Advice please!

[LPGC]

As far as I can ascertain my petrol fuel trims haven't deviated to the point of being problematic (it starts promptly on petrol) and I don't want to lose that.
If utilisation of the grey wire(s) is helping this should I consider just tagging the AFR meter to one of the O2 sensor wires?

This:''The mixer system may hold lambda voltage above 0.5v above 4000rpm,'' is a feature of mine. If not the 4000 rpm threshold then a (configurable) TPS threshold. That is one of the things I want to monitor (and tweak) with the AFR meter.

Worth considering leaving the grey wires connected but connecting the LPG system in the 'normal' way without the grey wires might not affect starting etc and even if it did you could easily go back to having the grey wires connected. Same regards the 4000 rpm threshold, the petrol ECU might switch to open loop before 4000rpm anyway... But if it remains closed loop above 4000rpm, and if grey wires are not used, then since the LPG system intentionally holds mixture rich above 4000rpm the petrol trims for above 4000rpm will eventually go fully lean (which may be better than fully rich but unless you've seen what trims are already and confirmed they've gone fully lean above 4000rpm already the car could drive a bit flat above 4000rpm on petrol if grey wires are disconnected).

However you wire in the LPG system, if you wire in an AFR meter it will need to connect directly to lambda signal voltage - But usually an AFR meter (at least if it's wide band) will need it's own (wide band) lambda sensor. Your truck's current lambda sensor must be narrow band if the mixer system connects to it. Some AFR meters allow you to replace the vehicle's narrow band sensor with AFR meter kit's wide band sensor, the AFR kit can then output a narrow band equivalent voltage for the petrol ECU and Millenium etc to connect to (as if a narrow band probe is still fitted).

[Pinger]

It'll be branded Millennium and will have Millennium firmware but if you look at the circuit board, it's still an AEB175, so quite what the other pair of wires are doing is anyone's guess. Can you pull the boot on the plug back and see where the extra pair of wires are connected?

Moulded on the back of the ECU box it has TIPO: AEB 175, Leonardo, Regmatic, Millenium. Previously, I identified it by the Regmatic and Millenium stickers on its front.

As for the wiring. There's a small bundle (within their own conduit) that goes to the four pins (in square formation) next to the smaller fuel level gauge/dash switch plug that has four wires coloured purple, white, orange and blue.
The other purple wire is in the middle of the plug and sits beside the only grey wire. Presumably then that grey wire is branched into two somewhere between the AEB ECU and the GM ECU but without taking the loom apart I can't confirm this.

[Pinger]

However you wire in the LPG system, if you wire in an AFR meter it will need to connect directly to lambda signal voltage - But usually an AFR meter (at least if it's wide band) will need it's own (wide band) lambda sensor. Your truck's current lambda sensor must be narrow band if the mixer system connects to it. Some AFR meters allow you to replace the vehicle's narrow band sensor with AFR meter kit's wide band sensor, the AFR kit can then output a narrow band equivalent voltage for the petrol ECU and Millenium etc to connect to (as if a narrow band probe is still fitted).

My AFR meter is designed to run from the existing narrowband probe.

Worth considering leaving the grey wires connected but connecting the LPG system in the 'normal' way without the grey wires might not affect starting etc and even if it did you could easily go back to having the grey wires connected. Same regards the 4000 rpm threshold, the petrol ECU might switch to open loop before 4000rpm anyway... But if it remains closed loop above 4000rpm, and if grey wires are not used, then since the LPG system intentionally holds mixture rich above 4000rpm the petrol trims for above 4000rpm will eventually go fully lean (which may be better than fully rich but unless you've seen what trims are already and confirmed they've gone fully lean above 4000rpm already the car could drive a bit flat above 4000rpm on petrol if grey wires are disconnected).

I think I want to leave the grey wires connected - not least as far as I can ascertain it has been OK configured thus (more on that in a moment).
GM fuelling strategy AFAIK is to go open loop not at pre-set rpm but at threshold TPS value after a period of time. Thus, sufficiently out of kilter to make preserving the grey wire worthwhile?

The AEB has both purple wires (but only one grey) and I'm assuming it better to leave it like that.
The obvious place to connect the AFR meter then is to tag it to one of the purple wires coming from a Lambda sensor heading for the AEB ECU. But for the concern any voltage emanating from the meter will be detected by the AEB ECU.
If I (initially) get the truck running and connect the lap top to it such that I can see the AEB activity ie, Lambda and stepper activity, if I then bring the AFR wire to the purple wire any untoward effects will immediately show on the Lambda and stepper activity - yes? If nothing untoward I can hard wire it. Any reason not to?

