SSC-32
SSC-32
by PaulP
Sat Mar 03, 2007 10:34 pm
Well. Received my SSC-32 today and began fitting it..
So far so good. Its a larger board but with a bit of jiggling its in place and connected.
BlueSmirf connected in with no hassle. Had to remove all the inversion stuff but the SSC comes with links that the smirf can connect directly to.
The board powers directly from the 10.8V and has a couple of spare 5.0V outputs so powered up smirf from there.
Now to learn how the Sequencer works....
So far so good. Its a larger board but with a bit of jiggling its in place and connected.
BlueSmirf connected in with no hassle. Had to remove all the inversion stuff but the SSC comes with links that the smirf can connect directly to.
The board powers directly from the 10.8V and has a couple of spare 5.0V outputs so powered up smirf from there.
Now to learn how the Sequencer works....
by PaulP
Sun Mar 04, 2007 12:36 am
Surprising well actually.
After a bit of trial and error I managed to save a reasonable home position and have now managed to reproduce a passable reproduction of the Swinging arms routines.
I was a bit dubious at first about the servos but I looked up the ranges, 180deg for the KRS788 and 270deg for the 4024's and set everything to not exceed those ranges.
It also allows me to further reduce the movement by marking the points at which the servo's / limbs may start to conflict with each other and prevent them from going beyond that point.
For my first real attempt, it all seems so far to be going reasonably well.
There are some things that are immediately lacking. The ability to teach a position for one but I didn't master that on the RCB till recently so I don't know how much I shall miss that.
I've yet to work out the speed aspect of things yet. I can't see an obvious way of making certain poses quick and others slower...Still its only day one...
After a bit of trial and error I managed to save a reasonable home position and have now managed to reproduce a passable reproduction of the Swinging arms routines.
I was a bit dubious at first about the servos but I looked up the ranges, 180deg for the KRS788 and 270deg for the 4024's and set everything to not exceed those ranges.
It also allows me to further reduce the movement by marking the points at which the servo's / limbs may start to conflict with each other and prevent them from going beyond that point.
For my first real attempt, it all seems so far to be going reasonably well.
There are some things that are immediately lacking. The ability to teach a position for one but I didn't master that on the RCB till recently so I don't know how much I shall miss that.
I've yet to work out the speed aspect of things yet. I can't see an obvious way of making certain poses quick and others slower...Still its only day one...
by PaulP
Sun Mar 04, 2007 5:28 pm
Wrote a file converter today to transfer the RCB files to the SSC.
Managed to get the bowing sample working quite well. Slight variations in posture mainly due to the difference in home position.
Slight variation in how the degrees are represented as well but single steeping thru the application highlighted which ones were causes issues. Should have guessed it would be the Shoulder rotate servo's
Difference in the way speeds are represented. KHR uses a speed value whereas SSC uses a time value. In other words, get from position A to position B in this amount of time. I suppose there is a formula somewhere to convert the two but for the moment its easy enough to adjust that.
The positions are the important bit.
Managed to get the bowing sample working quite well. Slight variations in posture mainly due to the difference in home position.
Slight variation in how the degrees are represented as well but single steeping thru the application highlighted which ones were causes issues. Should have guessed it would be the Shoulder rotate servo's
Difference in the way speeds are represented. KHR uses a speed value whereas SSC uses a time value. In other words, get from position A to position B in this amount of time. I suppose there is a formula somewhere to convert the two but for the moment its easy enough to adjust that.
The positions are the important bit.
by shsan
Mon Mar 05, 2007 12:00 am
PaulP.
Just in case you missed it (I doubt it),
But to convert from position to angle on the RCB3, you need to multiply by:
0.346 for 788HV and 2350HV
0.52 for 4014HV
0.5 for 4024HV
And I am not convinced that Speed is really Speed in the H2H either. It might be time too.
Just in case you missed it (I doubt it),
But to convert from position to angle on the RCB3, you need to multiply by:
0.346 for 788HV and 2350HV
0.52 for 4014HV
0.5 for 4024HV
And I am not convinced that Speed is really Speed in the H2H either. It might be time too.
by PaulP
Mon Mar 05, 2007 1:07 am
Thanks shsan, I did spot that but it has odd effects when compared to the data stored in the files. I think it may only apply to those joints that arent set to zero when you assemble the robot initially.
Most of the joints have a positive and a negative aspect to their movement but channels 3 and 7 (upper arms) dont so that caught me as well. Instead of -90 to +90 they are 0 to 180. Still its all learning.
By trial and error I managed to spot those angles that needed to be halved in value and those that didnt.
I wasnt expecting to be able to play them straight away on the SSC, I expected one or two adjustments.
Got them across now and alls fine (ish) for the moment.
Most of the joints have a positive and a negative aspect to their movement but channels 3 and 7 (upper arms) dont so that caught me as well. Instead of -90 to +90 they are 0 to 180. Still its all learning.
By trial and error I managed to spot those angles that needed to be halved in value and those that didnt.
I wasnt expecting to be able to play them straight away on the SSC, I expected one or two adjustments.
Got them across now and alls fine (ish) for the moment.
by PaulP
Mon Mar 05, 2007 1:10 am
BTW the speed / time thing almost solved itself.
If its time then the lower the time, the quicker it has to move, If its a speed then its inverted so the lower the number the quicker it has to move.
Either way, all I had to do was put a scalar factor in to relate to something SSC understands. 100 on the RCB equates to approx 1000 on the SSC so scaled by 10 and it came out about right
If its time then the lower the time, the quicker it has to move, If its a speed then its inverted so the lower the number the quicker it has to move.
Either way, all I had to do was put a scalar factor in to relate to something SSC understands. 100 on the RCB equates to approx 1000 on the SSC so scaled by 10 and it came out about right
by shsan
Mon Mar 05, 2007 1:17 am
Well good to hear, so it's "time" then.
What is the SCC expecting for that value (what is the doc saying?) ?
Because I had one issue when I was doing the simulation in Webots with that same value and I am not sure that I coded the simulation properly because of that.
Somehow It's working but it seems to be a bit slower than expected.
Also I found that when I was doing some measurements with that value it was not a linear field but more an exponential one.
So you might have issue if the SCC is expecting a linear value.
What is the SCC expecting for that value (what is the doc saying?) ?
Because I had one issue when I was doing the simulation in Webots with that same value and I am not sure that I coded the simulation properly because of that.
Somehow It's working but it seems to be a bit slower than expected.
Also I found that when I was doing some measurements with that value it was not a linear field but more an exponential one.
So you might have issue if the SCC is expecting a linear value.
by PaulP
Mon Mar 05, 2007 1:32 am
On the SSC you allocate a time period for the servo to complete its motion.
Generally, similar to RCB, you have a sequence of poses. For each pose you allocate the time say 1000 for it to reach that position. The SSC will then move all the servos so that they arrive at the same time.
You can override each individual one if you want. If for some reason, one servo has to finish quicker (or slower) than the others then you can set a different value for that individual.
It has all sorts of grouping functions as well that allow you to say that this particular group all act together etc.
Generally, similar to RCB, you have a sequence of poses. For each pose you allocate the time say 1000 for it to reach that position. The SSC will then move all the servos so that they arrive at the same time.
You can override each individual one if you want. If for some reason, one servo has to finish quicker (or slower) than the others then you can set a different value for that individual.
It has all sorts of grouping functions as well that allow you to say that this particular group all act together etc.