No Crank Trigger / RPM - New Installation Configuration

[MTRacing]

Thanks guys!  I'll try these changes over the weekend and report back.

Take care.

Mike

[16Victor]

Some updates:

Last night Mike and I grabbed the scope and chased the PC synthesized 60-2 waveform from the laptop headphone jack all the way to the LM1815. 

Result:  the output of the LM1815 (pin 10 (late note: I was checking Pin 10 based on the LM1815's datasheet, now that I found the V3.3 schematic I see I should have chosen pin 12..doh)) can be seen here (signal looks overdriven but that's a result of the camera's exposure settings...still, 2Vpp clearly visible) 

Some notes:
- Found a bad RCA patch cable was attenuating the waveform from the laptop to the VR sensor cable.  Replaced it.  Now, Megatune shows an erratic RPM only when the PC volume is way low, like at 2/10 or so.  Above that, RPM reads 25550 with no ign or inj LED activity.

If anyone can offer any input regarding the components on the board, there are some high res pics here:
 http://scirocco.dhs.org/cheapassron/mt/4-18-09%20pics/

Hypothesis: I think we have a config problem!  Why?  Because with the input amplitude very low, the LM1815 output wigs out (erratic output on pin 10), and the VEMS processor reacts accordingly by firing inj and ign outputs in some related sequence...so a signal is getting there...now onto telling the processor how to handle it correctly...

Observations, comments, redirects, etc all very welcome.

Thx for the continued support,

Ron
(MT's helper boy)

[16Victor]

...i removed R30,R57 and put short instead of C31.That cured all of my triggering problems....

Thanks - this eliminates the one-shot circuit, right?  (and the VR pullup).  How did your pin 12 signal look before and after this?

Thx

Ron

[rob@vems.co.uk]

Okay, try this:

http://www.vems.co.uk/VEMSInstalls/VemsMT1.0.73.60-2.zip

It works fine on a 60-2 trigger, its 4-cylinder though, but good for cranking tests.
Download the firmware AND the configs and tables from here in the usual way, then test.  It might be an idea if it works to extend it to work on an 8cylinder, and to patch your wideband setup, required fuel etc.

Good luck

Rob

[16Victor]

Here's where we got last night:

- After borrowing a known good VEMS with a known good 60-2 config, we determined that the hardware was OK.  Got the same results on both units after transferring config and tables to Mike's unit. 

- Both triggered fine when driven by the Signal Gen program.

- Neither triggered when we cranked the engine (Electromotive VR sender and wheel, 1.2Vpp sine-ish wave output, measured at the ECU).

Noticed odd DC bias on the EC36 pin for Trigger 01, 2.8VDC on Mike's unit (with nothing connected and with VR connected), compared to ~.5Vpp on the borrowed unit...was wondering if this is preventing a 0V crossover thus no LM1815 output???  (Odd that when inputting the Signal Gen waveform, the signal crossed 0V therefore no DC bias - the PC was probably forcing the issue).  Is this a red herring?

We won't get to work on it for a week or so but could try a few caps to absorb the DC voltage.  Or go to a VW sender as recommended by GRMRacer and others....

More thoughts welcome!

Ron
helper boy

[rob@vems.co.uk]

Using a cap to filter DC is not good as the signal will get screwed, the changing frequency of the trigger will make stable triggering impossible.

A slight pull-up on the VR input is not unusual with noisy triggers we usually increase the pull-up voltage, you are seeing a negative voltage on that line when you 'scope it don't you?

[16Victor]

No, on Mike's unit the whole wave is above 0V.  The borrowed unit (Brian's) has less of a DC bias, but we saw no triggering with it either.  I'd have guessed that it'd cross 0V, but we didn't scope it. 

BTW since we got the borrowed unit, we went with the proven config in it vs following the letter of your suggestion above.  We had that on tap but ran out of time.  I figure we got the gist of it though.

Thx again,

Ron

[16Victor]

Mike sourced another VR sensor (spec'd for mid-90's VW Jetta III, ABA block), got 2Vpp and and a nice looking waveform, but no triggering on his unit.

