AX-12 Settings
AX-12 Settings
by JonHylands
Thu Jan 18, 2007 4:49 am
Hi everyone,
I'm having a bit of trouble with one of the servos on MicroRaptor. Specifically, if you look at the following image:
http://www.bioloid.info/tiki/tiki-browse_image.php?imageId=100
The joint I have circled in red is the one with the problem - it holds up the neck. I'm hoping this is just a setting problem, but I've tried a pile of different settings, and the servo basically chatters when it is holding the weight of the neck anywhere close to horizontal.
I've got the compliance slope set to 200, the compliance margin set to 2, the speed set to 500, and the punch set to 200. I've tried varying most of those values, and they have different effects. If I drop the punch back down to the default 32, the servo stops bouncing and just squeals. If I set it higher, it just bounces worse.
I'm hoping there is just some combination of servo parameters I have missed...
Anyone have any ideas?
- Jon
I'm having a bit of trouble with one of the servos on MicroRaptor. Specifically, if you look at the following image:
http://www.bioloid.info/tiki/tiki-browse_image.php?imageId=100
The joint I have circled in red is the one with the problem - it holds up the neck. I'm hoping this is just a setting problem, but I've tried a pile of different settings, and the servo basically chatters when it is holding the weight of the neck anywhere close to horizontal.
I've got the compliance slope set to 200, the compliance margin set to 2, the speed set to 500, and the punch set to 200. I've tried varying most of those values, and they have different effects. If I drop the punch back down to the default 32, the servo stops bouncing and just squeals. If I set it higher, it just bounces worse.
I'm hoping there is just some combination of servo parameters I have missed...
Anyone have any ideas?
- Jon
by BillB
Thu Jan 18, 2007 12:09 pm
Micoraptor is looking fantastic. I hope you are getting it working, I look forward to more videos.
Just a few thoughts - the neck looks quite long and perhaps when it is close to the horizontal (parallel to ground) there too much torque (due to gravity) on the servo. Especially since it is holding aloft 3 servos and 7 brackets.
My bioloid is currently in the King Spider configuration with a similar number of servos and brackets on each leg - but they are never in a completely horizontal position and the servo closest to the base rotates on a plane perpendicular to the ground (sonsequently which does not need much torque).
So I suggest that you first try and shorten the length of the neck and see if the servo performs any better - if it does perform petter then try and change the axis of motion of the servo you have highlighted in red.
Hope this helps.
Good Luck
Just a few thoughts - the neck looks quite long and perhaps when it is close to the horizontal (parallel to ground) there too much torque (due to gravity) on the servo. Especially since it is holding aloft 3 servos and 7 brackets.
My bioloid is currently in the King Spider configuration with a similar number of servos and brackets on each leg - but they are never in a completely horizontal position and the servo closest to the base rotates on a plane perpendicular to the ground (sonsequently which does not need much torque).
So I suggest that you first try and shorten the length of the neck and see if the servo performs any better - if it does perform petter then try and change the axis of motion of the servo you have highlighted in red.
Hope this helps.
Good Luck
by JonHylands
Thu Jan 18, 2007 3:01 pm
BillB wrote:Micoraptor is looking fantastic. I hope you are getting it working, I look forward to more videos.
Thanks. Its moving along, slowly, but its a big task.
BillB wrote:Just a few thoughts - the neck looks quite long and perhaps when it is close to the horizontal (parallel to ground) there too much torque (due to gravity) on the servo. Especially since it is holding aloft 3 servos and 7 brackets.
Its actually only holding two servos. I've got the same configuration on the neck mount that the hip and foot use, with two servos giving two axis of rotation. In this case, pitch and yaw of the neck. The yaw servo works fine, its the pitch servo that has the problems. I'm 100% sure its a torque problem, but these are 16 kg-cm servos, I would think they should be able to hold that much weight. MicroRaptor's head is going to get a bunch heavier when I add the stereo cameras, so I'm clearly going to have to either change or just get rid of that axis of movement.
Thanks,
Jon
by JonHylands
Thu Jan 18, 2007 3:04 pm
Bullit wrote:I'd try compliance CW/CCW slope of about 20 and punch of 1. The more punch the more chatter.
I hope this helps.
Bullit,
That definitely gets rid of the chatter, but with those settings it just whines...
