2GR-FE 07 Camry

[Rokassileika]

Tomorrow is a past now, but really didn't think I could have so much trouble fitting these injectors .. most of the trouble was trying to find some metal pieces and cutting, fitting it, modeling, drilling holes so the injector fit an exact position from my visions. Anyway, one of the hardest parts is almost done, Just a few more fixings and it should be ready to calibrate. The #2 AEB injector is an obvious culprit now that it is dissasembled and I see the nozzle was dark and flooded with LPG

I came up with a few questions while installing:

1. Are these rubber bumpers under the rails a mandatory? I only have 2 and I would need 4 of them. (The 2 I had has been installed) Do vibrations do anything bad for such injectors?

2. The engine itself is working as a (-). And all these metal fittings touching the Barracudas steel body. Is it going to interrupt functionality of injectors? The AEBs were plastic.

Thanks!

[LPGC]

Would everything fit better if you made the pipe between the 2 rails shorter?

Rubber 'bumpers' are not necessary and it won't affect the operation of the injectors but, as you've found out, if the injectors touch something solid it can act as a sound board so you'll hear the injectors ticking away louder. If you have any more of that rubber pipe spare you could cut strips to put between the injectors and engine, if you make that pipe between rails shorter you'll have a bit spare anyway. The AEB's are a rather quiet injector, the Barracuda's are a bit louder but are nowhere near as loud as some injectors.

Are the injector nozzles correct size?

[Rokassileika]

I had to make the pipe the exact lenght it is now, since the nozzle pipes would be too short to reach and there is too little space between the inlet manifold and wiring harness body so tried to fit them as low as possible (bonnet would probably touch them if they were higher i think). The "V" form that is seen there also is made for hose connection under the PCV hose, there is very little space as well I did have a lot of trouble finding the 2.7 mm drill, but I have it in my hands today, so going to drill the nozzles and finish the installation. Is it bad if the hose is a little longer than it could be? Or no difference?

[LPGC]

The hose length won't make much difference, I was only talking about making it physically fit.

I'm sure you already know it'll need pipe clamps on that pipe or it will definitely leak and you need to remove the nozzles to drill them.

[Rokassileika]

Okay, so the mechanical part is probably finished, if there are no leaks that i'm going to check for later. While installing secondary filter I had to cut the hose a little bit, and nearly cut a spiral that i believe is mounted to see the LPG temperatures. I had no choice but to push the filter with the spiral a little, I hope this modification issue will not mix up real temperature. Would like to hear some insights about what's good or bad with my installation

[Rokassileika]

Everything went really good I would say- no leaks, engine is working much more stable, on highest rpms and load the gas pressure didn't go below 1.05 or so. And there was a setting for minimum injector opening time- it was set to 0 ms. The photo I added is from the calibration process. I'm a little confused. The blue line finishes at ~13.4 ms. And on almost WOT while driving on LPG the dot just hangs in the yellow circle zone that I painted. I predict that it is running too rich by the given blue dots on the map at such load (pinj should be from 10 to 12 ms and with LPG it goes down to ~9.8. But then again- we have a lower load situation where the pinj should be 8-12 ms. But in that load on LPG pinj goes from 7 to 9.2 ms. The first idea that comes to my mind is RPM correction. Would it be right to do so? I still don't understand what values should be entered in RPM correction screen and where. And how the heck a maximum load has lower pinj numbers than other loads? And maybe its the nozzle size issue? Or maybe I can call it an issue when the injectors will start overlapping and all I have to try now is set the multiplier scene?

[LPGC]

To answer your post #45, OK it looks a bit untidy but pipes are not kinked and it's working OK, there's nothing technically any worse than when you had the AEB's fitted.

Your post #46...

The reason the dot doesn't move at full throttle is because it's running in open loop mode at full throttle, in open loop mode regardless of how lean or rich the mixture is the petrol ECU won't increase or decrease pinj to compensate for mixture.

