Aeb auto tune just going rich

[minimutly]

Finally got around to getting to grips with the mgzt turbo.
When the car was bought it had been converted some 4 years previous, had done 90k miles and had a top end rebuild.
The car was running poorly on petrol, a huge weak spot on idle to boost building. Once on boost it was excellent, my lambda reader showing exactly this. This was after replacing a faulty primary lambda.
The system fitted is an aeb, which mint sold me a comms lead for, and I downloaded the risn software.
I've done numerous battery off resets, each time the petrol running has improved, but I suspect the rich gas fueling was driving the corrections weak.
So today I had a look with the laptop.
1st problem, gas pressure. I failed to drop the pressure, it was wound all the way out. I wound it in 2 whole turns before stopping, the pressure reading stayed at 1.35 ish.
2nd problem, while attempting an auto tune, the gas times just carried on going up, over 10ms, enough to almost stop the thing.
Anyway, I reset the whole thing, and somehow it seems to have reverted to sensible settings. With the petrol timings still way low, the gas was just as bad. I tweeked the left hand columns in the map way up, but left the maximum at. 18(feels high to me). It now drives reasonably well on. Both fuels, and will be driven on petrol until the engine check light goes out, we'll have another look then.
So, do I need a new regulator/rebuild kit, and what's going on withe the calibrate on the ecu?
Thanks for taking a look,
Huw

[Tinleytech]

This problem with the gas injection timing climbing and climbing can be caused if the gas injector/petrol injector pairings are mixed up so definitely worth checking this.

Regards.

Tinley Tech Ltd.

[minimutly]

Ah, thanks for that TT, the install has been tinkered with by an "auto electrician", before I got my hands on it. (Presumably the one who managed to bend one of the pins on the gas injector plugs).
I will have to check through the loom. K series mgzt should be sequential I guess.
Does 18ms at 1.4 bar sound reasonable injector durations? The mgzt runs around 7 psi boost I believe.
Ta,
Huw

[Gilbertd]

Does 18ms at 1.4 bar sound reasonable injector durations?

At what revs and load? The LPG durations want to be 1.2 - 1.5 times the petrol durations. So if your petrol durations at the time are 15mS, then yes, otherwise, no.

[classicswede]

First off AEB does not have an Autotune but it does have an autocalibrate to help with getting the mapping started.

The most common cause of what you are describing is overly high flowing injectors. As the autocal can not see lean it just keeps going rich rich rich until the engine stalls. Just manually calibrating saves that bother. If the injectors are showing towards too big on the scale then chances are that is why it is doing it

[minimutly]

Ok, I stand corrected auto calibrate, not autotune, thanks. Apologies for any confusion. As usual when you look at something new for the first time it takes a while to get to know what it's trying to do, especially when it's not doing it at all, or at least not very well.
So in autocallibrate, is it just trying to match a calculated curve (according to pressure), to the correct lambda at a set point in the rpm range, according to lambda?
Does it know what to set the left hand figures to? Or does it do a calc for this?
And what use is the too rich/ good/ too weak indicator on the calibrate page? How do I use it?
Thanks for answering,
Huw

[classicswede]

What you call the calibration tool is not important but what is important is to know it is only a toll to help with the very first step of calibrating the map.

The calibration does not use lambda or anything else like that but on works of petrol injection times.

The best way to do a calibration is by going on the map page and matching the petrol injection times at idle. For example your tickover petrol injection time at idel with hot engine but fan off is lets say swinging between 2.10ms and 2.20ms. You switch to LPG and the petrol time goes upto 2.40ms, you then know to select the full map and add 5 at a time until the petrol times come back to the 2.1-2.2ms. The only thing you need to watch is the gas injection at idle needs to be over 2.5ms (gas not petrol)

[minimutly]

Mmm, I have a problem straight away, the petrol running is way weak on idle, unless is picked up since I had the car last. Hopefully, by the weekend it will be there, if not I'll have to buy a scan tool to see what's going on.
But I see what you mean about matching lambda, but what do I adjust? Surely not the individual cell?
Huw

[classicswede]

you can select all the cells and add the same amount to all of them at the same time. For fine tuning after you have the idle sorted you can select smaller areas to adjust or even just single boxes.

Don't worry about the lambda to start with just look at injection times. Just use the lambda after to make sure you do not have lean spots but bear in mind under cruise it should switch rich, lean, rich lean (unless it is a wideband)

[minimutly]

How do I select a range and correct all?

[hitman]

Highlight all the boxes you want to change

press enter

select percentage or whatever way you want to increase or decrease

add your number in the box (use a minus sign in front of the number if you need to decrease)

[LPGC]

What injectors are fitted and what injectors are selected in the software?

