Let say these guys are giving you 300 oz/in.
So a motor around $30
HS-805BB around $30
And there, for $60 you have a pot, control board and gearhead motor. I think it will even fit into an HS-805 case.
Only problem is these motors run on 12V and the HS-805 board will not handle that. Or the Amps that the motor will pull at stall so the weak link is the H-bridge and controller board. And I guess that’s where Jim’s evil genious comes in.
Now I’m going to go see if these motors fit into an HS-805 case. It’s bugging me now.
If nothing else, these could easily be fitted to work with Jim’s Mega SES set and ran on an over-volted 14.4V system. They would be some very impressive joints.
I would assume (or at least really like to see) that there will be a separate power line for motor power, making it easy to adapt it to uses other than its original one.
Adding serial and I2C control and a heavy-duty H-bridge to the usual complement of servo features, along with the motor’s performance specs, makes this a formidable beast to begin with. Considering the “hackability” of such a setup (replacing the motor with something else to be driven by that H-bridge, or playing with the feedback loop, for example), as well as having built-in SES hardpoints, all I can say is that this is going to be fun!
I knew this had “open servo” written all over it when I read your first post about this. Now, I am glad I have been delayed at converting W.A.L.T.E.R. (formerly known as Octabot) to use motors for locomotion.
The Open Servo control boards are not conversions of existing servo control boards. They are complete new designs with many added features, custom software. It’s all Open Source too!
Think digital servo on sterioids…
You should go over to Open Servo and read the forums to see exactly what capabilities are being worked on there. It is indeed very exciting! Yep, I think W.A.L.T.E.R. will get a couple of these for locomotion and Walk 'N Roll is just begging for this sort of thing to roll around on.
Yeah, chalk it up to the hazards of having multiple conversations going on in the same thread. The above response was actually a reply in which I thought a “normal” servo was being referred to, rather than something running an OpenServo controller.
I’ve since figured out where the conversation is pointing.
I got one of these cordless screwdrivers some time back to tinker with. It is 6v and has pretty good torque. The last I looked walmart had them in the tool section for ~$9.
I’m designing it for 180° range. To get more would require a totally different aproach, and it might not be as strong. This design is going to be strong enough to handle the torque available from the motor.
I dont know what kind of price point you are shooting for but i would imagine you will fall considerably below some of these 100 to 200 dollars servos. If this is the case i would take the opportunity to really make something unique. I dont know about the rest of the robot enthusiasts but i would rather pay a bit more and get something that outperforms everything else and is still open source.
i have a couple ideas but i could really use with some more information on those gearheads so i can model them up in SW. Also some pics of the internals would help too. Actually i tell you what, how much would it cost me to get one of those bad boys overnighted? Basically im wondering if i can hack this planatery gearbox into a new arrangement. I will try to throw a basic concept together before the end of the day. Though what i would really like to see is instead of a pot or encoder, use a magnetic rotation sensor. (We use them here at work for a spatially correspondent manipulator for subsea ROVs and they have preformed flawlessly)
It’s early to quote pricing, but I expect it’s going to be well under $100.00 and hopefully under $70.00. I only have a couple of the motors right now, so I can’t let 'em go. I have looked at the magnetic rotation sensor from US Digital. But they want $29.00 for them. Do you have a cheaper alternative? Is the IC they use a standard part? Stray magnetic fields must interfere with their operation eh? I am currently investigating a totally new approach to contactless rotation feedback. I can’t go into any more details for now…
why do you always have to leave us with such tantalizing words. Honestly some better clues would be nice.
my guess is absolute encoders but, most of the ones i have used are either delicate or have too many wires for small scale projects.
the magnetic rotation sensors we use are standard 8 pin smd ICs i think from honeywell but i dont have any details, I will see what I can dig up. Having worked in an environment that at best can be considered harsh i would imagine our R&D department has come up with a couple of solutions. I will look for a pic but our smd sensor is put over a 1/4" hole put into the board. the board is mounted to a thick plate with a matching hole where a shaft with a pre-installed magnet is placed in very close proximity to the sensor. The plate doubles as a structural component of the manipulator and the whole setup added less than 1/4’ to the width at the pivot point. all structural components where non-magnetic and the massive motors had minimal shielding. (the water tight motor housing is entirely out of al. and only the wires are truly shielded).
I’m not usually so secretive, but I need to do more testing before I go blabbing on about it. It uses light…
I just received my samples from US Digital. They look real nice, but if the IC is inexpensive I can’t see where the $29.00 is going. Here is an image of the progress. I received part “A” but am still waiting on part “B” to close the loop.
Maybe im wrong but that looks really long, could you give us a round about number for length. I think italian is right, it would work great for arms but im not so sure about legs and the like.
Im wondering if you could mount the motor right beside the gearbox. with a small plate to link the two, two plastic gears (they will encounter minimal load) and a small shaft the same diameter as the motor shaft. but thats just overly complicated.
Perhaps what you have and then a different version that uses bevel gears. those multi leg guys would go nuts with something that could lift all kinds of weight.
And just in case you haven’t, you may want to look into what the lateral load rating is on those rotary sensor housings.
The length is about 4.0" so it’s not that bad. Like I said earlier, I’m designing this specifically for arms. This is the smaller of the two I am planning. It can be used for the shoulder for an average sized servo arm, or for the elbow on a larger arm using the larger version of this for the shoulder. This arrangment is simple to implement and will do fine as a prototype for testing. After I have all the bugs worked out I can easily change it into different configurations.
The sensor spec sheet lists 2lbs as the lateral load rating. That may be plenty as is. However, in this configuration the motor output shaft is so close to the end of the sensors shaft, I think once the missing “B” part is installed it will make the arangment much stronger.
You said for the shoulder and elbow of an arm. Would it be possible to make a complete arm out of this? If the setup is too heavy, would it be possible to implement two ,motors on the heavier load lifting part like the shoulder?
This is an image of what I would like to build in the future, but about 1/3 of the size and much lighter. It would be for lifting 2-3lbs of matter.
It’s bugging me now.
Now, I am glad I have been delayed at converting W.A.L.T.E.R. (formerly known as Octabot) to use motors for locomotion.