When you're driving around on petrol the map plots blue points in positions that relate to manifold pressure versus pinj. When you're driving around on LPG the map plots green points in positions that relate to manifold pressure versus pinj. Note that it is pinj that is plotted in both cases, it doesn't plot ginj when you're running on LPG, it still plots pinj. The idea is that if for a given manifold pressure pinj is higher when you're running on LPG than when you're running on petrol then the plotted green points will be higher on the chart than the plotted blue points, so if all other things are the same (especially rpm) then wherever you see green points on the map above blue points it could mean that LPg calibration is lean at that position (at that pinj point) on the map. Obviously the opposite is also true, so if plotted green points are lower than plotted blue points it could be that the multiplier at that point on the map is too rich (the multiplier at that point on the map is too high). We/I call this method of calibration 'the join the dots method', it can be useful but it isn't totally accurate even if the blue and green plotted points at any given manifold pressure were created at the same rpm. It becomes even less accurate if plotted points were created at different rpms because at any given manifold pressure the engine has different volumertic efficiency at various rpms, i.e. the engine has an rpm at which it produces most torque and is 'on cam', at that point volumetric efficiency is highest, which means it takes bigger gulps of air into the cylinders and if it takes bigger gulps or air it will also need more fuel for mixture to remain correct. I do like to see the green dots on top of the blue dots but 'the proof is in the pudding' and it is more important to get fuel trims correct than to join the dots.

I would dial in some rpm correction before starting calibration. Don't worry about setting different numbers for different loads, just set rpm correction for any load to be zero at idle rpm rising to around +1% at 1500rpm, +2% at 2000rpm, +3% at 2500rpm, +4% at 3000rpm, etc, so you should have +10% at 6000rpm with +1% increases at every 500rpm. If your rpm correction settings are a table then at any given rpm you'll have dialled in the same percentage for all loads at any given rpm. If you don't dial in rpm correction then you could find that during some test drives when you're only reaching (say) 3000rpm but at higher engine loads mixture is correct but then if you later put your foot down mixture will go lean at that same pinj. Of course it is also possible to get it wrong and add too much rpm correction, in which case it could go too rich at high rpm, but the figures I've mentioned are typically about right and you can always fine tune rpm correction later.

I would also slightly adjust temp compensation and manifold pressure compensation before doing calibration, where possible. Iirc Stag300 doesn't feature manifold pressure correction but it does feature gas vapour temp correction. In the gas vapour temp correction screen the standard straight line doesn't mean that no temp correction is added/subtracted, any adjustments to the straight line are adjustments to the standard temp correction. The standard temp correction tends to be excessive so subtracts too much percentage for cold temps and adds too much percentage for high temps. I would adjust the line adding an adjustment point at 0degC and adding 2% there, add an adjustment point at 90degC and subtracting 2% there. Some of the later (box etc) Stag ECU's do allow manifold pressure correction, on an engine that features a returnless fuel system or constant petrol pressure at the fuel rail (not an old-skool mechanical pressure regulator on the petrol rail) an LPG ECU's ability to compensate for manifold presssure can be useful, not for compensating for manifold pressure but for indirectly compensating for effective differences in petrol pressure... An engine that has constant petrol pressure of (say) 4bar at the fuel rail has effective (rail minus manifold) 4 bar petrol pressure at the rail only when the manifold is at atmospheric pressure (either when the engine is off or when the driver has their foot to the floor on the accelerator), at part loads when the manifold is under vacuum the effective petrol pressure is that 4bar plus the manifold vacuum, so at idle effective petrol pressure would be around 4.7bar. The amount of fuel an injector will flow depends on it's own physcial properties and also on effective fuel pressure, any given injector has it's flow effected by changes in pressure and the flow rate for a given pressure can calculated because flow depends on the square root of the pressure. If you double the presure flow won't increase by 2X it will only increase by the square root of 2 (pythagoras constant) which is around 1.414 times. But the difference in flow between fuel pressure at 4.8 bar (engine during over-run) and 4 bar (engine at full load) works out to be around 8%. If the LPG ECU allows setting compensation for manifold pressure then it allows setting for compensation of effective fuel pressure and we can 'dial in' perentage compensation figures for effective different fuel pressures. The petrol ECU map is already set up to account for effective changing fuel pressure at different manifold pressures, the petrol ECU will account for the 4.7bar effective fuel pressure at idle and part loads by decreasing pinj, while the LPG ECU uses pinj as primary input on which it bases ginj. A petrol pulse length of X milliseconds will deliver a different amount of fuel to the engine depending on manifold pressure, while the LPG ECU by default always delivers the same amount of LPG for a given pinj pulse length. If an LPG ECU doesn't feature manifold pressure correction we can still in a way account for pinj meaning a different amount of fuel will be delivered at different manifold pressures by adjusting the multiplier map and/or rpm correction but it won't be as good a way as accounting directly for effective fuel pressure by dialling in manifold pressure correction, the shape of the map won't be as smooth (it will be more peaky around the low load area) and although it can still deliver good results the map/rpm correction method isn't as good during transient/changing conditions such as if/when the driver suddenlt changes throttle position because during those conditions the petrol ECU might intend on giving (say) an immediate 20% increase in fuelling in anticipation of increasing airflow into the engine but the LPG ECU will see that immediate 20% increase in pinj and calculate a new ginj to try to mirror that 20% increase but since the manifold pressure has not yet had time to increase the new ginj won't totally mirror the 20% extra fuelling that the petrol ECU map intended, the ginj fuelling could be as far momentraily incorrect as +-8% depending on whether the driver/engine was accelerating/decelerating. But don't worry about the compensation for manifold/effective fuel pressure lol, if your engine doesn't have a fuel return you'll just map it as normal and the map will seem to have more of a peak at the low load area. Do adjust temp correction like I said if possible though.

