rwyatt365

Ok, end of March my trusty old CX Racing churbo bit the dust - the oil seal on the hot side failed and it started spewing oil like a sumbitch. I thought I could salvage it with a rebuild kit but I botched that job (long story, not relevant) so I ended up buying another EBay churbo (NOT CX Racing).

After tweaking several tables to fit the characteristics of the new turbo, everything was good until I went to a track day and discovered that I was backfiring when the MAP got above 190kPa (ish). Below that, everything was fine. Above that it was chitty-chitty-bang-bang. At first I thought it was because I was hitting the rev-limiter and the fuel cut was causing a lean condition. But I found that I could duplicate the backfiring on the street, well below the rev limit. It seemed to be when the MAP exceeds 190 kPa, and only then.

I poured through the setting to see if I had inadvertently changed something that I shouldn't have, but couldn't find anything. I also looked through the log to see if I could find anything that would correlate to the backfiring. Then I saw this;

I see that;
- Just before the backfire, the MAP is at 191 kPa
- and there is a mysterious spike in the "On/Off Outputs Status Bits" channel that correlated with the peak MAP value, which only seems to be present when MAP > 190 kPa

Just AFTER the peak MAP value, and the mysterious output status bit peak;
- The PW spikes
- The Duty Cycle spikes
- The RPMs fall off (this is the backfiring happening)
...and things don't recover until I get out of the throttle, then it's business as usual

I was hoping that I could find out what a value of "8" in the On/Off Output Status Bits channel meant, but I couldn't find an answer here, or on msextra.com. My thought was, "If I knew what that meant, I could find and fix the problem". No luck.

My car is a '99, with a built bottom end and the aformentioned churbo on a Kraken manifold. I'm running GM coils so I don't think they're breaking up. And I have a MSPNP2 controlling this. I've attached my MSQ and the log that the screen shot came from (there's a mark at the backfire event).

Any suggestions are welcome. (happy hunting!)

rwyatt365

Ok...my quest to figure this out continues.

I'm seeing that at the point just before where the backfire/misfire occurs, the EAE value spikes low. I don't know if this is the cause of the problem, or a result of the problem. So, when EAE peaks low, it "pulls down" the PW, and results in the AFR spiking lean. I still don't have a reason for any of this, just trying to correlate the data.

Mudflap

google: tunerstudio status1. document opens, search "status1"

8=RPM fully synced

if you are seeing that twitch around and lose sync you've got something mechanically wrong. Like your cam or crank angle sensor is offset.

but I'm not seeing status1 budge in your log.

Your fuel just spikes to 24ms out of nowhere. That will generate the backfire and loss of power. Are you running a traction control strategy? Nearly every other status change would trigger a PW=0.

rwyatt365

Thanks for the info. I figured it was one of the status indicators, but couldn't figure out which one.

Yeah, I saw that spike in the fuel and have no idea where it's coming from. And, no, I'm not running any traction control.

I'm wondering if (somehow) AE is causing this behavior, specifically the "TPS WOT curve" which I did change in between turbos.

Before (w/ old turbo);


Then, I thought I was being "smart", so I changed that curve to;

My thought process was, "I don't want AE to be more sensitive at lower throttle openings, I want it to be less sensitive, so I'll raise the TPS% at lower RPMs and that'll do the trick!". Now that I think about it, maybe I'm causing AE (and maybe EAE) to keep "in play" for longer than I want it to. So I changed the curve to;

It's close to the old curve, but different. I wan't able to get a real WOT run (traffic in ATL, you know) this evening, but here's what I saw (an overlay with the backfilre event);

Everything is "less than: (RPM, MAP, TPS, etc...). But what I find interesting is that the EAE % doesn't spike like it did with the "old" TPS/WOT curve. I think I might be onto something!

Tomorrow morning I'll get a full-on WOT event and see what happens.

curly

iTrader: (25)

I think it's your boost tolerance. Turn overboost protection on, turn boost tolerance off, then turn overboost protection back off. I'm guessing overboost protection is off for diagnostic reasons, I would suggest eventually having it on.

rwyatt365

Done...results in the morning tomorrow.
Exactly right. As soon as I get this resolved, overboost goes back on.

rwyatt365

No love. I "eased into" the throttle instead of a quick stab. The results were the same - pops n bangs above 190kPa.

The one thing I'm noticing is that as the MAP approaches the Boost Target, the problem starts occurring - and that's with overboost protection TURNED OFF! Just for ***** and giggles, I'm going to bump up the Boost Target at the higher RPMs and see what happens. MAYBE the MS2 is ignoring the Overboost on/off and cutting off everything - which would be a bad thing!

curly

iTrader: (25)

My next suggestion would be to go to open loop boost control, again with overboost protection turned off, and see what happens.

rwyatt365

First I'll increase the Boost Target in the the affected area (just in case my suspicion about ignoring the Overboost Protection is correct). I'll do that in the morning tomorrow.

Then I back that out, and go to open-loop control and see if there's a difference.

Data will follow tomorrow evening.

VcrMiata

WRT the On/Off Outputs Status Bits - you have Output2 set to become active at 5200 - it is just coincidence it triggered around 190KPA.

What is it controlling?

curly

iTrader: (25)

Vics

rwyatt365

Yes, VICS

rwyatt365

Same thing; PW spikes, pops 'n bangs ensue.


Below is a run at WOT with CL Boost control. What's interesting is that just before the backfire starts, the EAE value stars to oscillate - which I think (and it's only a theory) is driving the PW way low. The system tries to compensate by spiking the PW and everything goes to hell in a handbasket until I get out of the throttle and things can settle down. Coincidentally, 190-200 kPa is at the tail end of the EAE SfW curve...am I seeing a "failure to interpolate" here?

With that being my new working theory, First I'm going to turn off EAE and see what happens. Next I'll turn EAE back on, and extend the AtW and SfW curves out to 250 kPa and see what happens. Turning off EAE will tell me if that is the culprit, and extending AtW & SfW beyond what the turbo can do will show me if the wall-fuel calculations have bit me in the butt.

Here's a screen shot that (I hope) will show you what I'm basing all of this conjecture on;

rwyatt365

EUREKA!!

While doing (more) research, I came across a post on msextra.com that stated "...if the EAE values are too high, that can lead to unstable operation..." (or something to that effect. So I took my AtW & SfW curves and cut everything by 50%. I just came back from a data logging run and - lo and behold - I was able to go WOT without any backfiring! Of course my AFR after decel are going lean, but I can try to figure that out later. I'm just glad to be able to FEEL THE POWAH again.

So. I'm going to get back to tweaking EAE to recover some of that decel lean condition, but I'm happy now.

PS - I would show a log of that, but my laptop decided to go to sleep before logging that WOT pull.

rwyatt365

In case anyone's interested, it seems that the limiting factor with EAE are the SfW values. I raised the AtW values to 75% of the original values and left the SfW values down at 50% of the original ones and was able to get to 235kPa @6100 RPM (see log below). The car felt good, but I need some work on the AFRs.


It's nice to be able to plant my foot on the floorboards again.