At my wits end... ANY IDEAS? LS3 with 24X trigger.

[CRSTune]

We've already been in contact before via email. He's about 3-3.5hrs south of me, not that I'm against going.

I've worked with more Pro units than the other MS versions.

[DLRacing]

Since I now have a 36-1 crank wheel and a half-moon cam signal, does anyone have any common noise filter parameters that should work properly on a motor that revs to 7,000 rpm, and that I'd like to get to idle down around 900-1000.

I've tried calculating tooth times at each rpm, and setting the filter interval to 1/2 that or a bit less, but it seems to cause more problems than it solves.

It just occurred to me that I calculated the time interval based on 36 teeth per revolution, but the tooth is only present for half that, so my times are double what they should be.
I did reduce them quite a bit during testing, but they still may be a little too long.
Barring any other suggestions, I'll recalculate them tomorrow and give the filters another try.

Thanks,
David

[nathaninwa]

It seems strange to need the noise filtering On all my installs so far none has been needed, especially with a Hall effect sensor

I can test your box the same day, 3 days to flat rate me, a day to test and flat rate back to you. My LS is 58x as well

[Raymond_B]

It seems strange to need the noise filtering On all my installs so far none has been needed, especially with a Hall effect sensor

I can test your box the same day, 3 days to flat rate me, a day to test and flat rate back to you. My LS is 58x as well

I agree on the noise filtering, DLRacing, are you using a really long USB cable or have something else generating a lot of EMI where you need the filtering? Maybe go without the USB or try Bluetooth for logging?

[DLRacing]

It seems strange to need the noise filtering On all my installs so far none has been needed, especially with a Hall effect sensor

I can test your box the same day, 3 days to flat rate me, a day to test and flat rate back to you. My LS is 58x as well

I really appreciate that. I'll let you know after I talk to Matt Cramer and see if he wants to check it out himself.

Thanks,
David

[DLRacing]

I reset the noise filter parameters, and for the first minute or three, it seemed to be working.
As it warmed up, the sync losses returned, and got worse the longer it ran. I noticed that yesterday as well.
It usually shows up when the RPM dips below about 1,000.

The .msl, even compressed, is too big, can't post.

[DLRacing]

I decided to scope the CKP input without the engine running.
With the signal low, I jumped the test input to turn on the fuel pump and coils, and the Wideband O2 sensor.
As soon as I did that, I saw over 750mv of high frequency noise spikes, with multiple intervals.

My first thought was the fuel pump, which is wired directly to the battery through a relay on the rear bulkhead.
I unplugged the relay and was surprised to see almost all the noise was still there. The pump was only producing a bit of very low level noise.

So my thoughts turned to the Innovate Dual Wideband. I pulled the O2 fuse on the MS fuse/relay panel, and the noise disappeared. Almost all this noise was coming from the WBO2 gauge.

I pulled the ECU to get to the WBO2 connectors, and double-checked that it was wired according to the MS3Pro manual.
I then ran a wire straight from the battery + to the load side of the O2 fuse, to eliminate noise being injected on the ECU power lead. The noise was only very slightly better.
Next, I considered using the Sensor ground for the WBO2, but considering its 6 amp current draw, I didn't want to put that much load on the sensor gnds. I disconnected the ground from the MS harness, and connected it directly to chassis ground, while still powering the O2 through the separate wire, and voila, the noise was almost totally gone.

Thinking I was finally onto something, I made a new ground wire for the WBO2, and cut the O2 relay power supply loose from the main MS harness, and fed it through an inline fuse. After putting everything back together, the noise level was still way down from where it had been. Nothing even close to logic high spikes were present, so I really felt this should make a big difference.

Well..... NO, it didn't.

I cut off all but the global noise filter and fired it up.
When I initially started it up, it lost sync once, then settled down and ran for a couple minutes or more before again starting to lose sync at low rpm. It got progressively worse and worse until the engine finally died completely.

Attached are the logs.

[DLRacing]

More logs.

[DLRacing]

While scoping the CKP signal, I noticed the signal was high at 5v, and the trigger wheel happened to be in the missing tooth area. There was a tiny protrusion at the bottom of the missing tooth slot, so I filed it smooth an rotated that area back to the sensor. This time it was low, so I taped a thin washer in the slot, and the sensor went high.
Suspecting the sensor was a bit too close, I increased the gap by about .020".

The car wouldn't even start, and the log showed it was missing teeth.

[DLRacing]

Next, I tightened up the gap to only about .005". I wouldn't want to run it this close on the road for fear of contact with the sensor.
This time it started, and the composite log looked good, but it was again losing sync within a short time.

I'm completely out of ideas!!!

