Official spec for various 2.4GHz servo frames

[pldaniels]

Is there an official spec source from the various mfg's of 2.4GHz gear for their servo pulse trains / frame rates?

While I'm able to test what I have at hand here with my scopes it's always preferrential to get specs from the "proper" sources.

Specifically seeking;

* Frame rate
* Pulse width range
* Pulse voltage

Would appreciate if anyone can help.

[pldaniels]

Forgot to mention, would be nice if there was an output impedence specification too.

While I realise that most are very similar to the existing PPM/PCM systems there does seem to be enough variances to cause some troubles now and then (eg, like having to move to a comparator input for pulse detection for pulses below 3V on a 5V uC).

Paul.

[AndyKunz]

With analog it was easy.

Pulse nominal 1.0 to 2.0ms. Min off time 6ms. Pulse voltage .75Vcc. The servos actually respond to pulses down to .25ms or lower, or over 3ms, but the mechanicals have a hard time (outside the physical travel range).

Digital servos and most ESCs accept faster pulses but the duty is still the same 1.0-2.0ms.

Real old radios had variable frame rates that could be anything from about 15ms on up. I had several old (but not ancient) Futaba's that were 16ms frame rate, fixed. Many modern computer radios use 22ms as standard (same as PCM), but high-speed ones like the DX7SE cut that in half (which puts them a little too fast for most analog servos).

If we knew what you were trying to accomplish, we might be able to help you more.

Andy

[pldaniels]

Andy,

This is an information gathering quest, not a specific application as such. Just been noticing across the selection of radios there's enough extreme cases (eg, the DX7SE) to cause grief if not documented and handled.

Regards,
Paul.

[JimDrew]

XPS -

Frame rate: variable as every transmitter is different
Pulse width range: 750us to 2250us (per PPM specification)
Puse voltage: 3.3v


Every transmitter uses a different PPM pulse stream. Graupner/JR transmitters use compression for PPM data when there are 9 or more channels and the gap between the last PPM channel and the first PPM channel is always no less than 5ms. Futaba uses a 22.5ms basic frame. Most aircraft transmitters are between 22.0ms and 22.5ms. Car radio frames are typically 14ms.

[pldaniels]

Interesting regarding the car-sets with their higher frame-rates, I suppose since they inevitably carry a lot less channels they can do so.

[vintage1]

There are no specs.

In addition, since the pulses are generated synthetically by the uP on the radio to drive 'legacy' servos., there is no reason why they shouldn't output them a lot faster, or all channels together.

And they are likely to be 3.3v not 5v on many sets as well.

[pldaniels]

Vintage,

I've heard that Spektrum receivers generate pulses as low as 2.7V, anyone else heard of this?

While I agree there's no "standard", it would still be useful to know what each set 'typically' produces. eg, I've heard a lot of anguish surrounding the use of Spektrum sets with producing shorter/lower-voltage pulses. Not having one on hand makes that a bit hard of course to validate and my google searching only brings up tenative details and nothing concrete.

Paul.

[FenceMagnet]

"I've heard that Spektrum receivers generate pulses as low as 2.7V, anyone else heard of this?"

That's because it's the LOW part of the pulse that counts... the pulse train is weakly pulled up by a current source to some voltage above the switching threshold, which one could arguably say is somewhere around maybe 0.5 volts.... :p

[JimDrew]

Spektrums use 3.3v. Futaba uses 2.7v on many of there systems. Servos that use 5v TTL logic are marginal with a 2.7v input for the HIGH side of the pulse.

There are specs for PPM that were published in the early 70s. That is what we based our system from. The spec calls for 750us to 2250us and a 380us header pulse for each output comming from the transmitter. The gap between the last channel and first channel must be no less than 2.75ms (2250us + 500us cushion).

Car radios typically have 3 channels, so they can easily have 14ms frames. Futaba's HRS mode uses 4.7ms to 5.3ms frames.

[David T]

Spektrum voltages are a moving target: AR9000/7000 3.3v, AR6100 3.1v, AR500 3.0v. Most receivers probably have 100ohm series resistor on outputs.

Paul, I've got some 2.4 frame info here http://www.flyelectric.ukgateway.net/24scanner.htm
dt.

[pldaniels]

David T,

Thanks for that, explains quite a bit (the voltage variances between the Spektrum sets). Any idea why they're doing this?

[FenceMagnet]

Thanks for that, explains quite a bit (the voltage variances between the Spektrum sets). Any idea why they're doing this?

That's because it's the LOW part of the pulse that counts... the pulse train is weakly pulled up by a current source to some voltage above the switching threshold, which one could arguably say is somewhere around maybe 0.5 volts....

deja-vu :p

[pldaniels]

Wait, I'm not sure I quite understand what you're saying Fence. What do you mean by it's the "LOW" part that count. I'm referring to the servo output pulses specifically. My understanding is that it's the "high" part of the ~50Hz pulsetrain that was more significant to the servo. Happy to have my thinking corrected, just not sure if I'm just misinterpreted (unless you're saying that the actual voltage of the pulse is insignificant, so long as it's greater than ~0.5V, in which case yes, that's understandable).

:confused:

Paul.

[Dan Baldwin]

I think FenceMagnet is talking about the PPM pulse stream that contains information for all servos. The statement "That's because it's the LOW part of the pulse that counts" is still not correct. What counts is the transition to transition time, not the low time. I assume that with PCM and spread spektrum systems, that pulse train doesn't exist.

Are you developing a new product that we should know about :D

Dan

[pldaniels]

Dan,

Nothing brand new today *cough* (maybe tomorrow ;) ).

I was also thinking perhaps FenceMagnet was talking about the PPM stream, though as you say, it's more the transition times that matter but the 2.4GHz systems don't even use PPM that I'm aware of (everything is just sent as bytes encoded in what ever modulation they use on the system). Can you clarify FenceMagnet?

Paul.

[rich smith]

Biggest frame rate issue is most servos fail at less than 3.5ms period. I have never seen voltage levels cause servo problems with any 72mhz RX.

[David T]

David T,

Thanks for that, explains quite a bit (the voltage variances between the Spektrum sets). Any idea why they're doing this?
The AR500 is the newest in that bunch and has the lowest reg voltage. The lower it is the more stable it will be because voltage dips won't reach the threshold as often. The technical reference material for their RF chips put considerable emphasis on voltage stability; probably the same for any RF chip. So it might just be them fine-tuning their design.
dt.