SD21 Servo Jitter

[Yessir]

I am using a Picaxe 18X mounted on an SD21 Servo Control board to read signals from a Futaba radio receiver using Pulsin, process the values and send the outputs to the connected servos.

Everything apears to be working OK except that the SD21 board seems to be sending interference to the servos which are still connected directly to the receiver, making them jitter. The problem only starts to happen when I have some servos connected directly to the receiver and some servos via the SD21. As soon as I remove the servos which are connected to the SD21 (but keep the SD21 connected to the receiver so that it still processes the inputs) the jitter on the servos which are connected directly to the receiver disappears.

I have checked cable connections, I am using one battery to power the receiver and one battery to power the SD21 with both negative leads connected together.

Does anybody have any suggestions?

 

[Andrew Cowan]

It sounds like the SD21 is sending servo interference back down the input lines. Have you tried putting a 100uF capacitor across the battery input of the SD21?

A

 

[Marcwolf]

Hi
I believe the SD21 has the capability of being powered by 2 supplies. One for the control logic, and the other for the servo's.

Have you tries powereing the reciever and the SD21 logic from one battery, and the other battery used for the servo's only.

Servos are inherantly noisy, and its advisable to use a lot of filter capacitors.

I use the SD21 and 18X in an animatronic controller.

Take Care
Dave

 

[Andrew Cowan]

I am using one battery to power the receiver and one battery to power the SD21 with both negative leads connected together.

Marcwolf - he says he is using two supplies, which is why this interference is so surprising.

I can only imagine that the problem is due to a lack of capacitors. More capacitors (both high and low values), equals a smoother supply.

A

 

[Marcwolf]

Andrew - In his original post he states

"I am using one battery to power the receiver and one battery to power the SD21 with both negative leads connected together."

Now the SD21 has a jumper on it that allows you to use one or two supplies on it.

One supply is used ONLY for the servo's, and the other for the logic side. However if all you have is the servo power then you can use the jumper to run the servo power through a on-board 7805 and power the 18X and controller PIC via that.

However - it does mean that any noise the servo's are generating can cause issues with the onboard 5v line. Hence my question re running the board controller logic from the reciever supply.

Hope this helps

Dave

 

[Wrenow]

Although power supply glitches can be common (and aoften cured wqith proper grounding and or some caps as indicated), do not forget the possibility pf RF interference from the SD21 controlled servos to the RX.

How is your antenna oriented? Well away from any servo cables and the SD21, I would hope? Are your power supply wires well separated from the servo leads? Have you tried running the servo leads through a ferrite core to dampen their RFI?

You can also get the jitters if the TX is a bit too close to the RX, creating a swamping issue = perhaps exacerbated by the rest of the system acting as an antenna.


Just a couple of thoughts.

Cheers,

Wreno

While you have your servos out, the latest issue of Servo Magazine has a CNC easter egg decorator based on 2 servos (personally I would have used 3 to allow pen lifting). Looks terribly Picaxable. In fact, with the AXE 024 or the SD21, you could make it sellf-cointained pretty readily for a demo/production unit.

 

[Yessir]

Thanks for all the replies.

Although the SD21 has the facility for 2 x power suppies, the problem is actually with servos plugged directly into the original receiver and not the SD 21 board, so I am not sure if this would make any difference.

I have tried some further tests to determine the source. With all inputs between the receiver and SD21 disconnected the jitter stops which rules out any Rf interference. I am actually reading inputs from 3 receiver channels and as I connect each input lead the jitter gets progressively worse (1 x lead connected = minor jitter, 2 x leads connected = more jitter and 3 x leads connected = significant jitter). Because I am using a Futaba 3 pin plug & lead for the receiver connections I effectively have 3 x negative and 3 x positive connections in parallel with each other. Would it help to disconnect some of these pins so that I just have 1 x negative and 3 x signal wires between the receiver & SD21?

I am not sure how much filtering is done by the SD21 board. Does anybody have a circuit diagram of the board ? - some of the components on the board are too small for me to identify!

