Lag on auto gear changes, omega v6 with stag kit.

[chrisgixer]

Evening all, wonder if you guys can help with a common issue we are having with v6 autos over on omegaowners.com

We are using stag 300 kit which seems to work fine on 4 cylinder manuals, but fit it to an auto, mostly on the v6 engines, and we are experiencing a lag, or jolt, when the auto box changes gear on some changes, there is also a slight lag on lift off from wide open throttle on some cars (mine included) with drive by wire throttle on later cars.

We have played with various vaporiser pressures, from 0.9 bar to 1.3 bar and re mapping while driving. This has an effect but only succeeds in moving the lag up and down the rev range at it's minimal point. Best setting seems to be 1.1 bar which leaves the lag at the top of the range where it can be driven around by lifting the throttle before the lag meaning the auto box changes up prior to. Or the ecu can be set to change back to petrol at say 5k rpm so the lag is avoided by running on petrol at that point in the rev range.

There is no issue with lack of flow as both banks of Lambdas never show a lean condition using the stag software.

Also when the tune is optimized, there always seems to be a tailing off and another calibration and remap is needed to settle it down again, usually within 3 or 4 months.


We are a tight bunch and simply changing parts and hoping for an improvement is not on our remit without proof of some sort that the issue will be cured. It has Bern suggested that the 3ohms red valtec injectors are not accurate enough, and the latest upgrade to the kit gives "magic" (?) injectors instead, which may be a clue.

Does any one have experience of this issue or know of a kit that works well on a 2000 on 3.2 auto v6 with drive by wire throttle?


I have just acquired a nice low miles example and need to gas it before the vat hike comes in so it's a bit urgent sadly.

Here's hoping,

Chris. Ta.

[junkcatcher]

Don't know if this helps I suspect you will have already tried most of this but -----
On Rovers (without drive by-wire or vg inlet) with the Stag-4 I encountered a hesitation problem or lack of throttle response when opening the throttle suddenly from idle.

The cure included drilling the nozzles much bigger than recommended by AC Stag; official Stag recommendation was 2.1 to 2.3 we ended up using 2.75mm at 1.2 bar but if doing again I would try 3mm at 1 bar . This is probably explained by the engine using fully sequential injection.

Other tricks was to get two accurate points at the bottom end of the scale. ie. Hand adjusting to get the petrol injector durations the same on petrol and lpg at no load idle with absolutely every engine load switched off,
Then exactly doing the same but with every possible load switched on --- AC, main beam, fog lights, wipers, heated rear window, cd player, heated seats........

After that it was just a matter of fine tuning the multipiers further up the scale by reading the the engine own ECU fuel trims via the OBD socket and minimising the change in the longterm fuel trims between LPG and petrol.

Doing this I get manage to very very close Zero deflection for the curve from idle conditions upward, but I not 100% convinced how good the curve fitting feature of the Stag software is.

For manifold pressure below normal idle vacum I increased the lpg injection multiplier setting above that at idle setting.

It took a long while to get it (almost) right but with the above settings the car is actually much more pleasant to drive in the very low RPM ranges on LPG than it ever has been on petrol, this was advantageous in recent weather conditions.

The Valtek Type 30 3 ohm injectors are working quite happliy with multipliers from about 0.8 at idle to around 1.3 at full load the 6,700 rpm red line. Petrol injection durration at idle is around 2.9ms so that would make the lpg injector duration at idle roughly 2.3ms.

[rossko]

There is no issue with lack of flow as both banks of Lambdas never show a lean condition using the stag software.

They're probably not fast enough acting to give you any clues about what goes on during this brief period, either about what the petrol does or what the gas does wrong.

You might be seeing the effect of simple mechanical lag; throttle-off and manifold vacuum increases, increasing differential pressure across the injectors (so inceasing fuel delivery per mS under throttle-off conditions). The mechanical regulator will adapt to its new MAP reference but (a) that takes time and (b) perhaps more important, there is no way to dump this new "excess pressure" in the gas rail except by using it through the injectors. It'll take a few injections to bleed the pressure back to nominal; and all this is happening just when injector times are short (power off) or nil (if the autobox commands a cut-off to the engine management to make a slicker gearchange). Could that be why it shows up on autos?

