Modification of HS-311

[Fleet]

Hello,

I'm having some problems modifying my HS-311. When I adjust the pothead, the gears stop but after a while, they start to spin again. This happens again and again. Am I doing something wrong?
I used this guide: http://flickr.com/photos/randomskk/2569969633/in/set-72157605559564231/
It's unfortunately not possible to replace the potentiometer with two 2.2K 1-2% tol. resistors (airman00's solution) as the motor is tied to the shield with a strong adhesive.

Best regards, Fleet

[Soeren]

Hi,

When I adjust the pothead, the gears stop but after a while, they start to spin again. This happens again and again. Am I doing something wrong?

What's the input to the servo under that condition?
Does the pot perhaps move with motor vibrations (a bit of paint can lock the shaft in position).
How long exactly is a "while"?

[Fleet]

Hello Soeren,

1: The mcu is running the hold_servo.hex file from this tutorial, which should make the servo hold at 0 degrees.
2: The motor didn't run initially when I had adjusted the pothead and glued it, so there were no vibrations from the motor but it stood on the table beside my pc, I don't know how "sensitive" it is. Suddenly it just started running fairly slowly.
3: A while = 7-10 mins.

Btw can you recommend a good way of cutting the top of the pothead of without damaging the "edge" (where a gear will rest when reassembled) of the potentiometer? I tried with a Stanley knife and a "skævbider" (danish word).

/Fleet

[Soeren]

Hi,

Not sure what (or who ) you call the (a) pothead, but save your side cutters for their intended purpose (unless you've got a large Bacho, that can chew a potmeter shell without even noticing) - If you have a Dremel or similar, a small cutting disk goes through just about anything (within reason).

However, I'd think you could manage to keep the pot and get the free ends of two joined resistors soldered to the outer terminals of the pot and just cut the middle wire and attach it to the common node of the resistors (use tape to avoid them shorting to the pot body). The resistors doesn't have to be 2.2k, it's just a value mitigated from the one person that first used it because he had a 4k7 pot to replace and didn't care to consult (or understand) the data sheet. Any two equal resistors in the range 1k to 10k would do fine and if it weren't for possible motor noise, you could easily go higher - the important thing is that they're the same value.
If you try that, you'll be able to tell if it was indeed the pot that did it, but did you consider temperature variations (eg. draught) causing drift?

If possible, measure the signal line with an oscilloscope, to see if the 1.5ms pulse changes width when the rotation start.

[Fleet]

I can't pull the pcb out of the case, so I don't have access to the wires unless I break the case. I don't have an oscilloscope (yet).

/Fleet

[Fleet]

This is a really noobish question: how does the value of the resistors used affect the servo and would these resistors produce good results?

[Soeren]

Hi,

This is a really noobish question: how does the value of the resistors used affect the servo

They just create a voltage divider with half the supply voltage out on the common node, this is how the pot will work when the wiper is in the center position.
It's a feedback to the control circuit, which is comparing the commanded value to the feedback from the pot and when in the middle, it equals 1.5ms pulses in.

There's a timing gap called dead band in the control electronics, to avoid the motor constantly hunting and if you initially set the pot close to one edge of the dead band value, a little temperature change will eventually make it drift.
You could try finding the two exact positions where it starts to move and place it right in the middle of them, but the two resistors will be easier.

and would these resistors produce good results?

If it is the center point adjustment that gives you a hard time they will.
The important part is not their exact value, but that their values are as close as possible - you may even wanna measure some and get two that's equal, but it shouldn't matter with 1% tolerances.
Why buy from Rapid (paying shipping as well) when you can get those in your nearest shop?

[Fleet]

Hello,

They just create a voltage divider with half the supply voltage out on the common node, this is how the pot will work when the wiper is in the center position.
It's a feedback to the control circuit, which is comparing the commanded value to the feedback from the pot and when in the middle, it equals 1.5ms pulses in.

There's a timing gap called dead band in the control electronics, to avoid the motor constantly hunting and if you initially set the pot close to one edge of the dead band value, a little temperature change will eventually make it drift.
You could try finding the two exact positions where it starts to move and place it right in the middle of them, but the two resistors will be easier.

Thank you.

Why buy from Rapid (paying shipping as well) when you can get those in your nearest shop?

Had to place an order there anyway so bought them as well.

/Fleet

[Fleet]

I have now replaced the pot with two resistors (1.6K resistors) , but when I run the hold_servo.hex program the servo just runs fast in one direction. Why would it do that?

/Fleet

[knossos]

If the two resistors actual resistance aren't close enough, it will affect your center point.  A software fix would be to adjust your timing until the servo comes to a halt.  You could also check the resistors with a multimeter and replace the resistors with two that are more closely matched.  Another option would be a resistor/trim pot combo so you could adjust the trim pot to get the correct center point.

[Soeren]

Hi,

Another possibility is if the zero position is sloppy in the servo hardware (but I'd think they use a controller, so shouldn't be that likely). Not all servos are created equal and while some (a few) keep to the 1ms to 2ms standard, a lot doesn't - I wouldn't be too shocked if the 1.5ms center position was off a bit, especially in cheap servos.

[Crunchy Theory]

I had the same problem that you described with the servo moving again after a little while. It turned out to be a simple matter of vibrations or the slightest movement when I was applying the super glue on the trimmed pot that would move it off-center. Unfortunately, I wouldn't discover that it had moved until after I applied the glue, so I ended up going through two servos (ruined) before finding the proper technique.

[Fleet]

I replaced the resistors with two 2.2K R resistors with equal resistance (measured beforehand), but when I run the hold_servo.hex supplied by admin, it just spins in one direction (fast). Isn't the modified servo supposed to stop when I run hold_servo.hex or should I adjust something?

Picture of servo:


/Fleet

[Soeren]

I replaced the resistors with two 2.2K R resistors with equal resistance (measured beforehand), but when I run the hold_servo.hex supplied by admin, it just spins in one direction (fast). Isn't the modified servo supposed to stop when I run hold_servo.hex or should I adjust something?

Picture of servo:


/Fleet

Please read the answers you get: http://www.societyofrobots.com/robotforum/index.php?topic=12554.msg94915#msg94915