1992 base model NA. DIYAutoTune MSPNP2 90-93. Stock 1.6L motor with GM IAT sensor, AFM removed. Flowforce 610 injectors. Wells 201 variable TPS. I've been tuning since March and I'm getting much better. Still running base spark map. VE table is tuned pretty well, although I'm still fine tuning idle cells and WUE's.

My attached log is with the car up to operating temperature, idling in the garage. When I step on the throttle, my TPSdot usually goes to 350-800. My AFR goes to full lean (22.4) about 0.5 seconds afterwards, then goes super rich (10's) at about 1 second. If you look at my time-based TPSdot AE's, you'll see that the FF610 injectors do not leave much room for negotiation. I'm at the lowest possible settings, 0.1-0.3ms. It seems to be adding an appropriate amount of fuel (it's actually overcorrecting, even at small pulsewidths), but the lag time from the TPS event until the AE correction is causing this oscillation.

There doesn't seem to be any setting to speed up the response time; only settings to affect the "tail" after the main AE event. Maybe MS3 has PID parameters here that my MS2 is lacking?

I have noticed that my DIY vTPS does not always reset to the same point at closed throttle, with about a +/- 2% error. There is probably some wiggle between the throttle shaft and the sensor. I'm thinking about squirting something in there to harden around the throttle shaft (modeling clay? drywall compound? dunno.) Nonetheless, it doesn't appear to be affecting my TPSdot readings or contributing to this issue.

Anyway, I could use some suggestions. Thanks

 

A pioneer of quantum mechanics can't fix his accel enrichments?
What firmware are you using?
This is a frustrating issue with no easy remedy, try to mix some mapdot in, try lowering your tps threshold a bit also.
I would say to enrich the cells you are hitting manually, but they seem rich already.

 

Any other ideas here?

 

Pics of cat?

 

We only have horses in my house.

 

Ah. For some reason I thought you had a cat.

I'll research TPSdot settings if they are still there when I look for them.

 

Lulz.

 

Found picture of my cat.

 

Don't know how helpful this will be, but here's what I learned wasting hours and hours on EAE:

forget tuning it per logs/numbers/data, and tune it by feel. does the car stumble? does it get rough? does it feel "wrong"?

it will always have that lean spike no matter what, it's a reactive system, there will always be a split second delay in fueling compensation. what I try to focus on now is how it compensates once it does compensate. if it dips like crazy to like 9 or 10 then obviously you're giving it way too much. if it stays lean for a while and hten tapers down then you're not adding enough.

after going bananas with it for days/weeks, I finally gave up. then I drove a stock Miata and realized that my ms settings were actually significantly smoother and nicer than OEM, and made the conclusion that I'm just obsessive and should spend time on more important things lol

 

I'm glad to hear that I'm not crazy, because I've basically arrived at the same place you have. Even with AE turned to minimum pulsewidths, I'm overshooting to AFR's in the 9's and 10's. I might as well just turn them off. There's a little hesitation coming onto the throttle, but I agree it's probably at least as good as stock.

I stuffed my crappy TPS/throttle shaft connection with epoxy putty the other day. If it works in taking out the wobble/ inaccuracy, I'm going to try a lower TPS threshold.

 

Just for giggles, try lowering the End Pulsewidth value and see what happens. What you have is telling the MS to set your PW to 2ms after the main enrichment happens. Lowering that value should reduce the over-rich condition you're experiencing (I'm using a value of 0ms). Also, you might want to consider raising the MAPdot Threshold value. If you're "stomp the go-pedal" MAPdot is greater than 300, then your value of 125 makes AE start happening at lower values, which means more fuel earlier - which seems NOT to be what you want.

 

Thanks Roger. I'll give that a try and report back.