ScottFW

iTrader: (4)

I've been idling a tad rich so I was adjusting some of my idle cells, and in one cell of the VE table I try to enter 34, and send it to ECU. But when I fetch the table back, that cell has 38 in it every time. I can't ever get it to store 34 in that one particular cell. However, that's the lesser problem at this point.

I tried to load an older .msq that had the values I wanted in the idle cells, but like an idiot I tried sending it to MS while the car was running. Usually I don't do that, maybe I was tired because it was late last night, whatever. I don't think it's unusual for the car to stumble a bit while you're sending a new .msq, but it stumbled to a stop and I got the "controller voltage too low" error. With the engine no longer running I successfully loaded a good .msq but then had difficulty starting the car. Crank crank crank, nothing. I figured I probably had flooded it by now, so I floored the gas pedal while cranking (flood clear mode) and it did start up, but now it idles like ****.

With my timing light I put the inductive pickup on each plug wire, and each of them triggered the light so I assume I'm getting good spark. Back on the #1 wire, I noticed my timing was off by several degrees. I verified that the spark settings in MT hadn't changed. I then told MS to give me a static 10 degrees and adjusted the trigger angle to sync up MS to the crank pulley again. I had to drop 14 points of trigger angle (from 82 to 68) to sync it up. I set it back to -10 to use the map, and the car is still idling like ****. It runs rough, wideband reads very lean (18-22:1 AFR). I activated EGO and gave it big authority but it still won't read richer than 16:1, and I can't get MAP any lower than 38 kPa. It doesn't want to rev above ~2500 rpm.

I'm thinking that when the car stumbled to a stop (due to my sending a .msq to a running engine ) it's possible that the timing belt slipped a tooth or three. I don't think this can be explained simply by an issue with the MS (firmwarez?) but I don't know. The timing belt has less than 1000 miles on it since I replaced it (Gates racing belt) but I did a track weekend on it and had no issues. Still, the fux0red timing and rough idle has me thinking it slipped. What do y'all think? Anything else I should look at before I start disassembling stuff?

Matt Cramer

Could be corrupted firmware. Try reloading the code and go back to a known good MSQ.

Skuzzle

i've had similar.
Only cured properly by reloading the firmware.
I managed to get the car working ok by playing with the msq, but never right.
Firmware all the way.

gospeed81

iTrader: (51)

+1...I've had serious firmware issues.

I'm going to stop loading .msqs while the car is running.

If you're lazy there is a way to just send your revised map instead of the whole .msq. This should reduce the chance of error.


EDIT:

To actually answer your question...

Look at the variables.

1. Your timing belt is proven, and it would be pretty damn hard to make it jump teeth unless your tensioner/tightening method was bad. Your trackday would have told you this since the forces on the valves & cams in compression braking are much more than a stumble when dying with no load.

2. Your mechanical timing shouldn't go anywhere. Your trigger angle easily could have been different between the two .msqs, or had a differention trigger angle addition...

So...

#3 is a winner...you were playing with MS when things went south. I'm 100% with Matt on this one. I've been there enough I relocated the MS to where I could easily pop the case open and jumper it for a reflash.

ScottFW

iTrader: (4)

Yeah, I'm really hoping it's the firmware so I can avoid doing what is essentially another timing belt job. I haven't had to reload it before, so I didn't know how corrupted firmware would manifest itself on a running car, but I imagine it could cause all kinds of weird stuff.

Usually I do only load the new VE table instead of the entire msq, and you can do that with the engine running and it causes no problems. But loading just the VE table didn't work as far as getting that one cell to store the value I wanted (another possible firmware issue) so I then tried to load the whole msq, but neglected to shut off the engine first and here we are.

I'll try reloading the firmware when I get home. Hopefully I won't need the boot jumper but we'll see. Getting physical access to the MS requires removing the race seat behind which the MS is securely mounted, although that would just be a mild inconvenience at this point.

gospeed81

iTrader: (51)

firmware issues defined

ScottFW

iTrader: (4)

I reflashed the firmware and loaded a known-good .msq, and the car still runs like crap.

Datalog and .msq attached. The engine is stock including the injectors and ignition coils. Immediately after starting the rpm quickly drops until it dies. That is at the very beginning of the datalog. After that, I was able to keep it running but it's very rough, you can see the MAP and rpm fluctuating.

At one point I pulled the intake to get a look at the throttle butterfly and make sure it's moving freely, and I noticed a little smoke in the intake tube, almost like exhaust is flowing back into the intake. The exhaust generally smells like there's a good bit of unburned fuel, not surprising considering how it's running.

Pulled the plugs and they're all blacker than normal. The #2 plug had some visible moisture on it but they all look pretty fouled to me. These are BKR6E-11s with <1000 miles on them.


My MS is a parallel install, and it wouldn't take long to hook up the MAF, remove the boomslang, etc and just run it on the stock ECU if I wanted to eliminate MS as the root of the problem. But that would have to wait until tomorrow.

ScottFW

iTrader: (4)

Well, I reverted to stock- removed MS from the system, hooked up the MAF, reinstalled the stock narrowband O2. Car started and idled okay. I went for a short drive and soon realized that even on the stock ECU, there's still a problem. Acceleration from 2K-3K is abnormally sluggish and it's not smooth, like there's a misfire or something. Flooring the pedal just causes it to bog instead of go.

Got it back into the garage, jumpered TEN and GND, and put the timing light on it, and it was at about 0 or maybe 1* BTDC. Attempted to rotate the CAS, and the most advanced I can get the timing before the CAS runs out of adjustment is 6* BTDC. Before installing the MS I was running 14* with the stock ECU.

It is getting real hard to ignore the possibility that the timing belt slipped. On the stock ECU when the CAS is out of adjustment range and the timing is still too retarded, really, is there anything else it could be?

mafoose

Do these cars have a "limp home" mode with the ignition system? It almost sounds like what EDIS modules do if something bad happens to them.

ScottFW

iTrader: (4)

Well, it was indeed a slipped timing belt. With the crank at TDC both of the cams were one tooth off. I put the belt back on correctly, just finished buttoning things up and it runs like normal on the stock ECU. Good throttle response, smooth accel with no stumbling. I also have enough range of adjustment on the CAS to set a reasonable base timing.

I'm gonna leave the stock ECU in for now, only because I'm due for a smog check pretty soon. I'm certain it will run fine when I reinstall the MS. The shitty idle with MS I believe was because it runs a MAP sensor rather than a MAF. With the cam timing all fucked up, the engine wasn't pulling as good of a vacuum at idle, which of course would confuse MS. With a MAF, the stock ECU would know that flow through the engine was impaired and compensate to keep the mixture more even than the MS. I think.

Bottom line, the timing belt slipped as a direct result of the misfire/stumble induced by sending a complete .msq while the engine was running. I strongly recommend not doing this.