VSS resetting to 0

[tobmag]

I only have problem with the vss, iss is clean....

Br tobmag

[gui67]

I first thought I had problems on my sensor or csensor cables, but your post made me take the gpio inhouse and try it with the function generator as you did:
I also have the same problem, using a square wave:
at around 65km/h, the speed drops:
you can see it on the log, at the same speed as tobmag, I have the speed dropping with the clean signal from function generator.

[Bernard Fife]

Guys,

I will have a look on my bench shortly (which I need to get set up again). What speed is 1900 OS rpm on your setup? Is the maximum speed set above this speed?

BTW, the ISS has no reset capability, so the fact that it doesn't reset doesn't necessarily mean the signal is clean.

Lance.

[gui67]

I don't use the VR1 circuit anymore:
at the beginning, I thought that maybe the circuit was not good ar I used not the right resistors, so I used the board Dual VR 2.1 from Jbperf with the same chip as microTCU: I thought that it might solve the problem.
but in fact it did not:
I have the output of the board directly linked to the PT0 port of the processor: so the problem does not come from the VR1 circuit, nor from the sensor or cable as the problem also occurs with the signal generator.

[gui67]

the maximum speed is set at 200km/h, well above the 65km/h

[Bernard Fife]

Guys,

I can run at any OS speed between 1900 and 2000 on my bench without issues (around 49mph with the default settings). I also tried 500 Hz (12 mph) and 1300 Hz (33mph) without issues. However, if I disconnect that signal while at those rpms, the speed (and OS rpm) act very erratically, and don't drop to zero the way they should.

So I need to find out why that is in the code (I suspect it has to do with removing some reset initializations a few versions back).

However, you might also check to make sure your signal path is perfect (I wouldn't connect the function generator directly to the processor port unless I was 100% sure it was clean, never more than 5.0 volts, never less than 0.0 volts, and there were no possible ground offsets between the function generator and the GPIO board), as the only way I could reproduce anything close to what you are seeing is to disconnect the signal entirely.

I will keep looking, of course...

Lance.

[gui67]

Lance,
here you can see a log with the same msq then above, but with the VSS input masking % set at 0% (i don't know if it can help)
There are also spikes to lower RPM, but not anymore at 1910 RPM: a bit lower.

about the code, I had the problem already with the 4.125, and with the 2.2xx codes, maybe even the 2.202 (I don't really remember)
can it be likned to the version of PCB? V2.10

[Bernard Fife]

Gui67,

It's not impossible that there is a problem with the PCB. It could be anything from residual flux to solder bridges to malformed traces.

If this was mine, I would wash it in 100% isopropyl alcohol, rinse it in hot water, and let it dry really well (overnight, or with compressed air, or anything else that dries it really good).

If that doesn't help, I would then get a good magnifying glass out and have a real close look at the board, especially around where the processor is. If you have a camera capable of macro pictures, you could post some shots of the processor and surrounding traces here and we could have a look (though solder bridges and flux can be hard to spot in a photo if the light isn't 'just right').

The VSS comes in on pin #1 (PT0). This pin is on the marked corner (circle) of the processor near the "U1" label, and the neighboring pin (#2) is PT1, which is used for Output 3. If this output is PWM'd and there is flux or a solder bridge, this could certainly cause issues.

I have seen solder bridges occasionally on some of our engine controller (usually alpha or beta versions), but I have not seen either on the GPIO board so far (though the quantity of GPIO boards out there is quite small compared to other B&G offerings).

Lance.

[tobmag]

I have done step 1 and 2 cleaning/scrubbing witha soft toothbrush and also looking on basically all solders I have done, didnt look around the processor yet, will do that.

I think I'm doing this test different then Gui67, I have my function generator connected to the +VSS and sensor ground as it is in the car so the signal is coming in via the pigtail/AMP connector and then over the board, I wanted everything that is normally involved to be in place.
I do measure the signal going into the processor at the 25pin header and assuming that even having noice coming in on the other side of the header the scope should pick it up?

For me looking at this it can hardly be hardware related since I have a really clean squarewave signal on the VR1 pin, I would rather suspect that the signal is processed in the wrong way by the code/filters....

I will during the weekend reload the code and I will also borrow a professional signal generator (Volvo engine R&D department so no garbage)

The 1300hz issue is the main thing since that would be around 60-70km/h this is basically the speed limit on many roads around the outskirts of the citys...
Swedish road department have now given us the following speed limits; 30, 40, 50, 60, 70, 80, 90, 100, 110 and 120km/h...NO Im not kidding!

BR//Tobmag

[Bernard Fife]

For me looking at this it can hardly be hardware related since I have a really clean squarewave signal on the VR1 pin, I would rather suspect that the signal is processed in the wrong way by the code/filters....

Tobmag,

That is certainly possible. But unless and until I can reproduce it on my bench, I have little chance of fixing it. So far I haven't been able to do that.

I do measure the signal going into the processor at the 25pin header and assuming that even having noice coming in on the other side of the header the scope should pick it up?

That depends on the scope and how it is set up to some degree. I wouldn't assume it to be always true, especially since we don't know what is going on here.

BTW, I'm sure you both know this, but the controller must always have good inputs, especially for the manual shift lever position inputs. If these inputs are in an invalid state (or floating), the controller will reset frequently, messing with the VSS and many functions. So you might want to check the other inputs to ensure they are connected properly.

Lance.