holy driver

So my car can hold an idle and all, and for most of this time I've had it tuned to 13.3:1AFR at idle (as DIY instructs) and the MAP can hold at 29-31kPa. However, I would occasionally get misfire under certain conditions, but I haven't quite figured out what the conditions are. I think it's when the CLT and AIT are hot, but I never really pinned it down.

I decided to try messing with it a bit this weekend and found that if I try to set the AFR to stoich (the way the stock ECU does), then I get consistent misfire at idle, and if I richen it all the way to 11.0:1 the idle is actually very smooth and gets to 28kPa. All of this is at 17deg timing in the idle cells. I tried fiddling with the timing in both directions with no real results.

For the last few days I've been idling at 11:1, but I'd really like to have a leaner idle if possible for fuel economy and eco-guilt reasons. Other than that, is there any reason idling that rich is bad? I guess it'll soot up my bumper? Is it likely to damage my wideband O2 sensor or my catalytic converter (high flow 2.5" metal core)?

Setup is as follows:
stock '95 motor
new NGK BKR7E plugs gapped at .035"
new NGK blue plug wires
MSPnP 94/95 w/ 10g HiRes and closed loop idle
Innovate LC-1
BEGI Stage 2 (GT2554R)
RX-7 460cc red top injectors

I know there's at least one member on this board (is it paul?) who runs a similarly rich idle and is comfortable with it... but I seem to recall the stock ECU idling just fine at stoich before I went MS.

hustler

iTrader: (6)

good grounds on the coil? Swap coils and see what happens.

Alta_Racer

iTrader: (4)

Is it possible you have a little too much bump in the timing?

ckkcmx5

I have the same situation here. It happens after hot restarting the car. Especially when IAT is high. Everything run fine if I start the car in cold, even until it run hot.

Actually the lean AFR can be tuned out with the "coolant related air density settings". But the AFR will be unstable and running stumble even if I have a good AFR (like misfiring) after hot restart.

I have tested unplugging the IAT sensor to start the car and re-plugged it. Then situtation improved.

I start suspecting if the MS take the initial IAT value to do any other calculation. Anyone have the same experience or could anyone explain this please?

My setup is silmilar as well:
rebuilt '94 motor with FM 9:1 piston
NGK BKR7E plugs gapped at .032"
MSPnP 94/95 w/ 10g HiRes and closed loop idle
AEM UEGO
RX-7 460cc injectors
FFS Supercharger

I have good ground on coil, MS and o2 sensor. Although I'm suspecting if I have a weak coil but it run fine when starting in cold.

evank

iTrader: (2)

First: it's IAT, not AIT.

I had a long IM chat with AbeFM about this same topic earlier tonight. My car, and many others, have the same problem.

The usual problem (grossly simplified): IAT sensor gets too hot, misreports an increased temperature to the MS, so MS pulls fuel, AFR goes too lean, causes the misfire.

One way to deal with this is to move the IAT sensor someplace where it doesn't get as much hot air from the engine, radiator, etc. .... for example move it as close as possible to the start of the intercooler outlet pipe. That's what worked for Paul.

Other people say the answer is to use MS's correction table. But that table is a bad design because it needs to be changed based on the weather, otherwise you run too rich (i.e., you tel it to add more fuel when the actual temp is X and the perceived temp is X+1, but what happens when the actual temp is X+1?)

On my car, despite moving the sensor down by the IC outlet, I still have this problem. I suspect, but I don't know for 100% certainty, that the heat is coming from the radiator. So I might try to build some kind of a sheild / shroud for the passenger side of the radiator, thereby preventing the heat from reaching the IAT sensor ... in theory.

Sadly, I think many people just say, "deal with it" and/or keep changing their correction tables.

Also: search before posting, this has been discussed before .....

evank

iTrader: (2)

It's also possible that heatsoak comes from the metal in which the IAT sensor is physically mounted, i.e., putting a metal-encased sensor into a hot metal pipe = the sensor picks up heat. Right?

So, maybe an answer is to devise a heat-resistant stub to go between the pipe and the sensor.

holy driver

The problem I'm describing is not IAT heatsoak. I do have that problem but have already compensated for it with the CLT related air density dialog and closed loop EGO correction.

Uncompensated, IAT heatsoak causes visible leaning of the AFR on the wideband gauge. What I'm describing is misfire even when the wideband reports the correct target AFR.

