Tartarini SGI has stopped working.

[stevejs]

My Tartarini LPG system has worked great for the last 60,000 miles on my 2004 BMW 4.4i with valtronic (320hp).

Then it started switching back to petrol if i lifted my foot off the throttle within 30 seconds of switchover. I lived with this for a year.

Then last week coming round the M25 it started switching back to petrol if i took my foot of the throttle too rapidly - it did this 3-4 times.

Then the next time out it switches back to petrol immediately after switching to gas. It does this every time now.

I had a look at the system under the bonnet - no leaks, everything looks fine. Plugged in the laptop, - everything looked fine except for the LPG pressure which was way too high at 2bar - so I adjusted this back down to 1.2 bar on the vapouriser. I had replaced the diaphragm in the vapouriser 6 months ago and the pressure had been fine at that time.

But the problem remains, switchover sequence looks normal but the system immediately switches back to petrol every time.

LPG trace.png (30.32 KiB) Viewed 4740 times

. The attached screenshot shows the switchover sequence at 4 seconds.

Is the problem that the gas injector time is too long as it seems to start at 4.8 ms where the petrol is 3.2 ms.?

The fact that the LPG pressure dips a little on switchover looks like the injectors did switch, but what would cause the too-long pulse?

Also there are no errors logged in the Tartarini error screen.

What can I check next?

Steve

[LPGC]

Valvetronics differ from normally aspirated engines, when running in Valvetronic mode they don't have any manifold vacuum. When your Tartarini ECU sees no petrol injection pulse (when you lift off the throttle on over-run) and no manifold vacuum it assumes your engine has stalled so switches back to petrol. The cure for this is to connect the brown rpm wire so that the ECU can see the engine hasn't stalled when there is no vacuum and no injection pulses. Usually the brown wire is connected to an ignition coil but Valvetronics can also stop firing ignition coils on over-run so on Valvetronics (which always have canbus tachometer rather than hardwired tach) is usually connected to a cam sensor. When you connect the brown wire you also have to change software settings to detect rpm by the brown wire instead of through frequency of injection pulses.

The fact that gas injectors pulse for 4.8ms when petrol injectors have been pulsing at 3.2ms doesn't meen anything in regards to your issues and may be OK (or as good as it gets) with the components you have fitted.

Symptoms may only have occurred recently if earlier your engine wasn't actually running much in Valvetronic mode... I'll come back and edit this thread with a link to a thread about one of my Valvetronic conversions. Various things can prevent Valvetronics running in Valvetronic mode including poor fuel trim readings in any mode (Fully Valvetronic, fully normally aspirated or any partial mode).

Should never adjust reducer pressure unless the engine is actually running on gas. Pressure reading will be higher at other times but is no indication of what pressure will be when actually running on gas.

[stevejs]

Yesterday I fitted new Crankcase "CPV" valves and found both the old ones had splits and were creating a manifold leak. This replacement has improved the smoothness at tickover - I guessed there was a problem when checking the fuel Trims - as the STFT's went up at tickover which since the fix they now don't do.

The LTFT's bank 1 9.4%, bank 2 14.1 % This hasn't changed even after driving maybe 250 miles on petrol or (so far) since doing the CPVs. The STFTs bounce around zero to +- 9% when driving, I thought that they looked more negative than positive.

As for Valvetronic - my LPG seems to have a very simple setup.
The pressure sensor is connected to the LPG rail pipe only.
The vaporizer (Tartarini RPG/05) has a vacuum pipe to the manifold.

This simple setup has worked perfectly for many years.

I installed new Valtek type 34 injectors in 2014 and from memory (may be a bit dodgy) pressure was 1.3 bar and injector timing just a little more than petrol.

What would you recommend as the next step to fault find this?

Steve

[stevejs]

I've had some more time to look at the my LPG system yesterday. The chart above shows "0" pressure (yellow line) that isn't right. Taking off the pipe from the pressure sensor to the gas manifold resulted in "oil" running out, it had filled the tube. The sensor looked oily and the tiny hole in the sensor for the pressure reading looked blocked.

Cleaning this out and reassembling the system it has started working, but the gas pressure was only 0.7 bar. turning the vapouriser adjuster I was able to increase the pressure but it is maxing out at 0.84 bar where I want to set it at 1.5bar. The Engine is running lean - STFT's at +25% when I use the throttle at all.