Also don't forget that if your engine/ECU goes open loop when you put your foot down the only way to calibrate open loop mixture is by watching pre-cat lambda readings (or being extremely familiar with mapping systems and able to accurately guestimate what the multiplier should be around peak inj based on the shape of the map from low to mid loads).

You may find that when you've added rpm correction the green and blue dots become closer anyway but add the rpm correction before going much further with calibration, and use an OBD scan tool to check fuel trims between running on petrol and running on LPG during calibration. Also remember that if you run on LPG for a while and the engine doesn't have a fuel return the petrol that is sat in the fuel rails will heat up, which reduces it's density, which means when you next run on petrol the fuel trims on petrol will be artificially slightly higher than they'd normally be if the petrol was being constantly used by the engine helping keep the petrol in the fuel rails a bit cooler. The opposite can happen if the engine is still warning up because under those conditions warm petrol evaporates better than cold petrol so in some situations less petrol is needed than if cold fuel were in the fuel rail of the mildly warmed up engine, so fuel trims on petrol might be artificially more negative under those conditions.

What size nozzles did you fit and with the multipliers you have already set in calibration, when the engine is running at 8ms pinj what is ginj? What is ginj when the engine is fully warmed up and idling?

[Rokassileika]

I fit (drilled original to) 2.7 mm nozzles. I don't have the car with me, but i made a screenshot after the latest calibrations (calibrated without rpm corrections) at 8 pinj on lpg it falls to 7.2 pinj, can't tell about ginj at the moment. Ginj is about 4.1-4.3 at 600 rpms as i notice in the last photo (the laptop died). I made a small example table about how the RPM corrections should look on my Stag 300-6. Could You tell me If I'm understanding this correct? As well with temperature correction screen. Thanks!

[LPGC]

Your temp correction is the wrong way around, I meant a positive correction at 0degC and a negative correction at 95degC. Cold gas is denser so actually needs a negative correction and vice/versa but the Stag temp correction chart doesn't show actual correction, it shows user corrections to underlying corrections that the ECU applies even if the user doesn't enter any corrections. The underlying corrections are usually a bit excessive at both ends of the scale, too much of a negative correction at low temps and too much positive correction at high temps, so by adding a positive correction for low temps and a negative correction at high temps we make the temperature correction back into line. Stag doesn't display the temp correction percentage it is applying but we can assume it uses corrections that have been calculated according to the laws of physics and gas density at given temperatures so on face value the standard temp correction it applies should be accurate, at least for an ideal injector for which temperature doesn't affect the performance of the injector. The gotcha here is that injectors are not ideal injectors, the performance is affected by temperature too, they might take a bit more or less time to open/close at varying temperatures even to the extent that the changing injector response affects mixture to a greater extent than the gas density change does...at least at low ginj's, but correct calibration at low ginj's is more critical than correct calibration at higher ginj's. There is also another reason why it's better to have a setup that is slightly richer with cold gas than with warm gas - if driving for long periods at high engine loads the gas temperature usually falls, it is better that the mixture gets a little richer under those conditions than a little leaner because we don't want the mixture to lean off during continued high engine loads.

You have the correct idea about the rpm corrections.

