Question regarding 76Hz RC PWM frequency

[jinx]

Hi ,

We were trying to control an RC motor controller with the ServoPod-USB. (SS4300 motor system from Novak -www.teamnovak.com)

The controller expects 50Hz PWM pulses from 1.1-2.1ms width. Basically , your standard RC servo pulses. Since the Servopod-USB cannot go below 76Hz on FULLSPEEDCPU (and we needed the extra speed for other things), we decided to give the 76Hz pulse and observe the effects.

Everything worked well the first time, where we could calibrate the controller and control the motor speed thru the pod. We tested this for around 15 minutes without incident.
But when we connected this the next morning, the controller basically shut down after 1 minute of use, even though we gave it no PMW (it was shut OFF). Now we cannot even get it to switch on again. It's gone, though it wasn't burnt nor damaged in any observable way (not warm, no burning smell, no plastic discolouration).

The two questions we have are:

-Has anyone on this forum tried giving 76Hz pulses to servos and RC controllers like this, without any incident? We have been using the 76Hz train to control normal Hitech ultra-torque servos and the Novak Super-Rooster without any problems. But after this (and because we may be shifting to digital servos soon), we want to make absolutely sure that going from 50-76Hz does not kill servos in the long run.

-Apart from going to HALFSPEEDCPU, is there any other way of giving a 50Hz pulse train to servos from the Pod? We need the speed and it seems unfortunate that the Servopod does not have a facility to have a regular RC PWM out at normal operating clock speeds.

thanks for any inputs on this,

regards,
Jinx.

[RMDumse]

Originally posted by jinx
-Has anyone on this forum tried giving 76Hz pulses to servos and RC controllers like this, without any incident?

If your failure was due to the 76Hz, it would be the very first we've heard of.

We have found some ESC's that were marginal, but at 50Hz! We were on the ragged edge of too slow for them. This is why we began experimenting with the higher 76Hz train and have never found an instance, or heard a report of it not working as well, or better, than 50Hz.

we want to make absolutely sure that going from 50-76Hz does not kill servos in the long run.

I would think getting Team Novak to analyze the failure would be a first priority then. Let's be sure it wasn't something like early failure due to lack of burn in or such.

-Apart from going to HALFSPEEDCPU, is there any other way of giving a 50Hz pulse train to servos from the Pod? We need the speed and it seems unfortunate that the Servopod does not have a facility to have a regular RC PWM out at normal operating clock speeds.

No problem. Use a timer instead of the PWM. Timers can generate PWM, and they have 16-bit periods, where the PWM has only 15-bit periods. So use a timer and get the same lower speed rate as would result from using PWM at halfspeed.

[blueeyedpop]

I have used The PWM at 76 hz on speed controls from a variety of manufacturers including a Novak Spy, with no issues.

How were you handling the output from the B.E.C. ? You should not tie the 5V out of the B.E.C. to the IsoPod in any way, though you should have a common ground.

Mike

[jinx]

Well, the way we were controlling it was that we were plugging the pwm input of the ESC to the pwm output of the Servopod. So in effect, we were tying the +5V of the pod and ESC together?

Sorry for the naive question, but why is that a no-no? The ESC was functioning well for over 20 minutes with this setup. The time it stopped working was when we connected the +5v and Gnd to the input of an oscilloscope to confirm the output reference voltages (we used an SS4300 system from Novak).

thanks in advance for your help on this.

-jinx.

[RMDumse]

Originally posted by jinx
The time it stopped working was when we connected the +5v and Gnd to the input of an oscilloscope to confirm the output reference voltages

Ut oh. There's your clue. The Oscilloscope has a built in Earth ground. Maybe your power supply does too. The result would be when the ESC put power on one side of the output, the scope made it a direct short to ground.

I blew a section of trace right off a PCB 20 years ago learning that one the hard way. Think that is likely what happened to your ESC.

Hooking the +5V's together isn't terrible, but it isn't great either. Which ever supply is slightly higher will wind up powering everything 5V on both sides of the connection. The higher supply might not be able to carry the increased current. Better to keep them separate. Depends if your ESC has an intermal 5V supply, or needs one supplied, as to whether you need to connect them or not.