Hi,

Seems like stadard PPM coding is very old, why are people still using it?
(RX do have shift registors in it, some more advanced a MCU,
but they can still be tricked to put servo arm in wrong position on radio loss,
because off old/too simple PPM protocol)

Why not make a "very simple serial protocol" instead with checksum,
so we don't get any servo flip?

(It have to be simple, so its will take no time to send a "packet",
then we have a chance to receive some "new sevo position" data with poor radio reception)

Some have added this to their PPM TX/RX with ex a PIC MCU?

Or what is the name?

BR
LoopForEver

Uhmm- its called PCM.

PPM is still in use because 99.99% of the time it just plain works.

Pat MacKenzie

>Uhmm- its called PCM.
I know PCM, but i pretty sure its do not include a checksum,
only checks for bad data, right?

>PPM is still in use because 99.99% of the time it just plain works.
Maybe also most because its a standard, and that what we always have used,
cheap to use because only simple chips is needed?

I do get alot off servoflips, when my radio connection is getting bad with PPM.


BR,
LoopForEver

How else to check for bad data but to use a checksum? Pretty much the standard way to do it.
Even the simplest parity bit is a form of a checksum, but PCM uses a more rigorous method.
Some details for Futaba PCM, including the checksum are here

Pat MacKenzie

>but PCM uses a more rigorous method.
Oh i see, it do have checksum calc :-)

Just me remeber wrong heh

So its near impossible to get a servo flip with PCM,
anybody with experince in real life?

LoopForEver

In my experience PCM never "glitches".
The worst that it is ever supposed to do is "lock out" and go to the failsafe positions or hold the servos, depending on how it has been programmed.

Not sure I have ever even had that happen. If it has, it was for too short a time for me to detect.

The Berg rcvrs ( and others) can do some smart things be looking at the PPM signal and making sure that it follows all the rules.
They even recognize your transmitter as far as number of channels and the sync pulse widths to prevent glitching. Poor man's PCM.

Pat MacKenzie

You can't "checksum" a PPM transmission because it doesn't transmit digital data, only pulse-lengths which are by their very nature an analog quantity.

If you turn those pulsewidth values into a binary value and transmit that value instead of the pulse itself *then* you could checksum (or even better CRC) the data -- but that's exactly what PCM does.

There's another fly in the ointment however -- it takes more bandwidth to send the binary value of a pulse than it takes to send the pulse itself so you'd end up with a slower frame rate or you'd have to apply compression to the datastream (which is what many PCM systems do).

So you see, you're just reinventing the wheel and coming up with PCM.

The most practical way to reduce glitching on PPM is to use the IPD/DSP techniques that a lot of manufacturers are now offering. This involves checking to see that the individual pulses (of variable width) are arriving within the anticipated time-windows and are within the expected normal range.

Small glitches can still get through but really bad noise is filtered right out -- resulting in the same kind of "lockout" as can be experienced with PCM.

Originally Posted by XJet

Y
There's another fly in the ointment however -- it takes more bandwidth to send the binary value of a pulse than it takes to send the pulse itself so you'd end up with a slower frame rate or you'd have to apply compression to the datastream (which is what many PCM systems do).

Another plus for PPM is its efficient use of bandwidth. Bandwidth on 72 is a scarce resource.
In a digital age we forget that good old analog signals carry huge amounts of information. Most of the time converting to digital means loss of information ( compression and/or digitization losses) and/or increased bandwidth.


Pat MacKenzie

Hi,

This idea is only ... 20 or 25 years old !!!

What was done by F.Thobois ?

1) send a digital checksum of infos + signature
2) send a classical PPM stream
3) If checksum applies to PPM stream ... then send it to decoder ... ELSE output the FailSafe stream ...

And the receiver could also work with "classical" Txs ... just a switch to toggle !!!

Coder was a 68HC11 and decoder a 68HC705 ...

Alain

...

