MS2extra resets and battery voltage increasing

[RoundelWrencher]

Hello,

I'm struggling through getting my '87 BMW 535is turbo up and running and have already run into several issues. Some issues are resolved and some are still there or parts of the MS are disabled until I know everything else is working correctly.
Running MS2extra 3.3.1a release firmware on a diy MS2 V3.0 wasted spark with relay board and harness from diyautotune.com.
Engine is a 6 cylinder BMW M30 hybrid (B34bottom end with B35 head/cam/intake) using the stock BMW B35 harmonic balancer trigger wheel and VR sensor.

It has ran and been semi-driveable a few times so far but rescaled the fuel table and now ran into this problem:

Vehicle is started cold and beginning to warm up, coolant is around 128F and it starts to cut out and I can hear the relays on the relay board clicking and then MS and the engine just shuts down. Up until this point after the initial fire up, the engine was running fairly normally except the in my VE table is all wrong due to rescaling the table and required fuel etc. Once I was able to get it running and adjust the fuel so it wasn't running lean, I let the engine begin to warm up and noticed the problem. I was running a datalog during this warmup as I had just rescaled the fuel and wasn't sure how it was going to act. I watched the log and noticed that the battery voltage was fairly steady at around 13.7volts most of the time. I have not yet done the battery voltage correction in Tuner Studio. I also noticed that during this short log there were 8 resets and on most of them, the rpm signal could be seen to drop about 2-300 rpm. The voltage reading by the MS begins to climb up quickly towards the end of the log up to about 18V right before shutting down. I used a meter to check the actual voltage while the engine is running and compared it to a gauge I set in TunerStudio and the voltage increasing is only happening in the MS, not the vehicle. I recently did replace the worn out voltage regulator in my alternator as it was only putting out around 12.5 while running confirmed with a meter on the battery. The voltage I am seeing at the battery or alternator with it running now is around 14V most of the time. My question is this, could the resets be causing this voltage increase issue? I took the MS out and performed all the 5V checks with the processor out hooked up to a JimStim with an AC power supply and all were ok so I don't think I have an issue with the 5V regulator. Only other thing of note is that I had changed the lag factor of the battery voltage reading higher the night before I had this issue up from 50 or 60 to 90 I think. When I look at the log, it looks like the coolant temp signal jumps up very slightly just before the MS voltage dips slightly, not sure if they are related but in several places it appears that way (I am by no means a datalog reading wizard though).

The relay board is fed 12V and 12V switched from the oem fuseboxs' 2 extra add on circuits and the ground is on a clean bare ground point on back of cylinder head. The crank sensor wiring is the shielded twisted pair that comes from diyautotune and is as far away as I could manage to get it from the spark plug wires. The ignition coil is the 6 cyl wasted spark coil pack that diy sells and I currently don't have a capacitor on the 12V coil wire (yet).

I have a screenshot of a small crank trigger log I can post if anyone would like to see it. Looks fairly clean but I didn't set the pots with a scope yet and have a 10K resistor on the VR input wire as per diyautotune.com recommendations.

Does anyone have any more input on this voltage increasing issue? Are the resets related to the voltage at all or possibly the coolant temp sensor voltage?

Thanks in advance and forgive my newbness!

I have my msq here:

7-31-2014 double req fuel cut VE by half for resolution.msq

Here is the datalog:

8-2-2014 AM starting rough_modified.zip

[RoundelWrencher]

I looked at a previous log before this voltage increase issue popped up and I have a fair amount of resets in that one as well.
Also worth noting that when the rpm signal dips, the signal on the air temp and coolant temp seem to spike slightly just before the rpm drops. I guess this is an indication that there is interference "noise" getting into one of or all of my sensors. Planning on re doing my air temp/coolant temp wiring and try to re-route everything as far away as possible from power wiring.

Another couple of questions I had was about using resistors. I saw a couple of threads where people had said they had used small caps on each of their spark output wires and that this had solved random misfires and possibly some noise as well. Does anyone know if this is a recommended option? The caps were fairly small if I remember but don't know what size/voltage I should use if this is recommended. Also, what would the recommended size of cap be for the power wire to the ignition coil? I see references to 20uF caps but no voltage ranges. I found a 22uF/35V DC cap at Radio Shack but I seems fairly small. Just want to make sure I'm using the right pieces and not having to re-do this stuff down the road.

[Matt Cramer]

A rising battery voltage can also indicate a short or failed component is pulling down VREF, causing the battery voltage to look higher in comparison. Eventually, VREF drops so low the processor shuts down.

