Shrinking Dwell

[grippo]

When rpm is below cranking rpm and you have selected trigger rise rather than calculated advance, the MT gauge advance is set to the trigger offset, since the code fires spark at the instant the sensor pulse arrives. After cranking it sets the advance to the advance calculated from your spark table, and this is what is used for spark timing. Is it possible you are using a high cranking rpm - like near Idle rpm ? If so this would cause the spark timing to change back and forth between craning timing and idle timing.

Does the timing that you measure on the strobe agree with that MT gauge timing when you rev up to say 2000 rpm ?, Does it agree if you measure while you are cranking (with injector wire removed) ?

[PJM38]

Thanks again for your help.
The answers to your questions:
-I always have used ‘calculated advance’.
-My cranking RPM is around 300 rpm, my idle is @850 rpm.
-When MT gauge and strobe measurement are not agree, it's true for all RPM .I do the test by setting all the spark table cells to 8°. I was reading with the strobe 32° and it was constant when i accelerate.
-I was not able to measure the cranking advance for the moment.



I did some oscilloscope measurements on Speed Sensor (Input1) @ VRin and VRout. I think the VRout signal is not correct. I was expected a 0/5v square signal but it's a +/-2v square signal that i measure.
Except measurement error from me it seems that i have to fix this signal issue. No?

Patrick

[grippo]

2 V is marginal, but what really counts is whether you are losing synch or not. You can tell this by either looking at your RPM gauge in MT while cranking and idling or datalogging it. If you ever see the rpm drop to 0 even for an instant then you are losing synch and you need to fix the signal. But if not, then you should be OK.

As far as your spark timing jumping around, I don't understand that, especially if you have all of the table values set to 8 deg. To go any further with this I would need your msq and a short datalog, if you can get one during which the problem occurs.

[Bruce Bowling]

I did some oscilloscope measurements on Speed Sensor (Input1) @ VRin and VRout. I think the VRout signal is not correct. I was expected a 0/5v square signal but it's a +/-2v square signal that i measure.
Except measurement error from me it seems that i have to fix this signal issue. No?

Patrick

What are you using for a ground reference? The output of +/- 2 volts is not possible since the on-board power supply is 0 to 5 volts, i.e. output cannot go below 0 volts....

- Bruce

[TheMonkey]

i'd guess that he actually has 0v - 4v, but the scope is set to AC instead of DC.

[PJM38]

Ok, i made a mistake with my scope AC vs DC!
Finally the 2 VR signals are clean, and i can not see any difference when the advance is good or when there is an offset.
I confirm now, the problem, a lot of time : I start the motor, MT gauge and strobe are not agree. The offset is constant and stable for all rpm, but is changing a little between 2 tests (25°+/5°). Motor running, I just have to open a menu in MT, without changing anything, do a 'burn to ecu' and the adavance read with the strobo come back in adequation with MT.
Except the advance measurement during cranking that i have already to do, is there other trial to do to troubleshoot this issue?
Thanks

Patrick

[PJM38]

Last Update : Advance mesure with strobe during Cranking is 41° (my setpoint is 8°). I just do the test one time.
I Hope that this information will be helpful.

Patrick

[robh]

Patrick,

I am running a very similar setup (87 944n/a with Wasted Spark COP's, stock VR's but using a Microsquirt). As of this morning I was running 0 skipped teeth and -305 trigger offset. Upon discovering this thread today I have updated skipped teeth and trigger offset to your new settings over lunch break and got my full dwell back! Thanks Al.

I'm closely following this thread as I noticed 2 occurences when my timing went unexpectantly high at idle. I will check tonight to confirm, hopefully I can help with some testing from my end.

Thanks for posting your work. By the way- did you get your stock Tach to display the right RPM?

[PJM38]

I fixed my problem!!!
Look at the scope of VRin and VRout signal of the crank sensor (input 2) during cranking : A lot of false triggers.
The VR board, i add, for the crank sensor is based on LM1815*. For this circuit, it's possible to choose different modes and different 'input arming threshold'. So i change from mode 1 to mode 2 (Threshold at 400 mV p-p). The signal is now clean during cranking and the advance is OK. I have to test it on the road but i'm very confident.

Thanks all for your advises

Patrick


(*) http://www.national.com/ds/LM/LM1815.pdf

[TheMonkey]

always nice to find an answer, nice work.

your mode 2 is 400 mv p-p, but your VR trace in looks like it's exactly at 400 mv during cranking, so is that cutting it close? how tight is mode 1? hard to believe that little hash is giving false triggers.