SOLVED - Resets - Closed Loop Idle RPM Error

[jsdevel]

I've been battling resets for months. Running MS2/3.57, low z injectors with P&H board. I have dedicated ground wires for the ECU & 02 sensor. Injectors & LS2 coils ground at the block.

I've tried everything. I've upgraded battery cables, ran dedicated power/ground. You name it.

You can see that everytime secL resets to 0 "Closed Loop Idle RPM Error" appears. Have I configured Closed Loop idle in a way that is bad? At one point some 8 months ago I had CL working quite well, so I'm honestly not sure what's happened. I'm guessing I've changed some settings trying to dial it in, and I failed to take a backup of my tune when it was last good.

Here's what it looks like in the logs:

cl.png

My spirits at this point are quite low, and I'd appreciate any help that anyone can give me. I've posted my current tune with accompanying data log. Do I have my settings wrong? Would starting my tune over at this point be advisable?

[jsdevel]

Ugh. So, going over my tune again and running the file through diff with previous tunes, I noticed that I had SPKA output set wrong. In the tune I uploaded, I had it set to JS10 when it should've been set to D14. At this point I'm guessing that my timing is off. What's strange is that it would idle for 10 seconds or so with that setting.

[rickb794]

It should not have ran at all with the spark set to JS10 if you have it wired to D14. Something weird there.

The image you posted shows a high battery voltage at 20v, check that.
If it is not high then it is likely something is pulling down the 5v Vref that powers the MS.
Which would cause resets...

[jsdevel]

If it is not high then it is likely something is pulling down the 5v Vref that powers the MS.
Which would cause resets...

Would a bad voltage regulator (U5) cause that?

[jsmcortina]

Most likely an external fault in the wiring or a sensor overloading TPS-VREF.

James

[jsdevel]

Most likely an external fault in the wiring or a sensor overloading TPS-VREF.

James

I checked my wiring, and connectors. The EV1 female connector on the TPS was pretty loose. I tried wiggling it to see if I could see my TPS changing in the logs, but it didn't seem to make a difference. Is it possible that this could be the problem and that my data logs wouldn't pick it up? Also, am I reading the logs correctly, that my TPS briefly spiked to 100%?

[jsdevel]

Changed the TPS connector. Nice and solid now. Still not working.

I think I may have fried my DSM CAS when I ran the car with SPKA on JS10. Would that be enough to fry it?

Here's the signal pattern coming from the CAS (acquired with DAQ AMI). You can see that it looks pretty irregular:

wheel.png

[jsdevel]

Also, I tested my MS2 on a Stim, and everything worked great. Didn't reset or anything. Would an irregular trigger pattern cause resets?

[jsdevel]

Changing my DSM CAS made a HUGE difference! I'm now getting fewer resets. You can see in the logs several times where I attempted to start my vehicle. It would start and idle for a few seconds before dying. The resets that I am getting now appear to be related to fluctuating battery voltage when cranking.

graph.png (41.55 KiB) Viewed 1003 times

Now my issue is getting this thing to idle consistently. I'm inclined to open a new thread for that.

Thanks for the help!

[jsdevel]

So, I'm still getting resets. At this point I'm not really sure where to check / what to do. I feel like I've tried just about everything. I suppose I could ground my LS2 coils directly at the battery to see if that'll make a difference. Anyone else have any ideas?

idles-then-reset.png (46.24 KiB) Viewed 987 times

[rickb794]

Are the LS coils mounted to an Un-grounded valve cover?

[jsdevel]

Are the LS coils mounted to an Un-grounded valve cover?

The bracket that they're mounted to is powder coated. I doubt it has much grounding. The valve cover may not have good grounding now that I think about it. Would that rebound lots of EMI from the coils?

Attached are photos of my setup. The sensor wires run through the firewall on the drivers side away from the coils and injectors. They then travel along the firewall over the the MS mounted on the passenger side.

