Top Stomach master/slave servo coupling

[Andreas]

Hi All,

I just finished printing the parts for Top Stomach and started the process of the assembly. While going through the instructions in detail I stumbled across the suggested servo modifications to couple them to a master/slave union. 

Is it wise to strip the complete servo controller from the slave and wiring the servo motor directly to the master? By that, the master servo controller has to handle twice the load of these quiet big HS805BB servos? Has anybody checked, if these miniature H-bridges in the servo controller can handle twice the power consumed?

Wouldn’t it be better, to only strip the potentiometer from the slave servo and feed the slider of the master potentiometer to the slave servo controller? With a splitted PWM signal as an input to both servos, the slave controller could power his motor for himself and avoid an output performance issue. Splitting of the PWM signal could be simply done by an Y adapter cable which comes with the servos.

Probably this would add more precision as well. I think that the linked video of the servo test (https://youtu.be/FAYjdWADEp8) shows a fast start of the left servo and an overrun on the right one.

What do you think?

Best Regards,

—Andreas.

[Morten Enholm]

I had the same thing, overrun on hk1529bb servos, what i did was to de solder the motor in on servo, turn it 180 degress and also swap the left and right pot wires on the same motor, then use a y splitt for the pwm :)

i have now both a programmer for the servo, so can probably do them same inside the program...

[Andreas]

Now that my servos arrived, I tried out my idea. 

***** BEWARE of that stupid idea! ***** 

It nearly killed my servos! The servos where a bit shaky first an then they sucked over 4 amps out of the battery. Thanks to my servo tester I saw that in time, before everything startet smelling.

After more thoughts I think, that the servo slider has to go to an AD converter, which got confused, when there where 2 of them in parallel. So there is now way to give feedback to 2 independent servo controllers with one potentiometer.

Today I made a second attempt and bought a "Stereo-Potentiometer". This is a 2 in one device as you can see in the picture. Now each servo can measure its own slider value and the servos behave fine. The problem is, the tolerance within the potentiometer. The total values differ about 10%. I tried to compensate that with resistors, but that seems to be a never ending task. 

In my experiment, one servo found its position and the other was still turning a bit. So this is not working again...

[Ray Edgley]

Just a warning on this,

What you are attempting is to have two independent control systems with separate feed backs that may not be exactly the same feeding in to the controllers that may not be the same doing the same thing while being physically linked.

Any error between the two control systems we result in the two systems fighting each other and if not spotted will result in the smelly blue smoke.....

You are far better of following the instructions created by Gael on the Inmoov site, that way uses only one control system and so there is no fighting and no blue smoke.

Ray

[Andreas]

Hi Ray,

you are probably right! I could identify the transistors on the servo electronics from the HS805 servos and checked the datasheets. They are supposed to be able to handle 7 amps. This should be enought for the 2 motors, so my concern is gone. On first thought I was only afraid that this H-bridge was overloaded by driving 2 motors in parallel. Thats why I came up with the idea to share the potentiometer for both servos...

Thanks for the advice.

--Andreas.

[Ray Edgley]

Andreas,

I love to change around the design of the electronic, that why my Inmoov, Fred, has no Arduino's fitted, instead i use Raspberry Pi 3 x 2 and a number of Adafruit I2C 16 Channel servo drivers.

But there are some things that just aren't worth changing.

Over that last six or so years, Gael has done a lot of experimenting with servos and all the electro-mechanical aspects of the Inmoov design. 

While some of it has been well engineered, other parts have been trial and error and all of it has a great level of artistic ascetics. 

I think for all of that side of it, we are better off following Gaels designs.

As for the control system, there is still a lot that can be done, and there will be a lot of trial and error there.

For anyone that is not confident with the diagnostics and lesser levels of support that comes with experimental designs, I strongly suggest following Gaels instructions, these have been tested as working and there is a lot of support for them.

If you want to live on the edge, then we will all want to hear of both your successes and failures. This is how we all learn as a group.  Even better if you do have a failure, if you explain what was the problem, it may lead someone down a different track that may well work...

You will find a number of the control system experimenters over at MyRobotLab.

Regards,

Ray