What's special about Pixhawks PWM outputs ? (arduplane, shaky servos)

[André Kjellstrup]

I upgraded a APM plane to Pixhawk.
All servos (Hitec) are shaking a little on certain small, random input.   Diffrent kinds, and different batches (dual servos for redundancy)
- and no- it's not a power problem.
Whan can be done about it ?

[André Kjellstrup]

Found it to be two small servos used on rudder that were unhappy with 3.3v PWM.

[Lorenz Meier]

Andre,

Are you sure about that conclusion? Because if it were true, those servos wouldn’t work on Spektrum, Futaba and FrSky receivers *and* would be incompatible to TTL voltage levels.

In case you’re interested, here is an in-depth coverage of what PWM interfaces are used in the RC industry:

http://pixhawk.org/users/actuators/pwm_escs_and_servos

I would be very interested to hear about the true root cause once you’ve found it.

-Lorenz

Am 24.05.2014 um 19:30 schrieb André Kjellstrup <andre.kj...@gmail.com>:

Found it to be two small servos used on rudder that were unhappy with 3.3v PWM.

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Randy Mackay]

     Yes, just carrying on with what Lorenz is saying, checking the board and servo voltages carefully might be good.  It may also depend upon how you're powering the Pixhawk.  For example if you use only a USB cable the voltage might be lower than if you power it using a battery+powermodule.  Also if those servos are being powered at 5.5V (or higher) it might make them unable to see the 3.3V (or less) coming out of the Pixhawk.

-Randy

[Lorenz Meier]

Randy,

With the same argument, it would mean that Hitec servos powered off 5.5V wouldn’t work with Futaba, FrSky and Spektrum receivers - although I have no proof to rule this out, I find it rather unlikely.

And the output voltage of Pixhawk is completely independent of the input power until you hit ~3.8V of your input supply. So neither the experimental data I gathered nor the schematic support your hypotheses.

@Andre: Are you using the power module? If not, you definitely should. The servo rail is often not stable enough alone to reliably power an autopilot.

-Lorenz

[Stefan Gofferje]

@Andre:
To my experience, many servos are quite unhappy when a 433MHz telemetry
radio is too close to them. I'm running OpenLRSNG on one of copters and
I had massive jitter in my camera tilt servo whenever I switched the
telemetry on. Until I moved the RX further away, to the back of the
copter. OpenLRS telemetry outputs 100mW, so if you are using 433MHz
telemetry on max. power, that could also be the reason.

@Lorenz:
I know at least one device which for sure does not like 3.3V PWM: The
Boscam HD-19 FPV camera. It has RC trigger functionality and it works
only with one of my 3 FrSky Receivers and with that also not reliably. I
measured all 3 and the one it works with outputs 3.42V peak while the
other 2 output around 3.25 - 3.35V.
With my OpenLRS RX (5V) it always works perfectly fine.

-S

--
(o_ Stefan Gofferje | SCLT, MCP, CCSA
//\ Reg'd Linux User #247167 | VCP #2263
V_/_ Heckler & Koch - the original point and click interface

On 05/25/2014 08:53 AM, Lorenz Meier wrote:
> Randy,
>
> With the same argument, it would mean that Hitec servos powered off 5.5V
> wouldn’t work with Futaba, FrSky and Spektrum receivers - although I
> have no proof to rule this out, I find it rather unlikely.
> And the output voltage of Pixhawk is completely independent of the input
> power until you hit ~3.8V of your input supply. So neither the
> experimental data I gathered nor the schematic support your hypotheses.
>
> @Andre: Are you using the power module? If not, you definitely should.
> The servo rail is often not stable enough alone to reliably power an
> autopilot.
>
> -Lorenz
>
>
> Am 25.05.2014 um 01:44 schrieb 'Randy Mackay' via drones-discuss

> <drones-...@googlegroups.com <mailto:drones-...@googlegroups.com>>:

>
>>
>> Yes, just carrying on with what Lorenz is saying, checking the
>> board and servo voltages carefully might be good. It may also depend
>> upon how you're powering the Pixhawk. For example if you use only a
>> USB cable the voltage might be lower than if you power it using a
>> battery+powermodule. Also if those servos are being powered at 5.5V
>> (or higher) it might make them unable to see the 3.3V (or less) coming
>> out of the Pixhawk.
>>
>> -Randy
>>
>>
>> On Sunday, May 25, 2014 2:48 AM, Lorenz Meier <l...@inf.ethz.ch

