V2Rocket

Did some more playing around today, still stumped, but did get it into the 0.9x lambda range idling.
I put a gauge together to check fuel pressure and FPR is in spec still at idle and under throttle blips.

Played with misc settings, played with reqfuel, played with VE table. Most current best-running config is attached, you will note the VE table is quite odd.
The left half is where I was playing because I am only focused on idle right now. The right half is untouched but gives you an idea the range where the left numbers started.
I had to reduce all the VE numbers to 1/4 of their original value (75% decrease) with minor adjustments afterwards, to get a steady idle in the 0.9x range.

Also, despite my reqfuel equation calculated value of 11 for gasoline (I was using 13 before since I have E85), I had to go down to 10 ms.

So all in all reqfuel is down ~30% and VE is down ~70% and I finally have lambda where it should be.
..
..
..
Bonus weirdness:
Since VE/reqfuel are somewhat inversely proportional to eachother in the AFR equation I also tried leaving VE alone and dropping reqfuel majorly - screenshot below, the car was idling very well like this, but would not restart again once shut off. Guess the pulsewidth is just too short like this.

Jim O.

I don't have any real experience with VEMS besides looking through their software, but I do have a lot of tuning experience with a number of standalones, so you might want to take this with a grain of salt...

Idling and start up are two different things. A lot of times you can get the idle right, but the car won't start - that's the role of the startup, afterstart, and warm up fueling parameters.

Looking at the screenshot, it looks like your injector pulse width is 1.2 msec. My guess is that you're around the minimum pulse width for those injectors at that fuel pressure and voltage. That's probably the reason that you can't make it any leaner at idle. Looking at the dead time values, it looks like the mechanics of these injectors are relatively slow to respond, so I wouldn't be surprised if you're well into the non-linear fuel delivery area of these injectors - you might have to get up past 1.5-2.0 msec until you reach the linear area.

I would recommend that you bring up the help screen for the "Injector Settings" dialog - there should be a procedure for refining the dead time settings for your injectors. Go through this procedure and see where you end up. Then, when the dead time is set correctly, try doing some data logging at idle, being sure to record lambda, injector pulsewidth, rpm, etc, while you play with your VE table. You can play that datalog through the "Injector Calibration" log file screen to see what is really going on. If you are at minimum pulse width, you may have to ignore idle lambda, and adjust reqfuel until the VE table makes sense everywhere else. Then trim VE at the idle area until you're either at minimum pulse width or at the desired lambda. VE at that area may look totally wrong, but if everywhere else makes sense, you might just have to live with it. Or, you might have to change to different injectors. Or you may have to play with fuel pressure. Or change to full-sequential fuel injection. Or???

V2Rocket

Thank you for your thoughts, it will give me something to think over.
You did help me to catch an error that I made that I cannot remember why I did it consciously; I took the published dead time values for these injectors at 3 and 3.5 bar to extrapolate for 2.5 bar, but I actually somehow ended up getting values for 3.25 bar ????

I will re-do that calculation and input the values and give it a shot and report back.

Raceboy

VEMS handles 944/951 ref sensor as secondary sensor (it also has 3rd, 4th and 5th sensor input).
Regarding fueling, it is just a tuning issue, if adjusting for new engine/injector setup, just start with reasonable config (where compensational tables have sane values) and no need to touch anything else before basics are set up (req_fuel+VE table). Also req_fuel is just a main fuel multiplier, it is not an absolute value so do not bother if it is few numbers off.
Regarding injector deadtimes, unless you know precise values for your injectors it is reasonable to keep the traditional approach (0, 400, 4000) and it works well.

V2Rocket

So I am using a blend of the base config I got with VEMS back in 2015 with much of the tuning done more recently.
But I think there may have been some kind of issue with my base setup the whole time I've owned it.
For example, reqfuel equation suggests a regular NA car with NA injectors should have a reqfuel around 17-18. But from 2015-2018 the car wouldn't even run until reqfuel was ~12. Then the VE table was screwy numbers but it ran well after some tinkering time.

So now with the bigger engine, injectors, and E85 I've been playing with it some more, driving around, checking lambda under various conditions, tuning, and doing this for hours. The VE table is still kind of weird (11-12 at idle but 80s-90s at high MAP) but lambda is in the ballpark most of the time. I haven't figured out logging yet as I don't even own any SD cards. I am fairly confident in my injector deadtime values now.

I think what I need now is a lot of time to drive around with auto-tune running.

..
..

Side note...Peep, is the VEMS ECU as you send it out built internally to work in a 944 application, or is the ECU a standard VEMS ECU with the EC18/36 pinouts routed to the right places on the DME plug? Could a 944-PnP ECU be used on a different, non-944 engine if the configs were changed?

