LPG Discussion Forum

Hello all.
I have a 1996 P38 Range Rover V8 4.6 HSE, 110,000 miles running on a Tartarini sequential LPG fitted by Autogas2000 about 5 years ago. The idle on LPG has been ropy for as long as i can remember now, i've been slowly nibbling away at a solution for the last two years.

The conversion was done at 73,000 miles which is significant as RR camshafts tend to drop dead anytime after 80,000, so the system may well have been set up on a half dead cam. Obviously, i didn't know this at the time!

At 109,000 miles i had an engine that stalled every time it dropped to tickover shortly after changeover. 5 sets of traffic lights inside 3 miles on my daily commute, stalled at every one. As soon as were on gas the idle would surge to 750rpm and fall in ever increasing cycles until it would fall below 300 rpm and then stall. I also had a engine that was beginning to leak more oil than i was burning petrol!

Investigating the oil leak in the valley gasket revealed a very poorly camshaft with some lobes missing up to a mm of metal, representing up to 20% of lobe lift, explaining the less than Spirited performance.

The long term fuel trims at this stage were both banks +100% but still closed circuit.

New Piper cam, followers and a set of Magnecore HT leads later and i have a completely different engine, but i still have LPG idle issues. It's better, a lot better, but not right. The long term fuel trims went from one extreme to another to -85%. A thousand miles later it's settled at -16%.

I have the Tartarini software which i have used to retune the LPG many times, without really understanding how it works, which may explain why i've made little headway, i may have also miss understood what the long term fuel trims are actually telling me.

What i understand so far is that the LPG map is fixed (unless you manually change it), and all it does is follow the petrol map and apply the LPG correction. When you re-tune the LPG you only re-tuning it to the Petrol map, if the petrol map is wrong because of the trims then you simply end up with an LPG map that is wrong as well.

I always thought of the (petrol) fuel trims as an overall % correction, i no longer believe this. I now understand the map is broken up into many (400?) blocks and each one can have a trim applied to it, the trim displayed is just an overall average. If the petrol ECU only ever sees tickover when the engine is cold or not quite up to temp then it will trim to it over time. Normally the engine would see tickover on petrol at running temp many times an no doubt balance out the trims, but mine never see's tickover at normal temp on petrol, hence the problem, it only ever trims one way.

I also now know that on a cold start the petrol runs on open circuit, what i don't know is what the criteria is to return to closed circuit, or if were still on open circuit when we change over.

What i now believe is that my petrol map at tickover is wrong and when i recal the LPG, which is also done at tickover, i end up with an LPG map based on the wrong criteria. What i also did when i re calibrated the LPG last was to reduce the engine load as much as possible thinking it would run leaner with less load. This meant physically disconnecting the AC to stop it tripping in and out, gearbox to neutral to reduce any drag, fully charging the battery to reduce alternator load. This did make a big improvement and my stalling is much improved, but it's still there.

I'm now beginning to think i need to run petrol for a few weeks and allow the Petrol ECU to re-map itself correctly over the whole rev range and then re-calibrate the LPG to it, but the truth is, i don't really know.

Any idea's?

Many thanks
Andy.

The Range Rover can need around 100 miles of driving on petrol for the ECU to sort itself out.

From memory the fuel trim is supposed to be -100

Once you have let the ECU reset and made a note of what the fuel trims are like you can then adjust the gas map so that the trims are then as cloe as posible when running on LPG

Thanks for your input classicswede. I've just today changed the spark plugs back to a new set of NGK BPR6ES from the Autolite Optima 64 plugs i had recently fitted as a general service part, which constantly gave rise to missfire. I drove around 10 miles around town and the LTFT has lurched back to -93. I'll give it 100 miles as you suggest and report back.

Thanks again.
Andy.

With Magnecores and NGKs we can rule out any ignition related problems, anything other than NGK is asking for trouble in a P38.

If the trims are now at -99 sounds like you are about there with. Do a few miles and see how it stabilises out

I would advise caution interpreting the LTFT from GEMS as reported by generic OBD gadgets - they get it wrong. GEMS seems to be "nearly" OBD compliant, but with quirks. Example, the air mass reading is probably a factor of ten out? My own reader usually says any random large LTFT number, different at each use.
STFT does report okay.

rossko wrote:I would advise caution interpreting the LTFT from GEMS as reported by generic OBD gadgets - they get it wrong.

I noted on my Easy OBD11 software that the o2 readings were rotating between 0.2V & 1.25V on what should have been 5V sensors. I found a procedure to test them properly at the ECU and all was well. I'm now doubting everything the software is telling me.

Thanks for your continued support.

My generic OBD code reader (U581) always shows the lambda output on my P38 as 1.024v, although a meter at the ECU says otherwise. It also tells me that my LTFT is always -24% too. The GEMS ECU is from before OBD became mandatory and I suspect that meeting the requirements from 2000 may have had a lot to do with it being ditched and replaced with the Bosch unit on the later Thor engine.

Hello all.
i've been running on petrol and the LTFT's are currently -96. I've also just changed the air filter which i had to wait for, odd size for my year apparently. One thing i didn't mention was that on start up the tickover is sometimes high, sometimes low. I've since learned this is classic sticky air by pass valve, i had cleaned this valve before, but i know understand it's not just the cone that gets dirty but the shaft and screw. I've found a procedure to remove the shaft and i can inspect it. I've also found a couple of good pieces on how the GEMS ECU works, which has been helpful.
I'll keep you posted on how i get on, but i now think i've got a mixture of age related issues, not just the LPG. Needless to say, no fault codes to be seen.

