MS3 PRO drops off at 5000 RPM, cant figure it out.

[1 fast z]

Ok, I have been having issues with my MS3 PRO. I loose communication with USB at anything over about 3500 RPM. I have switched to serial and it seems to be better but freezes up over about 3800 and then when you bring the RPM back down it will un-freeze. A bit about my setup:

8, IGN-1A coils from DIY, MS3 PRO.

I have placed a lead sheet over the coils to help eliminate RF noise. ECU is away from the plugs and coils. The engine has 8 pyros over the CAN BUS system.

Now for the real problem, on the dyno, above 5000 RPM the motor ends up surging, meaning wont go up any higher and tires to go to zero, and then right back up again. Seems in the datalog it shows it go to zero, but the motor does not physically go to that. I am running a HALL sensor on a 60-2 wheel.

[slow_hemi6]

Sync Loss, you get a bunch of it while cranking and then again around 5000rpm. Sync Loss count is viewable in your log.

[1 fast z]

I am no expert. How do I correct it. Or better yet what's causing it?

[jeffmarsh750]

Your Hall sensor will not work well at high RPM with 62 teeth. Either use less teeth 32 max or install a VR sensor and a resistor with the 62 tooth wheel.

[1 fast z]

Ok, so I put a 1 k pullup between the sensing wire and the 12 volt power going to the sensor. I then used shielded wire going all the way from the MS harness to the sensor. I got a pull up to about 6400 RPM, seem to be sputtery, but looks like was also pig rich. Take a look at the log and see what you see.

What is the MAX RPM with a 60-2 wheel with a hall sensor?

I was told by others this combo would work, but like I said, Im no expert.

[jeffmarsh750]

If you are using a Honeywell GT 101 sensor here are the specs. It needs a 1K pull up. Make sure your target is soft ferrous material. If it is work hardened steel you need to anneal it to soften it. Iron is best.With a 4" diameter your tooth limit will be 25 for a reliable 4,000 rpm limit. So 62 will not work !Note the tooth spacing and depth and target width. The DIY sensor does not need a pull up and acts very similar. The Honeywell will take more heat. I know DIY tested these but on a test bench they perform much better. Add some heat, crank deflection, vibrations and missed triggers are the result. I use an 8 tooth crank trigger and have reved to 12500rpm and no missed trigger signals unless the tire breaks loose and the rate of accelleration on the engine goes way up then the sensor gets pissed off.

[jeffmarsh750]

I might add that Motec reccomends a 12 tooth wheel with these sensors.

[1 fast z]

Ok, this sorta sucks. I get told that this setup will work with a 60-2 wheel, and am now finding out it wont. The diameter of the wheel is 6.5" with a 3/16" thick steel wheel. 60 teeth, missing, 2 teeth and the teeth are .250" tall. So I guess the easiest thing to do, would be to switch to a VR sensor. What are the benefits to using a hall rather than a VR? Or is there none?

[jeffmarsh750]

A VR sensor is more susceptable to noise. Easier to hook up.Make sure the harness stays away from all coils, plugs wires and uses the factory ms shielded cable. I would Install the resistor that came with the MS3 pro between the pos and neg while you are at it. A diesel tach sensor works well with a .020 gap. We use one from a 1990 ford 7.3 diesel tach sensor. They are short, threaded and have a good magnet. Use the hall sensor for your cam trigger unless using waste spark.

[subwoofer]

I found that the 1GT101 needed quite a lot of distance from the toothed wheel to work reliably in my setup (36-1), on the order of 4mm.

[1 fast z]

Ok, so this is what I did. First, placing the 1k between the power and the signal on the Hall sensor (pullup) seem to fix the drop offs at 5k, as shown in my previous datalog. Then today on the engine dyno I would get spark erratic at around 4k with these 8, IGN-1A coils being fired through Megasquirt even after I installed a VR sensor on a 36-1 wheel. Also shows the lost sync count, not sure why. See datalog below. named VR 1, VR2, and VR Sync.

So after that I am thinking that there has to be something wrong with these coils, or how MS is firing them, etc. The dwell is at 4.0 and spark duration of 1.8, which should work. But you can hear on the engine dyno where it is definitely spark breaking up.

So, I then install a EDIS system (since the 36-1 wheel and the VR sensor were now mounted) a long with 2, 4 cylinder coil packs. Now the motor runs smooth throughout the whole RPM range.


Anyone have any suggestions as to why the EDIS system would work great, but the IGN-1A coils would start to break up around 3500-4500???

[jeffmarsh750]

All of your data logs show crank sync loss on start up and then no more. One time on your VR2 log you had to restart the engine or it almost stalled and set a sync loss with reason 2 which means it didn't pick up the cam sensor first. Make sure your cam triggers at least a few teeth before your missing tooth. You can also try a different edge. During your pulls or running up through the gears you do not have a sync loss. The IGN1A coils with 4ms dwell should fire anything you throw at it. Resistor plugs and wires ? You have the dwell cranked way up in VR2 doesn't need to be that much. That will cause problems. Also when you tried to make a pull the AFR went to 10.1 , battery voltage is so low it is amazing it even runs. Maybe that's why the dwell is so high, it has battery compensation.

[1 fast z]

Well went to go do some more testing and MS is not coming on, no communication, no V ref, Fuse F1 is good and has voltage. Ground is good, power is good. Not sure what to do now.

[Matt Cramer]

It sounds like you may have been overdwelling the coils - 4.0 ms is a LOT of dwell on a coil that hot.

As for the current no communication issue: Try opening the MS3 case, putting an aligator clip on the boot jumper, and try reloading the firmware. Let me know what you get.

[knightrous]

DIYAutoTune posted a video back in 2011 of a 36-1 trigger wheel spun at 19000+rpm and the GT1 sensor was picking this up easily.
[EDIT: As per below post, DIYAuto were not using a GT1 sensors, this post may/may not be relavent now]

https://www.youtube.com/watch?v=dD4vuNSO8cg

At 19000rpm, the sensor is detecting 684000 trigger events per minute (1 tooth every 0.087 seconds).
A 60-2 wheel at 8000rpm is only 480000 trigger events per minute (1 tooth ever 0.125 seconds).
A 60-2 wheel should be able to do 11400rpm (same trigger events per minute as 36-1) without issues.

[Matt Cramer]

DIYAutoTune posted a video back in 2011 of a 36-1 trigger wheel spun at 19000+rpm and the GT1 sensor was picking this up easily.

Just to clarify - our sensor is not a GT1. Similar looking sensor from a different manufacturer.

[ms3rx7]

I am battling similar issues, ms3pro hall sensor 36-1 with ign1a direct fire. I get sync loss reason 2 anywhere between 4800-5000 rpm. last one I got was at 5000 rpm on the dot under WOT, high load is the only time i get this sync loss. did you ever resolve yours?

[Matt Cramer]

I am battling similar issues, ms3pro hall sensor 36-1 with ign1a direct fire. I get sync loss reason 2 anywhere between 4800-5000 rpm. last one I got was at 5000 rpm on the dot under WOT, high load is the only time i get this sync loss. did you ever resolve yours?

Can you post an MSQ and tooth log?

[ms3rx7]

Unfortunately I do not have a tooth log, Do have a datalog and msq to provide. see attachment.

[Matt Cramer]

Change the tach period rejection time to zero. I suspect it's filtering out real pulses.