If I were calibrating the setup I would have multiplier correction points (the yellow dots) at 2, 2.5, 3, 3.5, 4, 5, 6, 8, 10 and 16ms, you need finer control (hence these points closer together) nearer the low load and idle area of the map than you do nearer the high load area of the map.

When you have entered the temp corrections and rpm corrections you should clear the LPG plotted points and go for another drive. As usual, switch between petrol and LPG during the drive whilst watching fuel trims. While you're driving on LPG the ECU will plot new green gas points on the multiplier map. If green plotted points are higher than blue plotted points it usually points to the map being lean (or vice/versa) at that point of the map, in which case you should increase the multiplier (yellow dot you added to the map) at that point (or vice/versa) but always check fuel trims between petrol and LPG at that point in the map before making any changes. **

When you've got calibration about right with the components set up as they are (nozzle sizes) we'll look at what multipliers were necessary on your map (and the related ginj versus pinj) to see whether or not it would be a good idea to increase nozzle size and recalibrate.

** In addition to comparing positions on blue and green plotted points on your multiplier map your Stag ECU might allow you to see a 'deviation table', a deviation table shows the difference in pinj between running on petrol and running on LPG with RPM on one axis and manifold pressure on the other axis so you can see in table form where it is leaner/richer running on LPG than on petrol. Obviously such table can only show information for combinations of rpm and manifold pressure that it has learned while you were driving the car, so there will only be entries in the table for combinations of RPM and manifold pressure where you have driven the car on both petrol and LPG. There might also be a table showing which areas it has learned while you were driving on petrol and another table showing which areas it has learned while you were driving on LPG.

[Rokassileika]

Small update. While trying to figure out how to change RPM Correction numbers (CTRL+ arrow keys) my laptop battery was draining quickly and, unfortunately, I couldn't make all of the changes- like adjusting yellow dot locations on the main map. But I managed to modify temperature corrections and RPM corrections. Took a screenshot of the lpg petrol maps as well. The green line on the map was closer to the blue one from the last time I had it screenshoted. I dont know if its a placebo effect or something else, but the engine power is now consistant, after corrections the engine feels smoother quieter and more powerful even with the fact that the only thing i was in time to do with this old dying laptop was temperature and rpm corrections. Will check the map on saturday again, change yellow dot locations and perform the final multiplier corrections while keeping constant RPM and looking out for STFTs and LTFTs. Do I have o keep the same RPM and change the engine load while switching from petrol to gas or will it be ok if RPMs fluctuate after such corrections? And for the last things.. That moment when engine works in open loop at WOT. It might sound stupid, but RPMS and load is high will I not damage the engine while switching from petrol to gas if the multiplier is off? Adding screenshots. P.S. why my installer didn't perform any corrections in the first place? The multiplier map looked like it was just an autocalibration performed.. How many installers do it that way lol? .. By the way with the right click on rpm correction table there was two options- turn off or on auto- adaptations. But even with it on I didn't notice any adaptations getting done.

[LPGC]

You haven't yet adjusted the multiplier but the rpm corrections have richened the mixture, made the blue and green lines closer together and narrowed the spread of green dots to be closer to the spread of blue dots. The RPM and temperature corrections were a good move. It won't really matter that you didn't extend the rpm corrections past 16ms because petrol injectors are likely never pulsed for longer than 16ms but it would still be best to extend them when your laptop is charged.

Looking at your latest map pic it seems OK until around 5.5ms then a bit lean after 5.5ms, which is what I would expect with the nozzles you're running and the shape of your map. The downward angle of the slope of yellow multiplier dots from 6ms onward should be less steep (more horizontal). I reckon if you raise the 6ms dot up just one click, raise the 12ms dot up to 1.4 and position the 9 and 10.5ms dots so the line is almost straight between 6ms and 12ms then the blue and green lines will line up more closely. From 12ms to 25.5ms increase the heights of the yellow dots so the line between them is almost horizontal.

However, having 1.4 multiplier at 13ms and 16ms means the map will calculate ginj to be be 18.2ms when pinj is only 13ms and at 6000rpm the +10% rpm correction would add an extra 1.8ms, so calculated ginj would be 19.8ms even if pressure doesn't fall. There is a risk of running out of window (period of opportunity for injection between engine intake strokes) at high rpm and high loads but we knew this when going for the 2.7mm nozzles and it was a thought out decision to go for the 2.7 nozzles at risk of running out of window because your reducer probably can't flow enough gas for the engine at full power anyway and because the smaller nozzles will make calibration a little easier for you (there won't need to be as much of a peak around the low load area compared to the rest of the map)... But if your gas pressure seems not to drop much at full throttle at (say) 5500rpm with rich mixture (look at lambda readings) then it might be a good idea to increase nozzle size and recalibrate again because if the reducer pressure seems OK then there's a good chance it would remain OK for when the engine is making it's full power.

