Logging crank twist using cam position input.

[Sean]

Hi all, I'm working on a problem with some engines dismantling themselves over time (main studs coming loose). I'm trying to log crank twist using a digital input configured as cam position, as per this thread viewtopic.php?f=11&t=1093&start=10 but the results I'm getting don't bear much relationship to reality.

I am using a Delco mag sensor via a DMC B to send a signal to the digital input. The signal looks like this when it's wired to the sync to capture it....

capture.jpg (91.09 KiB) Viewed 20058 times

The digital input is configured thus...

ecuconfig.jpg (54.24 KiB) Viewed 20058 times

And this is the log......

logfile.jpg (126.25 KiB) Viewed 20058 times

As you can see, the logged data clearly isn't representative of what is actually happening. Has anyone any suggestions as to what I am doing wrong here?

Thanks, Sean

[AdamW]

Have you confirmed the VR sensor connected to the DMC is giving the correct falling edge signal? You might want to post a bit more info about the set up, it seems you have an 8 tooth wheel at one end of the crank for "cam position", what is your ref sensor at the other end? What ref/sync mode are you using?

[Sean]

You are an ideas man Adam, I have just beeped out the wiring to the DMC B, the sensor was pinned backwards I assume this can mess up the signal conversion at the DMC?

How do I choose which edge to set in the ecu, as far as I see the signal from the DMC has both rise and fall?

4 tooth on the front of the crank, gives notionally 8 teeth for one rotation of the cam.

front.jpg (50.42 KiB) Viewed 20043 times

Mode 5, Delco magnetic with 40-1 trigger machined in to the flywheel for ref.

flywheel.jpg (35.07 KiB) Viewed 20043 times

I have the option to fit a 40-1 to the front of the crank as well as another sensor to the rear to use the ACL/ADL phase measurement method, but its a major fight to dig in to the vim wiring to get to the inputs. Do you think the phase measurement method will serve better for us?

[AdamW]

You are an ideas man Adam, I have just beeped out the wiring to the DMC B, the sensor was pinned backwards I assume this can mess up the signal conversion at the DMC?

"Not so much "mess up the signal conversion", but if the signal is "upside down" you will be triggering off a point in the waveform that varies in position relative to RPM/amplitude. Sorry I'm not quite sure how to explain that more clearly - I might see if I cant find a picture that explains it better somewhere later. The net effect is you will see false advance or retard of the "cam position" as RPM increases.

How do I choose which edge to set in the ecu, as far as I see the signal from the DMC has both rise and fall?

The DMC datasheet says you must use falling edge. Normally that is just because one edge has a more "vertical" edge.

I have the option to fit a 40-1 to the front of the crank as well as another sensor to the rear to use the ACL/ADL phase measurement method, but its a major fight to dig in to the vim wiring to get to the inputs. Do you think the phase measurement method will serve better for us?

I have no experience with an ACL and don't have much torsional vibration theory. I think the problem you have is the crank doesn't just gently "wind-up" in one direction due to load - it probably "twists" from one extreme to the other many times per cycle. In contrast to a "cam position" that is just slowly moving in one direction over many cycles. So I think the resolution with just 4 teeth might never be good enough to show a trend that maybe completely reversed every 180°?

[Sean]

Thanks Adam, that all makes sense, even the bit about the reversed waveform The new DMC datasheet does mention the falling edge I see, now I have looked at it, the little data sheet that came with the DMC didn't

I take your point with the relative accuracy of the cam measurement versus measuring the phasing difference front to back over many teeth, likely I will verify the existing setup is working correctly, and then have a crack at the ACL method. The torsional vibration problems we are chasing are fairly unusual, they do occur in some 410 sprintcar engines as well. Having very small flywheel/clutch mass exacerbates the problem, we have just made a new flywheel over twice the weight of previous, which is helping. The whole driveline has a series of torsion bars built in, axles, driveshaft, gearbox shafts, input shaft which doesn't help... Even changes in chassis setup have an effect

[AdamW]

Sean, have you seen this article before: http://www.motoiq.com/MagazineArticles/ ... pment.aspx
The mathematics are probably a little over my head but it seems in this case they are measuring torsional vibration using just one high res trigger wheel...

[Sean]

Thanks to Adam W , it's all working! Still need to work on logging and can transfer rates, but its making some sense...


Here is the result, for those who interested. Given what we are seeing in the engine, all makes sense.