LPG Discussion Forum

The RPM correction figures are in addition to the multiplier.

The figures I suggested were solely rpm based but the rpm table is really a rpm x load table with 13 rows of rpm and 26 columns of pinj, so it is possible to set 338 individual adjustments for combinations of rpm x load.

Instead of using rpm correction to make fuelling correct it is a better idea to get the shape of the multiplier correct (to have fuelling as correct as possible without rpm correction). As injector pulse frequency increases the pulse length normally needs increasing for the injector to deliver the same amount of fuel for a given pulse length compared to lower rpms, which is why I suggested the rpm compensation figures I did. They won't be absolutely correct but will be a good approximation of near correct rpm (injector pulse frequency) compensation.

The learned (ISA2 automatically entered) figures in your screenshot suggest that the shape of the multiplier map isn't quite correct because the figures are all positive, they suggest that if the multiplier is increased the figures in rpm correction could be lower.

If I were calibrating this from scratch I'd use the maximum number of user settable multiplier points allowed (10 points I defined) and I'd start by setting those points at 1.5ms, 2.5, 3.0, 3.5, 4.0, 4.5, 6.0, 8.0, 10.0, 12.0 and 16.0, (the points at 0ms and 26ms are always present). You can add the extra points by right clicking on screen or delete a point by selecting it and pressing delete. With the info you've provided I'd suggest starting with a multiplier map that looks like the one I've posted below and setting the rpm correction back to how I advised. You could leave ISA2 enabled to see how it changes rpm correction figures with the new multiplier map... Also remember that if your petrol map was collected (blue petrol dots and line) when you had the lambda probe issues the blue points and line will be incorrect in which case you should click on 'delete petrol map' and drive around on petrol for a while until the petrol map is collected again - only then switch to gas and see how the dots/lines/rpm correction figures compare.



I don't remember whether your engine has a mechanical fuel return (petrol pressure regulator on the fuel rail with a pressure bleed pipe running back to the petrol tank)? If it doesn't have a petrol pressure regulator on the fuel rail I will suggest entering some different figures in rpm correction and a different shape map.

If you drive the car flat out on petrol what is the highest pinj figure you see? Maybe ideally the install could benefit from some quicker injectors or bigger nozzle size and lower pressure but depending on some of the answers if you keep with the current injectors (you can't change their speed or increase because speed is due to the design of them and you can't increase nozzle size because the V30s have a 2.5mm internal restriction) the install might benefit from increasing pressure... Yet to be confirmed anyway. Speed of injectors (how quick to open) mostly affects the shape of the multiplier chart at low pinj's (this mostly affects where the peak point of the chart is, quicker injectors move the peak in the map shape to the left), nozzle size affects multiplier more than pressure at low pinj points but pressure does affect low pinj points, combination of nozzle size and pressure affects the whole map. If injectors are too slow or nozzle size isn't big enough the multiplier at low pinj's can be excessive and this can be a negative for accuracy of fuelling particularly when accounting for temperature and pressure differences. By changing aspects such as make/type of injectors, nozzle size, pressure (and by compensating by for any differences in petrol pressure when engine load changes if a mechanical fuel return isn't fitted) we can manipulate the overall slope (mult/gamma) of the map (difference of mutliplier between peak point and lowest point), the 'lead in' (upward slope of the map at very low loads), the shape of the curve of the map and overall height (average multiplier) of the map.

That is an awesome amount of information, thank you! Helps me understand things a lot better and explains other things I have read on the forums but didn't understand why.

I erased the old petrol map and collected from new after changing oxygen sensor, over a range of driving conditions

This car does have a fuel return line from pressure regulator. Was surprised reading on the forums here that some cars don't have this.

Maximum petrol injection time-I did push the car a bit but I'm hesitant to go flogging the hell out of it too much and giving myself more things to fix and rebuild (how we change as we get older hey!). Maximum I got was just above the longest duration on the most recent image, maybe around 12ms going up a hill, lower gear, foot flat to the floor and trying to get it to rev out more. Maybe got to around 4000rpm. I'm guessing fully loaded or towing a trailer might get me higher.

I'll pop out one of the injector rails later on and check what size I drilled them out to. I've got some oil leaking from rocker cover gasket so got to get in there at that time regardless.

Thanks again, really appreciate your willingness to help me and others out on here, and giving what I imagine is a fair amount of time typing up responses (and I've noticed some people don't even bother coming back to let you know how they have gone or to even thank you).

