Timing belt measuring bench

[krista]

Even reading the pitch with a CMM was a struggle. I ended up clamping a bar of aluminum at the edge of the belt and carefully laying the probe in the corner of a tooth pressing straight against the bar. It took about 4 or 5 tries to get data that I committed (which I showed in the drawing).

I recall setting up the sprockets and belt and measuring the perceived straight ends of the belts, but I think I did that before I turned the sprocket bolts on the lathe. I feel OK enough with the accuracy of the CMM to work about one sprocket to revisit the whole "diameter" issue, so I may revisit this. I do recall feeling really confident about the 55.70 mm number. Enough that with the latest version of the test bench Arduino code, it actually isn't adjustable anymore (while before it was with the up or down joystick input). I've been ignoring other obligations at work obsessing with this, so I may let the nuance of diameter go for now because like you said, it is apples to apples, so this is a great tool.

Chris

[krista]

A picture with another view of the test bench which shows the DRO scale and control panel mounting. And my hack photoshop job of making the background stuff darker. The plastic box next to the control panel is also bolted to the bench and holds the various springs and "blocks" used to tension the belts of different lengths. Conveniently, the power pack (or wall wart) for the bench fits in there, too!

[koko64]

The man is into it.
Maniacal committment 
That's why I buy his stuff.

[ducatiz]

Amazing drugs they have in Cali.

[krista]

I realized today as I was wrapping up the project that I needed to make the instruction sheet. And it was then I realized that the test bench DIDN'T have any feedback about if a belt is too long.

So I took the bench home with me to play with its software one more time... I added a bit to display if I think a belt is good or not. I uploaded the changes to Github and here's the commit verbage:

Added visual feedback to operator about lengths of belts in measurement vs history of belts we have measured. The bench calls a belt good if it is within -0.007" / +0.010" of our current "control" ExactFit belts for each model.

Do note that the global beltMax[] array has the maximum observed c2c values of used OEM belts (when available). The OEM 900 belt is allowed 0.020" longer c2c than we call good and the OEM996 belt has been observed at an astounding 0.040" longer center to center.

Updated the code in calibrate() in the upload section to remind you to unplug the wall wart for uploading. This way we don't have to unseat the 7166 IC, which is a PITA.

This is part of why I'm on the call out to get "good used OEM belts" from people or shops. If you've got a used OEM belt and you didn't think anything was wrong when taking it off, I'd like to get it please! We're good on the 900 and 996 belts, but I'd like more data on all the other models about what a "good used" OEM belt's length is. This way I can edit the software in the tool to be more forgiving of belts I measure and be less strict.

As always, we're having to research extra hard to see how wide of a variance is considered normal. 

Chris

[Dave R]

sorry to cause you headaches    once again  

this is why I have always been a staunch supporter of Chris and the business he has built!  His dedication to the Ducati brand and fellow Ducatisti through thick and thin is very admirable.  A wonderful example of how a dealer/vendor/service provider should perform, especially within the motorcycle community  

At SUB we choose to stock CA-Cycleworks belts exclusively.

Chris, watch for a pile of lovely OEM ones coming your way.... we take care of lots of red bikes these days  

[krista]

Dave, FDR, F*king Ducati Rider,

You're the best. You were the ONE DEALER that would listen to me pregnant dog about my Ducati and then still care!! Don't worry sir, you're only HALF the problem/reason all this happened. There was another shop that got two 1198 belts; one was so tight that at full slack it tensioned 150 Hz and the other one was OK.

Between the two of y'all, I knew I needed to be better informed about this weird concept of "belt length". I really appreciate your commitment to US and supporting us as ardently as we support you.

I can't wait til we meet again so I can buy you a beer. BTW, we'll be at world superbike in laguna.

Thanks!
Chris

PS - Dave Roosevelt was GM of Ducati Seattle circa 1999, when I called up and we first started talking shop. + Dave was the 1st to GAF about a certain 1992 black 900SS with a broken frame. Now he's running SUB

[krista]

Ok, so if you liked this thread, here's a preview of the TPS test bench. I take OCD batshittery to a whole new level...

Why?

Because YOU doing makes us very very happy.

And then WE can either or instead of because it really hurts us when you are and not and/or .

Chris

[jaxduc]

Chris

 

[ducatiz]

I still have no clue wtf you're doing this for.

What are the dimensions at issue that you need to test for?  Tooth shape/diameter/width?

[krista]

ducatiz, total belt length.

[suzyj]

Ok, so if you liked this thread, here's a preview of what's next. I take OCD batshittery to a whole new level...

Why?

Because YOU doing makes us very very happy.

And then WE can either or instead of because it really hurts us when you are and not and/or .

Chris

Hi Chris,

I've done a few automated test systems for my work that are kinda similar to what you've done (albeit with a lot of RF test gear bolted in). I usually code this stuff using National Instruments Labwindows, which is a visual C compiler with heaps of machinery and test widgets, like dials and gauges, and the ability to talk to most all electronics test equipment.

The data acquisition part is often done with National Instruments DAQ boards, which just plug into a PC via USB, and let you monitor voltages, and switch things on and of, or control stuff etc.

Anyway it's all good gear, and probably a bit easier to work with than those teeny little screens.

Check it out: http://www.ni.com/lwcvi/

The DAQ boards: http://sine.ni.com/nips/cds/view/p/lang/en/nid/201986

Some guys at work also like Labview, which uses all the same widgets as Labwindows, but is programmed entirely graphically: http://www.ni.com/labview/

[krista]

suzyj : whoa that stuff is spendy! If we were that rich, I'd hire an engineer to do it for me.

[koko64]

Chris, your OCD is to our benefit!
Cheers .

[krista]

Thanks, koko!! Very good to hear it's appreciated. BTW, you'll crap yourself when you read about the TPS test bench I just finished.

Oh, and if there's a v2.0 of the belt measuring bench, it's so full of batshittery, I'll need a bigger forklift... We're talking linear stages and load cells, folks. Stuff that took me weeks of googling to learn how to google them. Lord help us if I find a way to dynamically print onto the belt surface!! That'll get me fired up like no tomorrow.