Orangebean wrote:

As a side issue here, but possibly related, how do you test an AEB025 MAP sensor to ensure it's actually feeding correct data to the ECU?

I mentioned one way above... When the engine isn't running the manifold reading should be very close to 1 bar (atmospheric pressure), when the engine is running and warmed up I'd expect to see around 0.4bar at idle on a P38 that runs well. Software shows gas pressure relative to manifold pressure so If there's only atmospheric pressure in the LPG pipe when the manifold pressure is reading 0.4 bar the gas pressure reading should be close to 0.6 bar and as manifold pressure increases gas pressure should drop correspondingly with manifold and gas pressure always adding to very close to 1 bar (and it's easy enough to take an LPG vapour pipe off to ensure pressure in the gas system is at atmospheric pressure), this check to be done with the engine running on petrol of course. It's only a basic check but is usually a good pointer. But some of us can just swap pressure sensors lol..

@Dhallworth
I believe the pressure sensor is a standard AEB025. If not I'd dare bet it'd be the same spec as OMVL Piro's AEB pressure sensor. If you look at the sensor from the other side where the pipes are connected, does the vacuum pipe or gas pipe attach closest to the wiring connector?
Reducer pressure adjustment is the screw in the middle of the sticky-out bit of the reducer
Valtek type 30 injectors would make for an easy and relatively inexpensive swap (but be sure to order them with 4mm outlet nozzles). Would concentrate on pressure for now, not buy injectors yet.

You could change injectors but you'd still have the issue of pressure being too high, would sort the pressure issue as the first step and then you might find injectors work OK.

Can you post pics of bits of your system (or find same bits on the net and link to them)? Would want a pic of the reducer, injectors and pressure sensor.

Reducer pressure is usually adjusted using an allen key, allen socket is accessible from outside the reducer (no need to remove anything to access it), inside the reducer the allen socket turns threads which screw a plate in/out to adjust compression of the spring which acts on the diaphragm. I'd aim for 1.5bar diff pressure as a start.Got to adjust reducer pressure with the engine running on gas.

Edit - Just noticed this crossed with Gilbert's post.
Gilbert raised another point (bit about 1.2 to 1.5 multiplier), to add to that, the LPG ECU compensates for pressure so where you have 2.3 bar where we might expect 1.5 bar we might expect your 5.5ms gas pulse for 4ms petrol pulse to rise to 8.4ms gas for 4ms petrol when pressure is only 1.5 bar [(2.3/1.5)x5.5 = 8.4], which would imply your injector nozzles are too small for 1.5bar (8.4/4 would mean multiplier of 2.1 which is outside the 1.2 to 1.5 multiplier range Gilbert mentioned), except AEB ECU's have a limited range of pressure compensation and it could be that your system's pressure compensation is maxed out (actually minned out since we're talking about compensating for over pressure). If pressure compensation is minned out it will mean the engine runs richer (or at least fuel trims will be more negative) when running on gas at 2.3 bar compared to running at a pressure that the ECU can compensate for. That's the basic theory but in practice nozzle size makes more difference at idle than pressure does because higher pressures also have the effect of slowing the speed of the injector opening.

What pressure is set in software? This is the pressure at which the ECU applies zero compensation for pressure, the range of pressures the ECU can compensate for is centred around this pressure. None of this affects the fact 2.3 bar is too much pressure for the injectors to cope with.

Diff pressure is the difference in pressure between gas pressure and manifold pressure. The fact that after engine has been off for a while your diff pressure and manifold pressure roughly add up to 1 bar suggests those readings are correct (and also suggests you may have a leak on the gas side, but a leak won't be relevant in this).

But your diff pressure goes up to 2.69 bar when the solenoids first open and to 2.49 bar when the engine is idling on gas, this is far too much pressure and because your pressure readings seem correct this suggests a problem with the reducer.

