PDA

View Full Version : [News/Announcement] Dynamixel MX-64 and MX-106



pauljurczak
11-02-2011, 10:15 PM
I just concluded a nice email exchange with Robotis and they informed me about release of Dynamixel MX-64 and MX-106, which they hope will happen in December this year. Both servos will have TTL interface in MX-28 style. Price is expected to be the same or only slightly higher.

Xevel
12-06-2011, 05:48 PM
By the way, for those who have not seen the official video: http://www.youtube.com/watch?v=LVb0L9CuNdM

One interesting point: MX-64 and EX-106 will have real current based torque control. I'm exited to see what will be possible to do with it :D

Janis
12-10-2011, 11:44 PM
Hey, I'm shopping around for 100kg servo. Will it (the MX-106) be available within a week or so? Or should I get the EX106? (it's December...)

lnxfergy
12-11-2011, 12:35 AM
Hey, I'm shopping around for 100kg servo. Will it (the MX-106) be available within a week or so? Or should I get the EX106? (it's December...)

I wouldn't hold my breath -- Robotis announced on their twitter account "The MX-64 and MX-106 scheduled for release by December in Korea. (might be delayed though!)" about a month ago. Seeing as they still haven't hit the market in Korea, I doubt you will get them outside of Korea before the end of the year.

-Fergs

portola40
03-13-2012, 01:38 PM
Do you know when these will be available?

Thanks,
Jake

lnxfergy
03-13-2012, 03:35 PM
I got some engineering prototypes to evaluate in mid-January. Based on previous product releases, it's typically been about 2-3mo after that before they are available from retailers in the US.

-Fergs

Th232
03-13-2012, 11:56 PM
Interesting, if you're not under an NDA or similar, how big a difference have you noticed in comparison to the RX-64 and EX-106?

lnxfergy
03-14-2012, 12:34 AM
Well, in my opinion, the biggest change for the MX-64 vs RX-64 is the increased resolution, which for things like arms is awesome, but I'm not sure it helps much with walking robots. As my robot runs off 12V SLA, the fact that all the MX-family is targeted at 12V is a big win, as I now get more torque than I used to (I was previously running EX-106/RX-64s at 12V, so they didn't get as much torque as the rated value).

-Fergs

Th232
03-14-2012, 02:02 AM
Thanks for the info, looks like this'll be very interesting to watch. One last query, is there any change in backlash between the MXs and RXs, or is the geartrain &c all the same?

DresnerRobotics
03-14-2012, 11:16 AM
The only difference is the electronic board.

Xevel
04-26-2012, 08:18 PM
Just saw this : http://www.robotis.com/xe/126727#0
(http://www.robotis.com/xe/126727#0)
The 64 and 106 both exist in two versions, one with TTL interface (MX-64T / MX-106T) and the other with RS485 intarface (MX-64R / MX-106R)

They are available from their Korean webstore, but not yet appear in the English one. All the R versions are a little more expensive (+$6 or +$7) than their TTL counterparts.

Similarly, the MX-28 have been renamed MX-28T (for TTL it seems), and they added the MX-28R with the RS485 interface.

Th232
04-26-2012, 08:31 PM
Interesting, more variety is always nice, especially with the MX-28R.

This only just pegged when I was reading the specs again, the MX series have a resolution of 0.088 degrees, while the EX-106 has a resolution of 0.06 deg. So the EX-106 will still be the best for accuracy unless there's something I've missed? A 360 degree control range is still looking very nice though, especially for turrets &c.

Bullit
04-26-2012, 09:01 PM
The EX-106 has a smaller 260 degree working angle and the MX-106 should be 360 degrees. The EX-106 although having a magnetic encoder on the motor still contains a Murata potentiometer on the final output. The MX-106 contains a contactless austriamicrosystems high resolution magnetic encoder on the final output. Because the gears in the EX and the MX are the same the mechanical backlash should be the same however, the capability to hold an accurate position should be significantly improved with the MX series as it is with the MX-28 vs the RX-28.

Th232
04-26-2012, 10:37 PM
Because the gears in the EX and the MX are the same the mechanical backlash should be the same however, the capability to hold an accurate position should be significantly improved with the MX series as it is with the MX-28 vs the RX-28.

Here're the minimum control angles from the Robotis website from the range of movement and the number of bits:

RX-28: 300/1024 = 0.29 deg
MX-28: 360/4096 = 0.088 deg
RX-64: 300/1024 = 0.29 deg
MX-64: 360/4096 = 0.088 deg
EX-106: 251/4096 = 0.06 deg
MX-106: 360/4096 = 0.088 deg

So for the MX-28 and MX-64 there's a clear improvement on the accuracy compared to their RX counterparts, but I don't see how you're saying there's an accuracy improvement for the MX-106 over the EX-106. The EX might have a pot on the final output as opposed to the MX's encoder, but from the website both the EX and MX-106 have the same number of bits (4096 positions) over their ranges (251 and 360 deg respectively). Unless their website is wrong and the 4096 is over 360 deg in both cases, in which case they'll have the same accuracy?

Bullit
04-26-2012, 10:55 PM
I've worked with EX-106/106+ for sometime. I don't believe ROBOTIS states accuracy on their website. Control angle does not equal accuracy. The EX uses a low resolution encoder on the motor combined with a poteniometer mounted on the output shaft. The potentiometer doesn't have the resolution alone and with the encoder being removed from the output shaft by many segments of the gear train there is loss in precision from the mechanical backlash.

On the flip side the EX-106+ is considerably faster than the MX-106.
EX-106+ No Load RPM ([email protected])
MX-106 No Load RPM ([email protected])

Th232
04-26-2012, 11:05 PM
Fair enough, thanks for clearing that up for me.

I guess the MX's reduced speed is due to a different motor since the gearbox is the same, or is there some other factor in play (alteration in the feedback loop or something)?

Bullit
04-27-2012, 05:57 AM
Looks like the gear ratio is different too.
EX-106+ is 184:1
MX-106 is 225:1
So perhaps the motor is the same.

lnxfergy
04-27-2012, 06:16 AM
I imagine the motor is different is as well, since all the new MX servos are designed for 12V (which should make integrating various size MX/AX servos a lot easier). The EX-106/RX-64 were 12-18.5V (recommended 14.8), whereas the MX are 10-14.8V with a recommendation of 12V. The torque ratings are also at different voltages with the EX-106 being rated 8N-m at 14.8V, and the MX-106 being 8.4N-m at 12V.

-Fergs

jwatte
04-27-2012, 03:03 PM
Just checking: even though it says 360 degrees, here, that means it still stops after one turn -- it's not infinite?

Xevel
04-27-2012, 03:40 PM
As far as I can tell, there is still no support for multiple turns in Joint Mode in the communication protocol, even though the hardware could be capable of doing it now...
In Wheel Mode, the servo spins indefinitly and you control its turn rate but not its position.
Custom firmware anyone?

jwatte
04-28-2012, 05:49 PM
Yeah, I saw the youtube animation, and it seemed to indicate continuous turning. Now, if you could get continuous rotation, AND position control (as well as sensing back,) that would be useful. Sadly, I'm not yet ready to sell my car to afford these servos in volume, though...

Th232
05-26-2012, 02:12 AM
So... these are now on back order at Trossen. Wonder how long before they're in stock.

DresnerRobotics
05-26-2012, 07:51 PM
They're backordered because they were added to the catalog before we received the physical inventory. Due next mid next week.