PDA

View Full Version : Big A' Servo?



Adrenalynn
06-01-2008, 02:52 PM
A while back Matt mentioned a servo that was pretty powerful. I can't seem to find it in a search.

I find myself in need of a beefy servo. Really beefy. I may end-up with a stepper, or with an encoded motor and drive.

Anyone have any linkage for beefy servos? Power and weight don't matter. Price matters a little. Like to keep it under a grand... [This is for my garbage bot that got a new platform last night courtesy of Lowes spring sale. ;) It's no longer a slip-skid, but rather has a scissor linkage turning mechanism.]

Sienna
06-01-2008, 02:59 PM
The most powerful servos I know:
Hitec HSR-5990TG: 330oz in, $125
Robotis RX-64: 800oz in, $290
Vantec SSPS105: 27 foot pounds, $590

Adrenalynn
06-01-2008, 03:26 PM
Vantec SSPS105: 27 foot pounds, $590


That's the badboy I was thinking about, thanks! http://www.vantec.com/ssps105.htm

Alas, it's slower than ... something really slow. .9deg/sec? 90 seconds to go 90 deg? Youch.

I've found some 5700 oz-in (about 30ft-lb) steppers that are a lot faster. That said, I'm starting to think my accuracy (just steering) doesn't need to be that high and a .5HP motor with an encoder might be the way to go.

Thanks for digging that up!

lnxfergy
06-01-2008, 03:53 PM
That's the badboy I was thinking about, thanks! http://www.vantec.com/ssps105.htm

Alas, it's slower than ... something really slow. .9deg/sec? 90 seconds to go 90 deg? Youch.

I've found some 5700 oz-in (about 30ft-lb) steppers that are a lot faster. That said, I'm starting to think my accuracy (just steering) doesn't need to be that high and a .5HP motor with an encoder might be the way to go.

Thanks for digging that up!

Isn't that .9 seconds to go 90 degrees? That's the way most servo speed readings are given, and would still be 1/5 of the speed of a typical servo (~0.18sec to go 90 degrees)

-Fergs

Adrenalynn
06-01-2008, 04:00 PM
Right you are! The first link I read said .9deg/sec. They really meant that it was .9 sec to travel the servo's max rotation of 90deg (+/- 45deg)

Ok - this looks a lot more attractive now!

4mem8
06-01-2008, 06:49 PM
Man, that's a meaty servo, One hell of a robot you are building Adrenalynn.

Adrenalynn
06-01-2008, 07:06 PM
:) As soon as they put the [literal] fires out at the datacenter - I'll post some pics of the new platform. I'm hoping in the next couple weeks I'll be able to build something to drive down the road. That servo is about all I'm lacking for round-one, basic "rolls down the road"...

Matt
06-01-2008, 09:58 PM
YAY! I have been wanting to see someone use these servos forever.

Adrenalynn
06-01-2008, 10:25 PM
:)

It was your comments that got me thinking that way. I'm still not certain it's the "right" solution. I'd like to be feeling closer to sure before I pull the trigger on it. That starts looking closer to "real money"... A whole heck of a lot easier than a stepper, though!

Droid Works
06-02-2008, 07:29 AM
I have to get a couple of those! I can think of 100's of cool projects for those:)

DresnerRobotics
06-02-2008, 09:42 AM
Isn't Hitec coming out with something akin to that servo in the short future?

JonHylands
06-02-2008, 10:01 AM
As a point of interest, in the latest Servo magazine, the Robotis ad mentions a new servo, the EX-106, which is the next step up from the RX-64. Assuming the name holds the same meaning, this one would have about 1400 oz-in of torque...

- Jon

Adrenalynn
06-02-2008, 10:25 AM
"Only" about 7.29 lb-ft... (1400/16/12) - not large enough for me, I fear...

JonHylands
06-02-2008, 11:14 AM
Yeah, I didn't think it would be, but thought I would throw it in, since the RX-64 was brought up...

