codrus

Something strange happened in the third log:



The RPMs decay away, the duty cycle goes down but the PW goes up. Huh? That doesn't make any sense to me. Then it cruises at 248 RPM for a while -- IIRC, from what Reverant says, "250 rpm" is what it shows for an engine that's having certain kinds of sync error problems. OTOH, the AFR stays pretty constant at 12-ish, which suggests there isn't something really crazy happening with the fuel.

Unfortunately, things like battery voltage, TPS, and VSS aren't in the log, so it's hard to say exactly why it's happening. Do you have a voltmeter gauge in the car? If not, I highly recommend hooking one, they're $30 from Summit and very useful for diagnosing weird alternator/battery problems.

--Ian

poormxdad

Ian,

Thanks. I had a spare alternator rebuilt by a reputable local company to try to give me some more amps. It's the alternator that's in there now. I've been tempted to swap in the other alternator, just to see what happens. It's free, but an aspain.

Any point in taking her out without the hood to monitor the fuel pressure?

codrus

Are you using laptop logging or SD card logging? I'm assuming SD card because of the limited fields, right? If so, and if you can easily reproduce the problem, then personally I'd say go log it with the laptop hooked up because that will capture a lot more fields than the SD card logging does. At the very least, that should give you battery voltage when these things happen.

Fuel pressure info is also interesting, if you can easily monitor the mechanical gauge while driving then sure. I hooked up an electronic fuel pressure sender in mine so that I could monitor it in the megasquirt logs along with everything else.

I think you're better off gathering more data than randomly throwing parts at the problem.

--Ian

poormxdad

I have an extension cable from the MS3 Basic that puts the SD card at the bottom of the dash above my right leg. I did not know the SD card collects less data.

I'm concerned the problem is getting worse in that there is more crappy running than normal now. I'd hate to take her out and have something fail completely and need a tow home.

codrus

In TunerStudio, under "Data Logging" and "SD Card Datalogging" you can configure what goes into the SD card logs. Different versions of the software allow different amounts of data.

As for a tow, there's no way to confirm a fix without actually driving it. Do you have AAA? If not, I'd recommend adding that, keeping your test drives within 5-6 miles of your house, and making sure you bring your cell phone and AAA card with you on test drives.

--Ian

poormxdad

I have USAA with towing. Funny, I called yesterday to confirm towing on all my vehicles...

poormxdad

I got up this morning to a slight smell of fuel in the garage. The Harbor Freight gauge was leaking at the gauge threads. I decided to remove and repair it (or throw it away) rather than driving with it installed. So, I listened to some audience recorded Grateful Dead and swapped out the alternator for a known good one. It ran well long enough that I thought she was fixed. Then, she **** the bed just like before. It also seemed to be running hot. It was in the 50s here today, but she was showing 190+ around town with the 180 degree thermostat. It just seemed hot, but it could be my imagination.

To recap, in the last couple of days I've swapped out the alternator, and replaced the fuel pump, fuel filter, spark plug wires and coil packs with new items. The crank position sensor is OEM Mazda and only has a couple of hundred miles on it. The battery is only a few months old. When it runs correctly, she's a beast. The intermittent nature of the problem has me completely boggled.

codrus

Did you get a good data log and/or composite log of the failure today?

--ian

poormxdad

I added (or thought I did) BattV to the data logging, but I don't see it. What would you like to see in the log file?

I did get these. 06 is sitting idling in front of the house. She sounded labored.

I will attempt to get an actual composite log today. There are a couple of other diagnostic logs. Should I get something else, too?

Thanks,

stefanst

iTrader: (1)

Any perceived problems during the recording of these logs? All I can see is that your tune is not all that great. Your AFRs are all over. And start logging with a laptop, so you get more helpful data.

poormxdad

When the engine is running properly, she runs fantastically. The bouncing AFRs are a symptom of the problem. When whatever is happening happens, she runs rough, makes no power, and the AFRs get lean, even at idle.

codrus

Can you please post:

a) a laptop log, these SD card ones are missing too much information
b) happening when the car randomly dies
c) ideally with a pointer to the location in the log where it happens.

--Ian

poormxdad

Roger. WILCO. Probably will have to wait till Friday.

poormxdad

Gents,

Please take a look at these composite logs. The one from 14:50:06 should have an instance where the car died. The other one is afterwards, in case I did not get the logging correct. The last time I had to take a composite log, all I had to do was hit Start and Stop and find the file. The new software has a different implementation and I'm not sure I got it right. Let me know if these are useful. I do have a handful of additional composite logs.

