LPG Discussion Forum

XJ8 calibration sorted by LPGC
Page 1 of 1

Author:  LPGC [ Sun Nov 13, 2016 8:45 pm ]
Post subject:  XJ8 calibration sorted by LPGC

Not going to mention which installer fitted this install and then messed up the setup/calibration, except that it was fitted not too long ago by an installer who doesn't install anymore but had a decent reputation.

Owner reported it had always run rough, particularly when fairly cold, particularly around idle and on overrun, particularly when throttle position was changed. It also changed back to petrol at high rpm with heavy right foot and the fuel gauge worked the wrong way around (i.e. showed full when empty and vice/versa). He had been back to the installer with these problems and (at the time) the additional problem of it using half petrol and half LPG (was actually using this much petrol as opposed to just petrol gauge shifting). On that occasion the installer improved the running a bit and prevented it using petrol, but he reckoned the fuel gauge problem was the norm and didn't even address it. Knowing this installer a bit, he'll have tried applying the (at the time) fairly recent introduced split fuelling abilities of the ECU, which led to it using so much petrol. Maybe split fuelling was also the reason he didn't feel it necessary to fit a valve lube system. The Emer ECU was in an old AEB2568D case which had had the sticker removed and case polished up, I don't know the reason but would guess the original ECU from that case went into a different vehicle inside the new Emer case.

The customer is from a fair distance South of me but had been holidaying in Scotland and opted to stay Sat overnight at a local pub (stroke guest house) and bring the car to me this Sunday morning. I half expected some nozzle/pressure issues knowing who'd converted it. I found the problem to be along those lines but with too high a combination of nozzle size/pressure rather than his usual too small combination (which is also a usual problem with done in a day installs)... I connected the laptop, saw the 1.6 bar pressure and map and immediately knew that no nozzles would be fitted in the OMVL injectors and that most problems would be due to this combination of nozzle size and pressure. Pressure from the OMVL HP reducer set at 1.65 bar, no nozzles fitted in OMVL injectors so effective nozzle size around 3mm... Way too high a combination of nozzle size / pressure to work properly on the XJ8. These points should have been immediately obvious to the installer. I would expect a working pressure of around 1bar on this engine with no nozzles in the OMVL injectors, 1.6 bar with no nozzles being way too much pressure.

Combination of too big nozzle/pressure combination results in gas injector pulse duration being less than petrol injector pulse duration and in severe cases results in gas injectors pulsing below the minimum duration they can work properly at. The best that can be hoped for in either of these situations is that the engine will run properly but fuelling accuracy will suffer due to diminished effective resolution of gas injector duration. I can only surmise that the installer received sub 2.5mm nozzles in the OMVL injectors from the supplier (you get random nozzle size with these injectors), didn't fancy drilling the fiddly little things, didn't have a stock of different sized OMVL injector nozzles and didn't know that to lower OMVL Dream HP reducer output pressure it's necessary to remove a spring from the reducer.

At idle with warm engine Pinj is about 3ms on this engine, with gas temp of 60c Ginj was only 2.8ms. At fast idle Pinj was 2.51ms, it would have been lower except the Emer ECU firmware sets 2.51ms as the minimum pulse duration for OMVL injectors. This meant that on fast idle the mixture would always be rich. This was with warm engine - With a cooler gas temp the situation would be worse, because the ECU has to compensate for gas temperature because temperature effects gas density almost as much as pressure does. With Ginj of 2.8ms for idle with warm gas, ginj might have needed to be lower than 2.51ms for idle with cooler gas but would be stuck at 2.51ms... this could account for rich mixture all the while with a cooler engine. OMVL injectors are better than quite a few budget injectors (and quite a few over-rated expensinve injectors) in terms of working well at low Ginj's but even knowing that I'd doubt the capabilities of working properly down to Emer's rating of 2.51ms for longer than just momentary conditions. I could have deduced all this even without even looking at the car, just given it's spec, nozzles and pressure. Would also have known that a map like the one pictured in 'before pics' below was blatantly wrong.

Instead of messing fitting new nozzles in the OMVL superlight injectors (awkward if someone has slightly damaged the plastic threads) I opted to remove the reducer spring. It still has one (usual narrow diameter) spring, I just removed the wider diameter spring that makes the OMVL Dream reducer HP pressure spec. This brought pressure down to 1bar, ideal. The fuel gauge was sorted as easily as changing the 'AEB' setting to '0-90' setting. I redid the calibration and numbers were intuitive, progressive and with multipliers above 1.

