wow, that’s a lot better than the planetary gear motor on the LM site, with 36 kilos is has 14 RPM 
What does Idk mean?
IDK means I dont know. lol
I may need some of these super servos for the legs of Mega Scout. 
8-Dale
Yep, the quiet, reserved type like me who never ever ever post coz we are just too shy to post coz we might get flamed… 
Sorry for the sarcasm.
In my experience its not always how much power a device can apply, you also need to recognise how often it can repeat the application.
A lot of servo’s have equivalent or better specs but fail because of poor implementation. Casings, brackets, ventilation etc usually cause failure long before the device can reach its maximum potential.
What Jim (apologies for he familiarity) appears to be proposing here is a device that has a very sturdy construction, is not going to fail because of an oversight in design and is aimed primarily at robotics rather than as a secondary application…
I could be wrong though, It is a trait of mine…
yeah, I went on a couple forums where evyone is bitching evryone a burning them, so when your new and ask a stupid (for them anyway) question, well instead of answearing, they just burn you. I like this forum because it realy isn’ like that, we answer to all questions best we can, thats all
First we answer, than we burn.
- hehehe
Jim,
Got a question that I should have asked earlier. What’s the motor shaft diameter and did you get it with an extension out the back side? I’d be tempted to buy the motor itself (I’m one of the dsPIC OpenServo hopefuls) and slap a CUI capacitive encoder on the backside.
Nothing says “I move you accurately” quite like a $30 encoder doing 2048 PPR.
Here’s the specs for the GHM-04 motor with the external shaft.
lynxmotion.com/images/data/ghm04.pdf
I looked at the specs you linked, the motor speed is like ~9000 rpm so make sure you don’t get so much resolution that the encoder can’t keep up.
Ok, so I just switch down to 1024 PPR on that then (and I was asking about your SES motor, but this will do for thoughts at least!). Hmmm… 50:1 ratio would give me a resolution of 51200 lines per rotation, or 204800 counts per rotation. Looks like I’ll have to build an extended servo algorithm. This is the equivalent of about 6.3 arc seconds of resolution, but then I’d have to deal with gearhead backlash… oh well…
I don’t design PCBs for a living, but I do work on variable frequency drives and industrial servos (service, system integration, machine retrofit). I’m asking about this both for my own use and you mentioned that there was an encoder in the possible future for this.
(Addendum)
Actually, I’d have liked to see a rear shafting option on your PGHM-20. Might have to buy one and see if I can extract the motor shaft.
Sorry I didn’t even look at what thread this was from. The only motor we have that has an extended motor shaft to accept an encoder is the GHM-04. That motor is too fast (at the final gear), and not strong enough to be used for a servo. The encoder I was referring to is attached to the final gear, not the motors shaft. It’s the USDigital MA3-A: 10-bit Analog output (2.6 kHz sampling rate) magnetic shaft encoder. This will work for this project much better than a quadrature encoder on the motors shaft. Sorry for the confusion.
Jim,
when can we expect an update on this cheap, no contact, super secret absolute encoder? this is becoming a habit I don’t enjoy, you peak our interest and then just leave us hanging there. 
nick_a
drop me a line if you want an MECE to take a look
I have been referring to the USDigital MA3-A in many of these threads. So that’s no secret. I did mention an idea I wanted to experiment with and was kinda secretive about. I don’t know if I will ever get to actually make it so I will tell you what my idea was. In the same style as a potentiometer, place an LED and a phototransistor in the light tight housing and separate them with a disk that has a narrow slot in it that goes almost the entire circumference. It would be good for up to 320° of rotation. It would require a constant current circuit to keep the LED brightness constant, and possibly some algorithm to convert the output to a linear response. I know how I would do it mechanically, but I just don’t know if I will ever have time to complete it.
I don’t really remember being secretive on anything else. If I have, I apologize for any hard feelings. If you can point me to anything I have left hanging I would be happy to explain it, resolve it, um or deny it. Sorry Nick! 
Thats the one, simple and cheap, I like it.
I forgive you Jim, its hard to stay mad at you
Now what can the community do to help this photentiometer project along, and what were the constraints that would have kept you from finishing?
nick_a
Going by what Jim has said about workloads etc, I think a 9 day week, 36 hour days and an extra 3 hands would be useful.
I have seen something similar before, but never had a chance to play with it, because it was an ultra-precise lab-grade type of transducer. A ball-bearing supported shaft had a glass plate with a neutral-density filter that varied from 0 to 100% through a rotation, with a light source and sensor, enclosed in a lightproof case. It was also available in a full-rotation version, with two tracks offset from each other, for a sine/cosine absolute position encoder. Both were essentially non-contact sensors, which imposed no more load on the shaft than that from the ball bearing supports.
Very slick, but very, very expensive - a version that would be accessible to hobbyists would be great.
I think that can be taken care of with the following exceptions:
Jim adds 3 LYNX 6’s to his back
Jim begins working on weekends
Jim builds a time machine
hm, i cant wait for these things to come out, any estimate on when it might be coming out? i wish to use it for the arm that gonna be on my rover 
if u dont know, then thats fine and sry for asking, im just so PUMPED!
I’d love to get a hold of that controller board of Jim’s.