>> <mailto:l...@inf.ethz.ch>> wrote:
>>
>>
>> Andre,
>>
>> Are you sure about that conclusion? Because if it were true, those
>> servos wouldn’t work on Spektrum, Futaba and FrSky receivers *and*
>> would be incompatible to TTL voltage levels.
>>
>> In case you’re interested, here is an in-depth coverage of what PWM
>> interfaces are used in the RC industry:
>> http://pixhawk.org/users/actuators/pwm_escs_and_servos
>>
>> I would be very interested to hear about the true root cause once
>> you’ve found it.
>>
>> -Lorenz
>>
>>
>> Am 24.05.2014 um 19:30 schrieb André Kjellstrup

>> <andre.kj...@gmail.com <mailto:andre.kj...@gmail.com>>:

>>
>> Found it to be two small servos used on rudder that were unhappy with
>> 3.3v PWM.
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "drones-discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send
>> an email to drones-discus...@googlegroups.com

>> <mailto:drones-discus...@googlegroups.com>.

>> For more options, visit https://groups.google.com/d/optout.
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "drones-discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send
>> an email to drones-discus...@googlegroups.com

>> <mailto:drones-discus...@googlegroups.com>.

>> For more options, visit https://groups.google.com/d/optout.
>>
>>
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "drones-discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send
>> an email to drones-discus...@googlegroups.com

>> <mailto:drones-discus...@googlegroups.com>.

>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google
> Groups "drones-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to drones-discus...@googlegroups.com

> <mailto:drones-discus...@googlegroups.com>.

[Robert Lefebvre]

FWIW, most of my helis use HV servos, running on 8.4V, and have no problem with the signal coming out of the PX4.

I've also had problems with telemetry causing problems, but this was with an Xbee, never a problem with 3DRadios.

[Stefan Gofferje]

On 05/25/2014 03:48 PM, Robert Lefebvre wrote:
> FWIW, most of my helis use HV servos, running on 8.4V, and have no
> problem with the signal coming out of the PX4.
>
> I've also had problems with telemetry causing problems, but this was
> with an Xbee, never a problem with 3DRadios.

You use 900MHz, right? I found 433MHz more disturbing than 900 MHz
(which is GSM in Europe) in many cases. I can hold a mobile phone (which
may blow out up to 2W in Europe) right beside my copter with no problems
but the 100mW 433MHz really messes with my servos.

[Robert Lefebvre]

On the Xbee, I think it was that it did not have a ground plane for the antenna, and used the entire ground system as the ground plane.  Something like that.  I'm not an expert on that sort of thing, luckily I didn't need to become one as the 3DRadios just work. ;)

[Lorenz Meier]

Stefan,

What you’re saying is that the Boscam is not compliant to how things are done in the industry for 30 years? Even with the maximum margin factored in anything above 2.4V is considered a high level. Are you confident its not something else? Did you connect signal ground?
Because that camera is almost certainly a 3.3V device.

I would be careful attributing these things to voltage levels - it seems so logical and an easy explanation, but industry standards and practices and the schematics of the devices involved actually in no case I looked at closer supported that theory.

Whenever you feel its the signal voltage I can only recommend looking closer at the setup, because something must be electrically messed up terribly if TTL levels are not working as they should. And this is why 3.3V signal levels make sense for FrSky, Futaba and Spektrum. they have enough margin so that any *properly wired* equipment will work great.

-Lorenz

> To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Stefan Gofferje]

Lorenz,

On 05/26/2014 08:28 AM, Lorenz Meier wrote:
> What you're saying is that the Boscam is not compliant to how things are done in the industry for 30 years? Even with the maximum margin factored in anything above 2.4V is considered a high level. Are you confident its not something else? Did you connect signal ground?
> Because that camera is almost certainly a 3.3V device.
>
> I would be careful attributing these things to voltage levels - it seems so logical and an easy explanation, but industry standards and practices and the schematics of the devices involved actually in no case I looked at closer supported that theory.
>
> Whenever you feel its the signal voltage I can only recommend looking closer at the setup, because something must be electrically messed up terribly if TTL levels are not working as they should. And this is why 3.3V signal levels make sense for FrSky, Futaba and Spektrum. they have enough margin so that any *properly wired* equipment will work great.

Of course I connected the ground :). I'm no RC person - I'm a
computer-guy :).
When I first encountered the problem, I even did a dedicated test setup
consisting of only a 12V PSU, a 5V BEC, the camera the RX and a
mini-breadboard to make the measuring easier. I did lots of measuring
and voltage was the only point that made sense. Especially, considering
that I tried with 2 RX of the same type (FrSky D8RII) and one worked and
the other didn't.

Could that the camera is a 3.3V device (although it doesn't work below
5V Vcc)... Maybe Boscam misdesigned an input resistor or something like
that. The camera had/has a few other bugs if your read throught the
corresponding RCG thread :).