Timing jumping around with VR input and direct coil control

[jamesawooten]

Hi All,

We are working on a Toyota Silvertop 4AGE using MSII and the V3.0 board. The NE VR signal is being fed to the MSII thru factory coax from the distributor which has had all but 4 of the NE wheel teeth ground off. The factory coil is being directly driven from MSII.

Initial timing was set to 10 degrees and the ignition table was set up for 16 degrees of advance everywhere below 2000 rpm (to ease in setting up the initial idle).

When a timing light is put on the motor at a steady 1000 rpm, I see the timing jittering anywhere from 10-20 degrees, not very steady at all.

Our initial setup of the zero crossing and histerisis pots on the V3.0 board were both fully CCW, but found that the zero crossing had to go CW several turns before cranking pulses were detected. Since then, we have not changed anything.

I'm assuming that there are some further setup steps with the pots, as well as some megatune settings that we may need to play with in order to get things to settle down.

Does anyone have any idea what those steps are?

Thanks,

Jim Wooten

[jsmcortina]

Try reversing the wires to the VR sensor.

James

[jamesawooten]

Hmmm, OK, I'll give that a try, although VR- is definitely going to ground right now.

Are there no proceedures for adjusting the 2 pots on the V3 board?

[Bernard Fife]

Are there no proceedures for adjusting the 2 pots on the V3 board?

jamesawooten,

The procedure depends critically on the signal of the VR sensor and any noise it pick-ups up, so it's hard to say how poeple should go about it, other than to say to adjust them for the 'best' input. That means it picks up all the true signals, but none of the noise. How you adjust the pots to do that is difficult to specify in advance. So it's a trial and error thing.

If you have an oscilloscope, and can look at both the input signal and the output to the processor at various engine speeds, this is a lot easier.

However, in general Bruce has said:

- For the VR circuit, when the VR sensor transitions from negative to positive, the crossing is zero crossing, and this is the intended polarity for the VR sensor trigger. The VR circuit inverts it so the neg-to-pos VR transistion becomes pos-to-neg digital output to the processor IRQ line, this is what triggers an IRQ. Now if someone is using another trigger source feeding the VR sensor (say the EDIS PIP or some Hall sensors) the polarity may need to be changed, this is when they use the other op-amp output instead - this polarity follows the VR sensor so a neg-to-pos VR sensor yields a neg-to-pos digital output.

- The VR circuit should be capable of triggering in the millivolt region, if it requires volts to trigger then I would suspect a component. I would trace the signal thru the input stage and thru the PNP transistor with a scope and compare with the set threshold transistor signal, an trace into the opamp.

- To adjust the screws for the pots for VR sensor operation, the pot on the transistor base (R56) needs to be turned to zero volts and the pot on the op-amp feedback (R52) needs to be turned to maximize resistance (i.e. give 100K feedback resistance). This should be with both screws turned fully counter-clockwise.

- The VR sensor settings for the two pots are both screwed all the way counterclockwise gives zero crossing detection and 100K of hysteresys.

- Try this on the car and if there is noise on the tach VR signal then start cranking the hysteresis pot (R52) clockwise until the noise vanishes. If you crank it too far then there may be too much hysteresis and cranking signal may be difficult.

- Now, on the other side, the hysteresis pot (R2), even cranked all the way counterclockwise, will introduce a little hysteresis. Removal of the pot will remove all histeresis and both side will trigger directly at zero crossing. This is the effect of the assymetric hysteresis - it only affects the pos-to-neg VR direction, not the neg-to-pos direction. The intent is to trigger on the neg-to-pos transistion direction, and the VR sensor needs to be wired such that this direction aligns up with the VR sensor pointing at a tooth on the reluctor.

- One can inprove the common-mode rejection of the circuit. Here is the first thing to try: take the VR return wire, the one you have been connecting to ground, and put in a series resistance, like 47K or so, and connect the other end of the resistor direct to the wiper of the zero-crossing potentiometer (R56), which also goes to the base of Q23. Next move the zero-crossing point potentometer setting off of zero volts to a point of, say 1 volt or even higher (measure the voltage at the wiper). Since the return point of the VR is also at this junction, the VR sensor will also be biased up the same amount, so the circuit will still sense zero crossing properly. In fact it does not matter where the zero-crossing pot is set, the return VR sensor will also be biased the same amount. But what this does is allow rejection of common noise introduced in the VR wires by the op-amp.

- Note that on the stim, the stimulator tach output does not go thru zero volts so it will never trigger the VR circuit. What you have to do is turn the zero-crossing pot (R56) clockwise until the threshold is raised enough (roughly 0.7 volts or so) to cause a trigger.

- Note that the VR circuit should also be able to take Hall inputs, EDIS (there is an inverted output available) and ignition coil primary, by setting the threshold and hysteresis trimmer pots to points yet to be determined. The existing MS V2.2 tach circuit w/ opto is also there for those who want to stick with this. Hopefully people will experiment with the VR input settings suh that we can find trimmer adjustment values which will work for EDIS, Hall, ignition primary, etc.

Lance.

[ChaosAD]

However, in general Bruce has said:

- For the VR circuit, when the VR sensor transitions from negative to positive, the crossing is zero crossing, and this is the intended polarity for the VR sensor trigger. The VR circuit inverts it so the neg-to-pos VR transistion becomes pos-to-neg digital output to the processor IRQ line, this is what triggers an IRQ. Now if someone is using another trigger source feeding the VR sensor (say the EDIS PIP or some Hall sensors) the polarity may need to be changed, this is when they use the other op-amp output instead - this polarity follows the VR sensor so a neg-to-pos VR sensor yields a neg-to-pos digital output.

- The VR circuit should be capable of triggering in the millivolt region, if it requires volts to trigger then I would suspect a component. I would trace the signal thru the input stage and thru the PNP transistor with a scope and compare with the set threshold transistor signal, an trace into the opamp.

- To adjust the screws for the pots for VR sensor operation, the pot on the transistor base (R56) needs to be turned to zero volts and the pot on the op-amp feedback (R52) needs to be turned to maximize resistance (i.e. give 100K feedback resistance). This should be with both screws turned fully counter-clockwise.

- The VR sensor settings for the two pots are both screwed all the way counterclockwise gives zero crossing detection and 100K of hysteresys.

- Try this on the car and if there is noise on the tach VR signal then start cranking the hysteresis pot (R52) clockwise until the noise vanishes. If you crank it too far then there may be too much hysteresis and cranking signal may be difficult.

- Now, on the other side, the hysteresis pot (R2), even cranked all the way counterclockwise, will introduce a little hysteresis. Removal of the pot will remove all histeresis and both side will trigger directly at zero crossing. This is the effect of the assymetric hysteresis - it only affects the pos-to-neg VR direction, not the neg-to-pos direction. The intent is to trigger on the neg-to-pos transistion direction, and the VR sensor needs to be wired such that this direction aligns up with the VR sensor pointing at a tooth on the reluctor.

So if the output of my distributor goes to positive them negative should I reverse the wires?