I built an MSII and just making sure the sensors are all good before moving on to fuel.

Everything appears fine, tach is perfectly steady at idle. I can raise it up slowly with the throttle and it is smooth, but as soon as I hit exactly 2k rpms, the tach signal disappears completely and the MS reads 0rpm until I let the actual rpms drop below 2k and then it immediately comes back.

I have the positive crank signal connected to the TachIn on the MS, and as per instructions I shielded it with a wire from pin 2 and connected that wire to the negative crank signal with a 10kohm resistor in between.

I actually used a 10k potentiometer since one set of instructions said to use a 6.8K and another said 10. Figured it doesn't matter to much, but I had one laying around.

Playing the the pot made no difference.

The I decided to remove the connection to the negative VR completely and ground the shield, which to me makes more sense.

That also had no affect. Still steady and accurate and then goes to 0 above 2k.

It doesn't seem to be a noise issue because it is very consistently at 2k rpms.
I also built a JimStim and it goes to 8k before things start getting weird.

I've tried various searches with no luck. Seems like it should be a simple fix.
Suggestions would be very much appreciated.

 

Does your crank wiring look like this?



Also what year is the car and where did you see to add a resistor to the crank signal negative?

 

I actually have a Mazdaspeed Protege, 2003. They share a lot in common with the Miatas and there is a lot more help here than on any Protege forum.

I'll double check my board, but I kind of figured if the JimStim worked, so would the car.

Thanks.

Oh and in answer to the resistor, that came from a write up someone had posted on mazdas247. But also pops up in a few places if you search for this problem with google.

One place I saw it was diyautotune's site, so it's probably not complete nonsense.

"You'll see this problem manifest as the RPM signal suddenly spiking downward at high RPM, and the car acting like it's hit a rev limiter. Putting a 10K to 20K resistor inline with the crankshaft position sensor wire generally clears this right up. Below is a tooth log showing this problem happening to a 36-1 trigger wheel."

 

He is using the VR circuit for the tach and not the opto isolator circuit.

 

I see that now - call me crazy but I was kind of figuring that he was working with a miata

 

Yeah, for some reason I thought the VR was the same for both cars.

Maybe I won't find much help here then. I am pretty sure the cam sensor was the same when I had the AEM FIC.

 

I was just doing a little digging around and maybe this could help you out a bit...

Megasquirt 2 EMS Guide for FS-DE 98+

 

Yep that's where I got the original instructions, which includes hooking up the shield/ground pin to the negative VR crank signal through a 10Kohm resistor.

I honestly don't see the point since I am running a parallel setup so my gauge cluster, obd2, etc still work. The crank sensor is still getting a ground to the negative side.

 

So I don't know if this applies to you guys at all, but I discovered that by playing with the trigger pot I am now able to get all the way to 4050 rpms before it drops.

So it is definitely a trigger problem. I am triggering on the falling edge, but changing to the rising edge didn't help.

The hysteresis pot doesn't seem to have any affect, so I don't think this has anything to do with noise.

It's almost like my tach signal is attenuating at higher rpms. Except in major increments, because minor adjustments of the trigger pot had no affect on the max rpms.

I'll borrow an oscilloscope from a friend tomorrow. I had it for a while during the build, but gave it back once I thought everything was working. That might shed some light.

 

Ok, I had to dig back in my memory to when I had issues with ms2 trying to read the signal from my cavaliers VR sensor. If I remember correctly I had to turn both of the VR pots (R52 & R56) counterclockwise a bunch of times (~15) and then turn R56 about 6 turns clockwise according to the manual. Then I had to wire the crank sensor a particular way, I didn't think you could screw it up but turns out I could. But this caused me headaches when it came to which edge I needed to trigger off of and was a choppy signal. The last thing I'd try would be to swap the VR sensors wires.
It sounds like you are getting a clear signal that isn't broken or intermittent but that the signal is getting so small at higher rpms that MS isn't picking it up. I'd try turning both R56 & R52 counterclockwise 12-15 times and give it a shot, then try turning R56 clockwise a small bit at a time and see if you can consistently get the engine to rev higher.

 

Thanks. That's pretty much what I've done. Most instructions say to turn the pots completely counterclockwise until they click, and then turn each clockwise until you have a clean signal.

There is a Tooth Logger feature in Megastudio that shows you what the MS is reading, with the Y axis being length of time between teeth.

So the missing tooth should be twice as long as the other 35 on my 36-1 wheel. What I am seeing is the tooth just prior to it is considerably longer than the rest.

As I rotate the trigger pot clockwise that tooth gets longer and longer until I lose sync, with sync error 2 which is missing tooth at wrong time.

If I continue turning the pot clockwise, the tooth increases in duration while the missing tooth diminishes, and they swap. I'll even gain sync back again after enough turns, but then it is lost almost at only 2k rpms.

My friend is out for the weekend, so I won't be able to get the oscilloscope until tomorrow night.

I need to figure out what is causing the MS to see two long gaps instead of just one.

Thanks for the help.

 

Great news.

So I left the shield connected to the negative VR wire with the 10k pot inline.
I then soldered in line a 10k resistor on the TachIn line going to the positive VR.

This actually caused my RPM threshold to drop from 4k to 3k. So I then put the shield back onto a ground line, and my threshold raised back to 4k. Although slightly less than the 4050 I saw last night.

So really the resistor didn't do much if anything. So I decided to try swapping the TachIn input to the negative VR signal before giving up until I can get a hold of an oscilloscope, and voila it works perfectly.

Can take it all the way to redline now without a problem. It was pretty jerky though until I turned off the Tooth Logger, so that was fortunately just a bandwidth issue.

Thanks for the suggestions, and now in case any of you guys somehow end up with a VR crank sensor, you'll know to use the negative.

I am not sure if the 10k resistor on the TachIn would have been necessary or not. It doesn't hurt, so I am just going to leave it in, since unsoldering it would be a pain.

The shield really doesn't seem to matter at all. I left it dangling and it didn't have an affect. I am going to leave it on the ground wire since that seems to make the most sense to me. Clearly the 10k pot is unnecessary but I'll just leave it in too I guess.

Here's a snapshot of the noisy signal I got with the negative side.
But now that I think about it, the wiring diagram for the car doesn't specifically say which wire is positive and which is negative. They are both AC signals just inverted from each other from my understanding, so it shouldn't matter.

I just followed the install guides and labeled pin 21 as positive and pin 22 as negative, it could very well be the other way around. There's no way to really tell.

Ok that's enough from me.