Which way should I be turning the adjuster on the vapouriser to increase pressure, RPG/05 ?? - The pressure goes up if I turn the adjuster anti clockwise - but comes to a limit at .85 bar.

I will also check the liquid filter today as I am wondering what has caused this upset.

Also the Tartarini software is causing much trouble and is very unreliable crashing my laptop every few minutes, (Version 6.01). This makes the whole job much harder especially with windows updates coming in half way through with yet another reboot in the middle of testing.

[LPGC]

Here's the thread I mentioned above https://www.lpgforum.co.uk/viewtopic.php?f=27&t=14183

If you read through that thread you'll see that your Valvetronic install is along the lines of what I call an old-skool Valvetronic conversion. Old-skool Valvetronic conversions can counter-intuitively give better results with injectors that are not as linear as other injectors.. Change or 'upgrade' injectors and you can upset some of the dynamics and your ECU has no way of compensating for such dynamics.

On this 'new' problem of reducer pressure not adjusting above 0.84 bar... On your setup reducer pressure is referenced to manifold pressure but your pressure sensor is referenced to atmospheric pressure, so when the engine is running in none-Valvetronic mode (throttled mode) and the engine is idling with manifold pressure of around 0.3bar (which is 0.7 bar below atmosphere) it will pull reducer pressure down by 0.7 bar. If you add that 0.7 bar to your 0.84 bar reading the reducer is really pushing 1.54 bar relative to manifold pressure. You've answered the question about which way to turn the adjuster to increase pressure yourself but it is likely the reducer is only adjustable to 1.54 bar. Mind you, you might want to try a new pressure sensor since you've said yours was full of oil... Although it seems to be working again now maybe it isn't working properly.

Unlikely the liquid filter is causing your problems.

A lot of software becomes unreliable while Windows is doing updates.

[stevejs]

Thanks LPGC

I've read the the article you have referenced, fantastically detailed.

I reckon I am most likely in non-Valvetronic mode as that explains the pressures. I did trigger a camshaft error (which I cleared) when I was investigating smoke from the exhaust when starting the car and was worried about "intermediate Lever" wear. Changing the CPV's fixed the smoking but I wonder if it's not yet gone back to Valvetronic. I will check the manifold pressure - swapping the manifold to vaporizer pipe to the manifold pressure sensor and reading it of the Tartarini display screen.

Assuming that's the case and I have no OBD errors will the system revert to using Valvetronic on its own?

Thanks again for your help.

Steve

[LPGC]

By all means move the vacuum pipe connections to facilitate reading manifold pressure via the LPG system but remember to put piping back to how it was before running on gas again or could expect it to throw calibration way off.

Reported cam sensor errors often turn out to be phantom errors on these engines but phantom errors can cause a switch back to throttled mode just the same a real error/sensor issue. Usually phantom errors will self clear if fuelling etc is correct but (like I said in the other thread) if the LPG system can only provide correct fuelling in one mode, say Valvetronic mode, the switch back to throttled mode then causes incorrect fuelling when running on LPG which then prevents the switch back to valvetronic mode so the situation is latched (perhaps for want of a better term but you get the gist). There would be a way to overcome the latch, run on petrol until the engine switches to Valvetronic mode again and then when you switch back to LPG fuelling is correct again (assuming the LPG system is setup and calibrated properly for valvetronic mode) so the engine will remain in valevtronic mode until the next phantom/real error arises or the ECU decides it's time to check throttled mode... but then you're back to the latch. The only way to prevent this is to have the LPG system capable and setup/calibrated to provide correct fuelling regardless of throttled/valvetronic mode.

It might swap back to Valvetronic if there are no OBD errors but the petrol ECU will want to see some of it's own behind the scenes tests passed before it will. Some OBD live data tools don't allow you to see the state of any of the tests (state as in not yet tested / testing / failed / passed), some allow you to see some of the tests, some allow you to see all the tests. There will be some different tests for throttled mode / partial valvetronic mode / full valvetronic mode - it could pass all the throttled mode tests so switch to semi/full valvetronic but fail the next set of tests so switch back to throttled mode. But rather than overthinking it or splashing out on an expensive OBD tool, if you get everything right on petrol it should switch to Valvetronic as normal. Then if you want it to run on LPG properly you just have to ensure that the LPG system fuelling matches petrol system fuelling under all conditions... but I expect this will be the problem bit given your setup.

Simon