speed signal issues

[Ap67et10]

I've been having issues with gpio dropping out an loosing connection with TS and MS3. I figured out that all my connection issues were from the OSS input to gpio. Here is my latest piece of info and where I'm at. The original circuitry for OSS was not even close, so instead of playing with it I put the VR input into an open spot on a dual VR board from Jean. Then brought the digital signal into GPIO and bypassed all circuits. This produced a very nice input, datalog looks great and follows vss2 that MS3 monitors perfectly. However once I rev the engine or get up to speed or even just let the car get heated up, it'll drop out and loose signal. I have verified this signal does not produce this issue when I put it into MS3 on a digital input. Something about GPIO is freaking out and I'm not sure what or why. I had an issue that was similar with a crank input through the same conditioner Jean uses on his board, so I am going to try some things today. However I can't figure out why GPIO has an issue with it and MS3 does not. Anybody have any ideas on this?


Andrew

[Bernard Fife]

Andrew,

What have you tried with the filtering options in TunerStudio?

Have you made sure there is no residual solder flux on the 25x2 header pins (or anywhere else in the VSS input path)?

Do you have a scope trace of the input signal? If so, could you post it.

Do you have a clear digital close-up photo of the VSS connections on the GPIO board that you could post?

If this was mine, I would also inspect the processor and surrounding area on the GPIO PCB very carefully to make sure no solder blobs, wire snipped off leads, or residual flux, etc. is bridging any of the CPU pins or circuits.

Lance.

[Ap67et10]

I inspected the vr board closely before sealing it off so I don't believe there is any issues. I also scoped the VR signal which was nice going into VR board. I have not scooped the signal coming out yet though. I just moved the signals around from ms3 to gpio and had no problems on ms3 and nothing but problems with the signal on gpio.

With filtering I have tried 0%, 20% (even though not recommended this still produces faster response than CAN seemed to give) 50%, 80%, 90% and 100%. Those were all tried with 4, 8 and 12 smoothing and all produced the same result. I did see a change in how quickly it would go down. I also would see some false trigger in park without much filtering.

I experienced an extremely sensitive VR on my MS3pro on my procharged 502. I actually solved this issues by putting a resistor inline on + of VR. Shunts of any kind did nothing to help. I could literally turn on the fuel pump on that setup and get a false RPM trigger from MSpro. I will likely try some of this on the VR side of the speed signal to make it work if I can. I would really prefer running speed into gpio.


Andrew

[Bernard Fife]

Andrew,

In addition to all the above, can tell us what with/how you sealed the board off?

Lance.

[Ap67et10]

The dual vr board is the one that is sealed, not the gpio. I just wanted to be clear as I reread what I posted and see it may not be. Gpio is not sealed at all. The dual vr board after everything was soldered was sealed in dual wall heat shrink. I got the idea from Jean to do it that way.


Andrew

[Ap67et10]

What I did for the digital input to gpio is bring it in through pin 14 for VR3 input, because I do not have any circuits on VR3. I then grabbed it on the board and jumped the wire right to pt0 (I believe thats the port number, but can't remember off the top of my head...whatever the jumper is for the OSS input on CPU next to vr1 header)


Andrew

[Bernard Fife]

Andrew,

Can you post the scope shots and photos I mentioned above?

In particular, scope shots of the input signal at the source and 25x2 jumper would be useful, a picture of the wiring done on the VR3 input, and a close-up photos of the processor could be helpful.

Also, have you inspected the board very, very carefully (including the 25x2 header on both side of the PCB) for residual solder flux? Has the board been washed in 99% isopropyl alcohol?

Lance.

[Bernard Fife]

Andrew,

Could you also post your MSQ? A datalog would be helpful too. This will help us to see if there is something going on in the code (like an over/under-flow in the speedo calcs, for example).

Lance.

[Ap67et10]

Well I got it fixed!

Based on what I thought it was doing I went ahead and put a 10k ohm resistor in. Originally I thought I would put it on the VR side, but decided to try it on the digital side inside the GPIO case before it went into PTO. This fixed everything....the signal is smooth, quick and accurate and no more signal loss problems!

Got to drive the truck around a bunch, and I need to do some more tweeking, but its coming along nicely. I did notice that if the TCC engages its good, but then if it disengages the trans feels like it is in neutral. I would have to stop, cycle the shifter to park and back, then its totally fine.........Now that I am typing this though, I was having signal issues with the PRNDL switch, and after adjustment on the transmission its all good. Maybe I will try the TCC again to see if it works the way it should now that that PRNDL switch is working as it should also.

Lance, is the 5.015 code the only one with TPS as a blendable load factor? I found while driving this could be useful when I wasn't full throttle but MAP went up quite a bit it would want to downshift into 1st, but I definitely wasn't wanting that. I could see that the blended load would be beneficial for that. (If that is even how it is suppose to work. I just saw it in the settings when I had 5.015 loaded yesterday, I'm on 4.139 now.)

Overall I am very happy with it though. I have many MS systems under my belt, but this is the first trans controller with it, and I'm liking it!



Andrew

[Bernard Fife]

Andrew,

I'm glad you got the VSS sorted out. This is often the hardest part of installing the MShift/GPIO, for a number of reasons.

Yes, the 5.xxx series of code is the only code with the blendable TPS/MAP/etc. You can read about the blended load here: http://www.msgpio.com/manuals/mshift/V5tune.html#gh The 4.1xx series of code is generally supposed to be largely frozen now, with new developments happening only on the 5.xxx code (so 5.xxx gets all the updates from 4.1xx, but not vice-versa).

I have a few ideas about the TCC issue (it may be unlocking if there is a gear lever error, so I need to check on that). I will work on it.

Lance.