If that works, I can periodically switch to petrol running with the AFR meter and see where the mixture is ie, fuel trims. If they are wayward, I can consider the alternative wiring arrangement (run the Lambdas direct to GM ECU, tag them for the AEB ECU, tag one of them for the AFR meter and delete the grey).

Does this make sense?

[LPGC]

The AFR meter can't show a very wide range of mixtures if uses a narrow band probe because narrow band probes can only measure a very narrow range of mixtures near stochiometric.

Whether you connect the grey wire or not shouldn't have any effect on how it runs on gas but can have an effect on how it runs on petrol when you've first switched back to petrol. You may think if it ain't broke on petrol don't change it.

The AFR meter obviously has to see raw lambda signal, which is the same signal the AEB175 has to see. The AFR meter and AEB175 shouldn't pull the voltage on the lambda signal wire (unless they go wrong) but some petrol ECU's do pull voltage on the lambda signal wire in certain conditions by design, this implies another potential positive for using the grey wire. For example, some (older) ECU's will pull their own lambda voltage input to 0.5v during open loop, fault conditions, or warm-up. It's easy to imagine the ECU designers thinking that they could program the ECU to ignore lambda readings during these conditions but that might take a bit of coding and the ECU might be a bit short of memory... Or they could just wire in a simple on/off output which when on (during open loop, warmup or fault conditions) connects 0.5V through (say) a 1K resistor to the ECU's own lambda signal input. Since the ECU is probably programmed to increase fuel trims for anything above 0.45v and decrease for anything below 0.55v holding voltage at 0.5v means the ECU doesn't change fuel trims. If we disconnect a lambda sensor from many vehicles that have 0>1v lambda sensors and measure the voltage on the ECU's lambda signal input wire there's around 0.4v, which might beg the question why not 0v.

Yes any untoward effects should be visible because you can use one reading to check the other. You could even use a digital multimeter or oscilloscope to check if connecting either/both AFR/AEB175 effects readings, all voltage readings should agree same as if you'd connected 10 digital multimeters / scopes and connecting/disconnecting such component shouldn't effect the reading.

Makes sense.

[Pinger]

Thanks LPGC. It will be done as described and we'll see if it works (or not).

[Pinger]

Bit of an update - but first the inevitable question!
Background first. Changed the inlet manifold gaskets and when I put it all back together it wouldn't start (on petrol) but soon found that it would if I opened the throttle very slightly. So, adjusted the throttle stop screw (normally a tamperproof screw but someone had been at it before me). It then started without throttle but preferred a tickle when cold. Then it threw a wobbler and almost didn't start. It was ultra sensitive to throttle opening. None and it would offer but not catch. A touch more and it wouldn't even offer. Finding the exact opening between those two points was hair-trigger. On the advice of the Chevy forum (wind the throttle stop screw out and let the IAC valve do its job) I now have a lower idle speed (welcomed) and good cold starting. But, the changeover to LPG is erratic (hot and cold).
Question: if I extend the 'injectors remain on time' (currently at 0.4s IIRC) will this help? I'm guessing the reduced throttle opening isn't enough for air and LPG vapour and the motor is dying before the LPG can get there (my changeover is on decell). Plausible?

Truck is running well. AFR gauge is working and swings around - the more so since.... the fuse for the O2 sensors and MAF sensor was found to be AWOL (during post inlet gasket change no-start) and re-instated. The improvement in O2 sensor is welcome but is nothing compared to the improvements in the transmission with a now functioning MAF sensor.
Previously it was slow to change up at big throttle inputs, schizoid kicking down (far too eager and compounded the hanging in gear issue) and suffered really hard part throttle upshifts. All remedied now. Up-shifts are smooth (now the ECU doesn't assume every up-shift is at WOT and ramp up the line pressure) and I can control shifting with my right foot and spend most of a day in top gear and let the V8 torque work its magic - bliss!
A bunch of other stuff also done but as not LPG related (ATF and PAS fluid changes, etc) I'll skip - unless anyone wants some practical advice on how-to while it's still fresh on my mind.

So, back to my question. Extend injector on-time and if so, by how much?
TIA - and thanks everyone for the help in getting my truck to a set-up that I'm really enjoying. Much of my concern about mixer size adversely upsetting transmission was in fact due to the MAF sensor being out of action. That concern has now gone.