Most installers don't add any rpm correction or adjust temp correction, I am the only one as far as I know to compensate for petrol pressure. Most installers do the minimum that seems to work and are not as fussy as me about correct calibration or achieving best driveability, most don't even ensure mixture is correct under high load open loop operation. Some rely too much on autocal / self tuning / self adaption facilities built into some ECU's but no such self tuning facility can do as good a job of setting fuelling correctly as doing a proper manual job of calibration.

I didn't say to activate any auto-adaption facility, if you've activated some of those then they could be offsetting results in the map screen. But you could show us info related to the self tuning facility you enabled and data showed in those could help me advise on adjustments to make to the multiplier. After making any changes I advise though you should clear the adaptive values and disable the facility at least until the manual calibration is completed.

[Rokassileika]

Looking at your latest map pic it seems OK until around 5.5ms then a bit lean after 5.5ms, which is what I would expect with the nozzles you're running and the shape of your map. The downward angle of the slope of yellow multiplier dots from 6ms onward should be less steep (more horizontal). I reckon if you raise the 6ms dot up just one click, raise the 12ms dot up to 1.4 and position the 9 and 10.5ms dots so the line is almost straight between 6ms and 12ms then the blue and green lines will line up more closely. From 12ms to 25.5ms increase the heights of the yellow dots so the line between them is almost horizontal.

- I'm a little confused. I thought that if the green dots are more to the left vs the blue ones it should mean that the mixture is a bit too rich and in that case we should lower the multiplier values?

Today i had time for some experiments, it did not go well, but I tried to lower the percentage of fuel enrichment at 2k RPMs and 9 ms. What I noticed was that on 2-2.5k rpms and 9ms the green dot would get a little too far to the left from the blue line and anything above 2-2.5k RPMs would seem normal, the green dot gets back near the blue petrol line. After such corrections and some map tweaking the map looks really off now. I have a save-file from old configurations so going to reupload them and retry everything without playing with rpm corrections. But before that I'm really confused if I understand the lines of petrol and gas maps correctly. To make sure if im on the right track- on higher loads and on LPG the pinj is lower than petrol so I should lower the multiplier, no?

Thanks, adding my failed experiment screenshots

[LPGC]

I'm a little confused. I thought that if the green dots are more to the left vs the blue ones it should mean that the mixture is a bit too rich and in that case we should lower the multiplier values?

Sorry, yes you're right, I was thinking of a different type of screen on a different system.

Still you seem to have realised that and have improved your map already

Don't forget though that when it's in open loop mode the position of plotted points won't change regardless of where you set the multiplier (height of yellow dots). If you adjust multiplier points in the open loop range you directly adjust mixture, so you need to set multiplier points in the open loop range by monitoring lambda readings.

[Rokassileika]

Reversed everything (my experiments), and corrected some multiplier points on high loads. Seems ok so far. I just noticed there is a thing called "Pressure correction". I was wondering what if i corrected something on lower pressure side like leaning it a bit? Since i have a little rich running issue on high loads? The lambda readings look fine tho, not sure if I should go any further. Just that situation - 2k RPMs and on 9ms (higher loads) it kinda feels like it's almost the only point that distorts the curve. I tried to correct it on the RPM correction table but it only got worse. It messed up low load area as well. What if I left 0's on ~2k rpms from the 9 ms point, but left the 2-3's on 0.9-8 ms? It kinda feels like it's not the right thing to do so

[LPGC]

If the setting is just a tick box tick it. If the setting allows you to change compensation settings for various LPG pressures it's usually best to use the default settings. This isn't the same as compensating for relative petrol pressure as discussed earlier. The default settings are usually correct because they are set to agree with the laws of physics - for a given injector open time if pressure is doubled flow increases by the square root of 2. That said, higher pressure does increase the time it takes an injector to open slightly but if you correct for that the correction will only be correct for a certain pulse duration and will make mixture incorrect at longer pulse durations.