Injector nozzles are at 2.2mm currently. That was what I thought they were but for the life of couldn't think how I drilled them to that size, only having 2 and 2.5 in all my drill bits. Dug through an old box of dremel bits and there was the 2.2 still with bits of brass on it, from the only time it's been used

In your first post you asked

disco stu wrote: Sun Mar 07, 2021 2:56 am Looking at the open times for petrol and gas injectors on the upper right hand side of images. This is taken at idle, and petrol is at around 3.4ms and LPG is at 5.9. I thought LPG open time was the petrol time multiplied the orange multiplier line. Which in this case would be 3.4 x 1.25 = 4.25ms, except the actual time is well over that. So what is the multiplier actually doing if it isn't multiplying the LPG injector time? I think that is what keeps confusing me about adjusting that line-if I knew exactly what it was doing I would understand what to do with it.
Stuart

You're right that on most systems ginj (gas injector pulse time) = (pinj x multiplier) + any corrections for pressure and temperatures. So you'd end up with 3.4 x 1.25 = 4.25.
But some systems are a bit different in that ginj = (pinj x multiiplier) + any corrections for pressure and temperatures + offset. In a way offset is supposed to account for the opening/closing time of the LPG injectors. Behind the scenes the offset being applied for the V30 injectors is probably around 1.2ms. If we bear this in mind and do the maths we have ginj = (3.4 x 1.25) + 1.2 = 5.45 which is still above what we'd expect but we haven't yet accounted for pressure and temperature. In your screenshots gas pressure less manifold pressure is 1.29 bar so is very similar to the 1.28 bar reference pressure set in software, so it won't be adjusting ginj for pressure... But your gas temp reading is showing 75C and it will be adding a few percent to ginj to account for the high temperature gas being less dense than cooler gas. So the maths may be 3.4 x 1.25 = 4.25, 4.25+10% = 4.675, 4.675 + 1.2 = almost 5.9.

Nevertheless the gas injectors need to pulse for 5.9ms to deliver enough fuel to the engine for idle (when pinj is 3.4ms) and the 5.9ms is too high.... At the other end of the engine load range (under high engine loads at high rpms) the injectors having to pulse for 5.9ms at idle will probably mean the injectors have to pulse for longer than 'window' (the amount of time available between engine 4 stroke cycles for the injectors to open, flow gas and close again in time for the next 4 stroke cycle). But we don't want ginj to fall too low either... The V30's are not a very quick injector and don't start to meter gas properly until they're pulsed for above around 4ms and we've got to remember that when gas temp is cold it will be more dense so the ginj will be lower still. The reducer may also give a bit more pressure when it is cold which will further reducer necessary ginj. And ginj is lower when the engine is at fast idle than at 700rpm idle. I'd expect to see ginj somewhere between 4 and 5ms at warm idle with these injectors on this engine with nozzles and pressure properly set up.

Depending on which of the V6's you have, the 3.0 or 3.5 and the bhp rating (anywhere between maybe 160bhp and 230+bhp) I would likely drill nozzles to 2.5mm. I would have fitted at least 2.5 nozzles by default if I were doing this as a new install... But then again I wouldn't have fitted the V30 injectors because I know they don't like sub 4ms pulse duration. Since you'll want to use the V30's we have to be careful not push ginj too low for this type of injector. What's the spec of your engine?

Another thing is that you were having problems with lambda sensors etc at the the start of the thread... so your original map may have been wrong (perhaps with the lambda issues fixed you'll find gas injectors don't have to pulse for 5.9 at idle?).

Just noticed you've written it's a 3.5 in the thread title.. I would definitely drill nozzles to 2.5mm.

Awesome-thanks!

Really helps me understand the why of it all, which is always what I like.

Engine produces something around 210hp. It's actually out of a Magna/verada/diamante like car, as pajero engines are so much more expensive here. Fits in fine with a few modifications, like mount holes drilled and tapped, plate fitted for cam sensor. Supposed to produce more hp than pajero engine, but it's got the pajero intake on it. I've been trying to work out the compression ratio, but can't work that out, and pajero engine produces 20hp less so I'm assuming it would be lower.

Anyway, will pop all the nozzles out and drill out to 2.5. I resisted the urge to do it yesterday, thought it best to wait for your opinion. I was thinking it would be required given the long idle times but was also conscious of Valtek 30 opening times that I've seen mentioned on here plenty.

