Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Ultegra 6800
(10-17-2020, 03:33 AM)Zaphod Wrote: So I got a smaller horn with 10mm radius. Not much luck with the 20kg servo, got another with 35kg & 270 deg travel. That one pulls the derailleur however the current during hold is quite high, I see like 820 mA on the largest sprocket, about 600mA on the 2nd smallest and like 9mA on the smallest (idle). That’s only the servo. I guess that can only come from the derailleur spring force and not any friction. Some pics attached.
Zaphod, sorry you have such a hard time getting good results.
There are some questions I have:
1. with the 35kg servo, does it fully use the 270deg travel it has? If it's not (say, it's only using 90deg to make the derailleur go through its range), you still need a smaller diameter servo horn. You could try with an in-between piece as I stated in this thread to see what the results are. That would save you to make new servo horns.
2. are you running the Shift4Me script?  In the script, there is code that does an over-pull and then slight return. This is to overcome cable tension. (you can see it happening, and with the effect it has on current consumption, in the video)

I've made a video where you can see the currents I'm getting. Derailleur is a Shimano Tourney (which has the strongest spring of the ones I had laying around), servo horn is a diameter 13 (radius 6.5). The servo is a PowerHD LF20MG, going through 160degrees. As you can see, peaks during shifting go to around 1.2 amps, holding in gear is usually 40mA (the over-pull-and-return is not always successful). I'm also pulling the derailleur back at some point, to simulate stronger spring tension.
Watch the video here on YouTube
Hi Jan, I used the ‘calibrator’ script as base with a button increasing the PWM by 50 ms each press, so similar what you have in the video just on the bike with the wheel. I added an INA226 chip to measure current/battery voltage and then stepped through all the gears while monitoring the current, so that’s not peak it’s the hold current. I think I can still reduce the radius somewhat by adding a groove. I’ll try that next. I may also find one of these luggage weight scales to measure the force on the cable.
Looking at your pics again, I see you have not installed a bowden cable. You definitely have to do that. So there has to be two “stops”  with the cable hose in between (as you see in my vid)
For the hold  force I don't see why it would make much of a difference. The servo is fixed on the frame near the point where the cable runs. The cable runs w/o hose near to the derailleur, somewhat close the derailleur there is a stop. Another stop when using normal shifter is quite high on the frame, the hose that goes into the shifter normally goes there. Anyhow I can try that too. Attached pic shows the derailleur, I think similar to what you had in the video just that my cable from the hose to the servo is somewhat longer.

Attached Files Thumbnail(s)
Ok, that test-setup indeed should work.
The Calibrate script (obviously) doesn't have the overpull-release. So currents could be good as it is. If you upload the Derailleur script and measure current again, you'll see a whole lot less current while holding gear.
That being said, are you using the full range of your servo to go through all gears?
i’m getting a new smaller horn with 5.5 radius, that should just cover the 11 gears range.

Forum Jump:

Users browsing this thread: 1 Guest(s)