Hi guys,

Perhaps you guys are able to help me out on this one.

Some background on my car:
- Reverent built MS3 basic
- AEM UEGO 30-0300
- Flow Force injectors
- GM IAT
- Otherwise stock 2005 EUDM MX5 (including returnless fuel system)

I'd been noticing that my VE auto tune values were different on different tuning runs. So yesterday, before tuning, I made sure my car was properly at operating temperature by driving it for an hour so first. Then, I tuned the VE with autotune. That gave me the tune that I've also attached. It idled pretty well with an EGO correction of about 105-106 (screenshot of log). However, a day later it starts filthy rich, so I set WUE to 100. The AFRs, even with WUE disabled are still very rich (11AFR) and it idles like ****. I drive around to get the car heated up - coolant at 90C (log attached). After the coolant is heated up, I take another look at the idle and see that it is decent with an ego correction of about 90 (screenshot of log). So that's a 15% ego correction difference between an engine that's been running for 1 hour+ and one that has been running for a few minutes (with about the same CLT).

My thinking is that this is related to injector heat soak. Would that be able to explain the 15% difference?

If so, what are my options? I was thinking:
- Replace the fuel system by a return style system (might cool the injectors a bit, and the lowered pressure might help as increased pulse widths reduce the impact of deadtime differences)
- Stop crying and go for <100% WUE to get it running decently during warmup, as a 10% correction with warm coolant / cold injectors is still acceptable

Any insights are greatly appreciated.

 

Autotune takes intake temperatures into account. It's crap either way. I would start manually adjusting the fuel cells. Interpolation is a great tool. Make sure you tune with the IAT's you usually drive at.

 

I have a similar problem, but with my ID1000 injectors. I also believe they are heat soaking and messing with resistances causing the injector to squirt a different PW than what is commanded from the computer. The issue is much worse with hot starts. It will crank and struggle to fire, normally takes me 2-4 attempts before the car runs. Once it does it's very lean (~17.0-18.0 AFR) and EGO has to add a lot of fuel to bring it down to 14.7 AFR. EGO correction becomes very minimal after a minute of running the engine. I assume it's because the fuel cools down the injectors. This is just my theory. I have tried many different things with no success. What I did as a band aid is add ASE and length of time that it's active when my coolant temps are above 170 F that way during a hot start the ASE helps with lean AFRs until the injectors cool down a bit and EGO takes over.

Sorry if I hijacked your thread, but I thought I'd add my experience with possible injector heat soak.

-David

 

What FW on MS3? On you cold start, the EGO is doing some strange stuff. Moving down and trying to control AFR, then jumping up to 100%. I don't understand that action at all.

I don't think your problem is the injectors.

Also, these logs are showing idle at up to 70 kPa. That seems not possible.

 

I kept the intake temperatures below 40C (or about 105F) consistently. Didn't tune idle while stationary but by depressing clutch while moving. I'm pretty sure the VE tables are ballpark accurate for a (through-and-through) warm engine

No worries. More data the better. I think my tune is relatively robust for hot starts now as the VE tables are based on injectors that are probably about the same temperature as the coolant. Downside being that tuning for that means that cold engines run rich.

Thanks so much for taking a better look than I did. I was focusing so much on the AFR's that I must've been missing some more. The idle was very rough. Thinking more about it, I think the following was going on during the cold start:
- With an already rich VE table, ASE added a ton more after cranking
- One or more of the spark plugs got soaked in fuel, not consistently firing leading to the rough idle
- When the spark plug doesn't fire, the fuel just gets sent to the exhaust, giving insanely high average AFR - even though it might not be all that bad in all cylinders.
- To compensate for this, the idle cold loop opens the idle valve even more - giving a pretty high MAP value
- In the meantime, EGO is making changes to fueling based on misfires - which results in weird stuff. Not sure why it resets now and then, but I guess that's not all that relevant if you are having misfires.

I'll install the Magnecor wires and new spark plugs I have lying around and reduce all fuel adders (WUE/ASE) for the time being.

However, despite all this - when the engine is warmed up it does run correctly on all 4 cylinders. So the heat soak difference is still about 15% between hot and cold. Would you say that's still manageable for EGO?

 

Rambling thoughts:

Raw fuel due to no burn in a cylinder will show lean, not rich due to un-consumed O2, which is what is actually measured by an O2 sensor. Maybe that is what you meant by "insanely high AFR" but I only saw rich, per my memory.