- Jon

Adrenalynn
06-02-2008, 12:03 PM
Absolutely. It's collecting resources like this that make future projects do-able! If it hadn't been for Matt mentioning it (and Sienna finding it again for me), I never would have known that Vantec was available. Who knows what any of us will need for some project we haven't even considered yet!

sthmck
06-02-2008, 12:17 PM
Have you thought of using a linear actuator? Sorry if I am repearing a question that has already been asked.

JonHylands
06-02-2008, 12:26 PM
Something like this (http://www.robotshop.ca/home/products/robot-parts/motors/actuators/firgelli-full-scale-actuators/standard-actuators/firgelli-24-400lb-linear-actuator.html) comes to mind...

More available here (http://www.robotshop.ca/home/products/robot-parts/motors/actuators/firgelli-full-scale-actuators/index.html)...


- Jon

Adrenalynn
06-02-2008, 12:51 PM
We did look at them. It seems that any that were reasonably priced were just too darned slow. I believe (unless I'm repeating past mistakes) that the ones Jon lists, for example, are 0.5"/sec. Even if we use two, one on each side, we'd still need to throw a good ten inches to preserve accuracy in turning. 20 seconds of turning to lock just isn't practical. Anything over a few seconds probably won't cut it.

Adrenalynn
06-02-2008, 01:09 PM
Before anyone asks - yes. We've looked at hydrolics, and that's still not out of the question. We'll likely use them for the lifter mechanism anyway. That said - I try to avoid 'em. What a complicated mess. So I leave that to my partner and try to stick with an electric solution for the steering and drive. ;)

JonHylands
06-02-2008, 01:13 PM
Yeah, I didn't look at the speed. I've been looking at linear actuators for a while, for my full-sized raptor project, but I'll be building those from scratch, and they will be series elastic actuators, and a whole lot faster than the ones from Robotshop...

- Jon

ooops
06-03-2008, 11:48 AM
series elastic actuators
Holy cow that is the coolest actuator I have seen yet!

JonHylands
06-03-2008, 12:00 PM
SEAs are really quite amazing. They are compliant, can hold position (more or less) with zero energy expenditure, and are force-driven. Setting the desired force to zero allows you to manipulate the actuator using external forces (like your hand), and it becomes a "limp" arm. This is a really useful property for limb-based robots that learn movement.

- Jon

MYKL
06-12-2008, 04:47 PM
I need to look up what an SEA is...

Why hasn't anyone with alot of money used liner motors to actuate a supermech?
http://www.trilogysystems.com/
I sure would give them a try...

When you say big 'A' servo Are you talking about the servos made by TONEGAWA SEIKO?

I've seen them use these on Mythbusters to steer full sized cars and boats... In fact they are used all over the Special Effects industry for R/C and costuming effects.

I've been gone too long, I've got to catch up with all of your projects...

SPECIFICATIONS
Operating Voltage:
Vmin-max: Dedicated 4.8 - 12.0 VDC, not shared w/receiver
Recommended Operating Voltage: Vin: 6.0 - 8.4 VDC Starting Torque: Ms 65.33 Kg/cm (907 oz/in) @ 6.0VDC
...91.50 Kg/cm (1271 oz/in) @ 8.4 VDC Speed: 0.29 sec/60 deg. at 6v, 0.21 sec/60 deg. at 8.4V Weight: 280 gr.(10 oz) Size: Length: 100mm(3.94") including mounting flange
..........75mm(2.95") body only
Width: 44mm(1.73")
Height: 93mm(3.66") w/ output arm installed
.........79mm(3.10") body onlyHere's the manufacturers website:
http://www.tonegawaseiko.co.jp/pro/e_pro02.html

JonHylands
06-12-2008, 05:26 PM
I need to look up what an SEA is...[/URL]

Check out [URL]http://www.yobotics.com (http://www.tonegawaseiko.co.jp/pro/e_pro02.html) for some information. They came out of research at MIT, and as far as I'm concerned they are the be-all and end-all of large actuators for large limbed robots.