Raceboy

944 uses VR-VR crank sensors, most cars use VR on crank and Hall on cam. You could use your VEMS on some other car with only crank sensor or modify VEMS for Hall secondary sensor.
You can still log without SD card Just log into laptop or your phone/tablet.

V2Rocket

Bump, still chasing my tail on this.

Looking through all possible configurations/tune areas and nothing is standing out.
Have to get the idle area VEs down to teens/low 20s to actually hit lambda target of 0.97 at idle...1.3-1.5ms pw.
Have taken reqfuel down to 5ms with only minor effect while running, lower than that car shuts off. Noticeably worse throttle response the lower reqfuel goes (blipping throttle), some hesitation.

These 42# injectors have similar deadtime values to the Siemens 80lb units that lots of folks have no trouble with, so I don't think it's that.

2.9L stroker, JME 8SR14 cam (so bigger than stock but not wild), E85 fuel (tuning in lambda, NOT AFR).

V2Rocket

Looking back into the early 2000s it looks like VEMS sort of branched off of Megasquirt development so I am assuming VEMS calculates PW similarly to MS...

PW = REQ_FUEL * VE * MAP * E + accel + Injector_open_time

(E = gamma_Enrich = (Warmup/100) * (O2_Closed Loop/100) * (AirCorr/100) * (BaroCorr/100))

http://www.megamanual.com/v22manual/mfuel.htm
..
..
..
At warm idle I have no warmup or other enrichment going on. No dTPS so shouldn't be any accel, and injector open time should be in the ballpark.

Droops83

Hey Spencer,

Looking at your screenshot above, the one figure that stands out is 24 degrees of spark advance at only 2% throttle opening and 1300 RPM---are you sure that the TPS is clocked properly and the throttle cable tension is good?

Also, it might be useful to completely disable the idle control when the engine is warm and see if the condition improves. I am at this point of tuning with my VEMS---it runs perfectly, except the idle RPM is high when warm, have never bothered to adjust it, I just disable idle control when smog is due and use the base idle setting via the adjustment screw . My next goal is to actually make the idle control work properly, if and when I have time!

V2Rocket

TPS works right, cable has some slack but it's closed at idle. It might be at 1 or 2 due to messing with throttle stop.
1300rpm is to give it some smoothness with the big camshaft and some room to "fall" when the monster Volvo radiator fan comes on.

Timing...this is something that has perplexed me too. IIRC stock 944 DME wants 8-12 degrees or so at idle, but even when I first got VEMS on a bone-stock NA engine it would barely run below 20 degrees at idle.
I've seen some other folks running similar...I have "Idle Control, Ign Advance" enabled as the ICV seems to struggle to keep idle up on its own.

Droops83

Interesting. In that case I would turn idle control off once the engine is warmed up and play around with the bypass screw and VE/base timing map settings to see if a reasonable configuration can be achieved (of course the radiator fan will mess with that). I know that if I actually bothered to delve into the idle control valve settings of my VEMS setup, it could be dialed in. The idling fuel requirements of a base 944 and a 944 Turbo are virtually identical, and even your 2.9L NA with a cam shouldn't be too far off.

Dave W.

This doesn't add up; you have 42# injectors on E85. The pulsewidth at idle should be much higher. Even on pump gas I'd expect the PW to be around 1.7 to 1.9 MS. Are you sure they're 42# injectors? Or are you running a 5 bar FPR?

V2Rocket

Yes, they are Bosch 0280155968 "green giants" ~42# give or take at 3 bar. I am running a stock 944 2.5 bar FPR, pressure verified to be working correctly with a gauge at idle and under load.

Yes, it *should* be higher. The calculated reqfuel value for my engine is supposed to be ~12ms on gasoline, add 30% or so for E85 and it would be 15-16?

Putting it much higher than 12 pegs the WB at full rich, it doesn't even read anymore.

So what's going on with a 1.5ms PW to make the thing idle at 0.75 lambda? WB02 was free-air calibrated a few months ago, and the exhaust smell is obviously quite rich and there is fuel in the oil.

DLS

Är you running traditional or simplified injector settings? Some have better luck with traditional.
Otherwise double check your deadtimes.
Maybe you have to pull some fuel off warmup table also.

The Req.fuel formula is not so accurate so you have to test yourself.

V2Rocket

Simplified, with input values specific to my injectors. Traditional did seem to make a difference though weirdly enough.

When car is fully warmed up there is no further enrichment going on but it is still running this rich. Today I will play with deadtimes and see what happens