Can also see that your manifold pressure goes from 0.41 bar when idling on petrol to 0.55 bar when idling on gas.. There are 3 possible reasons for that:

1. Sometimes when the ECU is supplying power to the solenoids and injectors it's reference voltages are affected due to the extra electrical load.
2. If the reducer leaks gas into it's vacuum reference port (vacuum at the engine side of the vac line, usually a line that's closed at both the reducer and pressure sensor, but if the reducer is blowing the other way and the sensor is in the middle it'll read some average between engine vacuum and reducer pressure.
3. If mixture is incorrect when idling on gas the engine has to draw more air to maintain idle speed (running less efficiently on incorrect mixture). Lambda readings reflect average mixture over a cylinder bank (in simple terms anyway), if average mixture is OK but the engine is less efficient it must be because some cylinders are getting richer mixture while other cylinders get lean mixture (if any cylinder gets lean or rich mixture the engine will be less efficient than if all cylinders get almost exactly correct mixture). Some cylinders getting rich while some get lean could point to injector problem(s) or could just be due to injectors struggling to dose accurately due to the high pressure.

Would expect readings from Testbook to be slightly different to readings from the LPG ECU anyway, not that there are many of the same type of readings to compare between the two. You won't be able to see gas pressure in Testbook.

Where's this.. I'd guess maybe near Tanhill or Holmfirth? I sometimes go for a drive around Tanhill and to the highest pub in England in summer.

Years ago I bought a BMW750 from Paignton, Devon and set off from Yorkshire to collect it in a Volvo with a rather large car trailer on the back. Stupidly I followed Satnav, which took me down some really tight country lanes with very sharp bends and a river ford. It was the early hours of a very cold morning when I got to the ford, had to take my shoes and socks off to check depth before crossing. Made the crossing then at the other side of the ford was a steep wet hill, front wheels kept spinning all the way up. Last time I'd use Satnav for a trip when towing and last time I'd want to pull something so heavy such a distance using the wrong type of vehicle lol.

Morat wrote:

that and you're not allowed to jump the queue ;)

I haven't forgotten you mate ;-)

@Dhallworth - The differential pressure shouldn't be as high as 2.47 bar, you will have problems if that reading is true. If the pressure is really as high as that it'll point to a problem with the reducer. If the pressure isn't really that high (yet it reads that high) it 'll point to a problem with the pressure sensor, loom or ECU.

What happens to manifold pressure when running on petrol? Were aircon, lights, etc on? In park or drive? I ask because most AEB systems on P38 V8's read just over 0.4bar manifold pressure at idle with warmed engine, if your manifold pressure drops lower when running on petrol it's because the engine is struggling more to run at idle on LPG (ICV opens more to keep rpm at idle). If that's the case it could be that even with half decent calibration and injectors that are in good working order the injectors are still giving inconsistent fuelling due to high pressure (as mentioned above). The pressure sensor is really two pressure sensors in one (manifold / gas), a problem on one sensor may or may not affect the other sensor, as sensors get old they sometimes show slightly incorrect readings but that doesn't always affect good results. As ECU's age the AD converters and/or reference voltage to sensors can see the ECU show slightly incorrect readings, again this doesn't always mean an immediate ECU replacement is necessary.

Yes I install and repair vehicle LPG systems for a living, could sort it if you like. But you could be closer to having fixed it yourself already than it seems, hopefully the new cable will work and point to an easy fix such as low reducer pressure, in which case we'd hope turning the pressure adjuster on the reducer will bring pressure back up. That said I wouldn't have bothered rebuilding a Tartarini reducer if the initial problem was low pressure, the kits will repair a leaking reducer but are less likely to repair a reducer that failed on pressure, they don't include all internal bits or the housing... and it seems whatever problem you had before the rebuild was the same problem you're having now?

Two vehicles here for repair at the moment: Supercharged Rangerover with a BRC system just came in which takes a lot of cranking when it's been running on gas, so I'll be looking for a leaking reducer diaphragm / petrol return which continues to dump petrol pressure when the engine is switched off / poor calibration on that one. Merc E500 with LPGTech, Antarctic reducer, V30 injectors with the same long cranking problem, already found the reducer on this one is leaking gas into the manifold and started fitting a new reducer.

V6 software will usually work on anything that V5 does, there was quite a change in terms of compatibility between V4.X and later versions with one of the biggest changes around V4.7 for most AEB stuff. 2005 Is quite old, I'd expect V6 to work but if it doesn't you might need pre V4.7 to work, maybe even V3.X. That said, Tartarini version numbers didn't always follow general AEB trends, the software works in a slightly different way too.

The interface setup you tried may have only 3 wires on the serial side?

Gilbertd wrote:

I can't remember if the pictures of the connections are of the plug on the cable or the plug on the ECU. Easiest way to check if you have a Leo/Millennium is to identify the 12V pin and work the others out from there.