[Gilbertd]

It may be you need to increase the changeover delay but how is it being controlled at the moment? The Leonardo has a pair of yellow wires connected to an internal relay which are closed when running on petrol and go open circuit when switched to gas, the idea being they can be used to disconnect the power to the petrol injectors. The switching is delayed and this is what is altered by the setting in the software. However, most petrol injection ECUs object to seeing open circuit petrol injectors so Pitagora injector emulators are used instead. These are wired in line with the petrol injectors and triggered by the blue wire (the one that switches the solenoids) to cut the circuit to the injectors but put resistors inline instead so the petrol ECU doesn't know that aren't there any longer. The Pitagora has it's own built in delay so you would need to adjust it there rather than through software.

[Pinger]

Thanks Gilbertd. It's as well I asked as by the looks, what I intended would have made no difference but I would have believed I'd made a change.
I have the emulators. Two of them, each of them dealing with four injectors IIRC. But I haven't seen (or noticed at least) any adjustability - what am I to look for?

Is the issue I've encountered the likely reason why the throttle stop screw has previously been adjusted? What I've been told on the GMT400 forum (by a very knowledgeable guy) is that these screws are fitted with tamperproof heads, never need adjusting, and when they have been moved it is usually to mask another problem. His advice to screw it back out and let the IAC do its works seems like good advice - and has improved the cold starting on petrol and reduced the idle speed to something more sensible already.

Would adjusting upwards the changeover rpm help? My technique is to blip to above changeover rpm and the changeover occurs as the rpm drops on a closed throttle. My thinking here is that it would give more time for the changeover before the rpm decayed to zero.

[Gilbertd]

Delay on the Pitagoras is adjusted with a pot that can be seen from the plug side between the two pushbutton switches (don't change them)

[Pinger]

Don't change the switches - correct?
How sensitive are the screws, in which direction is extending the overlap? How much should I turn them at a first attempt?

[Gilbertd]

No idea, according to the AEB documentation, they are initially set for 0.4 seconds so one way or the other. I would suspect, and this is purely a guess, clockwise would increase the delay. But it may be anti-clockwise......

[Pinger]

Thanks Gilbertd. I found the screws and decided that before I go messing with them I'd give it another chance as the changeover thing was last week immediately after I'd adjusted the throttle stop on the driveway after a longish run.
Started up and switched over OK today - but the idle on LPG was a bit iffy. After a short run it seems to have sorted itself out. All of this I suspect is rooted in the IAC valve making adjustments but it seems to have found a setting (the inlet manifold gasket change put its nose out of joint for some reason). I'll know better next week after a 150 mile work trip.
At the very least, I now know that emulators can be tweaked - which I didn't before. Cheers!

[Pinger]

More advice please – and I thank you up-front for your patience in reading this!
Since I arrived at the set-up I've been running with the AFR gauge has always showed mixture at the rich end of the scale. In the last few months though I've noticed it give leaner readings on initial throttle openings but usually head back towards rich. At the same time as this, the actuator default has shifted from what was a consistent 144 but is now consistently at 174 (it was at 190 one occasion) when checked with the truck parked up. Puzzling (given nothing has been altered) but not alarming.

Yesterday however it had an episode of massive inlet backfires and the AFR gauge immediately showed extreme lean (eg,AFR = 19.8:1) thereafter - with the exception that if I do go hardish (but probably not as far as trip the enrichment set for 4.0V at the TPS) the mixture will then show mid-range values but light throttle is (very nearly) off the scale lean. This would seem reasonable (it seems to drive fine at full lean - just a little less forcibly) - but for the inlet backfire incident. Definitely inlet backfires - they blew the lid of the air filter box and were ferociously loud – gun shot loud.

But why the backfires? They occurred during hard cornering. And I do mean hard. I'd pitched it hard into a tight turn and jumped on the throttle and the banging started. I'd filled the tank to the maximum a few minutes/miles earlier. Is it possible with a full tank for the fuel to surge enough to starve the pick-up? Tank is a 120 litre cylinder laid transversely behind the rearmost seat with the fittings biased well to the LHS and rearwards (it was in a 90 degree LH bend it occurred at around 50-60mph). Or was that just a coincidence and what the AFR gauge is showing me was the cause - but it never happened again in the remaining 70 or so miles of the journey. Not even when I got brave enough to provoke it with more throttle.