Before pics.
1600rpm changeover set.. Note reference pressure 0.8bar but working pressure was actually 1.6bar, not good.

'oh dear' map looks like a dog's dinner

good job sequential changeover was disabled before too... I was able to enable it after setting everything else up properly.

After pics. 400rpm changeover set with sequential changeover. 0.9 bar reference with reducer actually doing 0.9 to 1 bar, much better.

and allows a proper map like this

fuel gauge sorted as easily as changing one setting

It remains that the modern day OMVL Dream reducer isn't really up to the job for a 4L V8 (the old OMVL Dream reducers of around 10 years ago certainly would be), so I set the system with a petrol addition above 12ms pinj, which only occurs when the engine is making most torque which with this engine is at fairly high rpm. Though reducers in general (maybe particularly the Dreams) can flow more gas when pressure setting is high, so my lowering it's pressure won't have helped the reducer cope at high power levels, even with the HP spring it was incapable of flowing enough gas for the 4L V8 anyway. My only alternatives in this case were to spend a fair bit more time digging out and fitting 2.5 OMVL Superlight nozzles (I'll have hundreds) or fitting a different reducer entirely. The customer opted for me to remove the reducer spring, which took 10mins, and recalibrate the system properly.

Now it runs properly, it will also run properly when it first swithces to gas on a cold day, the fuel gauge will work properly, it doesn't beep back to petrol when you boot it, there will be no more problems due to incorrect fuelling. I advised the customer he should have a valve lube fitted, which he says he'll return for one day soon. Though everything works fine now, this setup isn't totally ideal because it remains the reducer isn't quite up to the job - but the customer is more than happy with the results today and isn't too concerned about the petrol addition aspect at high engine loads because he seldom works the engine hard. He did comment that he doesn't like the intake roar of the engine when working it hard anyway, because the installer had curiously fitted the reducer in place of the ducting between air filter and front air intake area of the car.

Another thing we can take from this is that ECU pressure correction doesn't work entirely correctly and/or ECU firmware applies incorrect offset for given injectors: The original installer had set reference pressure to 0.8 bar and had actual pressure of 1.6 bar. My changes see reference pressure set at 0.9 bar and actual pressure at 0.9 bar, yet even though the original reference pressure was set lower than my reference pressure setting and numbers in the original map were around half the value of the numbers in my may, both maps provided fueling that saw fuel trims near correct in the engine's mid load range - If pressure correction worked entirely correctly the original map should have meant very much leaner fuelling than it (and my map) provide. ECU pressure correction errors and issues that can be foreseen just from knowing a few basics such as nozzle size, pressure and multiplier (at any point in a map) such as injectors pulsing for shorter duration than petrol injectors, or shorter duration than minimum spec of injectors, or longer than available injection window at high load/high rpm and issues with other correction factors such as tempertaure correction with too low/high multipliers are all points I have discussed on forum before, even tried impressing on the original installer. They are points that might be just about got away with on a simple low demand install but certainly won't be got away with on a more techy/demanding install. None of this is shadow boxing.


Author:  classicswede [ Tue Nov 15, 2016 5:35 pm ]
Post subject:  Re: XJ8 calibration sorted by LPGC

With the key that opens more features you can see the pressure correction and it only allows a small window from base pressure hence not coping with the actual pressure being way over the base pressure

Author:  LPGC [ Tue Nov 15, 2016 6:02 pm ]
Post subject:  Re: XJ8 calibration sorted by LPGC

classicswede wrote:
With the key that opens more features you can see the pressure correction and it only allows a small window from base pressure hence not coping with the actual pressure being way over the base pressure

Agreed. I think I've mentioned on forum before that AEB pressure correction works better than on most systems but maxes out, phrased the same point in a different way above.

You and I will have keys/dongles that open up some of the usually inaccessible settings etc for systems such as OMVL/Bigas in software versions up to around 5 but they're not generally available, or available at all for some systems, and don't work on recent systems using v6 software. On a few occasions I've dug out an old parallel port dongle and used V5 software on a desktop PC just to adjust such as temp corrections on OMVL installs.

Reckon we'll also agree on the main points above - Installer should have known the combination of 1.6bar with no nozzles was too much without even trying them, having tried that combination the numbers in the map tell that story anyway, so does minimum pulse duration hanging at 2.51ms and ginj being far smaller than pinj at any engine load. Should also have known about removing the wider diameter of the 2 reducer springs to bring pressure down, or fit smaller nozzles, or different reducer or different injectors.


Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group