What about using free mix of classic transmitter, to compute "checksum" and sending it as PPM last servo signal? (even analog one can compute summary, using few resistors)
It is better than simple parity bit for noisy environment. Yes it is not any self-repairing code, but which RX really does more than Ok/drop testing of checksum?
Anyway, some functions are better if controlled than if in predefined position or even Hold state.
Strict Checksumming is not so welcome.. PCM can save, but also can kill the plane.. hopefully only the plane. At least the motor have to cut off.
For example PPM, MPX IPD - does floating average of signal. That may save the plane even in case of bad noise..
Nopw it is too late to develop PPM.
2.4GHz systems are here and there is everything a bit different. And far above simple checksum.

Originally Posted by pmackenzie

Another plus for PPM is its efficient use of bandwidth. Bandwidth on 72 is a scarce resource.

Eh?
PCM on 72 MHz takes the same bandwidth as PPM on 72 MHz. I do believe you generally pay for PCM with a somewhat lower signalling rate, however. Is it worth it? Probably -- but PCM still won't magically cut through the interference of somebody else turning on on your channel -- it'll just make your plane deal with not hearing from you better.

The PPM that we use isn't particularly efficient with bandwidth -- even our AM R/C radios are better -- but it's good enough, and there's not much benefit in trying to use less. And perhaps more importantly, the gear is cheap, readily available and light.

But give it a few more years (ten perhaps?), and I imagine that PPM for R/C use will only be used for the really cheap toys, and most of the serious R/C users will be using spread spectrum. Sure, people will still be flying 72 MHz (or whatever their local equivalent is) but mostly just because they already own the gear and it still works -- anything new will probably be spread spectrum.

In a digital age we forget that good old analog signals carry huge amounts of information. Most of the time converting to digital means loss of information ( compression and/or digitization losses) and/or increased bandwidth.

Well, when you convert to digital, you also gain the ability to do things like checksums, error correction, retries, compression, encryption, etc. And often the compression allows you to use less bandwidth than even the straight analog signal (even using SSB, but that's rarely used for consumer gear because the quality is gear) would use. That's probably the #1 reason why cell phones are all digital now, for example.

Another advantage of the current system is that with signals only being sent 1/50th of the time then far less electric is being used than if it had to send a signal almost continuously, which is a big plus with battery power. I don’t want to have to charge and carry 5x tx batteries around with me.

Originally Posted by Rufe0

Another advantage of the current system is that with signals only being sent 1/50th of the time then far less electric is being used than if it had to send a signal almost continuously, which is a big plus with battery power. I don’t want to have to charge and carry 5x tx batteries around with me.

Actually, FM PPM transmitters transmit 100% of the time -- they bouce back and forth between two frequencies 5 (I think) KHz apart. It's pretty much just FSK but we call it FM anyways, probably because I think we just send our PPM signal into a FM transmitter stage -- which emits FSK.

AM (it's not really standard AM either, but we call it that) transmitters only transmit on one frequency, and it turns on and off -- but it's on far more than 2% of the time.

PCM is ordinary FM transmitter, just instead of transmitting one pulse with various length for each channel like PPM, it transmits many shorter pulses that represents binary number. It is allmost the same.
Well, PCM uses more than 12 bits = pulses to transmit a number (depend on number of steps = precision of movement, checksum and another infos size), equal to duration of single PPM pulse..
With PPM, duration can be transmitted and received by any precision (depend only on carrier frequency, e.g. 35Mhz vs 72MHz)..
The maximum precision of PCM depend on which one of many incompatible PCMs is used, but have to fit within carrier acceptable bandwith.

Well, PPM has single, but VERY important characteristic: compatibility.
I would bet that servo signal standard, basicaly one channel part, directly splitted out of PPM, will survive ages that I will not.

AM is very outdated and gets allmost unusable within todays noisy air.. even if 2.4Ghz is a bit likely to AM.

The current consumed by transmitter is neglible, in case that procesor inside is not wasting energy, or the display backlight of some new, nice transmiters with Windows CE .
But allowed transmitting power is 1W here in europe, and in real transmitters are using far lower power.