I’ve just finished checking the induction system – no leaks, damage to mixer etc, or anything else I can see that would cause the step change in mixture strength after the backfires – or cause the backfires. There were new plug leads and distributor cap fitted recently. It hasn’t done more than 3500 miles since a new reducer and all pipe work and filter were replaced and in any case, it will go rich under load – so not fuel starved between tank and reducer. A manifold vacuum leak would possibly account for the trend toward leanness but it idles serenely and as it did before.

Chicken or egg? Did the backfire cause the step change in mixture strength (the shift was occurring anyway) or was the cause of the backfires the mixture strength going too weak as part of that trend?
I’m doubting the latter as it drives fine since. In a sense, it could be considered ideal – lean at light load and richening at higher power outputs. If it wasn’t for the threat of a backfire that is.
Or, was the hard cornering the cause? Is it possible for fuel surge to have starved it or, for the reducer (orientated north/south with reference side facing offside to kerb) to be affected? Bear in mind, when cranked into a corner as it was, there’s a significant amount of roll as well. It’s never done this before though despite routinely being cornered hard.

Any ideas and/or advice? It goes without saying my main concern is a repeat of those inlet backfires. If someone could pinpoint the cause of them... the tendency towards lean probably isn’t of too much concern – possibly to be welcomed. The backfires though, have me absolutely paranoid of a repeat. They are so violent I fear for real damage to occur.

Pic of my tank below.

[Gilbertd]

Lean mixture will cause inlet backfires and I would have said it could have been caused by hard cornering left uncovering the pickup in the tank. But that would only happen if you were getting very low not just after filling up.

[Pinger]

Thanks Gilbertd - that's kind of what I thought.
A few more thoughts have occurred since my earlier post.
Such a violent backfire must reach all the way back to the reducer (I read on one of the threads how even pressure oscillations from individual cylinder draw can affect reducer operation). Could this account for the step change in mixture strength at low load (from circa 16:1 to 19.8:1? The internal calibration of the reducer knocked slightly askew?
Re this low load mixture - is it actually a problem? I don't think I am getting or got the flashing yellow light indicating 'excessively lean event(s)' though it's possible I missed it on Thursday. I don't think so though.
A possible cause of the backfires could have been wiring failure to the solenoids (broken or chaffed/short-circuiting). My Suburban is body on frame and they are known to flex - esp during cornering. Maybe a problem there?
It is perplexing. Not least as it's not as if Thursday's bout of hard cornering was anything new to the truck - all part of what I read as a kid in a book about enduro riding described as 'husbanding momentum'.

[Brian_H]

I've not long had issues with EML codes coming up for lean mixture for no obvious cause at the time with +25% trims coming up. Pressure seemed a little unstable as well so ran it on petrol for a bit to confirm, it seemed ok at that, so I intended to look at it the next day (which ended up not happening!) A week later went to join the M1 to be greeted by beeping - Low pressure. Wouldn't switch at all. Some investigation at the next Services came to the conclusion that the rear solenoid coil had failed. Not sure what had happened to it, but the post was actually bent on further investigation (might have been from removal of the tank from the original vehicle that valve was in) and nearly impossible to pull the coil off. Removed the guts from the post and reassembled it, Could then run on gas and the fuel trims back to normal levels and the pressure much more stable than it was, so I'd guess the solenoid had been failing for some time and not fully opening.

Bit harder to tell with a mixer as you've no visibility of the pressure, but it can certainly cause a problem if the solenoid isn't kept open as it should be. And out of any of the components the rear solenoid coil seems one of the most likely to fail, I've had at least 3 of mine go, particually when they are externally mounted (I know yours isn't from the photo above).

Given this started with inlet problems it sounds like the lean issue is where that started in any case? If you suspect some sort of issue with the rear solenoid then removing the plunger from the inside would bypass it for testing purposes. Given you have a multivalve on the tank, hopefully the shut off on the tank works which makes that a fairly easy job to do. If that stops it happening, then I'd think you either have a wiring problem there, or a coil thats playing up when moved about.

[Pinger]

Thanks Brian_H.
It's really puzzling as to the extent the two issues are linked - if at all.
The lean low load running just sort of happened and continued then immediately worsened after the backfires but as it is only in one operating regime (light throttle) and richer mixtures are on tap with more throttle, it is at least consistent and any starvation (including electrically induced) wouldn't permit that. When I say 'just sort of happened' it might have been after a couple of mild inlet backfires during cold starting (a while back) that I never connected. I'm still curious if an inlet backfire can knock the reducer's internal calibration slightly askew.

