Frsky problem.

[Whiteprop]

I got so curious that i had to test in power reduction mode right away.
No real difference that i could spot.

As others have said though, if i add weight to the servo, it moves a little smoother, but i can still feel the individual steps though.
Also the servo's are incredibly noisy in "neutral". I guess the Rx is switching back and forth between two "steps". Very seldom do i manage to put any of the sticks so that the servo does not make any noise.

I added my DIY module to an old radio. The servos jittered at neutral and the motion was choppy. I scoped the +9.6vdc supply and the PPM signal line. Both had approx. 300mv of noise riding on supply and signal. Going to power reduction mode lowered the noise to 100mv and the system seemed to work better. The problem was a defective 9.6vdc nicad pack. Wired in a new pack...problem went away and servos are quiet and smooth. Scoped the same lines and they are completely clean and flat. Apparently my problem was loading on a bad power supply. Going to reduced power obviously lowered the load and the system appeared to work better. Scope your power supply to see if there is any noise at all on it.

[takilara]

Hmm i'll take a look. The battery in the radio is fairly old (approx 4 years old) so that could be the reason.. (never had issues when running FM though), but i see the voltage meter in the display of the radio "jittering" more with the FrSky unit than with the FM unit.

[Whiteprop]

Hmm i'll take a look. The battery in the radio is fairly old (approx 4 years old) so that could be the reason.. (never had issues when running FM though), but i see the voltage meter in the display of the radio "jittering" more with the FrSky unit than with the FM unit.

Same with me. The old AM 72mhz. system worked fine. Seems like the FrSky 2.4 system may be less tolerent to power supply issues. You say your front panel volt meter is jittering also?....not a good indication.

[RCModelReviews]

It's hard to tell (without testing) whether the problem is in the DIY module or the transmitter encoder.

Unlike the old FM/PCM systems, the 2.4GHz RF section in our radios only transmits for short periods of time, then they turn off until they're ready to send more data. This means that the current they draw goes up and down quite markedly -- going from just a few mA to over 100mA many times in a second -- whereas the old FM/PCM drew a constant, steady current all the time because it was always transmitting.

The effect of this on some radios is to introduce "noise" on the main supply voltage and that noise can affect the encoder so as to introduce some irregular timing into the pulses that are generated.

In theory, the regulators used in the transmitter to drop its 9.6V battery down to 5V should remove the noise but in reality it doesn't always do so adequately -- especially if the radio is getting old (which could mean a weak battery and dry capacitors).

I'll be looking at ways of reducing this noise and they'll probably involve the use of an extra capacitor and a diode to further isolate the RF part from the logic part.

I'm also working on a test-rig that will accurately quantify the resolution and jitter associated with various module/receiver combinations.

[takilara]

I scoped the battery while using the FrSky module. I saw no noise at all. I didn't get around to scoping any of the pins to the module.
I also tested an analog servo, and that was less "choppy". Other things that might have an effect, i run 6V to the Rx.

I scoped the servo pulses, and they looked quite good, a little bit jitter, but not much, and the pulse seemed to increase very nicely.
However, i was puzzled that the amplitude of the pulse is only 3.3V (this might be correct though)

The jitter i see on the display on the Tx, is only visible with the FrSky module, probably due to the cyclic load generated by 2.4GHz system as Bruce explains. I was unable to see this jitter when scoping on the battery directly.

[Whiteprop]

I scoped the battery while using the FrSky module. I saw no noise at all. I didn't get around to scoping any of the pins to the module.
I also tested an analog servo, and that was less "choppy". Other things that might have an effect, i run 6V to the Rx.

I scoped the servo pulses, and they looked quite good, a little bit jitter, but not much, and the pulse seemed to increase very nicely.
However, i was puzzled that the amplitude of the pulse is only 3.3V (this might be correct though)

The jitter i see on the display on the Tx, is only visible with the FrSky module, probably due to the cyclic load generated by 2.4GHz system as Bruce explains. I was unable to see this jitter when scoping on the battery directly.

Get on your *9.6 vdc line...select AC coupling on scope....50mv/div...no noise?

[RCModelReviews]

A little update on the servo jitter/resolution issue.

I didn't have much time today but I connected a low-cost Savox digital servo and did notice what appeared to be a drop in resolution. (Note: I was using a FlySky 9X radio which has somewhat less than 1024-step resolution and, when high-amounts of expo are used, definitely shows lower resolution around center)

I also tried an XCore coreless digital servo (tighter deadband) and the resolution definitely improved.

The Savox required 8 clicks of trim per step of the servo arm, the XCore required just 5.

Surprisingly, a standard JR non-digital servo responded at average once every 1.5 clicks.

Very interesting!

How can a standard servo seem to have more resolution than fancy digitals?

I suspect it's down to the fact that the analog doesn't position itself as precisely as the digitals. ie: the digital won't move until the change of pulsewidth exceeds its programmed deadband level but the analog will follow a much smaller change in pulsewidth -- even though it is supposed to have a wider deadband (especially if lightly loaded) because it doesn't have a hard cutoff for that deadband.

As a matter of comparison, I notice this same effect when I use Hitec's QPCM and JR's PCM systems on 72MHz with analog and digital servos. Using a JR9XII (9303) with JR PCM, I get movement of the standard servo with each trim-click. The digitals still take several clicks of trim before they take a step.

Right now I'm programming up my test rig to do some quantitative and definitive tests to see just what the true resolution of the transmitter, the RF link and the servos are so we can better understand what's going on here.

[takilara]

Get on your *9.6 vdc line...select AC coupling on scope....50mv/div...no noise?

Is it ok to connect in parallel in the battery plug?

I used DC mode, but i don't think i can use 50mv/div (at least not for DC mode). My scope wont let me set GND lvl outside the display, i'll do some more tests though. If I get more data from AC mode..

(i have a cheap USB scope from USB Instruments http://www.usb-instruments.com/oscillo_stingray.html)

[mr.sneezy]

Right now I'm programming up my test rig to do some quantitative and definitive tests to see just what the true resolution of the transmitter, the RF link and the servos are so we can better understand what's going on here.

Bruce, if it's PIC based test rig I'd love to build one as well...
Martin

[RCModelReviews]

No problems, as with all the stuff I build, I'm happy to share with anyone who might find it interesting.

Just wish I had more time to dedicate to details like documenting and writing project pages.