Yes you can leave 0's in rpm correction for 2000rpm 9ms. But first check you're not confusing a point at which it switches to open loop fuelling.

[Rokassileika]

Sorry, forgot to answer about auto-adaptation facility. It is found only in RPM correction screen. It works on single numbers from the given table with a right mouse click i can turn it off, the numbers go gray, or turn it on and that selected number goes black again. My other question. I did notice that on very high loads the STFT goes 0 aka open loop mode if I'm correct. But how do I correct the multiplier if I can't see the STFTs? And besides all of this I can finally say that the problem was in injectors and tuning the mixtures. I've been driving the car for days now, and the power is very very consistant and strong, consumption has improved significantly. I just can't believe how I had been driving it with such low and random power for 5 years. The installer could not do anything about it from day 1. And by the biggest help of Yours (I can confidently call You an LPG scientist with all this meticulous knowledge with hundreds of questions and answers in my posts it is nearly fixed (nearly because i'm having some final touches getting done soon). I learned a lot and I'm willing to learn more later. I'm amazed to find such dedicated people at their thing knowing it is 2026 and what most people care now is how to rip you off and get away asap. Well at least, I would say, 95% of companys, individuals do that in my country. World needs such good people

[Rokassileika]

Okay, this is really really getting out of control. The lines match perfectly, but the inconsistancy is getting back .. I connected diag (DS150E Autocom) on idle LTFT on one bank -18, on other -16... It looks like the inconsistancy begins when the weather changes. One day I re-check the maps p and g., lets say i make it perfect. On lpg switch there is no feeling of vibrations or anything- butter smooth switch between fuels. STFT does not fluctuate at all. The other day I switch from petrol to LPG in same conditions- STFTs are off, engine vibrates for a second after switching fuel types.. I did some research and mostly from what i have read that it simply could be the MAF sensor. That would explain why on warm/ hot weather the car mostly drives very well and strong. I mean maybe the sensor thinks that the engine gets more air than it actually is getting or a temperature sensor on the MAF is out of spec. What do You think about it?

[LPGC]

At high loads when it's running in open loop mode you set the multiplier so the lambda readings show suitably rich mixture.

I doubt the MAF is at fault. More likely some other condition is different, such as temperatures like you mentioned. No doubt when you were spending some time switching between fuels and setting the multiplier the engine had been running for some time and temperatures would be what we'd expect for a fully warmed up engine.. We're talking temperature of the engine itself, petrol feeding the petrol fuel rail and petrol injectors, LPG pressure reducer, LPG vapour temp sensor, LPG injectors, etc. There is a point during engine warmup when the petrol system is running in closed loop mode but the engine temperature and petrol temperature haven't yet warmed up to normal operation temperatures, at this point the petrol injectors may need to pulse for some extra duration to compensate for some percentage of the petrol not actually burning in the engine because it doesn't fully atomise - the engine is running in closed loop mode but the petrol ECU is still applying some 'choke' effect (the extra pulse duration). LPG is a vapour when it is injected and doesn't need any choke effect, if all other things are around the same (such as correction for temperatures / pressures / fuel density, injectors response, etc) then any 'choke effect' that is applied to petrol injector pulse duration will cause mixture to be richer when running on LPG. Or in other words if 10% of injected petrol wouldn't burn inside the engine because the engine or the petrol is too cold or petrol droplet size is too large then the petrol ECU will need to inject 10% more petrol to account for the 10% of injected petrol that the engine cannot burn, but all injected LPG can burn because it is injected as a vapour not in liquid form, the vapour doesn't have a droplet size.. So since the LPG system used petrol injector pulse duration as primary input, it it reads petrol injection has increased by 10% it will cause LPG dose to increase by 10%, when running on petrol the 10% increase would correct what would otherwise be a lean mixture on petrol, when running on LPG the 10% increase makes what would otherwise be a correct mixture on LPG a 10% too rich mixture on LPG. This is probably the biggest cause of the effect you're seeing but there can be other causes too, such as during this 'choke' condition engine load is also higher because engine tolerances are tighter and during the 'choke effect' the engine ECU might hold rpm higher than idle rpm for a fully warmed up engine, the temperature compensation factors also need to be correct, the LPG injectors may respond slightly differently when they're cold compared to response when fully warmed up. The petrol ECU may be using a different set of fuel trims for the warmup conditions. The response of LPG injectors can change slightly from being brand new and newly fitted to having been used for a few hours or even after being newly fitted then sat overnight pressurised with gas. If you've got the shape of the map correct and compensation factors (temperature and pressure) and if there is no choke effect being applied then sometimes different fuel trims during warmup can be an indicator that the shape of your map is wrong around the transition area between warm idle no load and just a bit higher loads.. Say if the engine usually warm idles with 3ms pinj but during warmup from cold conditions after choke effect has been completely turned off but the engine is still a little tight idle pinj might be 4ms, you might not often pay much attention to the 4ms point during calibration but during the warmup stage it might hand around 4ms for quite a while.