These injectors came with the kit when I bought it second hand unused. That guy was going to put it on an Audi v8 but changed his mind for some reason. If I ever replace injectors I'll be sure to check specs closely

So this is where I'm at currently. Drilled out the nozzles to 2.5mm, ran auto calibration again for base settings. Tried it on autoadaptation and it kept changing the rpm correction map, not the multiplier. Turned off autoadaptation and wiped all of its changes, put it back to what LPGC recommended.

I've now just been adjusting multiplier manually based on the join the dots line of LPG vs petrol. It will get close, then on a drive it will drop back down in those lean parts a bit, but its getting a lot closer than previously.

I did try getting my daughter to help me with watching injection times on petrol then changing to lpg, but she is only young so will get "it just changed!" when I put my foot down and haven't switched between petrol and LPG, and its hard finding a piece of straight quiet road with a hill or similar to put various loads on the car to compare, without some other variable coming to change things. I figure the map drawn is based on the same times so should give me a decent way of getting it close to optimum.

Autocal won't change the multiplier, will only change RPM correction table. You were right to set the table as I said and change the multiplier.
Edit - initial autocal will change the multiplier but ISA2 etc (autocal 'as you drive') will only change the RPM correction table.

The multiplier shouldn't go up between 10ms and 12ms, it should be a slight decline all the way across to 25ms... If the multiplier doesn't slope down enough at the engine's high load range it will likely run too rich at high engine loads. Under high load conditions the plotted points won't change regardless of where you set the multiplier because the engine will be running in open loop mode (which is why autocal cannot set correct mixture at high engine loads). The way to set multiplier for high load open loop conditions is to watch lambda sensor voltage.

Don't forget that plotted points are not just related to manifold pressure versus pinj/ginj, the height of plotted points will change for RPM also - Pinj/ginj are how much fuel the engine needs to match the amount of air it is using, for a given manifold pressure it will use a different amount of air at different rpms.. for any given manifold pressure the height of plotted points will be highest when the engine is most 'on cam' (when it's volumetric efficiency is highest / when it's using the most air per 4 stroke cycle). On some systems you can limit the range of rpms under which plotted points are collected, this helps to make sure the installer is comparing apples with apples rather than comparing points that were plotted at different rpms.

When the engine is under load but still running closed loop (at say 8ms pinj at 2500rpm), what are ginj and the pressure readings?

Thanks.

I upped the multiplier in 10-12ms as it was showing lean there. Its now showing rich. I'll put it back to where it was. With what you're saying about open loop multiplier, are you saying that we should ignore what the petrol injection time is in open loop and set the LPG multiplier so the gas time in that range is at stoichiometric rather than running rich like it would in petrol?

I'll try going for a drive with my son who is a bit more literate with technology and try and get some more info.

Thanks again

The join the dots method doesn't work in open loop mode because pinj will stay the same regardless of mixture, the petrol ECU doesn't try to correct mixture in open loop mode.

You still want mixture to go rich on LPG under high load open loop conditions but not quite as rich as it goes on petrol. If the vehicle has a narrow band lambda sensor the mixture will probably be rich enough on petrol that the lambda voltage signal is maxed out. On LPG aim for just a little lower voltage than the voltage on petrol but make sure it never goes lean under sustained high load conditions (when gas temperature falls and the ECU reduces ginj to compensate for the lower gas temps) and never goes lean under open loop not quite full throttle conditions.

Beauty, thanks

Haven't looked at high load yet, just trying to get mixture right in the area up to 8ms (majority of driving). I keep bumping it up about 2 clicks at a time, but it keeps showing lean to around the same amount. Is there some sort of guide like you should move the multiplier around the same amount as the gap between petrol and LPG lines or anything similar?

It has started becoming jumpy again in the range of 1000-2000ish rpm when cruising in lower gear and I change throttle position. I'll see if it improves once things get closer, but it wasn't happening at all when LPG was a bit leaner than currently

Heh- I got my threads crossed and replied thinking this was the Jag thread doh! I'll come back to this later...

There's a long answer to your 'is there some sort of guide' question, on some systems there's kind of a way of knowing roughly how far you should move the multiplier but the best bet usually is to move it a bit if there's a small gap between petrol and LPG plotted points and move it further if there's a bigger gap - as you've been doing.