Your diagram shows pinout looking from the rear of the interface cable plug, or to put it another way the diagram shows pinout looking at the LPG ECU's plug from the front socket end. I just compared to one of the AEB interface adaptors I linked to above.

As OB said you'll know when you can see readings but it'll probably need the reducer output pressure turning up.

Early software might need low com port numbers, later software usually works on just about any com port. Early software may need com port selection.

Depending on age of ECU you might need early / late software. Latest software is version 6 and is the most compatible version (works with most year ECUs unlike say V5 which is less likely to work on very early or very late ECUs) so V6 software is a good bet.

Pulling the LPG ECU fuse can sometimes sort connection failures (after putting the fuse back in of course). Molex pins soon get corroded if exposed and can even snap off sometimes leaving the snapped pin in the female end. You'll have noticed in Gilbert's diagram that the Tartarini Molex is pictured the other way up to the other system Molex. Some interfaces (mostly out of those that use USB to serial dongles before the actual interface but also some USB ones such as ESGI's own brand cable) which work OK with Leonardo's won't work with sequential systems, these tend to have 3 wires to the serial connector at some point in the chain as opposed to 4 wires.

OK thanks Gilbert.

I wonder who converted it, someone must be sweating! No worries, I always make triple sure with fuel returns lol!

On The Sun link there's talk of the interior looking burned out

Can anyone tell if this will be a BMW or Jag engine'd model?

If Jag engine'd I wonder if under bonnet petrol return piping has come off..

Hehe I thought Gilbert might post with wiring diagrams, nice one ;-)

Can see on the diagram how it's possible to wire a single Molex plug to work with both the usual AEB connection and the Tartarini connection, you just need the 2 additional pins on the 2nd row connecting to work with Tartarini.

If you bought the AEB cable you'd need the adaptor too.

As said, you can use your cable but you'd need to make the mods I talked about. If you look at your interface Molex and your LPG system Molex the pinouts don't match up, you'll have a couple of wires that connect when you plug the Molex ends together but will need to (at least temporarily) hardwire the couple of wires that the Molex connection doesn't make (because Tartarini Molex pinout is different to other AEB system Molex pinout).

Or you could buy another 8 pin Molex connector and wire it to your LPG system, that way you'd wire the new Molex plug the same as your Leonardo system's Molex plug so and it would work with your existing cable.

Clue to me is in where you said your cable works perfectly with your Leonardo and Millenium ECU's - Those ECU's will have the 8 pin (only a few pins used) Molex type connector as opposed to the later AEB type 4 pin Supaseal connector. Tartarini systems that use the Molex connector have a different pinout on the connector (some of the pins on the second row are used as opposed to other AEB systems that use pins all on the same row).

Your cable will work if you just chop into 2 wires on your interface (the wires that are not connected by the Molex connection) and the relative same wires on your LPG system's interface connector cable - just make a temporary patch using bits of wire to make the connections that your Molex connection isn't making. You should be able to compare wire colours to your Leonardo / Millenium system. Or instead of messing with USB to serial dongles, chopping into wiring or buying a cheap supposedly Tartarini specific cable you might prefer to buy a proper AEB USB to Supaseal cable and a AEB Supaseal to Molex interface adaptor. The Molex connector on the adaptor is wired for both normal AEB type Molex connectors and Tartarini type Molex connectors.

Morat will be right, it seems the software you're using expects to see a USB dongle plugged in to allow the software to work. USB dongle has nothing to do with your interface cable. Which begs the question - What Tartarini software are you using and on what version of Windows? The correct Tartarini software for your system shouldn't need a USB dongle to run (though a dongle could be plugged in to open up a few more options such as temperature correction) but even correct software can do funny things if running on the wrong version of Windows, sometimes if running on certain versions of 64bit Windows.

If nobody else beats me to it, later on this evening I'll post a Dropbox link to software that should work.

As OB said the LPG map shouldn't have changed, if it has changed it'll probably be because there's a problem with the LPG ECU.. They can lose mapping info when the battery is disconnected when they're old, so if the problem's as simple as that a remap may be an almost total cure (save for the fact the map would be lost next time battery is disconnected). The ECU losing mapping info is unlikely though, far more likely there's an issue with the reducer, injectors, pressure sensors, temp sensors, in which case mapping won't fix it but the interface will allow you to see readings and (likely) allow you to switch individual cylinders back to petrol (to find problems with individual LPG injectors).