Erratic RPM

[nzgt6man]

hi,

Now I've succesfully got one car running, I'm onto my second MJLJ (EDIS6) implementation. I've got the car running but the tickover is not completely smooth with the engine missing occasionally, perhaps 4 or 5 times in a minute. When I hook up the MJLJ I notice that when the engine misses the software is being told the revs are not what they should be so it advances the timing to the value it thinks the revs are. As an example when I'm running at tickover the display will jump from the correct value (800) up to anything from 1600 to 5500! Obviously this plays havoc with the timing. when it jumps to >4000 the engine cuts out as the timing is advanced to about 33 degrees when the actual speed is still 800. any ideas?

I've replaced the EDIS module with one I know works and I get the same thing. I've used shielded cable throughout gfor the pip, saw and VR sensor without joins. I still get a miss when the MJLJ is disconnected but as the timing remains static at 10 degress the engine does not cut out.

Any pointers gratefully appreciated,

Rgds,

Robert Proud

[brentp]

One thing to do is connect the ground and power connectors for both the EDIS controller and the MJLJ to the same spot- this has helped in the past with jumpy RPM signals. Basically what is happening is that it's detecting noise on the PIP line and falsely calculating RPM based on that, and as a result, improperly triggering the SAW signal.

If that doesn't work for you, try running the EDIS controller and MJLJ off of it's own power supply- i.e. another battery, not part of the auto's power supply. Of course this is only a diagnostic test.

Let us know what you find out.
Brent

[Oliver Sedlacek]

My Megajolt also suffers from this. Looking at the PIP signal, the problem seems to be that this signal does not give a good clean low signal, as it swings between 2V and 12V. The input buffer and attenuator doesn't really clean it up, and a the spark event sometimes triggers the PIP interrupt on the micro. I've tried fiddling with the input circuit, but no success yet.

[brentp]

The firmare is configured to detect *falling edges*- see the fimware source code for reference. So- I'm not sure if the PIP line not going to 2V is a factor here. Since you seem to have an oscilliscope, you might want to look for noise when the PIP signal is high- this might trigger a false interrupt.

Did you try grounding and powering the MJLJ and EDIS from the same points? I've seen cases where the grounding and power were derived from different places, problems like this resulted.

Brent

[Oliver Sedlacek]

Brent,
The Megajolt and EDIS are earthed to the same point. As I said, the PIP low signal is about 2V, which translates to about 0.7V at the micro. This should be good for a solid low, but the micro picks up an interrupt about half the time around the time of the spark. I've tried a small cap on the IRQ so far.

[brentp]

Thanks for the update.

Have you tried powering the EDIS and MJLJ from an ultra clean power source, like a seperate battery? This way you can isolate the car's electrical system- for testing/debugging purposes, of course.

Brent

[Oliver Sedlacek]

I've tried a 3 terminal ferrite on the 12V feed to the regulator, without making a noticeable difference. I've also tried replacing R2 with an 820R resistor in series with two diodes, which did make things better. I'm going to try a circuit with a bit of hysterisis tomorrow.

[brentp]

Good info. thanks for the update.

Can you actually see the noise on the PIP line with an oscilliscope?

What is it do you think that makes the problem *not* happen on your other engine? i.e. what is different about this setup?

On the software side, I can investigate if triggering the PIP on a low instead of a falling edge would be better. I need to re-setup my test bench anyway for the next version of the firmware.

Brent

[Oliver Sedlacek]

After measuring the PIP signal, I decided the circuit would never be completely reliable over all component tolerance. I've modded the circuit to add hysteresis and it now works cleanly. You can see the revised schematic on <a href="http://community.webshots.com/album/187745001srPbmU" title="Webshots">webshots megajolt pics</a>

[brentp]

Excellent, thanks for you contribution.

Anything special about the bcx53 transistor- would the 2n3906 devices work as well? I do see the voltage ratings are different between the two devices.

Thanks,
Brent

[brentp]

Oliver, I'm going to compile a version of the firmware that may reject the noise you're seeing, using the original circuit. Would you be able to test it for me on your system? You can use the bootloder software to upload it to your processor:

http://picasso.org/mjlj/?q=node/view/4

[nzgt6man]

Hi Brent,

Sorry for the delay in getting back to you I've been away on business.

I've tried what you suggested (shared power supply and separate battery) and I still get this problem. I've no idea what is causing this. If I remove the components (mjlj, coil and vr sensor) and put them on my other car it runs fine! Only difference is that my GT6 doesn't have much electrical equipment (runs fine) compared to my 2500TC which does. The engines are the same but the alternators are different. Could a "noisy" (electrically) alternator cause a problem?

Rgds,

Robert Proud

[Oliver Sedlacek]

Brent,
it's potted up now, which makes it inconvenient to go back to the old circuit. However, if you want a tester for interpolation code, I'd give that a go. There's a bunch of us installing Megajolts and you can read about it on <a href="http://www.blatchat.com/t.asp?Id=60816" title="Blatchat">blatchat</a>. A guy called Bill Shurvington is coordinating the process.

[Oliver Sedlacek]

It's what I had in stock, nothing special.

[smithste]

Hi there,
A bit late now, but I've come across this problem in an A series, and found that the plugs I was using were non-resistor plugs. Swapped these to resistor type and all my problems went away. I think all ecu equiped engines use resistor plugs to eliminate RF frequency interference, so if you are upgrading an old points style ignition its worth checking out which plugs you use before swapping to the edis unit.