That is, using EGO correction at idle, I can set my target idle AFR to 13.3, and see 13.0-13.5 on the gauge, and still get occasional misfire. Or I can set it to 11, see 11, and get no misfire.... or set it to 14.7, see 14.7 and get misfire consistently.

evank

iTrader: (2)

I must be nearing insanity, replying to my own reply ....

Google informed me of the patently obvious: good insulating materials include epoxy, fiberglass, polycarbonate (good old "lexan"), PVC, and silicone rubber.

So I might try drilling a hole in the pipe, 1/2-inch wider diameter than the sensor thread, then securely cover the hole with one of the above materials and re-thread it. Or life might be much easier by simply chopping a whole slice from the pipe and replacing that slice with silicone rubber tubing, that way there's no awkward-shaped patch to mount; then thread the sensor into that part of the hose (or use PVC instead of silicone; PVC is easier to thread I presume...)

What do you guys think? Crazy or ingenious?

holy driver

I believe the actual sensor element is already insulated from the metal based on the construction of the sensor. Anyway, like I said, that problem can easily be solved by using EGO correction in addition to the table, as long as you trust your WBO2 sensor.

evank

iTrader: (2)

But perhaps not enough.

I plead ignorance to EGO. I assume you DON'T mean when people are always telling me, "Evan, you really need to correct your ego."

holy driver

The problem of sensor heatsoak is as you described... the sensor gets too hot and throws off the readings. It reports the air as hotter than it actually is, hence the calculation for the number of molecules of oxygen is lower than the number of molecules of air entering the combustion chamber. Thus the ECU doesn't supply enough fuel for the actual quantity of air, causing a lean condition.

Your AEM UEGO WBO2 mentioned in your sig will report this lean condition. If you've hooked up a wideband signal to the megasquirt then the megasquirt can make a correction based on the reading. It will see that the O2 reading is learner than your preset target and add more fuel to compensate (within a given authority that you've set). I believe some people insist on running an open loop idle (not IAC, they keep that closed loop, but AFR wise), but I'm not sure why that is.

evank

iTrader: (2)

I don't pretend to understand that sentence.

But .... I spent the past hour-plus reading as many IAT / heatsoak threads as I could find here, and also on msefi.com .... I think the most intelligent comment is what Braineack wrote in post #9 here: https://www.miataturbo.net/forum/t24976/#post297555 (for those who don't feel like clicking, he said: "If it were me, I'd start corrections closer to 90*F ... I don't see any reason to have corrections active doing non-heatsoaked conditions.") .... they don't call him "Brain" for nothing!

Is there any reason why his idea does NOT make sense? I assume (yeah I know, people shouldn't assume) that when the IAT sensor gets heatsoaked, it's by a very wide margin, not by 10 or 15 measly degrees. Meanwhile, here in New Jersey, I'm not doing much driving in temps above 100(-ish) degrees outside anyhow. So the worst that could happen is the car will run slightly too rich when I'm driving in what's locally considered sweltering weather ... a small price to pay in return for avoiding heatsoaking in the vast majority of conditions.

There is a bucket full of precedent for me being oblivious to obvious errors in logic. But right now, at 2:42 AM on Sunday night, this seems like the best "compromise" answer to the problem.

Looking forward to the replies!! I hate having to wait until morning when the sane people are awake again.

SKMetalworks

iTrader: (7)

sounds good. Once i get my cars exhaust done i think i might have this problem. Would WI prevent this problem? Or does that just get sprayed when you hit your boost threshold?

evank

iTrader: (2)

We're talking about something that happens at idle, so no, I don't think WI would help ... unless there's some way to instantly water-cool the sensor when it heatsoaks! Coming from a computer background, I find watercooling to be quite exciting. Unfortunately the Megasquirt is not a Cray supercomputer.

SKMetalworks

iTrader: (7)

So are you going to correct the temps at 90? Or are you waiting for somebody else whose up at 2 am ( your time ) It only gets in the 90's a couple days here in Seattle. I Cant wait to run into the next batch of problems associated with my car. A never ending quest that feeds your adrenaline withdraws

evank

iTrader: (2)

I'm going to wait for one of the resident Megasquirt gurus to tell me, "Yes, that's a good idea," or "No, you dumbass" before I change anything.

Also it's 3:30 now and my neighbors would be very unhappy if I start the car .... 'tis loud.

Instead, I'm off to hunt elusive wild ZZZs.

holy driver

Welp, I guess this thread has been successfully derailed.

Evan, do exactly what Braineack says in the post you linked and you'll be fine.

Anyway, my problem is what he alludes to in that same post... I want my car to idle at stoich!