Re: [inmoov_notification] Controlling elbow server with Pi and servo controller

26 views
Skip to first unread message

gael langevin

unread,
May 12, 2024, 1:58:33 AMMay 12
to Stewart Hutton, InMoov
Hello,
The pot is giving the driver board the feedback of the current position. Once the pot is placed in the the small gear of the elbow the driver board will make the motor to reach the desired position. The pot is connected to it’s original wire placement on the servo driver board and is only giving feedback.


From my iPhone, Gael

Le 10 mai 2024 à 16:56, Stewart Hutton <ste...@stewarthutton.co.uk> a écrit :

First time poster, super slow builder :)

Looking for help as I have a hole in my understanding that I have not been able to fill myself!

I'm building my 'bot very slowly from the fingers up :) and now adding the bicep to a (working) forearm and hand assembly.

I'm using a Pi Zero and a 16 channel PCA9685-based servo control board.

Having extracted and extended the servo potentiometer, everything still works - in that on the bench the servo turns and with the adafruit servokit library I can change it's direction. And I can adjust the pot by hand to get the servo to stop (though this is very sensitive - maybe with the weight of the forearm hanging off it there will be more tolerance).
I can imagine how by matching that up with a specific bicep:forearm orientation that will stop the servo turning. But that only takes care of one end of the travel, let's say arm closed - how do I get it to stop at the other end of its travel, when fully open? Without reading the pot directly, I can't see how to stop the servo when arm is fully extended.

Is this a Pi issue? I've read the Arduino has analog pins not on the Pi, are they being used in the default instructions?
Should I connect the pot to the Pi via an ADC so I can read it and use that to set both limits?

I should add that I haven't looked past controlling all this via my own Python code, so haven't looked at MRL, RobotLab etc etc.

Thanks in advance :)

Stewart

--
You received this message because you are subscribed to the Google Groups "InMoov" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inmoov+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/inmoov/2bb4c1d0-0aa3-407d-a3fc-8b6b301f178fn%40googlegroups.com.

Stewart Hutton

unread,
May 12, 2024, 8:35:41 AMMay 12
to InMoov
Thanks Gael - understood.

The main problem here I think is that the feedback from the pot is analogue and the Pi has no analogue pins to understand its feedback on.

I have a multichannel ADC chip on order and *think* my best plan will be to connect the pot to the pi via to read the pot values. That might actually work better for me as I could stop the servo at each end of its travel based on a slightly larger range of pot values than the direct-to-servo connection seems to use. 

Stewart 

gael langevin

unread,
May 13, 2024, 3:49:40 AMMay 13
to Stewart Hutton, InMoov
You do not need pins for the pot on your PI board. The pot is coming out from the servo case which has a driver board. So the pins of the pot are connected directly to that driver board. The only thing you need is to make a three cables extension between the driver board and the pot.

From my iPhone, Gael

Le 12 mai 2024 à 14:35, Stewart Hutton <ste...@stewarthutton.co.uk> a écrit :

Thanks Gael - understood.

Stewart Hutton

unread,
May 13, 2024, 6:38:52 AMMay 13
to InMoov
Aha - got it last :)

Thanks - and sorry for the confusion. I'll explain my mistake here in case it helps someone later ...

Because we remove the physical stop inside the servo to allow it to turn continuously, I was treating it as a "continuous servo" in the software - one that does not know what angle it is at.
But I see now (doh) that because the pot is still providing the angle info to the servo (it's just measuring it from a different place) we still use code that directs the servo to specific angles. I've simulated that now with the servo outside the bicep structure so at last am confident to proceed with minimal chance of exploding any printed parts!

Thanks for your help and patience :)

Stewart 

Reply all
Reply to author
Forward
0 new messages