Weird problem with losing spark on test engine

[MrDomino]

We're using a MS-II and I'm not sure what code version. If anyone can help me figure this out it'd be much appreciated.

Anyways, I've posted here a few times before about issues with our test engine and you guys were really helpful so I'm hoping you can help me with our latest issue.

We were running our single-cylinder test engine with gasoline last night for our lab class and during the lab the engine just cut out. Fuel is still being injected the engine is producing any torque narrowing it down to the ignition system. We're using a Bosch coil hooked up to the power source for the fuel pump and then the ground is hooked up to the IGN pin on the relay board.

Basically the issue is that we're intermittently getting spark. The engine will run fine for anywhere between 5 and 20 minutes. Looking at the timing light it produces a nice bright consistent light. Eventually the signal will all of the sudden get weaker (the timing light will be much dimmer) and sometimes there is no signal at all (the timing light doesn't illuminate). We can keep the engine spinning because of the dynamometer and sometimes it'll start sparking for a few seconds and then the signal disappears again.

I checked the coil and it was getting 14.5 V on both positive and negative terminals and the resistances looked good. I even swapped in another similar coil and the same thing happened. I also replaced the fuel injector just to be sure it wasn't a fuel related issue (fuel pressure is good).

I'm not sure what to check mainly due to the fact that the engine has run fine for several years. Any ideas? Is it potentially an ECU issue?

[Matt Cramer]

What sort of ignition module or internal coil driver are you using for the spark?

[MrDomino]

I didn't put this mess together but from what it looks like there is just a connection from the IGN terminal of the relay board to the negative terminal on the coil and then the fuel pump power (14.5 V) goes to the positive side. The fuel pump power comes from a power supply and then I'm assuming the relay board grounds the coil every time it wants to send a pulse. We're using a shaft encoder to input RPM into the ECU.

I'm not sure what you mean by "internal coil driver." Is this something on the ECU? I have it open so I can check quickly if needed. I'm not too familiar with the inner workings of the ECU since I didn't put it together.

Thanks.

[MrDomino]

Update: We just ran it again to see if we could narrow it down and it ran fine at an intake MAP of 70 kPa for 40 minutes. The second we increased the intake MAP to around 92-95 kPa the engine began to sputter out and the timing light showed that there was no pulse being sent to the spark plug. The engine would mostly just motor but every few seconds it would fire for a second and then sputter out again. The second we decreased the MAP to below 90 kPa the engine would fire right back up and run fine.

Wat

[DonTZ125]

Can you post your .msq (MT or TS tuning file)? Something very weird is happening; I almost want to say it sounds like you've got a table switch triggering to an un- (or badly) programmed secondary spark table. (scratching head)

[Matt Cramer]

I'm not sure what you mean by "internal coil driver." Is this something on the ECU? I have it open so I can check quickly if needed. I'm not too familiar with the inner workings of the ECU since I didn't put it together.

Thanks.

Yep. Normally it's in the Q16 slot. See if you can get a picture of what you have there and post it.

[MrDomino]

Ok, so I spent some time searching around other threads and haven't yet found anyone with a similar issue so the troubleshooting continues.

1. I've attached a copy of the maps we use to run the engine although it's been the same one since August 28th (I made some minor fuel corrections a few days ago for the low MAP sections since we ran it for the class there due to the spark issue). I don't think the file has any problems.

2. I opened up the ECU and checked out the ignition coil driver (Q16). I think we may have a winner. Check out the pictures. It looks like it's a little toasty. I already scheduled the electronics workshop here at the university to swap it out with a spare tomorrow after the lab class is over (I hate soldering things).

Also, I couldn't really read what was on the transistor (I kept going cross-eyed) but found some spares along with the sheet specifying what the driver was.

VB291 High Current Ignition Coil Driver 350V 7.5A ISOWATT220 (P/N 497-2716-5-ND)

[MrDomino]

.MSQ File attached (limited to 3 attachments on the previous post)

[Matt Cramer]

That VB921 looks like it's overheated at some point from running too much dwell - I can tell by the brown goop around the nylon nut. They can start to flake out when you do that to them, and eventually die. The Bosch BIP373 we use in newer kits is a lot tougher.

[MrDomino]

We're going to solder in another VB921 just to see if it solves the issue but would it be possible to swap in a BIP373 in the near future?

[Matt Cramer]

We're going to solder in another VB921 just to see if it solves the issue but would it be possible to swap in a BIP373 in the near future?

They're a direct drop in. You will need to use a mica insulator with them, but that's the only change needed.

[MrDomino]

Ok so I had the new coil driver soldered in and when I started up I got the following error (see picture). I know this is something to do with the firmware but I'm not sure exactly what I should do and would like to avoid turning the ECU into a brick. What should I do?

edit: I was checking different ports to see if the wrong port was selected and now it's MT is stuck on COM4 which gave what seemed like erroneous readings and now when I open up MT the main interface never appears but the program is still running in the background.

[Matt Cramer]

Sounds like MegaTune is open off the screen...

[MrDomino]

MegaTune doesn't appear on the taskbar, just in the processes window.

I may try uninstalling everything and reinstalling MT. Do I need to back up anything other than the project files? (.msq files and the .ini files)

Also, what firmware should I install? Do I need 2.88 or can I use the most recent version? I just don't want to have to spend the weekend reconfiguring everything if I can avoid it.

