I have been following this for awhile. And I just created a new account. After searching for a cheaper servo, I believe I may have found it. Has anyone tried the Robostar SRS-3216htg? They may be a good substitute for the MX-28. There is also the SRS-5314Htg as well. They both can be found at the Hobbyking web site. Here is a link: http://www.hobbyking.com/hobbyking/store/__64645__RoboStar_SBRS_5314HTG_280_Digital_Metal_Gear_High_Voltage_Robot_Servo_53_1kg_0_14Sec_81g.html
Dynamixel AX 12 or AX 18 are budget servos but there is a servo that has magnetic encoder instead of potentiometer and ARM Cortex M3 @ 72Mez/32Bit instead of Atmel AVR microcontroller, contactless absolute encoder & PID ie MX 12W. Unfortunately its torque is very less for robotic applications involving torque, made for wheel mode only.
So…I was thinking if we change electronics of AX 12 or AX 18 with MX 12W then we can have AX torque with PID gain in it if possible. It this idea good … can someone give some insides?
Found a guy tried with the reverse it but not sure about the results.
What do you say.
Actually MX-12 has a very low torque and are made for wheel. It would require to change the gearbox so you can use them as position control servo.
I never tried this motor but it looks interesting
I bought this servo SCS15 from Feetech for my first robot, it is 15kg-cm torque, with feedback, daisy chained.and the price is really attractive. this lady helps me a lot.
Rosita,Skype: sales018feetechrc,Email: firstname.lastname@example.org(agreed)
Just starting to get interested project. This might be a something that already has been talked about, but is there a reason why the project needs high-resolution 360° rotational servos to start with? I get the torque requirements but very few human joints need more then 180° and it doesn’t seem really that necessary to have .00* accuracy when 1/4 of a degree is plenty high for self stabilization.
it seams that we could pretty easily trade unusable degree of movement and over precision for decreased cost and increased torque and maybe faster reactions.
Poppy is made for research on walking but thinking about passive walking. The servo used are unique not because they have high precision or 360 ability. It is because they can be completely compliant (the torque is actively set to zero when disactivated.
Well… As “completely” compliant as any coreless dc motor with gears can get
Actually there is an internal active friction compensation so it is indeed completely compliant.
So the torque is not actively set to zero. Plus -I suppose- that for that friction compensation, first must be sensed some kind of movement (or current), therefore the static friction is there, at least for the very beginning. But that’s out of topic (sorry) and I agree that Dynamixel servo are the best choice for Poppy so far.
I read this post and i didn’t found any solution which will be the best to replace MX-28 and will not require to change a structure of poppy body. Is there anyone who create whole robot with different servos or is sure some of solutions?
There are several solutions to the servo motor issue. I’ve designed a case for standard servos that fit the poppy chassis. You have to remove the bottom and modify the stock mounting points of the standard servos but everything lines up nice. I’m using arduino and lynx motion ssc32 controller and unity to do the programming of body movements. Everything is pretty easy. Unity even writes the arduino code for you one the animations are done. John Choi and his halley ambassador bot is where I took inspiration from and mob Bob project. My robot is still basically a really cool unteathered marionette but it’s a fun little project. Working on using me old Samsung s4 as screen for eye animations and some higher function aspects of the robot. Somewhere I have the .stl posted feel free to download and print them.
You have any updates on your robot or videos (if completed). We are also looking at poppy but the dynamixels
are still very pricey…
i haven’t made any new progress with my robot aside from creating a case that turns standard servos into dynamixel servos. i did find some 15kg servos online but if yet to allocate funds to purchase them. the little 4kg servos did get the upper body to wave and make a really cool zombie crawl across the grass at halloween last year. other than that i really havent done much with the robot. its still in pieces. i’m a team of one and sort of lost interest in the robot for the time being. programming is not my strongsuit and going my own route with arduino code and the ssc 32 proved to be beyond my capabilities and interest level at the moment. Definately interested in a collaboration or providing an stl of the servo case if its not still up on the forum.
Hi everybody, new to this forum.
I’ve seen that some links are broken because this topic is a littlebit outdated. But my question:
Can someone tell me if it’s a good solution using this adapter with this servo motor specified to 25kg orque? Does it need a special protocol to communicate or “servo” standard? Thanks. Chris
welcome in this forum.
The AX-12 case can get a servo of size 40.7 x 20.5 x 39.5 mm.
The DS3325 servo size is 40 x 40.5 x 20 (when you cut the two screw plates of the sides). It may be forcing but you can modify the case so that it enters.
To control this kind of servo, there is no need of protocol. It is a standard servo you can control with a servo controller (radio), or an arduino. You can also control with a Raspberry Pi but there is a jitter if you want to make fine control.
thanks for the answer. 2 more questions, is there a source file for the case I can adapt for the servo I will choose? With the AX-12, I’ve seen they are connected in serial and communicate via specific Bus protocol, right? If I use the servos option, is there a documentation that could help me to understand how to connect to the Odroid and how to adapt your core program to address the servos? Or is there a fork of the project that implements this? Thanks, Chris