1997 Range Rover P38 4.6 HSE Gems V8 DIY install.

[MrGorsky]

Okay,

First up, I've upgraded my ECU firmware to the latest version (4.9.1 I think) and am now using the dongle free BiGas software with the fuel mapping option, rather than just carb modifying.

I've adjusted the reducer pressure to 1.25bar. Re-Calibrated, and the software is giving gas inj timings of around 5.5ms at tickover. Also it's still telling me I've got slightly too big injectors. Although it is running fine, the tickover is much better, and it drives really really well. Loads of power, no stuttering or anything like that.

Should I up the pressure a bit to get the Tinj times down to nearer 4ms? or is it okay. And should I worry about the software's opinion on my nozzle size? I have a double output reducer, and the vap pipes run through the injector rails in big circle. With the fuel pressure T half way round the circle.

I hope to adjust the fuel map this weekend, I need to do some more test driving (hopefully this weekend), but last night petrrol inj timings went from about 10ms on petrol to 13ms on gas (at 70mph) so I'm running a bit lean right? And need to up the number in the map.

Cheers for all your help so far!

Al.

[rossko]

Also it's still telling me I've got slightly too big injectors.

It's just a guide, don't put too much faith in it.
"Long" idle times the exact opposite of "too big", and are no problem of themselves - so long as they are not then too long at full power. Too-short idle times cause roughness or hunting.

but last night petrrol inj timings went from about 10ms on petrol to 13ms on gas (at 70mph) so I'm running a bit lean right?

Probably. It's so much easier to do if you can get OBD access to the petrol system and monitor the trim %, but it is possible if you keep a sharp eye on petrol inj times.
Switching gas->petrol usually produces best effect, because the switchover process in that direction is cleaner.

Neither times nor trims can tell you what happens at full throttle (open loop), which is where monitoring lambda sensor outputs is useful.

[MrGorsky]

Neither times nor trims can tell you what happens at full throttle (open loop), which is where monitoring lambda sensor outputs is useful.

I've tried monitoring the lambdas, but they swing around up and down so much I'm not really sure what I'm seeing! (Or what I should be seeing)

I do know that at full throttle, the lambda reading backgrounds on the software go green, and stay green. The probes output 0Vish. How I use that info to improve my fuel mapping is a bit beyond me! I've read and read and read about it but I still don't really understand what monitoring the lambda tells you! Or at least, I'm not sure how interpret what I'm seeing.

I'll leave the pressure alone then. It really is as smooth as it is on petrol, and is very very drivable. I'll have go at tweaking the mapping this weekend and try to line up the Petrol Tinj reading on both petrol and gas and probably leave it at that. I'm getting a bit beyond my understanding of it all now.

Thanks so much for everyones help though, I can afford once more to run a car!

[rossko]

Even with incomplete understanding, it is possible to see what lambda sensors do when driving on petrol and compare with LPG. Worth doing for the full throttle case.

[mat_fenwick]

Below WOT, the petrol ECU will monitor lambda readings and adjust fuelling till it's exactly right (stoichiometric) which was why you see the readings rapidly switching from lean to rich. This happenes even if the LPG system isn't set up perfectly, which is why the petrol inj. times alter. At WOT, the petrol ECU will ignore lambda feedback and run slightly on the rich side (running lean can cause valves to burn out) and the petrol inj. times will not change. Make sure the LPG mapping is such that it's also slightly on the rich side.

On the Lucas 14CUX system, WOT is defined as something like over 75% throttle and or 3500 rpm. Yours may well be different being a later management system.

[MrGorsky]

Hello Again all,

I've just replaced my vapour filters (it's one year since my install) and done a quick recalibration.

I was re-reading this thread to check my notes as it were, and Rossko you mention it being useful to be able to read the short term trim % to help with getting the map just so.

Well I now have a Range Rover diagnostics computer able to give me long and short term trims for both banks of the engine.

How do I use this knowledge to get the map perfect? I'm thinking I might start again from scratch.