Next steps are to run the same test with the borrowed unit, then (depending on the result), run a 'cheater' cable to provide the signal to the ECU but bypassing the premade harness.

Seems the DC bias may be an issue after all.

Will keep posting updates for posterity.

Ron

[16Victor]

Things weren't making sense.  Both the borrowed unit and Mikes would trigger just fine on the new config and PC speaker synthesized .wav file input, neither would work with Electromotive or VW sensor.  So we went back to basics and created a cheater wire that bypassed the entire harness supplied with the unit.

Presto, signal.

Dissected the harness a bit and found what I think is some very questionable wiring, in that the VR signal (-) and shield were one and the same.  Therefore, either VR sender was hooked up to a wire going nowhere.  Look closely at pic, the black and blue wires are cut off flush.  Maybe it rang out OK because the shield contacted the cut-off wire?  Just a guess, but it did ohm out fine, hence the head scratching.

Why it worked with the PC speaker and not the VR sender I have no idea.



We'll get that mess cleaned up and begin configuring!

Thanks for all the help so far,

Ron

[rob@vems.co.uk]

That's good, but exercise caution, while I've never been over happy with the shield being the negative wire its never been a problem before.  And make sure that you tie the grounds in correctly - the signal ground is connected to the ground cluster directly through that shield.

[gunni]

Shielding connected to signal ground is common practice on OEM vehicles. But of course they must end in a wire going to sensor ground on the ECU. Ending nowhere is no good

[MTRacing]

UPDATE:

After reworking the factory VEMS harness to connect the ground wires with the shield wires at the ECU end, it seems to have solved the problem! VEMS can see the crank trigger signal and I'm getting an RPM reading on Megatune.

I'm moving on configuring the injectors and coils. Look for a post soon.

Thanks for all of your help!

Mike

[Sanjuro]

First I would like to say HELLO EVERYBODY my name is Richard! It's my first post here

MTracing I had similar problem to your,and after hours spent on searching in the internet and playing with scope,i removed R30,R57 and put short instead of C31.That cured all of my triggering problems.Hope it will help to you.

cheers Richard

Apologies for digging up an old thread, but what does R57 and C31 do?  The wiki only had us remove R30 for v3.3 primary VR, but didn't mention the next 2...

[SamSpade]

They set the timing (pulse width) for the rpm output of the LM1815 (U11) Variable Reluctance Amplifier chip, what's called an RC timing circuit in electronics.  The LM1815 filters the VR signal then sends clean pulses to the Atmel microcontroller.  So basically, the RC combination sets the length of the pulse that the Atmel microcontroller sees for every tooth of the crank wheel.

According to the LM1815 datasheet, the output pulse width is:
PW = 0.673 x R x C

Which means in the schematic:
PW = 0.673 x 75kΩ x 1nF = 0.050 ms = 50 us

But in my 3.5 board R57 is 33kΩ, so:
PW = 0.673 x 33kΩ x 1nF = 0.022 ms = 22 us

Ideally, to minimize spurious triggering, you set the pulse width to shortest time that one tooth actually passes the VR sensor.
For 60-2 trigger wheel and 10,000 max rpm:
Tooth pulse time = (1 / 10000 rpm) * (60 sec/min) * (1 rot / 120 'teeth') = 50 us

For 36-1 trigger wheel and 7,000 max rpm:
Tooth pulse time = (1 / 7000 rpm) * (60 sec/min) * (1 rot / 72 'teeth') = 120 us

If you're seeing a noisy rpm value (jumps up), increasing the resistor value of R57 so that the PW fits your max rpm may improve it. But the datasheet recommends only setting it to a max 150 kΩ.

You can download the LM1815 datasheet here: http://www.national.com/ds/LM/LM1815.pdf

[Sanjuro]

Wow that is much more detailed than what I was expecting!  Thank you so much for the explanation.  I do see the ocassional RPM spikes, although I very much need to set up the audio plug bench test to make sure it is not my VR hardware.  Good thing rescue kit 1 comes with a few 7502s in case I need to apply a fix (I have 60-2 with max rpm ~7500)