WinALDL O2 data for LPG help needed.

[kz1000a]

Hi, I have a problem with my 91' VP V6 calais stalling at idle whilst running on LPG. It only happens after the engine has reached normal operating temperature. When it does start to stall, I turn on the A/C to raise the engine revs (doesnt stall as much- but still wants to), or I flick the engine over to petrol (no problems whatsoever on petrol).

I am asking if anyone here has O2 data for Win ALDL on a LPG engine.

The O2 data that I log whilst on petrol shows the O2 sensor fluctuating 0.2 volt -> 0.9 volt. I believe this to be normal.

The O2 data that I log whilst on LPG shows the O2 sensor remaining at approx 0.8 volt. It does not fluctuate at all. Is this unusual?

Have changed the coolant temp -> ECU sensor and changed the thermostat as well, to no avail.

As said previously, the engine runs fine on petrol.

Any thoughts appreciated.

Cheers, Mark

 

[Cheap6]

The O2 sensor data on LPG should be on average the same as it is on petrol. If you import the data for each fuel into a spread sheet and average them, they should match at around 450mV/0.45V. More data is better and the ECM doesn't always idle in closed loop, certainly on petrol and possibly on LPG (depends on the LPG controller), so it's best to avoid using data obtained when the engine is idling. It may be obvious, but data with the sensor cold i.e. not working won't reveal anything.

The nature of closed loop control on LPG (non-SVI systems) is such that the rate of change of the signal is much slower than it is when on petrol but it should still change.

The WinALDL data is at the factory output rate of 160 Baud so you are not getting a full picture of the petrol O2 readings either. The spreadsheet averaging gets around that somewhat. An LED mixture meter will illustrate what is happening better than will WinALDL.

It depends on what LPG system you have fitted, but if it's an Impco (very common and the type which was factory fitted) or a few others, it will default rich.

It is leaned back by applying some mixer venturi vacuum onto the converter diaphram, which is otherwise referenced to atmospheric pressure and determines the pressure of the gas being fed into the mixer. Atmospheric pressure as reference = 'high' gas pressure = rich, more vacuum = lower gas pressure = leaner.

The amount of vacuum applied is controlled using a solenoid valve that is opened and closed by the LPG controller.

If you are getting a steady 0.8V (rich), one possibility is that the vacuum solenoid is not operating. This could be because the solenoid has failed or the controller is failing to switch it, either is not uncommon.

It is also possible that a fault in the converter or mixer is resulting in the mixture being richened to the point where the closed loop control is unable to correct it. A leaking (internally) converter will do that.

I can give you WinALDL data but it probably won't help much because it varies so much with what I was testing at the time I gathered it: solenoid disconnected, different bleed orifice for vacuum solenoid, different mixer shims, different idle screw adjustment, different maximum gas flows on the mixer etc. With the likely state of tune of many LPG systems, I would suggest that any logs from other vehicles are going to vary a lot.

 

[kz1000a]

Thanks for your input Cheap6.

QUOTE "ECM doesn't always idle in closed loop, certainly on petrol and possibly on LPG (depends on the LPG controller), so it's best to avoid using data obtained when the engine is idling."

At present, I can only gather data while the car is stationary via my home PC.
I bought a 2nd hand lappy (vista) with USB ports ( no serial ) and also a Belkin F5U409 usb - serial adapter. The belkin adapter works fine when tested with hyperterminal and an external device @ 2400 baud, but I cannot get it to work with WinALDL. i.e. WinALDL recognises that COM3 is OK, laptop device manager is set to COM3, but its as if nothing is connected to ALDL connector. Strangely enough though, hyperterminal will recognise the 160 bit stream as a series of repetitive characters that look like, IMO, a 'house' with no windows or doors.

Once this issue is sorted, I hope to obtain data under load conditions, hwy, hills etc.

Is there a foolproof way to check the convertor for internal leakage?

Cheers, Mark

P.S. I have just stared to notice the 'rotten egg smell' when I stop at lights.

