Hello Everyone,
I've been sorting out problems one at a time with our Megasquirt, but this one has me stumped. I can start the car up and let it idle until it gets warm just fine. I let it idle for a while and it runs fine (I have the "idle" set at 1700rpms right now, because if I get anywhere near about 1000rpms it just wants to die). The screen cap from my datalog shows the car running fine for several minutes, and then the rpms start to die off and it just stalls. I am not touching ANYTHING, and the rpms are set with the screw and the throttle position does not change.

http://www.n1xrr.com/tuning/randomstall.jpg

At this point, I cannot get the car to start or even try to start. RPM's jump up maybe 30-50 rpms while cranking, but it doesn't even try to catch. The plugs are fine.

One thing that I noticed is that as the rpms go down and the kpa goes up, the pulse width goes up but the air fuel ratio starts getting leaner. Thats exactly opposite from what I expect. I would expect the AFR to get richer as rpms decrease and pulse width increases.

After letting the car cool down for a while and starting it again, we got to the point of the car stalling slowly again, so this time we sprayed the LM1815 chips with some electronics cooler in a can and the rpms came back up. I did this about 6 times or so, and by the 6th time the chip was very cold, and then the Megasquirt just took a crap and the car died.

http://www.n1xrr.com/tuning/cooledchips.jpg

Any ideas?

 

30-50rpms when cranking sounds to slow... Is the motor cranking slow? Maybe the LM1815's are faulty or you've got a ground loop in the VR wiring. I had a ground loop when i first wired my setup and the rpm's would read in the 30-50 range when cranking and it wouldn't start. Sounds like the LM1815's are dead now though :-(.

Kris

 

I also see some noise on your rpm and MAP signals... I'd want to see the coolant and air in the datalog too....

That noise looks like AC getting into your electrical system.

 

I haven't fixed my single-rotor map signal yet. Thats on the schedule for this weekend. I'm also going to make a new harness, just in case.

http://www.n1xrr.com/tuning/datalog1.xls
http://www.n1xrr.com/tuning/datalog2.xls

Those are the two datalogs. You can see in the 2nd one where the computer just gives up in the end.

 

The noise I'm talking about shows up on the rpm signal too... it doesn't look like a good, smooth rpm signal, the oscillations on both of them are too fast to be anything but noise.

The MAP signal would oscillate but not this quickly from being taken from a single rotor.

 

I see. Hmmm...

Maybe I need to swap out to those Electrolytic caps on the VR board...I'll give it a check this afternoon.

 

I have news!

I hooked the VR conditioner board up to a function generator today and found something interesting.

First off, I made another VR board based off of the "EasyVR" design on the GPIO website. Everything there seems to be reading correct. When the wave goes into positive voltage, it triggers low. When it goes into negative voltage, it triggers high. Fine.

Hooked up the function generator to my board on the 2-tooth part of the circuit, everything looks fine. Hooked it up to the 24-tooth part and oh boy. Its only outputting .6v to the Megasquirt chip. Hmm...that sounds like a problem to you?

 

yeah, it should be a +5v square wave going into the MS.

Ken

 

Well, looked at it again and it was actually outputting .6v to U4, and then the 5v output to the pin on the MS chip. Oh well, I thought I had it there.

The waveform output from U4, is that supposed to be a nice, clean square wave?

My U4 kinda "ramps up" to the signal going to the Megasquirt.

 

the ramp up might be because the input to U4 is only .6v... the input to U4 should be a +5v square wave....

U4 just inverts the output coming from the lm1815 so that when the tooth is crossing the sensor, the CPU sees a falling edge instead of a rising edge.

Are you sure you hooked to the right pin on U4? If you accidentally hooked it to the wrong pin, you might be triggering U4 on the LED input instead of the transistor's base input.... maybe that could be it? You should be connected to U4 pin 6, which is directly across from pin 1.

I guess if it doesn't ramp up fast enough, the CPU might not watch for the falling edge or something... In the lm1815 circuit, there's a pullup to 5 volts... it's a 4.7k resistor hooked from +5v to pin 12 on the lm1815. If something is wrong there, or something is wrong with the lm1815 itself, I think you could get less than +5v to U4 from there.

 

Yeah, I actually suspected the pullup also. I tested it and it was fine, but I still replaced it. If I disconnect the LM1815 circuit from the Megasquirt, the output is a nice +5v square wave. With it connected its a .6v square wave.

I also tripple checked the pinout of the opto-isolator. Pin 1 is the one with the dot next to it. Looks like that pins 4-6 is actually a transistor thats activated by light (hense the opto-isolator name). Is this not correct? If so, why exactly are we triggering pin 6?

http://www.n1xrr.com/Megasquirt/optoisolator.jpg

I think the reason why its only registering .6v is because there is a pulldown somewhere on the trace that leads to pin 6. I actually double checked this one my other v2.2 Megasquirt and it has the same result.

Does this make any sense?