If I set my reducer to 1.2 bar and perform an auto calibration, I usually have to increase the injector times by 10% in the 1000 rpm zone to get the tick-over nice and smooth. I've also found increasing the "enrichment on acceleration" by a couple of clicks makes the car better to drive, and less hesitant when pulling away.

As an example, if I look at my short term fuel trims and they say -12.5%, does this mean I should increase the whole of my map by 12.5% to compensate? Is it as easy as that?

I'd love to get some help with this. It's been running very well for the last year, but I'd love to get it perfect if I can. I'm a bit OCD like that!

[Tubbs]

In your example -12.5 is 12.5 % too rich. So adding 12.5% would make it a larger - fuel trim, so taking away say -10% on those surrounding blocks where the dot floats at will give you a smaller fuel trim. But -10% does not correlate to taking away 10% on the fuel trim. If you take -10% from the -12.5% trim it will come out at a -8% or so.
You need to concentrate on getting the map with steady correction figures with no huge jumps. So one block doesnt say 130 and the next 255.
Thats the key to good mapping.

[rossko]

Before leaping in, remember there is only one fuelling map in the petrol ECU. Trim values of themselves tell you nothing about gas performance; the difference in trims between running on petrol and gas is the interesting part.
Step one then is establishing the petrol norm, drive around on petrol and see that LTFT are sensible under various conditions. GEMS has more than one LTFT depending on engine load etc. GEMS is a slow leaner, so ideally you'd drive round on petrol only for a couple of days to allow it settle down with appropriate trims not influenced by gas.

[Tubbs]

Also worth remembering the gems won't learn anything under a 1/4 tank of petrol.

[MrGorsky]

Thanks for your help. Sorry I'm a bit late getting back to you, I updated my email address and was deactivated until a mod classed me fit to fly!

So,

I've reset the adaptive values in the GEMS ECU using my Nanocom diagnostics tool, I will spend a week on petrol to get a good base to start from however....

I did a few trips up and down the M3 at various speeds yesterday to do a bit of experimental mapping, and I was able to get the Short Term Fuel Trim (STFT) to vary by no more than 2.5% in the end. To get to within this though, I generally had to increase the auto calibrate 'K' number from about 65 to 81.

The pressure on my BiGas twin reducer is 1.4bar and I've got 3mm injectors.

The lambda signal was very lean when auto calibrated, and once I'd upped the number the lambda was good, swinging from rich to lean as it does on petrol.

I had to increase the idle setting up to 84 from 65 to get it to idle smoothly. Again lambda was good, and the STFT were at 0%, and similar to the idle Tinj on petrol.

Do I put this down to a bad auto calibrate? Or am I better off just ignoring the auto calibrate figure (now that I can) and just manually tweaking the whole map to fit with what the STFT and Lambda probes are telling me? I wasn't expecting the auto calibrate figure to be so far out is all, or do these K numbers represent very small steps in actual fact.

Also, as standard, my map goes from 1000rpm up to 6000rpm on one axis, and from 1.5ms to 15ms on the other. Is there any milage in me changing it to 4ms up to 24ms and say 1000rpm up to 3500rpm to get a better range of settings for the typical driving my car does. I don't mind setting it to return to petrol over 35000rpm, as I'd only ever rev that high in an emergency, or when my brother is borrowing my car...perhaps about twice a year. It's seems a waste to have so much of the map not used! Will this give my map more "resolution" or doesn't it work like that?

Thanks again for your help so far, I realise that by not letting the GEMS system learn a bit on petrol I'm messing with an imperfect "factory" fuel map, but it does already seem to be spritelier than it was before, and I was excited to have a go!

[rossko]

I was able to get the Short Term Fuel Trim (STFT) to vary by no more than 2.5% in the end.

That's good. I really wouldn't obsess about it any more than that.
Check that all is well i.e. rich on the sensors, at full tilt (where trims are meaningless).

The autocalibrate can be better or worse depending how quirky the host car is. Perfection is not expected, just a good first approximation. It was never intended for amateur plug-n-go, some manual tweaking is perfectly normal.