Manoi AT01 vs. Robotis Bioloid. You make the Call...

[NullARC]

I originally posted a similar question like this on another forum. Thought I'd try it here and get a few more opinions.

Originally I included the new KHR-1HV in this 'shootout' but I've narrowed it down to just the AT01 and the Bioloid.

First I decided on the Bioloid, then I jumped over to the AT01. :rolleyes:
I want the best kit that suits me and the things I would like to do. If I hade more 'disposable' income I wouldn't even ask. I'd just buy both!! :D

I guess what is most important to me is to be able to make a biped that can be upgraded to around 23 DOF. Something I can hack and upgrade. (more torque etc.) Can execute smooth, precise movement sequences.

Here are the pro's and con's as I see it:

----------------
Robotis Bioloid:
----------------
Pro's:

1. Very versitile construction system. Customization should be fairly simple.
2. Additional servos only roughly $50.00 USD.
3. Servos have 300 degree rotation range.
4. 22 DOF robot already built though not tested yet. Look here:http://forums.tribotix.info/forum1/54.html
5. Comes with 19 servos out of the box.
6. Great price Approx. $900.00 USD.
7. Servos created for robotic use, with many more features than other standard biped kits.
How useful will this feature be to me? More useful in the future?

Con's:

1. Currently no gyro support for added stability. Big liability??
2. Relitively low torque servos (16.5 kg/cm). Though higher than AT-01.
3. Higher torque servos (DX-117's 28 kg/cm) are approx. $195.00 USD. ouch!
5. Would need M2 controller if upgrading to 14.4v DX-117 servos. ?? not sure ??
This is a concern.



------------
Manoi AT-01:
------------
Pro's:

1. Can be upgraded to 20 DOF fairly easily. (shoulder/waist mods in manual)
2. Has gyro and accelerometer support.
3. Additional KRS-4024 HV servos only roughly $60.00 USD.
4. Can be easily upgraded to KRS-4014HV servos for a massive 40 kg/cm torque!
5. Servos have 260 degree rotation range.
6. Widely touted as the most advanced biped.

Con's:

1. KRS-4014 servos cost approx. $145.00 USD. Again, ouch! (though less than DX-117's)
2. Servos only have a torque rating of 10.5 kg/cm. Is that enough??
3. Only comes with 17 servos out of the box.
4. $1,500.00 USD Price.

Please add your pro's/con's as you see them and your opinion on what would you think would be the best option for me.

Thanks for your help!

[JonHylands]

For me, the biggie with the Bioloid is the servos are serial devices on a bus, and they are extremely configurable/programmable. I'm going to be building my own controller, so the CM-5's capabilites (or lack thereof) don't concern me at all.

I'm going to be using my Bioloid kit for research and development in AI, so having the torque feedback and compliance control that the AX-12 servos give me is vital.

- Jon

http://www.huv.com/blog for details...

[NullARC]

For me, the biggie with the Bioloid is the servos are serial devices on a bus, and they are extremely configurable/programmable. I'm going to be building my own controller, so the CM-5's capabilites (or lack thereof) don't concern me at all.

Thanks for your reply Jon! This certainly is a very powerful tool. However I was questioning whether I would use those capabilities. I was afraid that it might be "overkill" for me. But in hindsight, if my abilities is this field progress far enough, having the resources that the Robotis servos provide makes the Bioloid more valuable. It makes the Bioloid a better platform for more advanced robotics.

Will you be sharing schematics and code? Following in your footsteps will help many of us who follow! Moving control of the robot from on-board to a remote site solves the problem of computing power to robot size ratio. I will be following your blog to see your accomplishments! :)

I'm going to be using my Bioloid kit for research and development in AI, so having the torque feedback and compliance control that the AX-12 servos give me is vital.

You're much farther along in this field than I am. At first, my biped will mostly be for "fun". Learning how to use the existing kit to walk and execute sets of moves. Making more "hardware" mods. Who knows, like I said, if I get far enough along I might tread where you are now walking!

Although the Bioloid is a better value (I think) because of it's price and capabilities. I still feel like for what I'm doing now, the AT01 might be the better choice. I can upgrade the servos for less, utilize existing hardware (gyros etc.) while still using the controller that comes with the kit.

