Frustration - Microsquirt RPM input death

I am super-frustrated w/ my ongoing MB 190E Microsquirt conversion right now, maybe someone can tell me where things are going wrong.  

Quick recap, fuel-only, intending to "use negative coil for RPM input" as it says in the manual, on forums, etc.  So, this being my first MS work, it took a little while for me to discover the "fuel-only" mode in the Ignition menu, and then turning on noise filtering.  Prior to those, I was having no luck starting it, and the RPM gauge seemed pretty whack (no surprise there).  

So, more research, and I finally discovered the correct starting points in the Ignition menu, ie setting it to Fuel Only, and turning on the noise filtering.  Once I did that, I tried again and it started almost immediately and ran for ~ 90s, even idling reasonably well considering it was the first time it ran.  Fortunately, I also datalogged during that run, and I was able to go back and see that I didn't shut it off, it just died.  I noticed that the RPM went from being correct to dropping ~600 rpm in one sample, then sitting there exactly for 6-7 samples, then dropping another ~600, sitting another 6-7, then finally dropping to 0.   After that, I tried to restart, and the RPM gauge in TS never budged off 0.  

Later I checked the coil signal being fed to the MS via scope, and it looked solid.  

I had some other issues as well and was in contact w/ Matt at DIY AutoTune, and to their credit even tho the unit was out of warranty (but also brand new out of the box) they gave me a discount on a new one, which I got installed today. 

My biggest frustration w/ this install so far is that there are at least 4 different, conflicting notes on how to wire up using the coil as the RPM input.  Having tried the first 3 previously, I discovered the 4th version a couple of days ago looking at some additional guides for the Microsquirt on the useasydocs.com.  So, I rewired today using that version, which says OPTOIN+ to the coil negative, and OPTOIN- to +12V.  It mentions "This uses the flyback spike off the non-CDI coil as a trigger (rather than the 12 Volt signal). Doing this reduces the energy dissipated in the circuit, prevent a thermal 'melt-down' of the components inside your MicroSquirt^® controller." Sensing that perhaps that is what had happened, this seemed like the Eureka moment.  

Today I got it all wired up as above, started a new tune from the ground up, got all the Ignition settings in, sensors looked good.  Pulled the fuel pump relay, cranked, RPM gauge in TS showed a nice ~250 rpm.  Excellent.  Plugged the FP relay back in, cranked, started almost instantly, then died.  Added in some fuel via the after-start enrichment, tried again, boom, started right up.  Ran for ~20s.  Then died.  Cranked again - no RPM showing.  Checked data log, same exact behavior.  Pulled the MS plug, checked the input signal, still clean.  The spikes are somewhere in the 20V range.   

So, I'm at a loss here.  I've yet to see any sort of requirement on what the voltages into the OPTOIN pins can or can't be, but apparently mine are too high.  This seems puzzling since it's a simple mid-80s Bosch coil that practically every European car in the 80s and half of the 90s used.  My circuit theory knowledge has atrophied pretty bad (and it was never great anyway), so I don't trust myself to read the circuit and make a pronouncement on what's wrong. 

At this point I'm very seriously considering scrapping using the coil as an input and fabricobbling up a trigger wheel setup.  

Yeesh, sorry for the book...

Have you thought about trying 4 pin HEI module off of your mag pickup (assuming it has one)

I had problems doing fuel only on an Rx7 with a Microsquirt. The problems aren't exactly the same as yours but I had problems when using the rpm input from the ignition coil. 

There was a resistor on the board that on the first start attempt failed, and I couldn't get an RPM signal. The resistor was replaced with a more HD resistor by DIY auto tune, this HD resistor worked for a while but failed (overheated, melted the solder and  fell off the board) at the first or second track day. I resoldered the fuse back on and installed a crank sensor, LS ignition coils, etc have not this issue since.  

I'm just not confident Microsquirt is robust enough to accept an RPM signal from the old cannister type ignition coils.

Yea, it sounds like the problem is the filter in the module to deal with the voltage input.  Assuming that you don't have an ignition box, would you entertain putting one in?  Most of them have a tach out signal, which is perfect for a MS system, a good example is the one V8Kid posted.

I had to add the set up V8 posted in parallel to the existing ignition module to clean the signal up on my MS1. Don't miss the little resistor in the diagram, it is the false coil for the module to "read".

Well I guess you guys are starting to confirm my suspicions, particularly bruceman's experience.  Right now I'm leaning towards putting on a trigger wheel.  The crank pulley is a 2-piece affair, where the actual pulley part bolts on to a separate damper-like thing that bolts to the crank.  

I think I have a couple of different options for adapting a trigger wheel to fit on the front of it (ie, rad side, not block side), which means I don't have to worry about spacing all the other pulleys out.  And there's a few bolts that hold other brackets and such that would be easy to use for a bracket for the Hall effect sensor.  

I wonder if I can track down a spare pulley on ebay or at the local JY...

I'm not super familiar with MB motors/models, but safe to assume you're adding the MS onto an already fuel injected motor, right? 

How does the factory fuel injection pick up the crank position? Seems like that may be a more productive avenue to explore. Does whatever is running the stock fuel injection also have input to the ignition system? If not, you could likely use the stock sensor in it's current place*. If not, it may be as simple as either a bracket for an additional hall/vr sensor, or bringing over whatever the stock crank position sensing setup is from a later version of this same motor. For examples, BMW had multiple ways over the generations of various motors for picking up the crank position, and there are fairly easy ways to swap them in between.

* Unless it's picking up the flywheel ring gear teeth. Some early BMW fuel injection setups were like this, and it's too many pulses for MS. They might have only picked up engine speed and not position (no missing tooth), I can't recall off the top of my head.

Edit: looks like you posted right after I started typing this, and are headed in that direction. that's probably wise. good luck!

bluej said:

How does the factory fuel injection pick up the crank position? Seems like that may be a more productive avenue to explore. Does whatever is running the stock fuel injection also have input to the ignition system? 

The original fuel injection and ignition systems are completely separate, and the original injection system is a continuous injection system, so as best I can tell there is no existing VR / Hall effect sensor that goes to the fuel injection side of things.  However, you got me thinking, and after a little research the X11 diagnostic socket in the engine bay appears to have the "TDC sensor" signal coming into it.  I'm pretty sure looking at the diagrams that the TDC sensor is a toothed-wheel type sensor on the distributor, so that might actually be a possibility...

TheV8Kid said:

That is basically what I use to drive the tach on my RX-7.

The distributor has two pickups.  One drives the tach, the other signals an MSD, and the MSD's tach-out is my Megasquirt's signal.  The MSD has a really awesome noise-free square wave that Megasquirt loves. 

(For rotary people who are looking at me like a confused dog: The MSD, in turn, drives two or three coils in parallel, either two through the distributor cap, or one through the cap for trailing and two for direct-fire leading.  I run zero split as a matter of necessity)

The input resistor on the OPTO circuit is known to go out with certain coils. I install a 1/4 watt through hole resistor in place of the SMD one and it has worked well for me on many installs. Generally these arent' used for fuel only so although the problem is predictable, it isn't well known or documented. 

If your distributor uses a hall effect sensor, using that to the VR input may be the easiest way to fix things. 

Generally all of this is avoided if you went "all in" and did the ignition too, but it seems I shouldn't kick when you're down ;)