 

[Toby]

Hi kz1000a,
Two things, A) I believe that a new o2 sensor would fix the problems described (Did for Me) Also a corectly installed gas system should not be able to show with winaldl, The EMC cannot be allowed to reset the system while running Gas as Once on this fuel all settings must be static.

 

[kz1000a]

Thanks Toby,
Will install a new O2 sensor this weekend. Have read in other posts with identical sounding problems to my own (i.e. runs fine on petrol) that there may be an air leak after the throttle body or to replace ECU and/or eeprom chip. Even the cat conv gets a mention here and there.

I will start replacing the basic stuff.

Thanks again all.

 

[Cheap6]

Is there a foolproof way to check the convertor for internal leakage?

1) Run the car on LPG, switch off and let sit for a few hours. Disconnect the fuel injectors and try a restart. If the converter is leaking the car will start straight away, if it's OK it will turn over like a Falcon taxi.

2) Smell. With the vapour hose off the converter, can you smell gas as it leaks out?

Hi kz1000a,
Two things, A) I believe that a new o2 sensor would fix the problems described (Did for Me) Also a corectly installed gas system should not be able to show with winaldl, The EMC cannot be allowed to reset the system while running Gas as Once on this fuel all settings must be static.

Maybe. If the O2 sensor is OK on petrol, it will also be so on LPG. The best way to tell is to look at the cross counts i.e. the number of times that the air:fuel mixture crosses stoichiometric. While the best way would be to log it fast enough to see that, an LED meter is better than WinALDL in that respect.

If the petrol MemCal, or the petrol map on the LPG MemCal, is being used, the output from the ALDL connector will be present. (Not every conversion uses an LPG MemCal).

Disconnecting the LPG input wire to the ECM and disabling the petrol system another way (pull the injector plugs off) will allow data logging from the ECM while running on LPG. (Although the ignition timing will then be the petrol map, the mixtures will be from the LPG system).

 

[kz1000a]

Replaced the O2 sensor and cleaned out the IAC valve while the T/B was off the engine. Valve was carboned up but the pintle was movable. Reset pintle to less than 28 mm as per the instruction from Mr Greogory.

Still runs fine on petrol.
Still has difficulties idling on gas.

Convertor does not appear to be leaking. Removed diaprahm and held up to sunlight. Cannot identify any pin hole leaks or minute tears.

The solenoid valve that is opened and closed by the (LPG controller - autotronic controls corp, model 4046) was definately not kosher. Replaced this and still not a good idle. New solenoid is definately pulsing on and off. I do not know how long old soloenoid has been inactive.

The engine now "stumbles" on take off from stationary. If accelerated hard, she will faulter and stall. Autotron website suggest that the plastic air / fuel mixture control elbow situated on the convertor may have too small an orifice (0.078 inch from OEM)

Should have said in OP that ALL the idle probs happened after replacing the main gas line from the convertor O/P to the mixer. Was driving home a month ago and notived the smell of LPG inside the car. Found that the main braided rubber gas hose from convertor O/P to the mixer was thouroghly perished and had major leaks at the mixer end.

Question time. Is it possible for the LPG controller to over compensate for a slowly deteriorating rubber hose, and then have some "memory issues" when a new hose is installed?

I will try investigating the 16 year old electrical splicing connections from gas contoller to ECM to find out if they are sound.

Apologies for mini novel.

Cheers,

Mark

 

[Cheap6]

Replaced the O2 sensor and cleaned out the IAC valve while the T/B was off the engine. Valve was carboned up but the pintle was movable. Reset pintle to less than 28 mm as per the instruction from Mr Greogory.

As an aside, the 28mm won't matter unless the pintle has been pulled out further than the position it was when it was removed.

Still runs fine on petrol.
Still has difficulties idling on gas.

Convertor does not appear to be leaking. Removed diaprahm and held up to sunlight. Cannot identify any pin hole leaks or minute tears.