[RoundelWrencher]

Matt,

Thanks for the reply. Are you saying a shorted component/wire that uses VREF or a short internally in the MS?
If it's internal in the MS, can you shed some light on what I can check for diagnosing what component(s) may be causing this?

[Matt Cramer]

Could be either one. I'd first try ruling out external parts by seeing if you can repeat the problem on a Stim.

[RoundelWrencher]

Ran a couple hour datalog on the stim tonight with the AC power supply. Voltage held steady the whole time between 10.9-11.0volts.
The log showed there were 3 resets and showed reason #2 (missing tooth at wrong time) which is strange as I don't have any rpm when hooked up to the stim since I'm using a VR sensor.

I'm re doing all of my sensor wiring and using shielded microphone wiring for the air and coolant temp along with the shielded crank sensor wiring from the diyautotune harness. Planning on making the TPS wiring twisted together and possibly wrapped with foil tape and running all the sensor wiring totally separate from the injector harness and power supply wiring. Making certain all of the sensor wiring is as far away from the plug wires as possible. Re doing both the ground to the relay board directly to the engine block on a clean surface with thicker gauge ground wire. Adding additional 2 gauge ground wire from block to body lug where battery grounds to body. I have enough 2 gauge wiring to re-do my nearly 30 year old power wiring between the alternator/starter and battery as well with all new terminals.

Any other ideas as to what could be causing this issue on the vehicle side of things or what to look for as far as the voltage suddenly increasing as seen by the MS?

[Matt Cramer]

Ran a couple hour datalog on the stim tonight with the AC power supply. Voltage held steady the whole time between 10.9-11.0volts.
The log showed there were 3 resets and showed reason #2 (missing tooth at wrong time) which is strange as I don't have any rpm when hooked up to the stim since I'm using a VR sensor.

Whoa... resets, or lost sync? They are totally different problems. A reset is when the processor shuts down and restarts, and cannot be caused by missing teeth at the wrong time. A loss of sync means it lost the RPM signal.

[RoundelWrencher]

Whoa... resets, or lost sync? They are totally different problems. A reset is when the processor shuts down and restarts, and cannot be caused by missing teeth at the wrong time. A loss of sync means it lost the RPM signal.

Oops, my bad!!... Lost sync count is what was happening. Maybe I was unclear on "lost sync" vs "reset" I guess looking back at the other datalogs from the engine running it was actually only losing sync and not actually resetting. Can losing sync be attributed to electrical noise or is it an indication that I don't have my VR pots adjusted correctly or both?

[Matt Cramer]

Either one; try capturing a tooth log of the loss of sync.

[RoundelWrencher]

Will do as soon as I get it back up and running.
Thanks for the help Matt!

[RoundelWrencher]

Well, I redid all of my sensor wiring and grounds. I ran shielded microphone cable wiring to the air temp and coolant temp sensors. Re wired the ground wire to the relay board directly to the engine block along with the LC2 grounds and all of the shield wires from the crank/air/coolant wiring. Ran an additional new 2 gauge ground wire from the engine block to the batt neg terminal along with another from neg terminal to the body. Sanded all paint from body ground point and cleaned all ground points on the block with sandpaper prior to installing wires.

Engine fired up and ran fine yesterday and was warming up. Had a DVOM hooked up at the MS power supply to the relay board and was seeing a steady 14.1volts which was .4-.5 volt higher than I was seeing before cleaning and adding additional grounds. While the engine was warming up for several minutes I did not have any sync loss errors and the battery voltage reading was steady. I calibrated the battery voltage with the vehicle running by watching my DVOM and adjusting the voltage in TunerStudio until they matched. Noted the engine seemed to run much smoother once the voltage was corrected. Saw no change in the batt voltage reading the MS was seeing while the engine was running for approx 5-10 minutes. Coolant temp got up to around 120 or so.
I looked at the trigger logger with the engine running and my VR signal looks exactly like a "normal" signal with a single tall pulse and all the others shorter and at an even level. I was only able to rev the engine up to just over 2k due to my fueling being off so don't know if I have high rpm issues yet but below 2k, it looked normal.


