Still losing tach; looking for new theories (Solved!)

[SQLGUY]

Hi folks,



I still have some resynch/tach loss issues with my 36-1 wheel. Moving the pickup further away from the wheel, and deepening the tooth troughs seemed to help initially, but I was still seeing tach loss at specific RPM's and only when the engine is warmed up. I have further cleaned up the teeth, further deepened and widened the troughs, and moved the pickup a bit further back, but the behavior is the same.

One strange thing is the the loss of tach seems only to occur around 1100 RPM, 3000RPM, and 6000RPM, and, again, only when warmed up. When cold I don't notice any tach loss.

When I lose tach, the trigger +/- increments, and the RPM's drop to 0 for a couple of milliseconds. When accellerating and hitting this at about 3K RPM, though, it's pretty annoying. I haven't data logged that (have to strap my laptop to the seat), but I will probably try it tomorrow to see if it gives any additional clues.

Does anyone have any other ideas about what could cause this? Could it be, perhaps, something related to dwell time and the VB921's getting hot, or their noise inside the uS case causing it to falsely "detect" an additional tooth? Any thoughts on the significance of this happening at specific RPM's?

Thanks,
Paul

[Bruce Bowling]

Does anyone have any other ideas about what could cause this? Could it be, perhaps, something related to dwell time and the VB921's getting hot, or their noise inside the uS case causing it to falsely "detect" an additional tooth? Any thoughts on the significance of this happening at specific RPM's?

At this point I would also think that ignition coil flyback noise is getting coupled back into the uS. The best way to determine this would be to move the ignition drive outside of the uS, using an external ignition amplifier. You can either convert the uS to logic-level or keep the drivers but use an external pullup resistor on the IGN lines (and invert the output drive signal) to give a logic output. The external Bosch driver that is documented on the uS site works very well for this setup, it will go to 13KRPM+ all day long with uS...

[old guy]

You might want to try some Mazda Miata coils as mentioned in this post.
http://www.microsquirt.com/viewtopic.php?f=89&t=23110

[SQLGUY]

Does anyone have any other ideas about what could cause this? Could it be, perhaps, something related to dwell time and the VB921's getting hot, or their noise inside the uS case causing it to falsely "detect" an additional tooth? Any thoughts on the significance of this happening at specific RPM's?

At this point I would also think that ignition coil flyback noise is getting coupled back into the uS. The best way to determine this would be to move the ignition drive outside of the uS, using an external ignition amplifier. You can either convert the uS to logic-level or keep the drivers but use an external pullup resistor on the IGN lines (and invert the output drive signal) to give a logic output. The external Bosch driver that is documented on the uS site works very well for this setup, it will go to 13KRPM+ all day long with uS...

OK, but why only when warm? And, if this is the case, what else could I do to improve noise immunity within uS? The ignition coil drive lines are coax as it is. I would assume that the VB921's have damper diodes built in. What about opto-coupling their drive?

Come to think of it, I did do some testing with freeze spray inside the uS and don't remember any improvement with cooling anything there. I'll have to get some more and try again, though, as that was before the initial improvements from tweaking the pickup and tooth troughs.

With the small amount of space I have, solutions that require additional external components are not my first choice, especially when the behavior seems to be just over the borderline from acceptable.

[SQLGUY]

You might want to try some Mazda Miata coils as mentioned in this post.
http://www.microsquirt.com/viewtopic.php?f=89&t=23110

Thanks.

I saw those. They look really cool, and I probably would have sought them out initially if I'd known about them. I'm rather loathe to rip and replace the work I've done now, though. Still, next time I'm at the U-pull-it yard, I'll definitely check to see whether they have any Miatas.

[grippo]

Way back when I was having similar problems with sensors I was told by an manufacturer, and observed it in my own car, that the VR sensors were temperature sensitive, but I can't remember in what direction or whether it affected just the peak voltage or the shape of the signal, much more likely just the voltage.

As far as the external ignition drivers - just a thought. This needn't be a permanent solution - just something to try and if not successfull, you have learned something. You can do it with junk drivers and duct tape temporarily. If it is successful, then you at least have an alternative that you know will work if you do take the trouble to install it properly.

[sportage4x4]

i've done several efi conversions on these older air cooled inline4 motorcycle engines. most are KZ kawasaki's tho.

i use a 2.5" 36-1 wheel mounted just like you have. i use the same style of VR sensor that you have as well. from experience, i think your problem is in the VR conditioner.

i would build a small external vr conditioner, and feed its output to the opto circuity in the uS. i've had great success with the NCV1124 from ON Semi. follow the datasheet with regards to selecting the two resistors and one capacitor. once setup correctly the vr conditioning is bullet proof. even accidentally moving the air gap, etc has no effect on its output.

and get rid of those copper core wires never had good luck with those.

[SQLGUY]

Just to give you guys an idea of how frustratingly close to OK this is. I did a datalog to try to catch the 3K RPM glitch (attached). The bike was already warm at the start and I kept it close to 3K the whole time from startup, moving back and forth a bit near the RPM that tends to trigger the glitch. Finally after 50 seconds or so, it glitched.

During my testing today, I saw Trigger +/- both increment and decrement when glitching. I also reduced my dwell a bit; this kept the VB921's from getting warm to the touch at all, but didn't seem to make any difference in glitching; neither did separating the wiring harness and moving the coax ignition wire further away from the rest of the bundle.

The one funny thing in the log is the negative spike for the CLT (oil temp, actually) reading that happens about 1s BEFORE the RPM glitch. Does this bring anything to mind?

Hey Sportage, I was thinking about a separate VR processor as well... I was looking at the LM1815, but I'll look into the NCV1124. I should note, though, that I've already modified the stock uS input circuit a bit: replaced D5, removed C30, and, recently, tried a 27K resistor in C30's place, which seemed to help things a bit. Moving the pickup back or forth a decent amount doesn't seem to make any difference. BTW, why don't you like the copper core wires? These are done with resistor caps.

[sportage4x4]

i see the resistor caps. i've always had problems when high secondary voltages are required (big bore, high compression, nitrous) with resets. i usually end up with spiral wound wires, resistor caps and resistor plugs. its over kill, i know, but cant have the rider complain about misfires down the 1/4mi.

i've tried the LM1815, and IMO, too hard to setup. the NCV1124 is a walk in the park. you could even, in theory, set it up with trim pots instead of fixed value resistors to allow for 'tuning' on the running engine. the Rinput pot would have to be of some decent wattage rating tho, i usually use 1/2w.

[SQLGUY]

I looked up the NCV. Seems pretty straightforward... the one thing I don't understand from their data sheet, though, is selection of C1. I don't understand why they're picking a cutoff frequency of 145Hz for the low pass. Wouldn't this be based on the expected tooth rate and not just on the 10KHz "guaranteed" voltage from the pickup?

For mine I'd expect an upper limit of 6000 or so Hz (teeth per second). So I would have expected I'd want the low pass filter set around 6000Hz, but they seem to be using the low pass filter to divide down the pickup voltage... and I am not clear where they defined the amount of attenuation they want to see here - specifically the .29V.

Do you use filter caps in your applications?

Thanks,
Paul