UHG!! I like them both, I still don't know!! Hmmm, maybe I could take out a second morgage... :rolleyes:

- Marc

[HarryBotter]

Hi Marc,

Looks like you've taken a close look at both kits and have the differences pretty well pegged.

I've got a Bioloid kit and can tell you that its met all my expectations. I'm looking at the Manoi also but for only one reason, to compete in ROBO-ONE. The Bioloid is better value for just about any other use. It may even be suitable for ROBO-ONE competition, it's just that nobody has entered one yet.

I'm also curious why you're so concerned about upgrading to stronger servos? The AX-12s are much stronger than the 4024s and probably suitable for almost anything you'd like to try.

A lot of people are looking at ways of adding sensors more easily to the Bioloid and that may happen soon. Have you taken a look at my Bioloid humanoid balancing using servos loads?

http://www.botmag.com/forum/showthread.php?t=313

Harry

[NullARC]

I've got a Bioloid kit and can tell you that its met all my expectations. I'm looking at the Manoi also but for only one reason, to compete in ROBO-ONE. The Bioloid is better value for just about any other use. It may even be suitable for ROBO-ONE competition, it's just that nobody has entered one yet.

Thanks for your input Harry. This sounds encouraging! For some reason I got the impression that the Bioloid would not be competitive in Robo-One comps. I'd certainly like to see someone give it a shot!


I'm also curious why you're so concerned about upgrading to stronger servos? The AX-12s are much stronger than the 4024s and probably suitable for almost anything you'd like to try.

The biggest reason why I had this concern was that I had read in some other forums that there were concerns that when extra servos and connectors were added the torque of the AX-12's would be insufficient to carry the extra weight. But form what you say this seems like it might be unsubstantiated. This might be a product of people not having enough experience with the Bioloid. What do you think?

A lot of people are looking at ways of adding sensors more easily to the Bioloid and that may happen soon.

I have heard that these attempts were being made. But I was afraid that these procedures would be too advanced for me to implement. I think that one of my major hangups with the Bioloid kit is that, for some reason, I perceive it as more advanced because of it's features and therefore beyond my abilities. Maybe this is not the case.

Have you taken a look at my Bioloid humanoid balancing using servos loads?

I actually have seen this video. It is pretty incredible and very intersting! But once again I was afraid these techniques would be too advanced for me for some reason. I think I might be underestimating myself. Either that or I'm overestimating the Bioloid kit. :) I'm very interested how this was accomplished. Could you explain how it was accomplished a little bit? Also, where might I look to learn more about what is going on in the Bioloid world?

I think you have successfully brought me back over to the "Bioloid camp".

Thanks for your help!!

- Marc

[HarryBotter]

Hi Marc,

The jury is still out on the Bioloid's suitability for ROBO-ONE although I'm really not certain why it can't be used. Modifying the frames may be more difficult and adding balancing sensors is not as easy but both can be accomplished. Of the two bots it may just be that the Manoi is more suitable out of the box.

Regarding servo strength the Robotis offering has more torque out of the box than any other manufacturer. This amount of torque is certainly adequate unless you want to build a 35 servo monster or hang a few extra pounds of sensors off the frame.

The Bioloid kit has a very simple programming language that should be easy for beginners to get on to. The balancing routine I put together simply checked the ankle servos for increasing loads. If it found any, it would move four of the bot's servos to reduce that load. Pretty basic actually.

One other point is that the Manoi instructions haven't been translated to English yet while the Bioloid kit's has.

My advise would be to pick up a Comprehensive Bioloid kit unless you specifically want to enter a ROBO-ONE contest and be somewhat competetive right away.

Harry

[JonHylands]

Will you be sharing schematics and code? Following in your footsteps will help many of us who follow! Moving control of the robot from on-board to a remote site solves the problem of computing power to robot size ratio. I will be following your blog to see your accomplishments! :)


I will be sharing schematics and code for all of the embedded level stuff I am doing, including any custom bus sensors/actuators I build. The high level (AI) stuff will likely remain private for now.