- Jon

Sienna
06-12-2008, 07:07 PM
SEAs are really quite amazing. They are compliant, can hold position (more or less) with zero energy expenditure, and are force-driven. Setting the desired force to zero allows you to manipulate the actuator using external forces (like your hand), and it becomes a "limp" arm. This is a really useful property for limb-based robots that learn movement.

- Jon

I have been thinking about this a while, and looking at the site, but I do not get your comment here.

From what I can tell, a SEA is nothing more then a linear actuator (such as a ball screw) which moves a pair of plates. These plates in turn have a third plate, mounted in between them, but separated from them by springs. The output shaft is connected to this third plate.

I understand the spring effect that this would give, and hence the compliance. However, I don't get where you say these are force driven. To me, it would seem that the range of output motion is the distance of the outer plates minus the lengths of the compressed springs, assuming that the central linear actuar is not moving. I do not see how you get a limp limb from this.

Can you explain any further, or point out where I am wrong?

Thanks!

Adrenalynn
06-12-2008, 07:46 PM
Hi MYKL,

Thanks for the recommendation!

The large Vantec servo detailed earlier seems to be about 4x more powerful than these. (1271 / 16 / 12 = ~6.62 lbft vs 27 lbft). The speed on the Vantec also seems a little higher. The operating voltage is a standard 12v which is probably desireable in my case.

Again, thanks! I'll keep them on my list for when I need a low voltage but reasonably strong servo.




I need to look up what an SEA is...

Why hasn't anyone with alot of money used liner motors to actuate a supermech?
http://www.trilogysystems.com/
I sure would give them a try...

When you say big 'A' servo Are you talking about the servos made by TONEGAWA SEIKO?

I've seen them use these on Mythbusters to steer full sized cars and boats... In fact they are used all over the Special Effects industry for R/C and costuming effects.

I've been gone too long, I've got to catch up with all of your projects...

SPECIFICATIONS
Operating Voltage:
Vmin-max: Dedicated 4.8 - 12.0 VDC, not shared w/receiver
Recommended Operating Voltage: Vin: 6.0 - 8.4 VDC Starting Torque: Ms 65.33 Kg/cm (907 oz/in) @ 6.0VDC
...91.50 Kg/cm (1271 oz/in) @ 8.4 VDC Speed: 0.29 sec/60 deg. at 6v, 0.21 sec/60 deg. at 8.4V Weight: 280 gr.(10 oz) Size: Length: 100mm(3.94") including mounting flange
..........75mm(2.95") body only
Width: 44mm(1.73")
Height: 93mm(3.66") w/ output arm installed
.........79mm(3.10") body onlyHere's the manufacturers website:
http://www.tonegawaseiko.co.jp/pro/e_pro02.html

JonHylands
06-12-2008, 08:53 PM
From what I can tell, a SEA is nothing more then a linear actuator (such as a ball screw) which moves a pair of plates. These plates in turn have a third plate, mounted in between them, but separated from them by springs. The output shaft is connected to this third plate.

You're missing one critical piece - there is a linear pot that is mounted to the two outer carriages, and the center carriage (which is attached to the ball screw) is attached to the pot slider.

With that, you can measure the amount of force being applied to the actuator. If you use that in the controller, you can imagine a feedback loop where the controller tries to keep the "force" measured (which is really the deflection of the springs) to zero.

Say you are using one of these SEAs to actuate an arm joint. The controller is instructed to make the limb go limp (maintain zero force), and you start to bend the joint. The springs on one side start to compress, and the pot measured the movement (which is interpreted as a force). The controller rotates the actuator enough to relieve the tension on the springs, until the pot is back in the center zero location. Thus you are "pulling" the arm, and it follows along, which is basically what a limp arm would do.

- Jon

Sienna
06-12-2008, 09:35 PM
Got it! Didn't realize there was a measurement device in the middle, but that makes sense then.

Has anyone made a SEA that the distance between the two outer plates is variable (controlled by another motor)? To me, it would seem doing this could dynamically change the joint's elasticity.