Sram eTap rear derailleur sporadically stops shifting - then performs all queued shifts at once?

[Jugi]

Also, when pairing the new rearm do I need to re-pair the front, or only the rear?

RD is the host of an eTap system. So, when introducing a new RD into the system, everything is paired to that.

[sfo423]

Quick Q:

When allaigning the rear der for first time with high - low cogs, do you:

(a) use high - low limit screw, then micro shift
(b) use micro shift, then set high - low screws

[Jugi]

Use microshift first to setup the position at the smallest sprocket. From the factory, limit screws should be wound out so far that it's possible to correctly microshift the RD in all dropouts.

Do that before attaching a chain. When RD's position is correct, attach a chain, setup low limit, move up and setup H limit. Move to bigger chainring and setup B screw.

[XCProMD]

Not sure if this change from flex-ribbon to one-piece PCB have something to do with the RD failing, if it is lack of ESD protection that kill some chip or as someone pointed before, the servo protecting itself when too much resistance is applied due to dust or grime on the pivots. It would be nice to have a look inside a failing RD

I'm starting to think this might be the core reason, ie. low voltage (or intermittently no voltage) going to the servo. There may be a mechanical aspect to this, as the issue may start or become more common after dirty riding conditions.

This is how my troublesome derailleur's (short cage) issues developed during last week (during preparation for and riding Tour TransAlp). I have been using it with an SRAM Red 22 chain and a XG-1190 11-28 cassette. Atleast on the outside, the RD was clean as there weren't any wet or dirty conditions encountered during the week:

June 19th, 40 km / 1h 22 min (last shakedown ride before packing the bike):
RD freezed two times during the ride. I could get same symptoms to appear afterwards on a workstand. Disappeared after a while, so decided to leave the RD on the bike. I had a properly working eTap RD one meter away from me on my TT bike, but decided not to swap them and packed the bike ahead of a flight to Innsbruck. I detached the RD and it got to fly in it's original packaging inside my bikebox.

June 22nd, 18 km and 30 m alt. gain / 54 min (transitional riding in Innsbruck):
No faults to report.

June 23rd, 90 km and 1140 m alt. gain / 2h 48 min (1st stage Innsbruck - Brixen):
Two freezes which happened almost consecutively. Both came up on a single, quite short (less than 10 minutes spent) uphill section.

June 24th, 142 km and 3330 m alt. gain / 6 h 1 min (2nd stage Brixen - Kaltern):
Two or three freezes. First during a long climb up to Pender Joch and once during final 10 kilometers on a slight uphill. Seemed to happen only on easier gears (bigger sprockets) but, bigger sprockets did see a lot of use all through TransAlp.

June 25th, 135 km and 3150 m alt. gain / 6 h 16 min (3rd stage Kaltern - near Bormio):
RD froze to the biggest sprocket for approximately 30 minutes. We were at a constant 6-8% grade all the way, so didn't really need bigger gears. Stopped and tried to manipulate it by hand (I thought I had set a limiter screw a bit too tight, therefore making the RD hit that and stop there). Miraculously just started to work again during riding, right when the grade eased to 3%. No faults during the rest of the stage, although gear shifts were few and far between as we were either bombing down a fast descent or climbing at 8% or harder grade. I observed the jam happening at the warmest time of the day. Morning hours went just fine and problems started in the afternoon. After the first stage, afternoon temperatures were well over 30°C for the rest of the week.

June 26th, 134 km and 3240 m alt. gain / 6 h 32 min (4th stage Bormio - Livigno):
First freeze after a kilometer of riding. Before noon it seemed like things were mostly fine when on the bigger chainring and freezes happened when on the smaller ring. During afternoon started to experience freezes on both chainrings and in all sections of the cassette. The general pattern seemed to go like this: 1. RD freezes while shifting to an easier gear. 2. After a waiting period of 5-20 seconds, the RD moves one step to a harder gear without any additional button pushes. 3. If trying to shift to an easier gear right after that ghost shift, I usually ended up on the gear I originally wanted. If not, sometimes an additional ghost shift happened dropping me to an even harder gear. I recall at least two occasions where I had three ghost shifts in a row, without shifts to the direction I wanted in between. That wasn't much fun.

Most freezes happened between the largest and 2nd largest sprocket. During final two climbs of the day, I resorted to moving the RD by hand a couple of times, as that delivered the gear I wanted immediately (but on the side of the road). My racing partner made a suggestion I should build a cable actuated system, which would nudge the cage into the wanted direction when it freezes. Can't beat dry humour at 2000 meters above sea level.

After the 4th stage I washed the RD with a pressure washer (to clean, lubricate with water and cool it as I suspected the relatively high temperatures) and didn't observe any changes in behaviour. Then decided to quit playing with it, first of all for other rider's safety (to avoid disruptions in a large pack if can't get a proper gear in). I resorted to a loaner bike for stages 5-7.

