Stim problems on the VR crank input

[Dennis_Zx7r]

MS3X with V3 board.
I generate rectangles (0V - 5V) using an Arduino to emulate a crank sensor (VR input mainboard, 24-1 wheel) and a cam signal (MS3X cam input).
This is a composite log of it working correctly and the ignition settings in the MSQ:


The cam signal works perfectly all the time. However, the crank signal only works occasionally and I can't figure out why.
Using an oscilloscope on TSEL, I can see the signal is not triggering the VR signal when it's not working. There's just nothing there. Changing the pots (R52/R56) does not change anything, and I left them fully "open" as recommended. If it does work (power cycling everything several times), it mirrors the input signal perfectly and works flawlessly until power cycling again.

Is there something I'm missing? I looked at other threads mentioning something like this, but from what I understood this is only a common problem with dual VR boards and not with the standard VR circuit.
Grounding the microcontroller on the MS and even powering it directly via the MS does also not seem to make a difference.

[jsmcortina]

Noise filtering settings perhaps? Try turning it all off.

James

[Dennis_Zx7r]

Noise filters are completely off. I didn't think this would affect the (nonexistent) signal at TSEL?

I previously forgot a picture of the crank signal not working.

[jsmcortina]

You are correct. The firmware settings will have no effect on that.

Just to check, with the pots set for a VR signal input, you need an input signal that goes negative.
Do you get a signal if you set the pots for a "hall" signal with the 2.5V at the test point?

James

[Dennis_Zx7r]

Well, that's a relief. With R56 turned back about 5 rev, and I now get a nice signal every time. It's kinda obvious in hindsight


Many thanks!
Is there also an easy way of processing the signal so that retuning the pots could be omitted?

[jsmcortina]

Simulating a "VR" you mean? The JimStim uses capacitive coupling and generates a signal that goes negative.

James

[Dennis_Zx7r]

I just googled cap. coupling and played around with some capacitors and my osci. I'm amazed again of how ignorant I am of some electronics basics.

For anyone else having this issue, put a capacitor in series (yes, in series!) between the oscillating 0V to 5V output from the stim (crank signal) and the crank signal input of the MS. I used a 1µF ElCo as this just shifted the DC bias to around zero and kept the rectangular signal. A 100pF ceramic cap also seemed to work, but averaged the signal quite a lot. Any ElCo from .22µF to 100µF seemed to work (220µF did not!).
+ = Stim output, - = MS crank signal pin.



However, this did NOT work when the MS was powered. There's a DC bias of ~3V at TSEL with the stim not connected. When the stim is connected without the cap, it's the 0-5V you'd expect. With the cap, there's the "AC-like" signal as in picture 2, but it now has the 3V offset.
Played around with cap sizes and found that a 1000µF was enough to make it work most of the time. So I fitted a 3300µF cap I once salvaged from an old radio, and this does the trick reliably even with the pots fully open like I usually have them.

---

TLDR: I put a huge 3300µF ElCap in series with the crank signal output from the stim (+ = Stim, - = MS crank signal pin). Works now even without fiddling with the pots (fully open)

[Dennis_Zx7r]

It's not working anymore even with the huge cap. If someone can explain to me on why this seems to have worked for some time and not anymore, I'm interested.
Decided to get rid of the cap and to turn the pots again for now, as this seems to be the only way for it to work every time.

[billr]

I was kind of surprised it worked with that big electrolytic cap, they are so leaky and the value is probably way mismatched to the rest of the circuit . Perhaps the cap has now "failed" because voltage across it was going into negative polarity?