Just went outside to turn the car around. Reversed out of the driveway, no problem, pulled forward, no problem, left it running while I put another car on the driveway, no problem, went to reverse into the driveway, moved about a foot and the fuse blew. Fitted new fuse, restarted the engine and reversed into the driveway. Definitely looks like I need to get under it to look at the wiring for the bank 2 lambda sensor but as it's raining again (and I'm supposed to be working from home), that'll have to wait.

Seems I owe Sloth a pint for being spot on. Went outside this morning to get to the bottom of this and found the problem. Got another one now but I'll get on to that in a minute.

Crawled underneath to look at the wiring and plug on the RH bank lambda sensor(s) and couldn't see anything wrong (for some reason I was convinced it was the RH bank one that was causing the problem). Looked across at the LH bank and noticed that the cable was very close to the exhaust downpipe. The outer sheath had a hole in it but the insulation on the wires wasn't touched. However, while moving it around to see why it was dangling against the exhaust I noticed it was also very close to the propshaft. When I'd looked before it had been from above and I could see the sensor and propshaft but the cable then just went towards the front of the car to the plug clipped to the side of the sump. At the front it had been just, only just, rubbed by the propshaft to the point where the insulation on the red (12V heater) wire had rubbed through. When going forwards the rotation of the prop had pushed it away from the prop but when in reverse.....

So, having taped up the dodgy bits, I re-ran the cable behind the transmission cooler pipes so it wasn't anywhere near the exhaust or propshaft, I decided to check that everything was working as it should. Rather than use the Nanocom, I plugged in my Launch C Reader as it can display live data as a graph with different coloured traces for different things. Started it up on petrol (as it now runs properly on petrol since I cleared the clogged injector) and at idle the lambda sensors were switching from 5 to 0V just as they should (although the C Reader displays it as 1 to 0V as that is what the OBD spec calls for). Drove off down the road and still switching. Stopped at a shop, switched off. I had intended checking that all was fine on petrol before switching to LPG to make sure it was holding the fuel trims correct. As the single point LPG system uses it's own separate lambda sensor, if there is any resistance in the feed, the controller will set the mixture slightly rich. So I wanted to make sure that the fuel trims stayed the same and the lambdas were still switching when on LPG too.

A couple of minutes later, got back in, restarted (still on petrol), rebooted the C Reader (as it disconnects as soon as the ignition is switched off) and both sensors were flatlined at 1V. Ran it for a mile or so in case they were still cold and still it showed a flatline on both. Unplugged the C Reader and plugged in the Nanocom. It showed bank 2 as running closed loop but the lambda sensor was only switching from 5.0V to about 4.6V while bank 1 showed OPEN FAULT with no switching at all. So why would it be switching perfectly until I switched it off for a couple of minutes and then not switching and showing a fault on one bank when restarted? Switching it off and restarting again and still the same. No fault codes have been logged either, weird. Sloth, what's your verdict this time?

Errrrr..... :)

Loose plug?

and so it continues..... Decided that for whatever reason the bank 1 sensor was going to sleep after a little while of running so tried it on the way to Summer Camp. Nanocom plugged in showed both banks to be switching correctly (running on LPG too) but after about 10 miles, bank 1 flatlined at 5.09V and up popped OPEN FAULT again. Switched off the Nano and forgot about it. The Nano GEMS documentation shows one error on that screen which is OPEN DUE TO A DETECTED FAULT which it says is caused by a faulty sensor so figured that's what it was telling me but the screen isn't big enough for all the words to be displayed.

Not done any more about it since until yesterday when the postman delivered a shiny new FAE lambda sensor so I crawled underneath and fitted it. Plugged in the Nano and started her up. Success! Both banks switching properly. For a mile or so then bank 1 went to OPEN FAULT again and sat at 5.09V.

Think I might have a loose plug......

Have a check for crispy cables inside the loom from sensor to ECU. Mine were shot for about the first foot from the sensor connector, and pretty manky where the cables connected to the individual pins in the connector...

Checked the cables and plug. Cables are fine but the plug wasn't. Tweaked the contacts and now bank 1 is staying closed loop and switching nicely. The same can't be said for bank 2 though as that is now going OPEN (but not Open Fault?). Looks like I'm going to be grovelling under it again.

From memory, the GEMS ones are standard 4 pin Econoseal connectors, so easy to get new connector pins if needs be...