What type of LPG system is it? I'm only really familiar with Impco converters but just one diaphram? The valves on Impco can also leak as the rubber seals slowly indent from being pushed against the seats. A primary seal leak will push open the secondary seal.

The solenoid valve that is opened and closed by the (LPG controller - autotronic controls corp, model 4046) was definately not kosher. Replaced this and still not a good idle. New solenoid is definately pulsing on and off. I do not know how long old soloenoid has been inactive.

The engine now "stumbles" on take off from stationary. If accelerated hard, she will faulter and stall. Autotron website suggest that the plastic air / fuel mixture control elbow situated on the convertor may have too small an orifice (0.078 inch from OEM)

The size of the orifice would depend on how the system is laid out and on the engine. I haven't found it to have a huge effect and I would leave that at this stage. It's more likely to be an issue witha failed component or a tune to work around one than an issue with the way the system was first set up. I would put the orifice size in the latter category.

Knowing the duty cycle of the valve would provide more information. Do you have a multimeter with a duty cyle function? Alternatively, an analogue (0-20V) voltmeter across the solenoid terminals will oscillate around a voltage indicative of the average duty cycle eg. 7V = 1/2 of 14V = ~50%, 9V = 3/4 of 14V = ~75%.

A high duty cycle would indicate a basically (excessively) rich mixture being pulled lean, and vice versa; it should be somewhere close to 50%. If it's very far away from 50%, the basic tune is off somewhere.

Should have said in OP that ALL the idle probs happened after replacing the main gas line from the convertor O/P to the mixer. Was driving home a month ago and notived the smell of LPG inside the car. Found that the main braided rubber gas hose from convertor O/P to the mixer was thouroghly perished and had major leaks at the mixer end.

Maybe the LPG has been 'tuned' with one or both of those faults present (the inactive solenoid valve and/or the leaking vapour hose).

Question time. Is it possible for the LPG controller to over compensate for a slowly deteriorating rubber hose, and then have some "memory issues" when a new hose is installed?

Some, but not all, controllers do have a learning/memory function. I don't know the controller you have.

 

[kz1000a]

Kudos to you Cheap6 !

If your CV has no mention of "Advanced Diagnostics Technician", well .... it should.

QUOTE (original post from Cheap6)
If you are getting a steady 0.8V (rich), one possibility is that the vacuum solenoid is not operating. This could be because the solenoid has failed or the controller is failing to switch it, either is not uncommon.

It is also possible that a fault in the converter or mixer is resulting in the mixture being richened to the point where the closed loop control is unable to correct it. A leaking (internally) converter will do that.
END QOUTE

I dropped into Regent Street Motors, Oakleigh,Vic (original LPG installer-1992) to see if they might take a quick looksee at me motor. Old hand Adrian took time out to have a squiz and immediately identified that the convertor was leaking. (i guess my olfactory sensors aint what she used to be) Definately on the rich side. No amount of closed loop control was going to correct it. Recommended that I replace my old B2 convertor ($180 new vs $90 rebuild kit) and also another electrical valve/solenoid that allows the liqiud gas into the convertor ($22), just for good measure. As said previously, the vacuum solenoid needed replacing too!

You are definately switched ON, Cheap6.

So I will pick up both parts on FRI and install on SAT.

Hope this cures my woes and as an add on bonus, I might even get better economy figures. Currently 350 k's from 60 odd litres at the servo pump- city driving.

For any one interested, here is info on 4046 closed loop gas controller. FTW Racing - Fabricated To Win
It basically elaborates on what Cheap6 was saying, although it is on a carburetted engine, same principle.
Also Downloads - Instructions

I am using an Impco 225 mixer.

Again, kudos to you Cheap6 for your time and effort.

Cheers,

Mark

 

[kz1000a]

Got my convertor earlier than expected. Replaced it during lunch today. Initial findings are all is well. Car finally idles on gas. I would imagine it should also be OK when I take it for a test drive. Thanks all for for your input, particularly Cheap6.

Cheers,

Mark