I'm assuming an absolute loss of gas pressure (pre-reducer) caused the violent backfires on Thursday and am increasingly inclined to believe a fault in the wiring - especially to the rear solenoid. Just had a quick look at the solenoid wiring in the engine bay and there's a bunch that runs under the pressed steel bracket supporting the ABS unit - chaffing a distinct possibility.

While I was there I noticed that the wiring to the reducer solenoid (and the now removed one that was on the old filter) is much thinner than that to the rear solenoid - which seems excessively heavy. What current do the solenoids draw? (so I can gauge (sorry) the gauge of wire required as unless I find a fault in the engine bay I'll likely re-do the wiring to the rear coil).

[LPGC]

With a mixer system if the reducer is fitted with the flat side of the diaphragm pointing sideways in the engine bay then cornering forces can affect the response of the diaphragm during cornering. Or on the flip side if you fit the reducer so the flat of the diaphragm is pointing forward/backward then acceleration/deceleration forces can affect it. Gforces acting on the diaphragm in the same direction as the diaphragm normally moves in response to tiny changes in vacuum can effect mixture... A strong enough Gforce could easily outweigh (kind of literally if you take my meaning) the very small vacuum changes. The hard cornering could have caused mixture to go lean which could have caused the initial backfire. Or even rich causing an ignition problem/misfire could then cause a backfire. The standard advice when fitting mixer system reducers was to fit them so the diaphragm faced sideways in the car because acceleration/deceleration forces can usually be greater than cornering forces.

It's rare but is possible for a backfire to damage or at least change the response of the reducer. A change of response might be fixed by adjusting the reducer settings again... But first make sure the backfire didn't cause any other sort of vacuum leaks etc.

Solenoids are usually between 11watts and 17watts, so they usually draw close to 1 amp. I doubt a solenoid issue is the cause of your problems.

If you had nearly a full tank when the backfire occured it's unlikely fuel slosh during cornering caused the problems.. with a full tank there wouldn't be enough headroom in the tank for fuel slosh to uncover the pickup pipe.

[Pinger]

Quick question before it gets forgotten - as before - and related to earlier posts about it faltering and dying on changeover to LPG on first start of the day.
When it dies (after the blip to 1500 rpm followed by the lift to initiate changeover) and I turn the key back and then go to restart it the solenoids click as soon as the cranking begins. This is fine when it's up to temp but when it's cold it can backfire through the inlet (which it has done a few times). I can get around it by manually switching to petrol (if I remember in time) but shouldn't it reset and not switch to LPG until the 'blip' procedure is enacted.
Leonardo Millennium system.

[Pinger]

Thanks LPGC.

Solenoids are usually between 11watts and 17watts, so they usually draw close to 1 amp. I doubt a solenoid issue is the cause of your problems.

I checked out the wiring (simple to do) and couldn't find anything amiss.

With a mixer system if the reducer is fitted with the flat side of the diaphragm pointing sideways in the engine bay then cornering forces can affect the response of the diaphragm during cornering. Or on the flip side if you fit the reducer so the flat of the diaphragm is pointing forward/backward then acceleration/deceleration forces can affect it. Gforces acting on the diaphragm in the same direction as the diaphragm normally moves in response to tiny changes in vacuum can effect mixture... A strong enough Gforce could easily outweigh (kind of literally if you take my meaning) the very small vacuum changes. The hard cornering could have caused mixture to go lean which could have caused the initial backfire. Or even rich causing an ignition problem/misfire could then cause a backfire. The standard advice when fitting mixer system reducers was to fit them so the diaphragm faced sideways in the car because acceleration/deceleration forces can usually be greater than cornering forces.

Mine is orientated as it should be - and I suspect the cornering tipped the balance. See below.

It's rare but is possible for a backfire to damage or at least change the response of the reducer. A change of response might be fixed by adjusting the reducer settings again... But first make sure the backfire didn't cause any other sort of vacuum leaks etc.

I was heading towards adjusting the reducer but thought a more concrete vacuum leak test worthwhile and found a massive one! See pics below.

Sequence of events then. Earlier cold start backfires damaged blow off valve (which led to weakened mixture at low load - at higher load, the blow off valve could close)
That weakened mixture/vacuum leak combined with reduced reducer sensitivity during hard cornering caused the big backfires which further damaged the blow off valve leading to even weaker low load mixture (but as before, OK under heavier load). Plausible?