Engine dies around 4500 rpm

[Blue]

Here's the most recent .msq file. I hope things go well at the meeting. The settings for EGO correction and acceleration were made to disable those features in an effort to eliminate variables. Thanks again for all your help.

[grippo]

Your Req Fuel is 56 - this seems way too high unless you have very small injectors. I see from one of the datalogs that you are briefly going over 100 % duty cycle - but I don't think this is long enough to cause a problem with overheating the injectors. It's probably a consequence of raising your ve to get a longer pw on the scope.

Also, I would turn off the AMC option until you get a rough tune on the engine or at least until you have no problems like this. Otherwise AMC can fight you - although I know you have disabled it by disabling ego.

Another thing is that tpsdot threshold is 20 %/sec - this seems too small. If you were trying to disable AE, as you said, this is the wrong direction. I would start with 200%/ sec and then bring it down till tip in stumbling stops.

That said, I don't think any of these is causing the problem, which appears to me to be electrical. I base this on the fact that I have never seen a problem like this and the calculated pw is mostly reasonable. The problem is that the scope doesn't agree with this. But, since the injectors work under stock ecu control, this points to the injector drivers being the problem, although we haven't seen any problems with them previously. Bruce has been swamped by action items from the meeting, but he will look at this when he gets a chance.

[Peter Florance]

Hi Brad

Chatted with Al and Bruce about this Friday night.
I'm unclear as to whether both injector channels exhibit this problem or not; that would tell me a lot.
Can you advise?

I've subscribed to this topic and I'll keep an eye on it.

Thanks Brad

[Blue]

Sorry it has taken me so long to implement your suggestions and respond. I tried using the second injector channel last night with no change in the results - the pulse width on the scope still does not match the predicted pulse width in MegaTune in the 4,000 - 4,500 RPM range. I also turned off AMC (Automatic Mixture Control?) with no change. I couldn't find the tpsdot (Throttle Position Sense dot?) setting in MegaTune, and I wasn't able to follow up on that suggestion.

I purchased a JimStim and a MicroSquirt harness from DIY AutoTune and hooked it up last night. I had the the MicroSquirt injector wire connected to the inj1 LED on the JimStim, so the load on the driver was very small. I connected the scope to the same injector wire to monitor the pulse width. The scope showed a nice square wave whose duration matched the predicted pulse width in Megatune at all values of RPM and MAP.

I am open to other suggestions, but I'm thinking of trying the following experiment. Connect my low impedance fuel injector to the injector wire from the MicroSquirt, using the JimStim for everything else. I would need to use a bigger power supply than the 12 V, 500 mA supply I'm using now. I was thinking of using a motorcycle battery, with a 5 Amp fuse to the JimStim, and connecting the MicroSquirt grounds, and power to the injector, directly to the negative and positive battery terminals respectively.

Since I first put the injector pulse on the scope several weeks ago, I have been concerned about the 70 -80 V spike that occurs when the injector closes. It seems like this might be too much noise for the MicroSquirt to accommodate. I realize that I may be getting ahead of myself, but would it make sense to use some sort of power transistor controlled by the MicroSquirt injector signal, perhaps with an injector resistor to limit the current? I have a Mechanical Engineering degree, but I'm no genius when it comes to electronics. Could someone provide a circuit that I could build up to try out?

As always, I am grateful for any advice or suggestions on this problem.

Thanks.

[Peter Florance]

Since it's a MicroSquirt unit, I have disabled PWM by setting the threshold time to 25.4 msec and current limit to 100%.

I was late to the discussion.
Are you running a ballast resistor with the injector?

If not, I can't see why you wouldn't run around 1.1msec for Time Threshold...

I wonder if that is an issue or am I missing something?

[Blue]

I'm not running a ballast resistor in series with the single low impedance injector. If I am interpreting it correctly, the main MicroSquirt web page (http://www.microsquirt.info/) indicates that PWM is not needed, nor is a ballast resistor. Here is the relevant section from that web page.

3. No pulse width modulation (PWM) to limit injector current. The fuel injector drives will max out at 5 amps each, enough to drive one low-impedance (or 4 high-impedance) injector per bank. To get everything to fit without lots of heat sinking, MicroSquirt uses the ST VND5N07 from STMicroelectronics to drive the injectors. This is not a 'peak and hold' driver, but it does clamp the current at 5 amps, so it can be used with one low-impedance injector per bank, however the close time may be a tad higher (or you can use resistors). For up to 4 high-impedance injectors per bank, it should work fine. For example, for motorcycle use MicroSquirt will be perfect!

Actually, I did use PWM initially, but later found out that it didn't seem to make any difference, so I disabled it.

[Peter Florance]

I'm not running a ballast resistor in series with the single low impedance injector. If I am interpreting it correctly, the main MicroSquirt web page (http://www.microsquirt.info/) indicates that PWM is not needed, nor is a ballast resistor. Here is the relevant section from that web page.

3. No pulse width modulation (PWM) to limit injector current. The fuel injector drives will max out at 5 amps each, enough to drive one low-impedance (or 4 high-impedance) injector per bank. To get everything to fit without lots of heat sinking, MicroSquirt uses the ST VND5N07 from STMicroelectronics to drive the injectors. This is not a 'peak and hold' driver, but it does clamp the current at 5 amps, so it can be used with one low-impedance injector per bank, however the close time may be a tad higher (or you can use resistors). For up to 4 high-impedance injectors per bank, it should work fine. For example, for motorcycle use MicroSquirt will be perfect!

Actually, I did use PWM initially, but later found out that it didn't seem to make any difference, so I disabled it.

Yeah I was missing something.
You should be fine

Did you scope the +12v line going to the injectors?

[Blue]

Actually, I've been putting the scope probe on the negative injector lead. I have the injector circuit wired so that +12 volts is supplied to one side of the injector when the key is turned on. The MicroSquirt pulls the other injector lead to ground to energize and open the injector via the injector 1 wire on pin 9.

[Peter Florance]

Actually, I've been putting the scope probe on the negative injector lead. I have the injector circuit wired so that +12 volts is supplied to one side of the injector when the key is turned on. The MicroSquirt pulls the other injector lead to ground to energize and open the injector via the injector 1 wire on pin 9.

Right, but I'd like to make sure the +12v is solid

Switch your scope to AC coupled and measure the amount of noise

Thanks!
Peter

[Peter Florance]

Since I first put the injector pulse on the scope several weeks ago, I have been concerned about the 70 -80 V spike that occurs when the injector closes. It seems like this might be too much noise for the MicroSquirt to accommodate. I realize that I may be getting ahead of myself, but would it make sense to use some sort of power transistor controlled by the MicroSquirt injector signal, perhaps with an injector resistor to limit the current? I have a Mechanical Engineering degree, but I'm no genius when it comes to electronics. Could someone provide a circuit that I could build up to try out?



Thanks.

The FET's have a built-in clamp at 70v so the spike is going to be larger.

You could make a power zener from TIP42C 270 ohm resistor and 36V zener to limit the spike.
But I don't that's your issue here.