If there's too much difference between pinj and ginj it can cause 'jumpiness', especially if ginj is smaller than pinj.

When pinj reads 3ms, 4ms, 5ms, 6ms, 7ms, 8ms, what does ginj read? Same question for both cylinder banks.

How long are the pipes between injectors and manifold? Where in the manifold are your points of gas entry? Have you set extra injection filtering? Have you got routing correct (if you switch each cylinder in turn back to petrol does the engine continue to run smoothly or do you get a misfire on 2 cylinders)?

Cheers. I'm assuming my injection times will be the same for both banks, engine only has one oxygen sensor after the pipes come together so I'm assuming it doesn't have a way of altering either bank separate from the other.

I'm best going for a drive while recording. This is assuming it records the whole screen rather than just the oscilloscope section. Then I can tell you more info on comparison at different times.

Can't remember hose lengths exactly, but I think it was about 25cm. Shortest I could get it due to only place I could install injectors without going horizontal. Manifold nozzles are just above the petrol injectors, maybe a cm or so. When you say extra injection filtering, I've got a filter after reducer if that's what you mean. Routing is all correct, checked again the other day after drilling out injector nozzles

Extra injection filtering would be a setting in the software. Some cars inject more than once per cycle on a single cylinder on petrol. The extra injection will be very short -too short to do on gas. The filtering would be to remove those from the map. I think its likely to be a tick box type option?

Thanks Brian. "Extra injection switch off threshold" is the option I have. Changed it to 0.9ms on instructions here from LPGC early on. I wasn't even sure what it was, so that explains it. Thanks for that

So, I'm holding off getting lpg fully dialed as I wanted to run on petrol for a bit to compare consumption now O2 sensor is working, plus get this timing advance fitted. Pretty sure I went through a full tank of lpg since sorting o2, and consumption was still around 18l/100km, maybe a touch under.

Just checked petrol consumption and it was 12.4l/100km, so that has dropped fairly significantly now (was 14.5).

That difference is much higher than expected. When I first got this on the road it was around 15% higher on lpg, now it's more like 70%. Maybe I have to go through another tank to make sure, but at this difference it's not worth lpg.

If something is majorly wrong what could it be? I've double checked all fittings, tightened things up and mixtures I thought were better than previously

I wouldn't further complicate matters by fitting the TAP at this stage, first I would get the LPG system calibrated / dialled in.

Cheers. It was pretty close to being sorted, as far as multiplier at least. I'm thinking I'll run another tank of lpg before being convinced that is the lpg fuel consumption as well. I'll try and get it tweaked on the drive home from work tomorrow.

Quite happy with the petrol consumption now. That was pretty much bang on what I was hoping to get. Just a few litres more per hundred than the Camry, for a much larger car

I just took a 6min recording driving home from beach, up the hill to my place. I wanted to trim the video to just get a few sections to look at but it wouldn't let me, so its full length, sorry. Start of the hill is at around 4min. If you would prefer I can grab a few screen shots from various places in the vid to compare things. A riveting watch for anyone else who happens along on this video!

http://www.youtube.com/watch?v=hsqiTCkK69M

Only if you're willing to watch it. I know its an ask, you guys are generous enough with your time and expertise. Not sure what's going on with the gas line above 10ms. Yesterday it was slightly high, so I moved it down in that area by 1 click only at the 12 and 16ms spots. Now it shows way low.

I also realised I need to take a screen recording where I switch regularly between LPG and petrol while at the same load so I can compare petrol times on each fuel. Hindsight

A few readings from various places to compare. Tried to grab these where it was approximately stable for a few frames, rather just on the verge of accelerating of decelerating sharply
Pinj------Ginj------RPM ------ Manifold pressure
7.1 ------- 10.3 ------- 1580 - ----- 0.60
4.7 ------ 7.3 ------ 1800 ------ 0.39
2.9 ------ 4.5 - ----- 760 ------ 0.26
6.8 ------ 10.2 - ----- 2060 ------ 0.53
10.0 ------ 12.7 ------ 2080 ------ 0.75
6.9 ------ 9.9 ------ 2280 ------ 0.53
8.5 ------ 11.6 ------ 2220 - ----- 0.65

Edit: sorry for weird format of the figures. Trying to separate the columns somehow so it's easier to read. Automatically removes the spaces I put in