Many thanks.

I've had high idle on my P38 before which was down to a sticky throttle position sensor. Slacken the two screws out that hold it in place, start the engine and if the idle is high, remove it from the throttle body but leaving the wires connected. If the idle drops as soon as it is removed, there's your problem.

Gilbertd wrote:I've had high idle on my P38 before which was down to a sticky throttle position sensor. Slacken the two screws out that hold it in place, start the engine and if the idle is high, remove it from the throttle body but leaving the wires connected. If the idle drops as soon as it is removed, there's your problem.

Unfortunately i'd took the valve out before i received this.

I removed the valve, left it connected electrically, turned the ignition on and the plunger screwed itself clean off! There was a considerable amount of waxy gunk on the shaft and thread which i cleaned off and lightly re-greased. I also placed a couple of drops of light oil into the servo motor as i'd read the back bearing can dry up and seize, at nearly £200 for a replacement it was worth a go. All back together and all seems well. We'll see how it goes over the next few days.

Many thanks.
Andy.

We saw some improvement after cleaning the IAC valve but then somebody passed me an article from Morgan about GEMS pointing out the piece about the idle mixture screw.

I had no idea what this mixture screw was, or even existed. I do now.

When I read the article I suddenly realised there was, what I recognised as a primary air circuit, which only serves the idle.

I thought that all the air for the idle passes through the IAC valve, it doesn't.

The IAC valve has a cone valve which is great for metering small amounts of air, the problem is that although the movement is linear, the flow increase is exponential. So ideally, you don't want the cone moving too far off its seat, the further away off its seat the more unreliable the metering.

The idle clearly has a known minimum air flow requirement, so what they've done is supply the minimum, less a bit, through a fixed jet (albeit adjustable) then the the IAC is used to add to the air supply to top it up with minimum IAC movement.

My fixed jet was all but blocked, forcing the IAC to screw further out to compensate, this potentially could make for an unstable idle as the metering is now less reliable, but it appears to be holding.

Q the LPG.

The LPG has a lower calorific value, so we need more fuel to hold the same level of RPM, which also means we need more air. The IAC valve is being asked to screw out even further, at best it can supply but the metering is no longer stable, at worst we run out of movement.

My theory is this.

According to the Morgan paper the ECU adopts a wait and see procedure. The O2 sensor calls for more fuel as soon as we switch to LPG, the IAC valve is already over extended and over compensates, the O2 sensor now reports over fuelling and the valve screws back in, but there's always a delay between sensor, action and result. Because we're asking the the IAC valve to work way outside its normal working parameters it can't cope, it gets caught in a loop because of the time delay, so hence the surging.

With the fixed air jet now clean and flowing air, the IAC valve is now working far closer to its seat, metering much smaller amounts of air, which is much easier to control and respond to, and the surging is eliminated.

Or so the way I see it.

Either way the surging has gone and all i've done is clean out this air circuit.

Result, the idle on LPG is the same as on Petrol. All that had happened was that oily air from a vent which was dumped into the air flow was being picked up by this air circuit, and slowly became more restricted as more and more dirt, possibly 15 years worth of it, got stuck and finally blocked up.

Two years but we sussed it in the end.

My thanks to all those that wrote in, on and off the board, this is what it's all about.

Thanks again.
Andy.

arudge wrote:The LPG has a lower calorific value, so we need more fuel to hold the same level of RPM, which also means we need more air.

No you don't, you need the same amount. For this reason, the LPG controller needs to supply more fuel than if running on petrol, hence the fiddle factor of increasing the injector duration.

Gilbertd wrote:

arudge wrote:The LPG has a lower calorific value, so we need more fuel to hold the same level of RPM, which also means we need more air.

No you don't, you need the same amount. For this reason, the LPG controller needs to supply more fuel than if running on petrol, hence the fiddle factor of increasing the injector duration.

I don't understand. If the volume of air required was the same then why did it idle ok on petrol but not on LPG? By clearing the air gallery I've only allowed more air, and arguably allowed the IAC valve to move back to where it should be, i havnt changed anything else. I also understood that LPG ran leaner than petrol at 15.5/1 as opposed to 14.7/1 for the petrol. So if we require a little more LPG we need even more air?
I'm certainly not criticising your comments, all comments are very welcome, I just like to understand what's going on. I thought I did, maybe I didn't.
Thanks again for your comments.
Andy.

It's the total volume that goes in, so, thinking about it, the amount of air is going to be slightly different but if anything I would have thought it will be less. You need more LPG as it has a lower calorific content but you need less as the stoichiometric ratio is different, hummm, getting too complicated after a day at work. I suppose it's possible that by cleaning out the idle air path you've moved the IAC valve to a more controllable part of it's travel.

We could probably theorise all night and, if we wanted to do lots of maths involving the relative weight of petrol vapour, LPG vapour and air, it might be possible to calculate the exact quantities of each, but it's now cured so why worry too much?

Gilbertd wrote:
We could probably theorise all night and, if we wanted to do lots of maths involving the relative weight of petrol vapour, LPG vapour and air, it might be possible to calculate the exact quantities of each, but it's now cured so why worry too much?

I don't want to launch into theory nor prove I'm right or wrong, but an understanding is important to me. But yes, it's cured and I'm chuffed to bits!

I've no idea what stoichiometric ratio is, but I will look it up, just to get a general understanding, always something to learn.

Thanks again.