In theory the gas system could compensate by means of its gas pressure sensor, in reality you don't know (I don't anyway) how fast the sensor responds, how often or quickly the gas software samples that pressure & MAP reading to correct delivery, or if there is any limit to how much correction it can apply.

I doubt the Stag is any better or worse than other systems at handling this but it might be! No experience with V6 auto meself.

Also when the tune is optimized, there always seems to be a tailing off and another calibration and remap is needed to settle it down again, usually within 3 or 4 months.

Probably the injectors bedding in, most likely the internal rubber bump stops squishing. Valtek (and many other) gas injectors are built like solenoid valves, not pintle designs like petrol injectors. Their responses change as the internal rubber bumpstops deform or deteriorate, as these are crucial to the delivery characteristics. If all is stable after 4 months, that's fine. If they keep drifting, the rubbers are their way to a sticky end...

I wouldn't expect "magic jet" brand injectors to be any better in this respect.

[chrisgixer]

Thanks for reply, can you recall the engine size on the Rover and rough bhp? And was it auto? Because we have nozzle size or 3.0mill for 3.2 with output of about 210 bhp. Although I am not aware of a pick up issue as you described with this kit/engine combination, by comparison the drilled nozzle size we have may not be correct perhaps? Poossibly bigger needed?


Would it be likely that the lag issue is a sign of a lean condition at that point in the rev range?


But certainly the problem is the other end of the rev range in purr case...?

[Kevin Wood]

Thanks for your responses so far.

I also have an Omega with the Stag 300 kit and have been working with chrisgixer and a couple of others to try to diagnose this problem.

junkcatcher:

The cure included drilling the nozzles much bigger than recommended by AC Stag; official Stag recommendation was 2.1 to 2.3 we ended up using 2.75mm at 1.2 bar but if doing again I would try 3mm at 1 bar . This is probably explained by the engine using fully sequential injection.

I think most of us are already using 3.0mm nozzles on the 3.2 V6s and have tried vapour pressure from 1.0 to 1.3 bar. I have monitored the petrol ECU fuel trims during driving and they don't wander around any more on LPG than they do on petrol. The Omega issue is only a wide open throttle, or closing the throttle from wide open, though, so the petrol ECU will be open loop at this point. Might be worth trying to get the lower parts of the map spot-on using the fuel trims and interpolate the shape of the curve into the higher load area, I suppose?

Petrol injection durration at idle is around 2.9ms so that would make the lpg injector duration at idle roughly 2.3ms.

I think we have a greater LPG duration than that but I'll have to double check. It did occur to me that these injectors are quite slow in terms of open/close time, and with only 20ms between engine cycles at 6,000 RPM that probably becomes significant.

rossko:

You might be seeing the effect of simple mechanical lag; throttle-off and manifold vacuum increases, increasing differential pressure across the injectors (so inceasing fuel delivery per mS under throttle-off conditions). The mechanical regulator will adapt to its new MAP reference but (a) that takes time and (b) perhaps more important, there is no way to dump this new "excess pressure" in the gas rail except by using it through the injectors. It'll take a few injections to bleed the pressure back to nominal; and all this is happening just when injector times are short (power off) or nil (if the autobox commands a cut-off to the engine management to make a slicker gearchange). Could that be why it shows up on autos?

This was my initial thought too but on the non drive-by-wire Omegas the (cable) throttle is evidently fully open through a gear change (and they still suffer lag) and I have observed that the DBW engines (via the OBDII throttle position data and also MAP during a gear change) also hold the throttle fully open during a shift, so the MAP should be at pretty much atmospheric pressure throughout.

The engines all pull well up to the red line, so I think fuelling at WOT is probably acceptable, just something goes wrong at the transient as you come off the throttle above about 5k RPM or the autobox changes up (Petrol ECU retards the ignition to reduce torque during the upshift). It's annoying because the logging from both the Stag ECU and the petrol ecu via a GM Tech2 both sample too slowly to spot anything amiss.

