bartturner

I purchased some of the DigiKey momentary switches, https://www.digikey.com/...t-detail/en/MTG72AD2 I originally used with Ultegra Di2 and they worked without a problem on my P4.

I recently purchased a P5 and built the bike with Di2 7970 and used the DigiKey switches. Everything is stock besides the cable splices on the front wire harness and the Digikey switches. The problem is that I can not get clean shifting in the rear. The shift will happen but not stop. The rear derailleur will continue to engage on it's own and make additional, random, shifts. I can hear it continue to engage when the Digikey switch is in an open position.

I am curious if anyone has run into this problem. I have not done much debugging on this yet. It is only a problem on the rear and not the front. It seems to me it could be a couple of problems

1) There is a short of some kind in my wiring causing the issue. I think this is most likely. But I have done many, many wire splices and never had a problem before.
2) These momentary switches work find with Ultegra Di2 but are not compatible with the old (7970) Dura Ace. The thinking being that the little black boxes that I harvested cleanup the signal from these switches to only send one shift to the rear derailleur with Ultegra.
3) The rear derailleur is defective. I actually believe this to be the most unlikely but you never know.

I am curious if anyone else has ever run into this issue? Any tips appreciated.

What I plan on doing

1) Disconnect the DigiKey switches and manually short the wires and see what happens with shifting. If this fixes it I plan on swapping the front and rear Digikey shifters and see if this helps. If this does not help then undo my wire splices and try the same thing with the wires directly out of the front wire harness. If still have issue with wires directly out of the front wire harness then I have a tough decision. Either replace everything with stock so purchase the Shimano front shifters and a new front wire harness that has not had the wires cut. Or just replace the rear derallieur. Thoughts?

2) The other option is just replace the 7970 with my Ultegra Di2 which I am not really using. But this would require me to re-cable the bike. Move the 7970 to different bike.

Thanks for any help in advance!

Continuity

Hmm - I don't suppose you have, or have access to an oscilloscope, do you?

It sounds like it would be interesting to see what the actuation of the originally-specced Shimano switch looks like plotted out...

wphamilton

Sounds like your switches need some de-bouncing, from a general electronics perspective.

adclark

This was my first thought too, but I don't know much about how the Di2 works. I can say from experience with a microcontroller though that if you don't have anything setup to debounce a switch, you will probably read several button presses every time you press it.

Looigi

Input can be edge sensed / interrupt driven or polled. Either way, it's not uncommon to debounce in software/firmware rather than use a hardware debounce. (e.g. 3 successive reads of closed for a make and 3 successive reads of open for a break). I don't know how the Di systems are set up. Have you disected or know what's inside the Shimano Di switches themselves, what type they are (SS or mechanical?), any additional components?

ThermionicScott 

This may not be relevant at all, but if I were to hack Di2, it would be to add some kind of analog/"friction" mode so that I could always adjust my shifts from the cockpit.

FBinNY 

Yo, This is a bike mechanics forum. You need an electrician.

As electronic shifting proliferates, and there more DIY done with it, I wouldn't be surprised if they broke out a separate electronic shift sub-forum.

Sorry I can't help you. I'm familar with the mechanical aspects of Di2, but have no idea of the electronic logic that controls it.

Looigi

or possibly a programmer... Bikes now comprise mechanics, electronics, and software...

Back in the day I used to be able to work on my car, then they became "computerized", but at least I could work on my motorcycle, but now they're computerized too.

wphamilton

Software debouncing is a little trickier than you'd think at first grasp. Code using the obvious delay-and-sample schemes might work fine after tweaking for the switches used but fail for other switches. If the DI2 is using sw debouncing I'd lay odds that this is the problem. Either way it's probably a lot easier to add a debouncing circuit than hacking the DI2 code.

Given the cost of Di2, I'd want it to trim the shifts automatically. A sensor of some kind to know precisely where the chain is. I'll bet it does move off a bit after a shift already though.