I spent over $2,000 on the MS3 and the Innovate WBO2, hundreds more on wiped out O2 sensors, another couple hundred on swapping to the 36-1 wheel setup, which makes my trigger configuration as common as dirt, and between $5,000 and $10,000 in time trying to get this thing going, and it still won't run!!!

I can't see this as anything other than a hardware issue in the MS3Pro.
I had high hopes for the MS3Pro, but I'm about ready to throw the whole thing in the trash and buy a Holley HP EFI system.

[ashford]

try triggering from the other edge. the last tooth before the missing looks too long compared to what comes before it and the missing is too short. a tooth time log will show this easier. if the trigger edge don't fix it i would go with a vr sensor or different resistor combos on the pullup

[Matt Cramer]

Ashford's idea is a good one. Also, can you post a tooth log instead of a composite log of an attempted start? The composite logs are good for the 24X pattern, but the tooth log is a lot easier for me to read reasons why something is going wrong with 36-1.

[billr]

The picture posted of your 36-1 wheel doesn't show any "protrusion" in the missing-tooth gap. Whatever you filed off there must have been very minor, I don't think that is relevant to your problem. Do I understand correctly, you are using the DIY 36-1 wheel and a GM 58X sensor? That may be a problem. The physical config of the wheel teeth and the sensor sensing target need to be somewhat matched. You may want to try using a DIY sensor that is well-known to work with that wheel.

Another suggestion is to take the wheel off the engine and mount it to some variable-speed motor, along with a bracket to hold the sensor. Then you can run the wheel endlessly and at various speeds, peacefully experimenting on-the-bench until there is absolutely no more sync issues. That may seem like another chore/cost to endure, but a bench setup for the wheel can actually be done fairly cheap and easy, certainly far less expensive and time-consuming than giving up on the MS.

If you are convinced this is a problem with your MS unit, I also suggest sending it out for test; either to DIY or one of the members who has offered. The cost and delay to do that is also trivial to starting over. Point is, if this is a problem with your install, not the MS unit, then you may well end up in the same situation after installing a different system.

[DLRacing]

Ashford's idea is a good one. Also, can you post a tooth log instead of a composite log of an attempted start? The composite logs are good for the 24X pattern, but the tooth log is a lot easier for me to read reasons why something is going wrong with 36-1.

I'll do both of those the next time I have time to work on it, probably Friday.
Thanks,
David

[DLRacing]

The picture posted of your 36-1 wheel doesn't show any "protrusion" in the missing-tooth gap. Whatever you filed off there must have been very minor, I don't think that is relevant to your problem. Do I understand correctly, you are using the DIY 36-1 wheel and a GM 58X sensor? That may be a problem. The physical config of the wheel teeth and the sensor sensing target need to be somewhat matched. You may want to try using a DIY sensor that is well-known to work with that wheel.

Another suggestion is to take the wheel off the engine and mount it to some variable-speed motor, along with a bracket to hold the sensor. Then you can run the wheel endlessly and at various speeds, peacefully experimenting on-the-bench until there is absolutely no more sync issues. That may seem like another chore/cost to endure, but a bench setup for the wheel can actually be done fairly cheap and easy, certainly far less expensive and time-consuming than giving up on the MS.

If you are convinced this is a problem with your MS unit, I also suggest sending it out for test; either to DIY or one of the members who has offered. The cost and delay to do that is also trivial to starting over. Point is, if this is a problem with your install, not the MS unit, then you may well end up in the same situation after installing a different system.

Yep, I'm using the 58x sensor, as it was designed for a single-row wheel. I wouldn't expect the wheel having fewer and wider teeth would present a problem, and I intend to switch to a 58x wheel when I tear the motor down to refresh it in the next year or two.

Thanks.

[DLRacing]

try triggering from the other edge. the last tooth before the missing looks too long compared to what comes before it and the missing is too short. a tooth time log will show this easier. if the trigger edge don't fix it i would go with a vr sensor or different resistor combos on the pullup

I don't want to rely on a VR sensor to mask this problem, as I'll have to deal with it again when I go back to the 58x sensor.
I'll try the other edge though.

Thanks.

[DLRacing]

Captured some tooth logs for Matt to look through.
Again, it ran fine for the first few minutes, then started losing sync again.

[DLRacing]

It took several logs to start getting errors.

[DLRacing]

Tried setting Ignition Input Capture to falling edge. The car wouldn't start, and I saw no teeth detected at all.

[DLRacing]

MSL for the above file.

Matt, please let me know if this gives you the info you need.

I'm at the point that I want to try another ECU.
I just can't believe it should be this hard to get it to work properly.
I'm going on 4 months trying to get this thing running.

Thanks,
David