If you run the car up to normal temperatures do fuel trims seem to come back to what they were when you set calbration?

[Rokassileika]

Long fuel trims keep pretty much the same. Today it was -15 and -11 LTFTs when I was about to start the car. But lets say if the car fully warms up, I can see that on idle the STFTs fluctuate whenever I change it from petrol to lpg. Sometimes it goes positive, sometimes negative. No matter how I set the multiplier at that point. What really got my eye today- the calculated engine load. At idle, in "Park" it was 37%. Lights off as everything else of course. While I was driving it ~25 MPH and barely touching the gas pedal at 1.9k RPM it was showing me 60-80% engine load. And I was on a straight road. Getting engine load to 100% i dont even need the WOT. Some people say it could be vac leak, I doubt it, since the LTFTs would go positive in that case? The LTFTs stayed about the same from cold to hot on gasoline or on lpg it didn't matter much. What else I noticed the IAT was showing +8°C and the Ambient temp was actually +5°C (before starting the car). My first question was is this really normal to see such high engine loads without even pressing the gas that much? It didn't get better when the car was at OT. 37% load at idle was measured in OT. But it is not the first time I notice worse performance in colder outside temperatures, but I wouldn't believe it could do such a big impact on the power the engine is creating. Maybe I should try to recalibrate the LPG map again? Or try and collect a new petrol map when the engine is fully warmed up and then continue on all other procedures?

[LPGC]

Are you reading engine load from LPG software or OBD? If it's from LPG software it probably points to something wrong (such as RPM reading etc). If it's from OBD it's probably just a case of your OBD reader interpreting the load data stream incorrectly or looking at the wrong data stream... unless by load it's showing manifold pressure...? Those readings mihgt be around correct for manifold pressure (map) - Map is like the engine load at a given rpm, while usually LPG software and OBD data streams show engine load as something closer to pinj x rpm but this is more of an abstract value.

When you first switch between fuels fuel trims will fluctuate a bit, it's turning off one set of injectors whilst turning on the other set of injectors, there are things that can be done to smooth the transission but you can reasonably expect some fuel trim fluctuation around that point. Especially when switching from petrol to LPG if it's run for a while on petrol before that point the lines between LPG injetors and inlet manifold will contain only air (not LPG) so if it switches immediately from petrol to LPG we might expect mixture to be momentarily a bit lean, but with some LPG ECU's we can counter this in various ways such as dialling in some crossover time during which petrol injectors and LPG injectors both fire momentarily at the same time, enough time to prime the gas lines between injectors and manifold. If such crossover time is set though, then if the lines are already primed with LPG (likely if the vehicle has recently run on LPG and engine load hasn't changed so manifold pressure hasn't changed) the fact that both petrol injectors and LPG injectors are delivering fuel at the same time might see mixture flick momentarily rich rather than lean. Crossover time is most beneficial when LPG injector lines are a bit long, if injectors were mounted directly on the manifold the lines would be zero length and a zero length pipe takes zero priming time.

If you mean trims are different even after this switchover period it could be because aspects like temperature readings don't respond to changes immediately, the temp sensors take time to change to the actual gas temperature themselves. Or because the LPG injectors respond slightly differently when first used than a little later after they've warmed up a bit. Or because pressure from the reducer is likely to be a bit higher at changeover than a little later. Or because the shape of the map is wrong (as said earlier). There are a lot of variables that can affect how consistently the LPG system provides same fuelling as the petrol system across the full range of pressures, temperatures, rpms, manifold pressures, etc.. But it should be possible to dial everything in so the fuelling is the same on LPG as on petrol across every combination of all those different ranges. This is the art hehe!