Wheel decoder minimum speed?

[subwoofer]

My car has always been a bit slow to start, but I have always been pushing other issues in front of me, so I really haven't bothered to look into it until now. It starts within a few revolutions when warm - which wouldn't be bad for a carburetted engine - but it gets increasingly difficult when really cold (and cranking is much slower). I then noticed that the RPM indication is jumpin to 0 a few times and that the lost sync count in logs seem to be at least 1 or 2. Wheel is a standard 36-1 with a Honeywell gear tooth sensor, once running it is absolutely rock solid.

Attached is a tooth log from a start from cold, does anyone spot any reason why it doesn't stay synced after first acquiring sync? It had a few coughs before it ran, you can see it almost dying after start due to pulling too much fuel from ASE. Unfortunately, I did not think to run a normal full log at the same time - if that can be done?

[subwoofer]

Ongoing investigation: logs, old and new, show lost sync reason 2. To the naked eye the tooth pattern looks ok, but this is a rather high compression engine (10.5:1 and a very mild cam) so it accelerates hard after TDC. I get more sync losses before it properly fires when it is really cold (-20°C) and the crank speed in down to 100rpm or so.

I would really like to get it to the point where it fires at a bump of the key like an OEM system.

[jsmcortina]

The large change in RPMs after start could cause a problem depending on the tooth#1 position. I'll check the logs a bit later.

You can run concurrent regular and diagnostic logs.

James

[subwoofer]

Tooth #1 is at 74°BTDC, that should be about where recommended by the docs. I'll probably be away for a few days, so don't expect new concurrent logs before the weekend.

[subwoofer]

I got some usable logs this morning, there is something weird going on here. The missing tooth looks erratic at low speeds, being a gear tooth sensor I don't really think the sensor is at fault. Since the missing tooth is in a "no compression" zone, it shouldn't be 4 times the length of the adjacent teeth. I think it should have synced within a single revolution, looking at the MAF signal it took at least twice that before initial sync, and it didn't fire until the third time it synced, I think.

Attached is a tooth log and the normal log. It fell on its face after initial start, thus the reset.

[subwoofer]

Forgot the .msq...

[subwoofer]

James, have you had a chance to look at the latest logs?

[jsmcortina]

I got some usable logs this morning, there is something weird going on here. The missing tooth looks erratic at low speeds, being a gear tooth sensor I don't really think the sensor is at fault.

What part number sensor?
Peter F reported that the 1GT101DC is not good with missing-tooth wheels. The problem sounds similar to your report here.

James

[subwoofer]

Indeed a 1GT101DC, sensor gap has to be a lot larger than I would expected from the datasheet, but smaller gaps caused sync loss at higher speeds, not lower. A gear tooth sensor works from DC, I would expect problems to be of the high frequency kind.

What I can't understand is that the tooth log shows a very long missing tooth at times, while the rest of the teeth look fine (and none are missing) - and the next time around the same tooth within 20% of the surrounding teeth. The physics of that simply does not add up, the starter should not be able to create the kind of acceleration/deceleration needed to stretch/compress only that tooth.

[jsmcortina]

Peter specifically said that he found problems using that sensor on a missing tooth wheel.

OEMs most frequently use VR sensors on the crank for more accurate triggering.

James

[subwoofer]

I understand what you are saying, but as I said, I have been using that sensor for 3+ years at the crank, it has been rock solid at speed all that time. That it selectively stretches and compresses one specific tooth while cranking does not seem likely to me.

But since the cranking is slow, is there an off chance the counter used for the tooth length is wrapping just at cranking speeds? I have had problems with race conditions in timer + overflow counter applications before.

[jsmcortina]

The software counter is 32bits x 1 microsecond, so that itself shouldn't be a problem.

Can you scope it and compare the scope trace to the composite log? If something weird is going on, that should show if it is the sensor, the electronics or the code.

James

[Peter Florance]

I found the sensor did some odd things at slow speeds.

This was measured with a scope.

Maybe some AGC built in?

Cherry and Allegro sensors in the same wheel worked perfectly.

One thing to check: try reducing the dwell correction at what ever your cranking voltage is, and below that. Some ignitions systems spark early if over dwelled. That can reduce your cranking speed.

[subwoofer]

I found the sensor did some odd things at slow speeds.

This was measured with a scope.

Maybe some AGC built in?

Cherry and Allegro sensors in the same wheel worked perfectly.

Good info, Peter! I may just have to order a different one from DIY and see. Do you have Farnell/Mouser/Digikey order codes for any known good sensors? At least two of these sell in NOKs, meaning I save 25$ in customs handling fees.

One problem is that it is getting warmer now, so the issue is less profound, probably due to higher cranking speeds when the battery is warmer and the oil less viscous.

One thing to check: try reducing the dwell correction at what ever your cranking voltage is, and below that. Some ignitions systems spark early if over dwelled. That can reduce your cranking speed.

My ignition is sort of old fashioned, still on a single dumb coil and dizzy. So I am pretty sure it doesn't go off early.

I can try to get a look at the signal with a scope one day, but since my ECU is somewhat hidden that entails a bit of yoga.

[Peter Florance]

I have part #'s for the threaded version.

[Peter Florance]

I'll post when I get to office

[subwoofer]

Thanks!

A quick rummage around Mouser came up with this one: http://www.mouser.com/ds/2/240/Littelfu ... 275379.pdf

[subwoofer]

Any comments on that Harlin part? Yay or nay?

Does anyone know what part DIY is selling, could that be the Harlin part? Pictures sure look similar...