PDA

View Full Version : More Powerful Servos (THIS is the edit!)



Bill Chennault
12-04-2006, 04:45 AM
All--

I am interested in beginning to experiment with non-Parallax supplied servos using the Parallax Servo Controller (because I have it and the support is great; but I COULD get something else, if necessary).

I have found this brute (HSR-8498HB HMI (http://www.servocity.com/html/hsr-8498hb_hmi_robot_servo.html); stall torque 103 oz-in) at Servo City. I wonder how many of these the ·Parallax Servo Controller will handle before meltdown? (Am I thinking correctly when I believe the stall torque and the free-running torque are important, but the TYPICAL torque is MOST important, and practically unknowable?)

Or, if attached to the PSC, will the PSC just·say, Huh?

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Post Edited (Bill Chennault) : 12/3/2006 10:11:09 PM GMT

allanlane5
12-04-2006, 04:58 AM
The important number here is the current the servo pulls at stall torque. The main current supply line passes pretty directly from your battery to the servo -- so that shouldn't melt the PSC. The 'control' line has very little current on it -- so that shouldn't be a problem either.

The only problem you might have is draining your battery pretty quickly.

And I believe the Servo control signal -- normally low, with a 1 mSec to 2 mSec high pulse, repeated every 20 to 50 mSec -- is pretty standard.

Bill Chennault
12-04-2006, 05:16 AM
[size=-]allanlane5 (http://forums.parallax.com/member.php?u=41268)--

[size=-]Thank you, once again.

[size=-]I am trying to figure out if a machine with, say, 40 motors would be cheaper to build using servos, dc motors with motor controllers, or stepper motors. Of course, a complex machine is likely to have a mix of all three. However, for the primary actuator, I keep coming to the conclusion that a servo with some kind of (external) encoder feedback might be the way to go.

[size=-]Clues from you?

[size=-]Of course, such a device, despite the chosen actuators, will require some sophistication using multiple, networked Stamps.

[size=-]--Bill

[size=-]

[size=-]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

allanlane5
12-04-2006, 05:55 AM
Tricky.

The chief benefit of Servo's is a small, inexpensive motor in a small, inexpensive package that gives you 50 to 150 oz-inches of torque. And that the control signal is easily generated by a BS2 with no additional hardware. This makes it idea for small robots, for which 50 oz-inches of torque is plenty.

Now, when you get to 40 motors, probably some of those motors will be moving some of the other motors. Even 150 oz-inches of torque is not much in these cases.

So, you'll need to figure into your design several issues:

1. How much torque will each motor need?

2. How much torque will each 'flavor' of motor generate?

3. If you need positioning accuracy, does that go along with full rotation? Servo's only give you 180 degrees of rotation with positioning accuracy. Steppers are quite good in this regard, but require fancier control signals.

Then, 4. How are you going to power all these motors? Current control and supply are the issues here.

Bill Chennault
12-04-2006, 06:26 AM
allanlane5 (http://forums.parallax.com/member.php?u=41268)--

The good news is that since joining this forum and buying a bunch of Parallax educational stuff, I can not only pose the questions, I can even understand some of the answers!

IN THE BEGINNING, I had hoped to standardize on a single motor-type. By doing so, I could stock replacements and quickly become expert in handling that type of motor. Soon, it became apparent that this would not be possible. Now, I realize that not only will the motors be of different SIZES, they will be of different TYPES.

There is no problem here that processing power will not solve. I am already thinking of graduating from the Stamp environment to the Propeller world. This may be especially attractive when Parallax introduces the Propeller with the Stamp pinout.

Currently, my thoughts are to dedicate a BS2px24 to each leg. Using stepper motors and a multiple (are THERE multiple stepper motor controllers) stepper motor controller would certainly reduce the number of Stamp pins necessary. Of course, I would use Stamp pins for encoder feedback, as well. (My current favorite technology is Hall-Effect.)

Yes, motors will have to lift motors. Therefore, the smaller the better. But, remember that at worst case, there will be eight motors doing the lifting. In the "best" case, there will be 24 motors doing the lifting. Too bad we cannot design for the best case!

Considering the worst case, with eight motors doing the lifting, it seems to me that I can get by with motors of lower capacity, since there will be eight of them. In fact, each motor would need 1/8th the capacity of a single motor tasked with the lifting.

I THINK servos (with external encoders for negative feedback correction) are a possibility for the primary lifting machines. The AX-12 (http://www.crustcrawler.com/products/AX12/index.php) servo looks perfect for most of my needs. However, its controller seems obscure, so far. Maybe, it is nothing special. However, the on-board AX-12 intelligence/capabilities seem (to this layman) far beyond the Parallax Servo Controller.

As always, I appreciate your thoughts as I learn and hope to return them when I have learned.

--Bill



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.