The ECU firmware may dictate a minimum LPG injector pulse duration - If this is the case then there can come a point (with injectors that flow too much for the engine with the set gas pressure) where lowering the numbers in the map will not decrease gas injector pulse duration, so will not lean the mixture while running on gas. If this is the case then you'd need to lower pressure at the reducer (or fit injectors with smaller nozzles). In any case, make sure the correct injectors are selected in software. In my experience, some Romano injectors are capable of high flow when they are open but take a relatively long time to open / close... this is one type of scenario that can cause rich running on gas at idle with a small turbo'd engine with an ECU that dictates minimum gas injector duration.

Above, Tinleytech advised a possible wiring issue, I agree, would also add that if the engine misfires due to being over-rich during autocal then petrol and therefore gas injector duration might also.

I would try Classicswede's tuning advice, maybe avoid autocal for now.. especially if (as you seem to imply) the engine is running open loop (weak mixture at idle on petrol)? If it is running open loop you might get gas tuning at idle something like correct (at least until you get it running closed loop again) by tuning according to engine manifold pressure (adjust calibration boxes until you see the lowest engine manifold pressure and then decrease the numbers just a little so the manifold pressure increases just a little)... By the way, what manifold pressure are you seeing at idle on petrol / on gas with a warm engine?

Simon

[minimutly]

That's some very useful advice, as well as some difficult questions without having the car running in front of me.
I'll fire up the laptop and see what injectors are selected - from memory they looked like black plastic block shapes with 4 tiny screws holding them together. The one I did strip had a coil spring and flat disc type valve - quite primitive looking I'd say.
Once I'd cancelled the autocallibrate the gas was weak, and I haven't added enough on idle/low load to make it register on the lambda, but then I don't really want to otherwise it will drive the corrections weak again/not allow the petrol times to come into the closed loop sector.
Interesting you suggest using manifold vac to get the mixture right at idle - I'll try it, but why would it be better than lambda?
Thanks for the tips, will report back when I know something new.
Huw

[rossko]

Setting up the gas in any conventional way requires the petrol system to be functioning well and in closed-loop mode. Doesn't sound like that is the case? You may want to shift priorities and 'repair' first.

Can you tell us what you are looking at conclude rich/weak, it's not clear what instrumentation you have access to,

[minimutly]

Rossko, you are correct the car is not running right in petrol mode. To date I've checked everything I can, and can find no fault with any item. However, when the car was bought, it had a weak misfire/ flat spot before coming on boost. I hooked on a lambda reader - just a barograph display for the narrow band in the manifold. It made no sense at all, so i changed the lambda for one I had off another rover. Immediately it showed the weak running range between idle and boost coming on.
That's how I know it was weak. (The old lambda read rich all the time).
Up until recently I haven't had acces to the gas ecu, so she ran it as it was, with the petrol showing weak (still with stumble until on boost, but fine after that). I had checked over the gas, it ran with one cylinder down, hence me stripping that injector, only to find someone had bent a pin in the plug - so easy fix. So it's been running on gas mostly since, and this was showing rich everywhere, and using quite a bit.
My assumption is that the initial lambda problem drove the petrol corrections weak, once the new lambda was fitted (and battery disconnected) the engine light went out. However using it on gas has driven the corrections weak again, so i can't properly calibrate the gas. However, the orders she (daughter) has had is to drive it on petrol for a while, to see if the engine light will go out again (after coming on while doing the auto calibrate thing on Sunday.
Btw, I have a scan tool, but it doesn't work on the zt for some reason, I will buy one that will.
Please ask if I've been unclear about anything.
Huw
Ps I think I have a problem withe the vaporizer, the pressure adjuster was all the way out, showing 1.4bar if I recal. Turning it in two turns made no difference. This May account for the originally rich gas mixture. Should the pressure drop to zero (running) withe the screw all the way out?

[rossko]

Adjuster only gives a limited range of adjustment, say 1 to 1.4 bar. I would imagine yours is at max and you need to screw in further.
Response to adjustment can only be seen when actually running on gas, not static. Usual action is to adjust whilst idling on gas.

In case of unexpected pressure readings, always check vacuum pipes to both reducer and pressure sensor, for leaks splits or blockage.

[classicswede]

some screen shots at idle on both Petrol and LPG would be a good idea.

As already pointed out it would be best to get it running right on petrol first. Have you ttied running it on petrol for a day or two?

[minimutly]

Ah, I assumed screwing in would adjust pressure up! And yes I understand it would need to be idling. Something to try there.
Screenshots impossible since the laptop is not connecting to the internet currently, but I'll try again.
Also, the car is back with my daughter, with instructions to use only on petrol.
Thanks,
Huw

[minimutly]

[img]C:\Documents%20and%20Settings\User\My%20Documents\My%20Pictures[/img]

1st attempt at a screenshot, this is an offline shot. pparently the mil light has now gone off on the car, so will have a tune on the weekend.
Huw
poo, I'll have to use photobucket then.

[Gilbertd]

poo, I'll have to use photobucket then.

Yup, we can't access your C:\ drive........