I do have a wideband lambda sensor which I could hook up to a car with a bit of work. That could perhaps be used to tune the mixture at WOT up to the red line. Maybe there's some mileage in trying that?

Kevin

[turbojo]

Hi, you might find the lag is down to reducer delay. You will notice a spike. Ideal pressure is 1.0 bar on most reducers.
Valteks don't really like an idle below 4ms, they cant react quick enough and you get a lumpy idle.
Also, the stag software will automatically up the multiplier if the gas pressure drops from the pressure specified in the settings page, so a poor reducer with a large pressure drop will cause an injector full open error even if the multi at that point is under 1:1.
I'm running a stag 300+ on a carlton gsi 3.0 24v which has been chipped, decatted, filter, exhaust and everything else you can think of and i'm only running 2.5mm nozzles at 1.0bar with a idle of 4.5ms on valteks 30's and hit 7k rpm with no leaning at all
I'd check your reducer pressure output and drop, you shouldn't need 2.9ms at idle.

[hardcorepit80]

Hi.
Possible problem is in red valtek's injectors since they been cheapest on market... they'r too slow for v6 engines.
Magic Jet or Matrix that's what You need.
Another thing , it's almost impossible for someone to solve the problem with any STAG installations without SCREENSHOTS after calibration or .osc file.
Post some screenshots and I try to help.

[turbojo]

There is nothing wrong with the Valtek's and they are not the cheapest on the market. Magicjets reliability is awfull especially if the car sees any dirty fuel. You don't need a 2ms injection time on a 3.0 v6. Im pushing 250bhp on the Carlton with valteks no problem with plenty of headroom and a silky smooth idle.

[lazydocker]

I'm one of the guys with the same issue and I'm visiting the supplier today to have a decent look at the problem.

We think it is, as turbojo says, reducer delay

[hardcorepit80]

There is nothing wrong with the Valtek's and they are not the cheapest on the market

They are cheapest everywhere but not in UK.
If You 're happy with them it means Your installation is set up perfectly, but most of the time running car on valtek's (red) its not easy...
All I say is: no valtek on your car, less problems with your installation

[turbojo]

It was nice to meet you!

[turbojo]

The copies are the cheapest and why would there be so many if they are no good? I'm not saying they are the best, not by any means. However, for the money they are good value and as far as I'm concerned Valteks are nice to setup. Awfull flow at idle, so nice high multiplier and idle duty time but improve as duty increases so you end up with a dropping multiplier. Gives a nice chunk of headroom topend and a good smooth idle. On both of my cars (other being a p38 rr 4.6 with a coscast block, 3.2mm jets @ 1bar) I start at a 1.1 multi and drop to 0.8. The only bug I've got with them is the need to flow test after a rebuild.

[hardcorepit80]

I start at a 1.1 multi and drop to 0.8.

Your jets are too big.Ideal multi between 1.2-1.4 adjusted by jets size.

[turbojo]

My petrol injectors are close to full cycle at 7k, anything over 1 actual will result in the lpg injectors being flat open so how can that work?

[lazydocker]

It was nice to meet you!

Ahh... I should have twigged really

Thanks for the help and I look forward to trying out the things we discussed. Perhaps I should have grabbed a timing processor while I was there for experimentation.

[turbojo]

Yeah, keep us updated with your tests.
The timing processor will help with a bit more power and/or more mpg depending how you drive, I wouldn't have thought it would do anything to help the lag.
However, saying that, after you left I had a think about it and possibly when the ignition is being retarded during gearchange maybe the mixture is being ignited so late in the cycle its not helping? Don't for get lpg takes longer to burn. When you have your tech2 session maybe keep an eye on the ignition timing during a change? If you let me know along with the rpm at the time I can work out when the main flame front hits.
I still think the injector spike is the root cause though.

[lpgelite]

I had a similar issue with my 2.5 Omega running the excellent ESGI II system.

I found I was getting what felt like fuel starvation at around 2400 rpm on moderate throttle at about the point the autobox would normally upshift on a light throttle.

I played around with pressure settings and got reasonable results with 1.3 bar and 2.5mm nozzles on the red Valtek injectors, but could not avoid this "flat spot" without consciously manipulating the gear changes with subtle throttle variations.

