Tri-tach install on a Porsche 911 3.2L with 129+1 tooth help

[julian944]

Your wiring seems to be correct according to what mark (notfalls) got out with in this thread here.

http://www.msextra.com/forums/viewtopic ... &hilit=944

Julian

[jsmcortina]

I looked at the comp. log and there doesn't seem to be any signal on the main tach input.

As a recap, the normal wiring is:

Hi-speed flywheel teeth -> VR conditioner -> PT4
Single flywheel pin -> mainboard tach VR input
Cam signal -> MS3X tach input

James

[BenGTT]

Thank you James and Julian for your confirmation. At least, I am sure that the wiring is correct.

Ben

[BenGTT]

tests of the day.

I did the tests with a spare Jean's Dual VR board 2.0 instead of the onboard LM1815 + Jean's V1.1 LM1815 board.
Same issue, no synch but some idea.

I have tried Log crank and cam mode and have also swapped the VR sensor to log them using the standard cran TSEL input.
It seems that I have both sensor ok.
Also, during cranking, I have the "protocol error" that is lighting on the dash. Is it related to my problem ?

Here are the datas

[BenGTT]

Also, here is some tooth log pics :

the 129 tooth that looks ok.



and the single tooth tooh log that does NOT look ok. It seems that the single pin vr sensor is also reading the high speed theeth . Am I right ??

[racingmini_mtl]

Is that single tooth log done using the V2.0 dual VR board? If so, that may be the issue because the MAX9926 is very sensitive. It might actually be too sensitive in this case. There is an adaptive threshold but since it has a time out of about 85ms, that will be too short for a single pin at low RPM (less than about 700RPM). That is probably the timeout period you're seeing in the log.

So if you were using the V2.0, try to make another single tooth log with the LM1815 circuit instead. Or you could try to use a shunt resistor (1k to 10k between VR+ and VR-).

Jean

[BenGTT]

Hello Jean,

yes it is with the v2.0 dual vr board. I'll try the shunt resistor tomorrow and if no success, I'll reconnect the 1815.
This car is a real school case , it is like running in a wall.

[BenGTT]

I have also added a 2 millimeters spacer under the VR bracket to give some clearance between the 129 teeth ring and the single pin VR sensor, but the issue remains the same.

[elaw]

and the single tooth tooh log that does NOT look ok. It seems that the single pin vr sensor is also reading the high speed theeth . Am I right ??

I'm not a Porsche guy but can tell you that _is_ a problem on Audis that use the same trigger setup. Later versions of the Audi ECUs actually would take the tooth signal, invert it, and mix a little of that inverted signal into the pin signal to cancel out the tooth signal that was appearing in the pin sensor's output.

I can tell you increasing the gap is the wrong way to solve the problem. Set the sensors up to the factory spec, then use resistors to attenuate the signal going into the VR board. If you attenuate the pin signal just enough, you should be able to get it to a point where the VR conditioner will trigger on the pin signal, but not on the tooth signal that that sensor is also seeing.

Jean: can this be accomplished by putting resistors in series, rather than parallel, with the sensors? Again this is Audi rather than Porsche wisdom, but I learned the hard way that if you load the sensors with more (ie lower resistance) than about 15K, it causes phase shifts that can mess up the relationship between the two signals. Even if you load both sensors equally, they're affected differently because of the nature of the teeth they're sensing.

Oh and one more thing: you'll notice the single tooth log has two long "teeth". That *could* be due to the gear signal bleeding through, but if you get rid of the gear signal and still see the two-long-tooth pattern, you might have to reverse the polarity on that sensor.

[sjf911]

Any possibility of modifying the single tooth sensor to hold a gear tooth type sensor instead of the stock VR?

[BenGTT]

Thank you Eric for your help. You are right about the gap. I tried, but since it doesn't work, I'll set the sensor as factory.

Ben

[BenGTT]

Any possibility of modifying the single tooth sensor to hold a gear tooth type sensor instead of the stock VR?

It should be possible, will take a look ! Thanks.

edit : http://sensing.honeywell.com/index.php/ ... 9/la_id/1/

[racingmini_mtl]

Jean: can this be accomplished by putting resistors in series, rather than parallel, with the sensors? Again this is Audi rather than Porsche wisdom, but I learned the hard way that if you load the sensors with more (ie lower resistance) than about 15K, it causes phase shifts that can mess up the relationship between the two signals. Even if you load both sensors equally, they're affected differently because of the nature of the teeth they're sensing.

Putting only series resistors will not attenuate the signal significantly, if at all. But what could be done is to use both so that the signal is lowered and the load on the sensor is not too much.

Jean

[BenGTT]

update, still using Jean's dual vr board v2.0, shunt with resistor value from 61K to 1K did not change anything, but 300 Ohms did, at starter speed.
Using log crank and cam mode, and the tooth logger, I am now seeing most of the time only 2 teeth, sometimes 1 (rarely).
Next test, I'll wire 2 resistors from VR+ t ground and VR- to ground and see if that help or not.

[elaw]

Here's what I'd suggest: put a 10K resistor in series with each wire coming from the VR sensor. Then try different resistor values across VR+ and VR-. Based on your prior results, you'll probably want something in the 1K to 3.9K range.

If you get it to the same point, where you see sometimes two teeth and sometimes one, try swapping the wires coming from the VR sensor and see if that gets you one tooth consistently.

You should also try to keep the circuitry for each sensor the same. If you find that the pin sensor works well with a certain combination of resistors and the wires reversed, do the same to the tooth sensor.

[BenGTT]

Thank you Eric, I'll try that this afternoon.

[BenGTT]

I have actually 2 10K resistor in series + a 2K2 resistor between VR+ and VR- and the result look promising ! 3 teeth in a 1000ms time lap, that sounds ok. This is with VR wires inverted.



but still no synch..let's continue..

[BenGTT]

here is a composite log

edit : pots were not adjusted in this log, they are now but still no synch, 0-150rpm-0-150-0-150

I'll double check if the cam sensor is reading soon enough.

[BenGTT]

Is it possible to use the tri-tach as "bi-tach" for full group/semi seq/single coil/wasted spark?

[jsmcortina]

With a code change perhaps. But once you have the flywheel tooth sensor and single pin sensor working ok, the cam sensor should be trivial?

James