MicroSquirt.com read only forum (Jan.15,2018)

evanevan wrote:...I found out that I've been struggling most of this time with bad wires off the VR sensor, when the trigger cover is torqued on us loses VR sync sporadically leave it loose and cranking stays steady. The tooth analyzer looks much better as well:

That TachRef plot looks better, and the polarity looks OK too. The issue with the VR sensor might be crankshaft end float, too large a wheel, or a non concentric wheel. I do know norzilla had issues with his hall effect cam sensor on VR2 and a VR crank sensor, but this was at a specific rpm, like 800vtwin had beyond his expected redline, and these were noise/AC spike related IIRC. I had problems with cranking rpms, due to inadequate cam sensor VR peak to peak voltages, because the sensor gap was too large and the cranking speed necessary to make the required voltages was outside the rpm of the starter. It is possible following on from my experiences, that compression loads slowing the crank down, also vary the peak to peak voltages, as well as the spacings (which you would expect) and so that is why you are seeing greater peak to peak variances....one way to see if this is the case, is pull the plugs out and do a cranking test, or record a datalog.

I also had issues with my shared grounds (pin31) which I altered with a diode

I would still try and do a datalog, as now you are getting cranking rpms, you should be able to record the other info, like trigger± and tachCount etc, which will be your next step in diagnosing your problems.

After looking into it, both norzilla's problem and mine was the cam sensor. A 5v pullup cured it.

I also run cranking pulse tolerance at 200%. This gives me a rock solid 1 rev start everytime.

I also run my cam ground off of engine ground with no side effects. This keeps the vr ac voltage out of the cam supply power. You could even run it off sensor ground.

Just to be sure, make sure you have the latest code and ini files. And NEVER use an msq from an older code/ini. Always make it from scratch. I had a lot of headaches and hard starting issues from that. Your wue problem could be a mismatched ini or corrupted msq.

Thanks for the help guys. I really appreciate it!

1) I tried re-downloading 2.890 software and starting the msq from scratch. That didn't help my WUE issue.

2) I tried grounding the cam sync to chassis ground. Still no difference.

I haven't tried the pull-up resistor. Should I try adding a 10kOhm inline with the VRIN2+ ? I'm not too savvy in electronics...

Below is what happens once it starts. Note that CLT, MAP, TPS are constant:




For lack of resources, I resumed starting the engine using only my trusty 16-1 Crank wheel input (no cam sync). I am running probably 5x the fuel I need during cranking and possibly after start. (It floods the engine in cranking and can run on the leftover fuel with the inj unplugged). It revs up to 2600rpm dumping black smoke after start. Can anyone with similar displacement verify enrichments with my msq attached below? (I'm using mostly default values for enrichment)


Thanks in advance!

That 16-1 wheel is looking a whole lot better, there are no trigger± changes, tachCount is steady with no resets. You might find it beneficial to put a flat advance curve in whilst you get a steady idle.

Black smoke is a sign of too much fuel, and you possibly have your cranking & priming pulses too high. I had the same using the defaults too on my GTS1000 on day 1. My WUE was also too heavy, required fuel and default VE table numbers were way out. I'll look at you msq later, hopefully somebody else will before, but the quickest way to mod this is just chop your req'd fuel down...until you get idling first and reliable starting. Then bring up your required fuel to what is suggested, and alter your VE numbers. Don't forget the calculated required fuel works for a VE number of 100, so you can halve the required fuel and run a VE of 200 IIRC.

PS Just looked at the raw .msq file and based on my experiences, with a default VE table your req'd fuel should be nearer 9 not 15.6, your crankCold starts at 9msec, mine was nearer 5..,my primePulseCold was 2 yours is 6, so I'd reduce the slope of your cold start enrichment, and warm up enrichment too.

Thanks! I realized that I didn't set the required flow correctly. Good info on your enrichment numbers! I must have another look once I find a new starter relay

Thanks again 24c. It's much better now but needs some tweaking on the warmup enrich slope to get consistent starts from cold to hot..

It's been 13 months since I've posted on this thread. I have just completed the build of a 2 seater offroad vehicle for this engine to go in, and I'm working on getting it to fire up again. I'm getting significant noise on my crank sensor's VR sensor output.

First thing's first, here is a picture of my 16-1 crank wheel:


It is oriented so that TDC of Cyl 1 occurs first tooth after missing tooth. The missing tooth synchronizes nicely with slowing rpm from the compression of the motor to make a visual distinction. See oscope trace below.
+/- 2.1V amplitude.
Compression tooth to tooth gap 22.5ms.
Regular tooth to tooth gap 7.75ms.
Missing tooth gap 45.5ms.

Am I missing something but how can I not register a tach signal during the Compression tooth to tooth gap with even 150% tolerance added to the Regular tooth to tooth gap?
(150/100*7.75)+7.75 = 19.375ms which is less than 22.5ms



However the noise/scatter creates extra pulses seen by the tach ref program. This lead to poor crank sync in megatune experimenting with 75%-150% cranking pulse tolerance.




I tried adding a capacitor across the VR+ and - but it did not clean up the noise at all. I attached my .msq and a data log.

Any recommendations? Very much appreciated.

evanevan wrote:...
Any recommendations? Very much appreciated.

If compression is slowing you down, just try without plugs first and a jump battery, while you take more logs etc.

Next try a fixed advance table, you might have too much advance under cranking, I did and it was stalling the motor and making things worse. By operating a fixed advance you'll be able to time the engine properly later, when you get a stable idle.

There have been some advances in the VR circuitry lately, have a nosey around the forum and check out these threads, as they'll have tips to deal with noise, and a better explanation of the workings.

PS You don't want a capacitor, maybe an inline resistor.

I think you are getting some starter noise during the missing tooth.

Ok, back to the basics then...

Removed all the sparkplugs and I'm still not getting a rock steady rpm sync during cranking. There still appears to be slight noise during the missing tooth, below is a scope trace. I wrapped the starter wire and the VR output wires with aluminum foil but that did nothing. How can I reduce this noise? The forum threads I have seen with circuit mods were for first version of the microsquirt correct? I still added a 10kpot inline of the VR+ output and it did not seem to attenuate the signal at all. What could I be doing wrong?



Crank pulse tolerance was set from anywhere from 50-100% with no real improvements.

I tried adding falling cam sync with the 16-1 crank wheel, and that wasn't any better.