When she died, she was running at a lower rpm than would be normal for the amount of gas pedal I was putting in. She sounded like Daa, Daa, Daa, Daa, Daa... like an old lawn mower trying to run after sitting for the winter, almost as if she was out of balance. I knew she was going to die when I let off the gas.

Thanks,

poormxdad

Gents,

I played around a bit. I know less than I did before.

Soooooooooo, I have four coils--two sets. I took out my timing light and hooked up the inductor to each of the plug wires. When the engine was running crappy, Plug 3 had the most erratic flashing of the timing light, so I swapped in one of the other coils on the driver's side. No real change, so I swapped the other coil in there for the driver's side. No real change. Then, I swapped sides with the coils that were in there. #1 had a more erratic flash, but the engine didn't sound any different, and #3 STILL had an erratic flash. Maybe it's a meaningless diagnostic. I'd like to hear from the experts.

I got a few more logs, this time from the laptop rather than the card. Check out the one from 14:22:36. WTF? 16:41:48 was on a short drive. The rest, one is a composite log, are at idle in the driveway.

Thanks,

poormxdad

Just some additional info.

The catalytic converter **** the bed a couple of events ago. I'm running with a gutted cat for right now.

I replaced the bottom end with one from a Japanese Domestic Market importer. I swapped in my rebuilt head, and did a timing belt/water pump job. Then drove to VIR. She's a beast when compared to the 143,000 mile bottom end that was in there.

Given that she ran so well, I want to assume I did the swap and timing belt properly. Is there anything under the cam cover that could have come loose that would cause her to intermittently run badly?

Thanks,

farpolemiddle

I am more inclined to think you had a bad ground or bad wiring and fixed it on accident by "touching it" when you swapped all this other stuff around.

poormxdad

Here's another log.

Since I had the timing light out, I checked base timing. It was a little off and I tweaked it back to 10 degrees.

I checked all the grounds that I know of, and also got under the dash and pulled and reset the connectors for the MegaSquirt, as well as checking my power shunts in the fuse box.

Towards the end of this log, she started to run rough. I checked the plugs firing again with the timing light. #3 showed dropouts often when compared to the other cylinders. Flash, flash, flash, flash, blank, flash, blank, blank, flash, etc., etc. There was no noticeable pattern. As I mentioned in a previous post, #3 did that even though I changed out individual coils. The plugs are brand new. Is there anything to that?

Thanks,

codrus

Back from Thanksgiving trip, I can look at logs now.

The composite logs form 11/25 and 11/26 don't show anything. There's a sync loss at the beginning of the 11/25 log, but that appears to just be from startup (it has to sync when you first crank it). The two composite logs from 11/25 are only 24 and 17 seconds long though -- did you really intend to only log that long, or is there something wrong with the logging setup? The composite log from 11/26 is 2 minutes long, but shows no loss of sync.

The two laptop logs from 11/26 don't show much of anything either -- no sync loss happening and AFAICT no stalling. The 14:22 log is 5 minutes long, and at 250 seconds in there's an odd glitch where VSS jumps from 0 to 3.4 stays there for a second, then goes back to zero. Did you roll it backwards down the driveway briefly or something? Nothing jumps out at me on the 16:41 log, but it's only 78 seconds long.

The 11/27 log shows VSS moving around quite a lot. Do you have it calibrated properly? TPS is zero -- was the car moving during the log or parked? If stopped was it idling up on jackstands with the wheels in the air, or on the ground with the parking brake on? The VSS is on the output shaft of the transmission, which means it's physically locked to the differential, so if both rear wheels are not moving, there's no way it can generate any kind of legit signal. If so, then I think there's a strong chance that you have ground issues.

The 11/27 log also shows a lot of spikes in the AFR, up to 18:1, starting around 18 seconds into the log, which is also when the RPM starts getting a lot more ragged. It goes from a nice stable 2.1ms of pulsewidth to something that's moving around a lot but averaging about 2.3 to 2.4. What's your injector deadtime? Assuming it's around a millisecond, that means it's seeing 20-30% more fuel on average, and yet the average AFR value hasn't changed. I'd say that either the car misfiring (and thus blowing unburned air & fuel into the exhaust, which fools the wideband sensor into thinking it's lean because it only looks at oxygen content, not fuel content) or the wideband is measuring incorrectly for other reasons (bad ground, bad sensor, something like that).

--Ian

farpolemiddle

How old are your plug wires?