Thread: Servo jitter

Tweet

Greetings,

Past winter I built a new control cabinet so I can safely hide all the control electronics I have for my haunt. Now I'm not the most efficient builder. Neither am I very electronically savvy but I've got stuff working in the past.

The outside:



The inside:




Servo hook up to Cat 5/6 cable:




I've only just started work on my first animatronic. So I haven't been able to test it all before. Problem is, I'm experiencing heavy servo jitter when I activate two or more servo's simultaneously. At first I thought it was the ground. As I'm running the servo's directly of a PSU the pololu boards I had didn't have a common ground. When I hooked that up the problem persisted however. I can control the servo's but they keep moving on their own. My guitar player ever breaks his elbow because the arm turns the other way.







The servo signal is being sent over 7 meter of cat 5 cable directly to the servo. The power is sent by a 7 meter cable to a box in which I divide it over the different servo's. I can control servo's but they're seem to be experiencing noise.

Any idea's how to fix this, or what the cause might be? The cable could be to long. Yet I can control the servo's. The cable itself is shielded cat 5. It might be the connectors between servo's. I've used different ways to connect cables, screw terminals, soldered and cable lugs (Connectors).

I've tried to attach most cable's out of the way of the signal cables to reduce noise and contact between them. (lest they work as an antenna)

I'm planning to pick up some ferrite rings tomorrow to try and dampen any interference. Any other idea's? Could it be the wiring from the pololu controllers to the cat 5 jack?

In my past experience with servos it came down to a power issue. I, like you, had the power in a remote location and the servos were going crazy. Start with the power supply. Make sure it is regulated for smooth power and can provide sufficient power for all the servos and controllers. Make sure that all circuits involved with the servos have a common ground. In my situation I ended up putting a large value (1000uf) capacitor across the power leads as close to the servos as possible. and the problem went away.

Something like this:
http://www.graupner-ifs-system.de/fi...B6r/7082.1.jpg

Good Luck!

I seem to recall some expert on this board explaining that when providing power to servo via an outside source you must still provide a connected ground to the the board. Something to so with the signal is still part of the circuit and it must have a stable ground to run against. On my 3 axis I ran power via the cat 5. But my unit is well balanced and the servo's draw low current. I did see drop off from my led's for the eyes. But I also got lucky, as just after Halloween I burned two traces on my interface board. I wish I could a better explanation, but I too believe this is power related or perhaps a better term would be unstable signal related. Try connecting the ground of the board to the ground of your power source. Without the positive there is no way to hurt the board as far as I can tell. I know a member from our group that is pretty sharp on this stuff, I will ask him.

Thanks for your replies so far. Some good suggestions.

With regard to the grounding. I have grounded my pololu controller boards to the power source. I've connected a wire from the minus on my pololu controllers to the minus on my power source.

I'm not sure if it's regulated although I think it is. I'm using this one: SP-240-5 Its capable of providing 5 volt 45 ampere. Enough current

Its got this nob I can twist to regulate between 4 and 6 volts. So far I've turned it all the way to 6 volt (I think) to cope with the voltage drop. I'm running 7 meter of cable so I'm sure I've got some voltage drop.

Upke,

Can you provide a diagram of the servo part of your system? Are you running the power for the servos in the same cable as the control pulse? 7m is a long way, and shielding only helps protect you from *outside* interference. If the interference is from the power wires going to the servos (because of electrically noisy motors), the noise can still couple to the signal wire from within the cable.

In order to verify whether you have a noise problem or not, try this:

1. Get the system to jitter
2. Remove one servo lead from your controller
3. Re-connect a servo directly (no long cable) at this port on the controller

If you still have problem with the locally-connected servo, then you probably didn't have a noise problem. If the servo works fine now, you probably did have a (cable-related) noise problem.

Spinman's right. The servos and the controller have to be operating on a common ground.

- Hook

First a power supply that offers 45 amps ? No doubt that is enough juice. If you can control the output voltage than I would bet it is regulated. But you could be seeing voltage drop. Servo's need voltage to maintain logic. I run my 3 axis with a 6 cell nicad pack. That is 7.2 volts. So the suggestions I would think is use short cables to confirm possible voltage drop and by all means route the signal cables away from the power. You could use a small voltage regulator near the prop and bring in a bit higher voltage to that point. But TBQH my cat 5 cable to my 3 axis is 20 feet long. I had no jitter and I was running 6 servos with just one cable. Again I am not putting heavy load on the servo's. And I did have to hardwire past my + and - traces in the RJ 45 connector on both ends.

I've just tried it with a capacitor in the circuit. Got it connected between + and -. It blew up after a couple of seconds.... Might have connected it wrong... >.< (it was a 10v 1000uF capacitor btw, should be enough)


Edit: Hooked up a new one. This time I did it right Didn't help though...still servo jitter...

I'll try hooking up a servo directly to the maestro controller.

Edit 2: Right I redid all the wiring, still jitter. I hooked one servo directly (the one that jitters the most of all) directly to the board. The jitter was gone.... This means its not a power issue right?

Help

I recommend checking the voltage as close to the servo as possible. I would also measure resistance across the ground wire. I also recommend using stranded CAT 5/6. The stranded wire will lay flat and it's possible that laying flat and being stranded, it is less likely to emit or pick up stray signals.

Thanks for all the replies. I've checked all connections again, even within my control center.

My last hope was shortening the cat 5 cable. So I cut off about 5 meters worth of cable. No jitters anymore. So the power cable is still 7 meters, the cat 5 cable is now only 2. No problems any more. I'm now going to see how long I can make that cable before the servo's begin to jitter.

All of my cat5 wires and there are 11 of them are over 5 meters and some are 15 meters.