I think I'll have to change the configuration, like I said in my previous response, because with the cameras on there its only going to get worse.
- Jon
by Pev
Fri Jan 19, 2007 12:11 am
Okay, I am not sure if the physics is right here but........
If the issue is a torque one that is not fix with a configuration/settings change, could you possibly put 2 AX-12s back to back and have a custom bracket with a driven horn at both ends.
As I say I am not sure about the physics but I think that should just about double the holding torque at that joint........
Well maybe
If the issue is a torque one that is not fix with a configuration/settings change, could you possibly put 2 AX-12s back to back and have a custom bracket with a driven horn at both ends.
As I say I am not sure about the physics but I think that should just about double the holding torque at that joint........
Well maybe
by billyzelsnack
Fri Jan 19, 2007 12:34 am
That's sometimes done in the RC world, but it's considered naughty because any two servos are not exactly the same performance
However!! Since we have feedback with the servos we can make sure they are mostly in sync.
Good idea. Might make for a good strong torso in a humanoid without having to buy the fancier servos.
However!! Since we have feedback with the servos we can make sure they are mostly in sync.
Good idea. Might make for a good strong torso in a humanoid without having to buy the fancier servos.
by JonHylands
Fri Jan 19, 2007 1:15 am
Its an interesting idea in general, but specifically for this robot it would be very difficult to fit an extra servo in place there. I would basically lose one of my yaw axis, and that is more important to me than the pitch axis.
The reason the yaw axes are important, in case you're wondering, is so that MicroRaptor can move his head left and right (to aid balance) while keeping his cameras pointing straight ahead.
- Jon
The reason the yaw axes are important, in case you're wondering, is so that MicroRaptor can move his head left and right (to aid balance) while keeping his cameras pointing straight ahead.
- Jon
by jonkster
Mon Jan 22, 2007 12:26 pm
may be way off track but... how much current are you drawing when this happening?
I have had a problem in past running a number of AX12's at high load with an inadequate power supply - some servo's would start oscillating. Fixed when switched to power supply better able to handle current draw.
just a thought
Jon
I have had a problem in past running a number of AX12's at high load with an inadequate power supply - some servo's would start oscillating. Fixed when switched to power supply better able to handle current draw.
just a thought
Jon
by JonHylands
Mon Jan 22, 2007 1:47 pm
I don't think the current is the issue. This happened even if that was the only servo under power. I measured the current draw when the robot is "at rest", with the servos switched on but not trying to position themselves, and it was around 650 mA. So even if that one servo was drawing an amp, we're still way under the limits of what either the battery or the power supply can provide.
In any case, its a moot point now, because I have removed that servo from the robot...
- Jon
In any case, its a moot point now, because I have removed that servo from the robot...
- Jon
by lnielsen
Tue Jan 23, 2007 5:17 am
Would you consider a DX-117? I know they are not cheap 
but it should give you the power you need.
I hope to start working on a Parallax Propeller controller that will support both types of servos for this exact reason. Currently I am planning on sharing a single communication bus so I don't have to worry about which servo I want to replace. The two things you need to deal with are combining the receive level converters and providing a different voltage for the servo power.
but it should give you the power you need.
I hope to start working on a Parallax Propeller controller that will support both types of servos for this exact reason. Currently I am planning on sharing a single communication bus so I don't have to worry about which servo I want to replace. The two things you need to deal with are combining the receive level converters and providing a different voltage for the servo power.
by JonHylands
Tue Jan 23, 2007 12:57 pm
Its an interesting idea, but it would require another power source, which I am not ready to do.
As an aside, I don't understand how you plan on "sharing" the bus, given that the AX-12 bus is one-wire serial, and the DX-117 bus is two-wire RS-485 (which is a fundamentally different protocol).
- Jon
As an aside, I don't understand how you plan on "sharing" the bus, given that the AX-12 bus is one-wire serial, and the DX-117 bus is two-wire RS-485 (which is a fundamentally different protocol).
- Jon
by Bullit
Tue Jan 23, 2007 4:59 pm
The CM2+ coming out from Robotis will have ports for all ax / dx & rx series servos including can.
http://www.robotis.com/html/sub.php?sub=2&menu=3
http://www.robotis.com/html/sub.php?sub=2&menu=3