[MrDomino]

Ok so I reinstalled the firmware and everything seems to be fixed. However, I think there is still an issue with how the spark is set up. I think the system may be set up incorrectly so it's firing once on the power stroke and then once 180 degrees out from that (end of exhaust/beginning of intake). I think the ECU thinks there are two cylinders instead of one. Anyways, that's something I can work out later. The biggest problem is that it seems like the signal from the coil is a lot lower at higher intake MAPs. At 50-60kPa the signal is pretty strong but higher up it starts to weaken. Any ideas what to do? At 100kPa the dwell is around 1.6 msec. I forgot to check it at the lower MAPs.

[grippo]

If it is sparking twice as often as needed, I would fix that first. Then, if still a weak spark, I would increase the dwell.

[MrDomino]

This isn't something I can easily verify. It's more of an inkling.

Let's start back with something that I've noticed was kind of botched together awhile back that was apparently never resolved. Under injection control, the parameters are set so that the MegaSquirt thinks there are two cylinders. I tried setting this back to 1 cylinder and then whenever the dyno was spinning at 1100 rpm, the RPM gauge would read 2200 rpm and the engine ran like crap. I'm not sure how the program reads in RPM. I know the ECU is wired up to the shaft encoder which provides a pulse every 0.25 deg crank angle and then another pulse at TDC but I'm not sure if it's seeing just the TDC pulse or what.

Anyways, does the input for ignition advance come from the RPM signal? If so then this may be part of the problem.

Something else I noticed is that when I had the ignition timing set at 20 BTDC (so 340 on our degree wheel), I'd see both 330 and 350. They'd move closer and then get further apart and then move closer and further apart. I'm not sure what is going on or why this is happening.

I'm also thinking of ordering a new coil just to be safe. If the coil driver burnt up does this mean that the coil could be messed up also? I checked resistances and it seemed alright but I'm using a pretty crappy multimeter. Is there any coil in particular that I should buy (read as: one where the dwell is known so I don't have to mess around with it)?

[grippo]

I think there are 2 issues here - one is hardware and has to do with the ignition coil and/or its driver, and the other with the configuration. You have specified 1 cylinder/ 4-stroke, so the shaft encoder needs to provide 1 pulse every 720 deg to get the right rpm and other parameters that may be dependent on it. If the shaft encoder is on the crank, which I suspect it is, then you are getting 2 pulses per 720 deg. The ECU can't be getting a pulse every .25 deg or it would go nuts, so there must be some intervening electronics that throws out pulses. Maybe you could scope the encoder output (that is, the trigger input to ECU). The mode you are in is meant for a distributor which provides a tach pulse per cylinder per 720 deg.

[MrDomino]

Ok, so I think I have the double spark issue figured out. Apparently the engine receives a signal at TDC from the shaft encoder. Since the ECU doesn't know which TDC (combustion or exhaust/intake) it's really at, wasted spark is used so it fires at both TDCs.

I just got the new coil today and it seems to have helped somewhat. I can now run at 100 kPa without issue but if I set spark timing below 20 deg BTDC at 1100 rpm, the engine cuts out and starts making horrible sounds. However, I think my timing light is crap since there were times when I was getting a stable spark but am only seeing a few blinks on the timing light. I'm going to see if I can buy a better timing light and then recheck to figure out what is going on. I really can't figure out when or even if the coil is firing when I only get a blink every 5 seconds.

Also, here's another example of something I can't figure out. 1100 rpm, 70 kPa.
Spark timing Engine response
20 Runs fine with good torque
15 No torque
10 No torque
5 No torque
0 Runs fine with good torque
-5 Runs fine with great torque output

What is this I don't even...

Also, last thing - dwell. Regardless of intake MAP, the gauge in MegaTune was reading a steady 1.5 msec. Shouldn't this be set higher for a direct coil controlled engine? I'm using a 1981-1988 Volvo 245 Ignition Coil (brand new).

[MrDomino]

I hooked an oscilloscope up to the ground of the coil and found that the issues with spark timing correlated to different traces on the oscilloscope. At 20 deg the scope showed three nice pulses which correlated to the ECU grounding the coil. At -5 deg the scope also showed the same thing. When I set the ignition timing to 10 deg, the scope showed a much different trace with each peak becoming significantly wider. Instead of nice peaks, they looked like mountains with part of the top chopped off so they were flat on the top. I then decided to mess with the dwell and spark duration parameters to see if I could fix this. Note: I found that the starting dwell was 2.0 msec and the starting spark duration was 2.5 msec. I have no idea why the parameters were set this way.

I found that if I increase the dwell to around 3.5 msec and decrease the spark duration to 1.8 msec, the issue disappeared for the most part. At 1400 RPM and 100 kPa, I still had an issue at 5 deg BTDC so I set the dwell to 4.0 msec and it's running fine.

So in conclusion, I changed the dwell from 2.0 msec to 4.0 msec and spark duration from 2.5 msec and 1.8 msec using the assistance of an oscilloscope and the issue seems to be fixed.

One question I have in the back of my mind is why this all of the sudden became an issue. It ran fine for the last year with no issues but then out of nowhere we started having problems.