My previous Omega had the (also excellent) Dream XX1n system and never suffered this annoying characteristic with the same engine and 'box combo. The earlier car had covered 110k miles more but was in good nick and had bags of power with that system fitted, so I knew it was possible to achieve good driveability on LPG. I did wonder at one point if the current car's autobox ECU had a later software version that didn't suit the LPG system, but soon discounted this.

I have now discovered and cured the problem, which to be fair was evident throughout, if I had paid more attention to what the ESGI software was telling me from the start.

The reducer I used with the kit was a single stage AGC P 250 which on paper could supply enough gas for the engine's output plus a reasonable margin, however when I observed the pressure readings whilst my other half was driving I saw a noticeable dip in pressure around this mid throttle trouble spot as the reducer struggled to keep up with relatively sudden but sustained heavy demand. This dip in pressure lasted only a few seconds, but that was enough to lean the air / fuel ratio enough to perhaps deceive the autobox into thinking it was time to shift up, hence the perceived flat spot, which I now put down to the engine being instructed by the 'box ECU to "ease off" for a phantom upshift.

Lifting the throttle slightly reduced demand and the pressure then overshot its datum 1.3 bar for a few seconds until it had bled down the excess gas. I knew the hot water supply to the reducer was good as this was confirmed by the reducer's temperature sensor showing a healthy 78 - 84 degrees under load. The supply incidentally is tee'd from the bridge at the rear of the engne - the only suitable connection on these engines.

I was sure the variation in gas pressure was causing the problem so I did a bit of googling and discovered the Dream XX1n is a 2 stage reducer based on similar principles to its stable mate, the much respected R90E, but with the addition of a manifold vacuum balancing connection.

This, AFAIK, is the only 2 stage reducer available for SGI systems and is particularly noted for its stable pressure output, which would have accounted for the superior performance of the older car.

I managed to obtain one for just over £70 including VAT and shipping from http://www.autogas-lpg.co.uk and after recalibrating from scratch the problem has gone, plus there is no noticeable hesitation when blipping the throttle from idle. Fuel economy has improved by around 3mpg, too.

I put this down entirely to the new reducer's almost constant pressure output no matter what demand is placed upon it. This is a vastly underrated piece of kit and was worth every penny. Its 2 stage design is what I believe gives it such superior performance over single stage reducers. I also read that Zavoli's zeta reducers are noted for stability of their output, but opted for the Dream XX1n because of previous good experience and the 2 chamber design.

I have since noticed the ESGI's LPG map has completely synchronised with my datum petrol map when I viewed it after a few hundred miles of self learning, which I never managed to achieve with the single stage reducer. My engine pulls as smooth as a gas turbine now and there is no difference in the autobox's behaviour between gas and petrol.

Swapping the reducer has transformed the performance of the ESGI II system which IMHO outperforms any others I have owned, including Dream XX1n and, dare I say, the Koltec system factory (Millbrook) fitted to my Astra.

I would definitely recommend the combination of ESGI II and Dream XX1n reducer based on my own experiences.

[mat_fenwick]

That's interesting, as it appears to contradict what Rossko has written here (6th post down). I'm not disagreeing with either of you here, but I wonder if there are other factors coming into play?

I have the Dream vaporiser, and do notice a drop under high demand - however I don't think it is the vap at fault.

[lazydocker]

Funny you should mention the LPG Pressure overshoot... It's exactly what we have now pinpointed as the most likely cause of the problem

[rossko]

That's interesting, as it appears to contradict what Rossko has written

Can only report wot I see ; in comparison to other makes of closely related system (Bigas, Zavoli, etc.) the steadiness of the D21 pressure under demand changes seems pants to me; but of course it works.
I do not think differences will be software related e.g. software slugging sensor response times and hiding poor shows, as these are all AEB based software.
Apparent differences between the same vap used on AEB or Polish systems could be down to software behaviour, though.

Excessive pressure overshoot is indicative of gummed up vap, poor design, or inadequate rating. Some variation is inevitable, just as it is on-load, what's acceptable or not is a judgement call.