Thanks.

 

[Yessir]

I have done some further investigation by writing a seperate test program and found that the servo jitter only happens when the picaxe gets to the PULSIN command. The test program starts with a PULSIN followed by a 5 second pause before looping back to the start again. The jitter happens at 5 second intervals. During the 5 second pause ther is no jitter.

Does anybody know if I can add some sort of filter (capacitor?) to the signal cable between the receiver and picaxe input to stop this problem?

 

[Wrenow]

Have you tried tying each of the inputs on the Picaxe low with, say. a 10K resistor to keep the input pins from floating? Not sure if that would help the jitters, but something nagging in the hidbrain says it might.

Cheers,

Wreno

 

[Wrenow]

I have tried some further tests to determine the source. With all inputs between the receiver and SD21 disconnected the jitter stops which rules out any Rf interference.

Doesn't necessarily rule out RFI interference, unless the SD21 is still hooked to servos and cycling them on its own while not being hooked to the RX.

Cheers,

Wreno

 

[TMeyer]

Of Note

You are clear in that you are using the PULSIN command.

The below *may* be applicable as a concept in troubleshooting.

I noticed when attaching servos and controlling them with the SERVOPOS command, the defaults in the manual are approximations and vary with each individual servo.

For instance- Page 162 of Man2:

init: servo 4,75 ‘ initialise servo
main: servopos 4,75 ‘ move servo to one end
pause 2000 ‘ wait 2 seconds
servopos 4,225 ‘ move servo to other end
pause 2000 ‘ wait 2 seconds
goto main ‘ loop back to start

Would cause jitter at the far CW stop, but not the full CCW stop.

Using trial and error, these observed limits were found:
init: servo 4,41 ‘ initialise servo
main: servopos 4,41 ‘ move servo to one end
pause 2000 ‘ wait 2 seconds
servopos 4,215 ‘ move servo to other end
pause 2000 ‘ wait 2 seconds
goto main ‘ loop back to start

41 provided an addition 20 degree swing from 75, but more importantly, the published default of 225 would jitter as it tried to push past it's internal "stop". This using an 18X on a AXE091 board.

The second servo I tested also had differing "stop" values.

Hope this helps.

 

[BeanieBots]

Does anybody know if I can add some sort of filter (capacitor?) to the signal cable between the receiver and picaxe input to stop this problem?

It's a good idea to put just a simple resistor between Rx and PICAXE input especially if your Rx has a different power supply to the PICAXE.
This is not for filtering but to protect the PICAXE from excess current which can occur if:-
a) The Rx is powered and the PICAXE is not.
b) The Rx supply voltage is higher than the PICAXE supply voltage
c) The PICAXE input is accidentally turned into an output.

1k would be a reasonable value to try to start with.
If it helps your problem (which I doubt), then consider going to a higher value.

 

[Yessir]

I have eventually solved the problem , although I am not quite sure why it was happening in the first place other than some sort of circulating current?

Essentially my original set up used standard Futaba servo leads to connect 3 x receiver channels to a small project board before connecting to the SD21 header. Each of the 3 x positive leads were soldered to a common rail, each of the 3 x negative leads were soldered to a common rail and each of the 3 x signal leads were soldered to an individual rail. The positive leads were just terminated on the project board and didn't do anything, the negative leads were connected to the SD21 negative and each of the signal leads were connected to the PICAXE input pins via the SD21 header.

To solve the problem I have disconnected 2 of the negative leads from the project board so that there is only one remaining negative lead between the receiver, project board and SD21 negative. This seems to have stopped the servo jitter that I was experiencing.

 

[BeanieBots]

Glad the problem has gone but I can't see how multiple 0v would have caused the issue. Admittedly, multiple grounds can cause problems in many other circumstances but I don't see it in this one. I'd guess that the simple disturbance of re-doing the connections put right whatever was really wrong.

The true test would be if the problem re-appears if you put the extra ground wires back. I'd be grateful if you could try it and report back.