Yes, EGO can handle a swing of 15% (you have 44% spread now, assuming your fueling table was set at the midpoint of actual fueling needs). But the EGO is not controlling your AFR. Outside of true hot restarts, you should not need that much, in my experience (after, say, 2 minutes of running).

My experience with converting from Non-return to Return was a MINOR change in hot restart conditions.

You are not getting more fuel with the same called for pulse widths. That is what cold injector coils would give you. Equal electrical pulses, different fuel amounts. IIRC, your cold start is sending much longer electrical pulses, than the steady state hot ones. The injectors are behaving properly and giving more fuel. That is why I think the issue is not injector, but rather control system.

Have you verified your MAP sensor settings? Although you may be right, that running of fewer cylinders could require a higher MAP. IDK?

I am not sure what your issue(s) are, but was hoping my input might spark some ideas from other members.

What FW version are you running on the MS3?

Maybe misfires are giving you AFR readings that are outside your set range for EGO control, but I don't recall seeing any that bad.

Also, I don't know anything about AEM AFR controllers.

 

One more thing. How did you load your msq and msl files? They downloaded better than any I have seen in a year. Maybe IB has fixed things? Seems so unlikely...

Thanks, DNM

 

First of all, thanks so much for taking the time to think about this. I just changed the spark plugs + leads and you are definitely on to something.

The ecu is reverent MS3, which is based on firmware 1.4.0. Not sure if I can go to the latest firmware with his hardware or whether the firmware is customized for the hardware. In regards to your other question: I've just uploaded the files as part of the forum built in uploader.

I've added three new logs:
- Cold to almost warm with same settings as before
- Ve corrected to make the AFR's about 14-15 - still kind of cold
- Ve corrected - warm

What seems most weird, I've outlined in this screenshot. When I press the throttle, the MAP goes down and RPM goes up. Then, when I release the throttle MAP goes up and RPM goes down. I don't really have an explanation for this. Perhaps I misunderstand. Do you have any idea?

The effect also seems to be there when warmed up, with about correct VE's. Although it seems the effects are smaller during warmup. In this screenshot you can see that TPS goes up, leading to increased MAP and RPM (makes sense). However, then MAP actually lowers below the idle MAP, while keeping the revs up high.

What I'll do at least is replace my vacuum line to the MAP- perhaps it's acting strange. Currently it's a silicone hose but I also have some braided rubber around. I'll let you know if that changes anything. Furthermore you point out it might be the control system - which I guess means the Megasquirt in this specific case. I'll reach out to Reverent to ask if he can supply me with a clean firmware. Perhaps one of the flashes went badly and something is in some corrupt state. Does that seem logical?

 

I won't be able to look at any of your logs until tomorrow evening.
1) 1.4.0 is good FW, and Rev preferred. His MS3-Basics use standard FW. I was fearful of 1.5.X.
2) By control system, I mean MS along with all of the sensors, and the FW, and the settings.
3) Take a picture, or describe where you have your MAP sensor line connected. Pic would be best.

 

Hey man. Good news! Due to your prodding in the right direction I was thinking last night, "What if the MAP sensor is right - could there be a reason why the MAP is high at low revs and low at high revs". I was thinking that with a very advanced VVT-cam, that might be the case. So i set the lowest column to -10 degrees, and it suddenly idled fine - also see the attached log.

I'll try to figure out what's going on with the VVT and retune my VE, and post back here if there's any actual heat soak issues... :-)

Thanks again. If you're ever in Amsterdam, let me know and I'll buy you a beer or two.

 

one of these days, we'll start incorporating the solenoid on the fuel rail into the MS -- so on hot starts you can remove vacuum from the FPR and increase fueling...

 

Quick update: Tried some things with VVT, got weird every time I tried it. Left it off for the time being. Noticed that as I was driving for a longer time with autotune, cells turned blue more and more. Still wondering if there's not also something else going on (heatsoak?). I'll try to make a log of the next cold start. In case anyone is interested, I've added the latest tune.

 

Cold start was entirely within 5% ego correction - seems like this is a VVT issue and not injector soak. Hot starts are a little lean, but the core of my issue was something else - as DNMakinson noticed straight away.

 

Did you get the issue fixed?
i seem to have similar issue- fueling goes off with heat or smth. Different tune in different days but same coolant temp.