Right now this particular RD is sitting on a shelf. When trying to shift without chain tension or other outer influences, it seems to complete 5-8 shifts into a particular direction and then freezes. After a short waiting period (some seconds) it continues to operate by user commands. Just for general interest, I will continue to play with it and try to recognize any patterns there may be to this. I won't be dismantling it as that goes above my know-how, but if anybody in Europe wants have a look at it, I would consider sending it over if we split the fees involved.

I'm the second owner and out of warranty. Already ordered a WifLi version to replace this, because I wanted a longer cage for future visits to the Alps. Although WifLi's seem to have been more prone to this issue, atleast I will have full warranty period to use & abuse it.

I think I have found one of the causes why what you so precisely describe happens. I don’t know if it’s the only reason why a SRAM eTap RD freezes though:

I had a SRAM X01 RD that became lazier at going into smaller cogs, particularly at the small end of the cassette. The obvious culprits were cable and casing, so I installed fresh ones.

Situation improved but not at all solved. Then I swapped bike with a friend and his exactly same setup (X01 shifter and RD) was noticeably more positive and accurate than mine.

My bike’s dropout is spot on, so I checked the RD itself for alignment (Shimano RD’s for instance are easy to bend at the parallelogram arms), and was also close to perfect. So next step was disassembling the parallelogram pins (circlips) and clean them.

Bang! Now that RD shifts perfect. Since I put new cable and casing I was thinking this and the eTaps I’ve seen failing could have the same problem, as mechanically they share a fair bit of stuff. So I tried cleaning the pins on an eTap that we had in the parts bin and that had been unresponsive for a while.

Again, bang. The derailleur shifts all the time now.

I think that SRAM derailleur pins have this brilliant possibility of being disassembled, but in the other hand, unlike Campag, they are sealed poorly. They can get sticky and that triggers the servo protection when the H-bridge is drawing too much current.




Skickat från min iPhone med Tapatalk

[Kesselflicker]

@XCProMD:

Can you take pictures of the pins?
would then better understand what you did exactly.

Thank you

[ChiZ01]

Should have stuck with Shimano.

Wise words

[XCProMD]

It’s those pins. This is another X01, but you get the idea.

The sealing in those can be ruined by solvents, for instance. Still, it’s true Campag derailleurs have a much better seal design and I’ve never seen one with broken seals that get sticky or allowing debris inside. Shimano’s can’t be disassembled.


Skickat från min iPhone med Tapatalk

[Calnago]

@XCProMD: I’ve never had occasion to disassemble a Campagnolo rear derailleur like that, but I’m contemplating a special project whereby I would refinish the rear derailleur to a custom spec. That would require complete disassembly to get a proper finish on everything. Have you done that before? It’s 2015 Campy 11sp I’m thinking of. I think it might be hard to get everything back together with the springs needing to be set while reinserting the pins, but just thought I’d ask as I don’t know anyone who’s done it personally. So before I start hacking away at it any words of wisdom? Other than “Don’t!” Of course, if that’s the correct answer then I will abide by it. Maybe.

[XCProMD]

Can be done as the spring is concentric to one of the pins and gets assembled at a fairly low tension. More often than not, it will stay in place due to the grey plastic sleeve it is concentric to:


Especially if you pop out the prong the longer end of the spring rests on:



Which is not difficult to do once the outer parallelogram likely is out as in my pics, pressing from the opposite side of the spring.

The thing has a flange but also a slot that allows to press it past the inner parallelogram link

Then the inner parallelogram pins are pressed in, but they can be pressed out without damaging the sealing or the interference fit at the knuckles.

The one in the picture is a “11s Revolution” model (2011+) but as far as I remember the same applies to the “11s Embrace” models (2015+), no matter if the knuckles are carbon fibre.


Skickat från min iPhone med Tapatalk

[Calnago]

Thx, mostly I just want total unencumbered access to the outer plate, with the graphics on it. Presume getting that off was the easy part then.
Or wait, is that the outer plate, still attached to the knuckles etc, just turned inside out. Little disoriented I guess. Where is the carbon plate with the graphics on it?

[XCProMD]

The carbon plate with the graphics on it is missing in the above pics. That’s super easy as the pins are just bolted in. All you need is a 2 mm Allen key (EDITED: it’s a Torx 20 that’s needed). Putting everything back together is not a problem at all, the spring won’t cause any problem.

You’ll see the quality of those hard anodised pins. Something to behold.


Skickat från min iPhone med Tapatalk

[Calnago]

Good to know, thanks!

[XCProMD]

My apologies, incorrect info above. What you need is a T-20 head to undo the pins:





Skickat från min iPhone med Tapatalk

[Kesselflicker]

Thank you XCProMD

[Nickldn]

Should have stuck with Shimano.

Wise words

I was told by LBS that Shimano can usually fix electronic problems on Di2 and it is possible to disassemble these units. Not so for SRAM etap.

These issues have really put me off SRAM as a company. I really like electronic shifting, but next time I'll go for Di2.