IMG_20170929_173619.jpg (618.95 KiB) Viewed 962 times

IMG_20170929_173733.jpg (652.78 KiB) Viewed 962 times

[rickb794]

If the valve cover is not grounded it will act like a broadcast antenna for the coil noise.

Star washers will provide grounding for powder coated parts.

On my Neon I ran a ground strap between the coil bracket/valve cover and the head.

Many of the newer motors have valve cover sealing that does not ground the cover, which is fine until you mount a coil on it....

[jsdevel]

If the valve cover is not grounded it will act like a broadcast antenna for the coil noise.

I think we can rule this out. I checked continuity and voltage from the bracket/valve cover/bolts to battery, and didn't see any drops or resistance.

The logs really don't show a voltage spike/drop, is this because MS resets as a protection measure?

I'm going to measure the voltage at the MS 12V/Ground to inspect.

[jsdevel]

OK. So I busted out my sound scope. Used 2 analog input channels to take some voltage readings.

• AI0 (red) connected + to 12V Pre 3 Amp Fuse for the MS, - to negative lead to the MS, goes directly to the battery.
• AI1 (blue) connected + to Fuel Pump 12V Post Relay, - to Fuel Injector Grounds, goes to the block

Here you can see that during priming (before starting), it looks like the coils slowly discharge as the falling edge has a curve:

priming.png (20.89 KiB) Viewed 955 times

Here you can see that the fuel pump power has some spikes fluctuating from 14 - 18 V:

coil-spiking.png (17.31 KiB) Viewed 955 times

Here you can see some voltage drops at the 12V feed for the MS:

voltage-drop.png (16.8 KiB) Viewed 955 times

Could these spikes / drops have an effect on resets?

This particular run lasted for about 3 minutes. From what I could tell, there were few resets. I usually hear the fuel pump relay switching when it resets and I didn't hear that happen once during the run. Promising!

When I look at the logs, BattV looks pretty stable; however, it does show a max of 20V although I can't see where it spikes to 20V in the logs. Would it actually be spiking to 20V in the logs if log viewer shows max of 20?

The only thing different this time was my laptop wasn't plugged in to my unshielded serial cable with USB adapter.

[jsdevel]

Well, I'm not exactly sure what happened. But I'm able to drive around now. I put a few ferrites on various power lines, and I installed a capacitor on the boot jumper of my 3.57 main board. I also tucked wires away from each other. I think I had one reset driving around for 30 minutes. Can't be absolutely certain because I wasn't logging.

[jsdevel]

Drove around some more. Still getting resets. At this point I think something is up with my wiring harness. I may have a short somewhere or something.

I brought it back in to double check timing, and noticed I was advanced by like 60 degrees. I set cranking RPM to 0, and adjusted my CAS. Increased Tooth #1 offset from 0 to 30. After doing this, it started resetting quite a bit. Moved Tooth #1 offset from 30 to 0 and the resets lessened in frequency, but they're still there.

[rickb794]

Are you using a USB cable longer than 6"?

I battled this early on in my MS education.
Some cars would run fine others would not make it around the block.

Switching to a RS232 9pin serial cable and an adapter with a very short or NO USB cable solved all that mess.

Also your logging computer should not be powered by the cars 12v it needs it's own power.
Or use a Cigarette lighter 110v convertor with the computer power supply to isolate the car from the computer.

[jsdevel]

Are you using a USB cable longer than 6"?

I'm using a 4-6' serial cable with a USB adapter.

Also your logging computer should not be powered by the cars 12v it needs it's own power.

Yea, I'm running on battery power or wall outlet.

[jsdevel]

So, I did some more driving today. This time I saved the output of the Composite Logger to a file and played it back with log viewer.

This looks like a normal 24/1 pattern for my setup:

good.png (21.79 KiB) Viewed 931 times

This does not:

bad.png (19.21 KiB) Viewed 931 times

Is it normal to see that kind of tooth loss? Would this cause resets?