- Jon

[NullARC]

My advise would be to pick up a Comprehensive Bioloid kit unless you specifically want to enter a ROBO-ONE contest and be somewhat competetive right away.

I think you're right. The Bioloid Comprehensive kit does seem like the way to go. I seriously doubt that I will be entering any ROBO-ONE contests any time soon! ;) It's a long swim to Japan! :)

Regarding servo strength the Robotis offering has more torque out of the box than any other manufacturer. This amount of torque is certainly adequate unless you want to build a 35 servo monster or hang a few extra pounds of sensors off the frame.

I'm VERY happy to hear that torque is not a problem. It did seem rather strange to me that there were torque questions (at least perceived by me) with servos that have 60% more torque than most other kits. Very encouraging!!

I will be sharing schematics and code for all of the embedded level stuff I am doing, including any custom bus sensors/actuators I build. The high level (AI) stuff will likely remain private for now.

Excellent! That was what I had the most interest in at this point. Thanks for sharing your experiences. Like I said, it is very valuable to have people ahead paving the way!

[NullARC]

Harry,

I just noticed something really funny. I was revisiting some of the web pages I have bookmarked pertaining to the Bioloid. While I was doing so I came to this page: http://forums.tribotix.info/forum1/54.html

If you look at the very last post in this topic, you'll notice one of the posts that got me questioning the AX-12's torque when adding more DOF. The writer of the post voiced such a concern. Now, here's the funny part... check out who made the post...................................HarryBotter !! :D :eek: :D

It seems you were one of the people who was scaring me off of the Bioloid!!
The post was dated October 28th. I guess you've changed your mind since then!! ;)

- Marc

[HarryBotter]

Good one, you've been doing your homework! :)

That was actually a reference to something odd happening to the ankle servos in one direction when the load starting increasing dramatically. I notified Robotis about it and they thought it might be a bug and were going to look into it. It wasn't actually related to the available torque of the servo.

Robotis suggests not going heavier than 2 Kg or higher than 350 mm with any robot. The robot that Peter is building in that thread you reference is higher than 350 mm.

Harry

[tempusmaster]

Robotis suggests not going heavier than 2 Kg or higher than 350 mm with any robot.

Given the respectable torque rating, those numbers are lower than I would have expected. Any idea why they're so conservative?

[HarryBotter]

Given the respectable torque rating, those numbers are lower than I would have expected. Any idea why they're so conservative?

I'm not really sure myself but guess that it's more of a support issue than anything else. The servos are certainly capable of handling more than that with a fully charged battery.

I've noticed a situation in the humanoid where some servos have periodically and momentarily "relaxed" when put under sustained high loads as when the humanoid balances on one foot. I've assumed it was related to lowering voltage levels in my battery pack but am not certain of that. I'd hoped to do some further checking on that condition when I had more time. Have you noticed anything similar with your humanoids?

[tempusmaster]

I'm not really sure myself but guess that it's more of a support issue than anything else. The servos are certainly capable of handling more than that with a fully charged battery.

Those parameters are at the low end of the typical ROBO-ONE, more like the ROBO-ONE J bots, yet the servos seem capable of much more from the specs and what I've seen posted.


I've noticed a situation in the humanoid where some servos have periodically and momentarily "relaxed" when put under sustained high loads as when the humanoid balances on one foot. I've assumed it was related to lowering voltage levels in my battery pack but am not certain of that. I'd hoped to do some further checking on that condition when I had more time. Have you noticed anything similar with your humanoids?
I haven't, but some of the Hitec folks have reported similar behaviors.

[NullARC]

Robotis suggests not going heavier than 2 Kg or higher than 350 mm with any robot. The robot that Peter is building in that thread you reference is higher than 350 mm.

I understand the weight restriction. But the height restriction has me a bit puzzled. Any clues as to why that restriction? The only thing I can think of is getting too much length on the end of a servo and creating too much leverage. Kind of like a fulcrum. Like putting a steel bar on the end of a wrench to get more leverage on it. Any weight on the other end would be compounded by the length of the connector.

[HarryBotter]

I think you understand the height restriction too. Or we're both wrong. :)