NEW ISSUES/ QUESTIONS....
I have my ecu and my relay board installed in my glovebox as this is where the stock DME was located in the BMW. I had my glovebox open so I could hook my DVOM up to the relay socket for the idle valve as I am currently running without a idle valve. One of the trace wires on the PCB of the relay board overheated and burnt while the engine was running and let the smoke out of it. I shut it down immediatley and removed the relay board to inspect. This now the second time this has happened to me on this car. The first instance, I figured I had a setting in TunerStudio set incorrectly and blamed it on my "learning curve". On the last instance this happened, the coil had been damaged from overheating from the dwell and battery voltage correction being set wrong as well. The battery voltage issue has been corrected and my dwell was set correctly (as far as I can tell) based on what I was told via email from diyautotune about their IGN6 coil. I was told the dwell was 2.5ms but now when I go onto the website, there is conflicting info on the coil. It now says "reference dwell 3.4ms" and says "arc duration 2.5ms" on their website. I have the dwell set to 2.5ms and the spark duration set to 1.0ms as this is the default setting in TunerStudio for the spark duration. Highest the dwell went on my datalog was 3.6ms and most of the time it was down around 2.5ms.

I bypassed the relay board connections for spark A,B, and C by running the wiring directly out of the ECU harness to the coil wiring and got the engine restarted. I noted that it was running rough but that it had taken several times to get it to start and thought it might have been loaded up on fuel or something. After a couple of minutes of trying to get it to clear out and run smoother, there was more smoke, this time from under the hood. Shut it down again and disconnected battery. Smoke was coming from the coil pack and the epoxy was dripping out of it. Now on my second smoked relay board and second smoked IGN6 coil. I pulled the ecu out and did some checks. I had previously (first time I smoke the relay board/coil) found that one of the BIP's mounted to the case of the ecu for spark C had continuity between the case and the transistor. I had re-installed that BIP and made sure it had no conitinuity to ground on either Spark B or C BIP's. Yesterday when I took it out and checked I again had continuity between the case of the ecu and one of the extra BIP's. Could this cause the transistor for one of the spark outputs to overload and burn up the relay board? This is my best guess...

I re-did the BIP mounting with an additional heat sink on the bottom of the board. Put tape on bottom and edges of the heatsink so it can't touch the board and mounted the BIP's and the transistors on top of the board using the long plastic screws and nuts to make SURE they can't short to the case or the board or anything else. I did use the mica insulators/heat sink compound and the plastic washers in the BIP mounting holes.

Does my theory of the transistors shorting to the case of the ecu and overloading the spark output wiring explain the smoked relay boards and coils?
I have $400 spent on 2 relay boards and coils so far and now will have to shell out for another coil. What should my dwell setting be and spark duration if I decide to continue to use the IGN6 coil from diyautotune

[Matt Cramer]

Yes, a short on the BIP373 tab can cause exactly the damage you describe.

[RoundelWrencher]

Ok, well at least I'm not going crazy. Feels like I've been chasing this down for quite some time now. Thanks!

As far as the dwell and the spark duration settings go Matt, what should I be setting this at? This is MS2extra if it matters. Should I go off the specs listed for the IGN6 coil on diyautotune for the dwell or should I use the 2.5ms that was recommended via email? What about spark duration?

[Matt Cramer]

2.5 ms is a play it safe number; 3.4 if you're going for more spark energy. Use the arc duration number (2.8 ms) for max spark duration.

[RoundelWrencher]

2.5 ms is a play it safe number; 3.4 if you're going for more spark energy. Use the arc duration number (2.8 ms) for max spark duration.

Ok, thanks again for the help! I'll order another coil and possibly a couple more BIP's and see how this goes. I won't add to this thread if the battery voltage climbing problem doesn't re-occur.

[freek corsa]

Yes, a short on the BIP373 tab can cause exactly the damage you describe.

I think my Megasquirt 2 V3.0 has exactly this problem.

How do I know what parts to replace?

[Matt Cramer]

I think my Megasquirt 2 V3.0 has exactly this problem.

How do I know what parts to replace?

If you have already fixed the short with a proper mica insulator under the coil driver, chances are the remaining damaged parts are going to have visible burn marks.

[freek corsa]

Thank you for your reply
I didn't find parts with burn marks on them.
I found in the v3 build guide the clock circuit and battery voltage are build in the same time.
Parts R3 and R6 are for battery voltage measuring?
I checked the resistance, these where within the 5% margin.

If I power up the megasquirt and look at the voltage gauge in tunerstudio, the value goes from 12 to 17
After a time ( lets say 5 minutes ) the reset/restart problem begins.

[Matt Cramer]

Check what the VREF is doing at the point when the battery voltage increase is reported.

[freek corsa]

I will do

I just remembered the Vref is dropping below 3 volt and then the megasquirt goes offline