Theoretical baud rate via RC link

[Hell-e-Guy]

Hi Guys,
I want to know what's the highest baud rate that I can pass through an RC transmitter module. I will connect the data source to the PPM pin in the transmitter module and extract it from the receiver before the shift register.

Whats the calculation?
what do you think....

Hell-e

[AndyKunz]

Unless you get into special modulation techniques (ie, non-standard RC gear), you are going to be limited to the delta of the frequencies for the FM modulation. If you are using a 27MHz system where channels are spaced 50kHz apart, you can get up to about 50kHz data through (38400 bps would be highest standard rate). If you are using modern 72/75MHz stuff where they are spaced at 20kHz, the max would be 19200. But you would be stepping on somebody's toes doing that, since there are other non-RC channels spaced in between ours and would really need to consider 9600 as the top speed.

http://webpages.charter.net/rcfu/HelpsHints/RFC.html

Andy

[Hell-e-Guy]

OK,
I just tried sending at 4800 and I don't receive the same character that I send.
I first thought the signal is reversed but the scope showed that on the receiving end the data is not nearly the same as the transmitter.

I am sampling the output of the comparator in the receiver just before the shift register.

could this be because of bandwidth problem?

Hell-e

[Peter Seddon]

IMHO if you are using one of the channels in your transmitter to try and send data you will not be able to send data at a faster rate than 50bits per sec. The reason is that the channel data is sent as a frame of 6 (or maybe 8) channels of PWM data, and the frame rate or repetition rate of each frame is 50frames per sec.

By directly modulating the carrier of the transmitter then you would be limited by the channel spacing as has been suggested.

But there again I could be wrong!

Peter

[vintage1]

Theoretically the baud rate is something like the bandwidth times the signal to noise ratio.

i.e. if you think in pure analogue AM terms, you could transmit a variable level signal at th highest frequency the channel could take, and detect its level between maximum amplitude and the noise floor, with about 1dB steps. so a 3Khz channel with 64dB S/N might give you 64 levels, which is 5 bits, times 3Khz is 15Kbaud.

very clever modulation can improve the differentiation..for example a 3KHz telephone line will do 56Kbaud.

The trouble is getting the clever modulation. PPM ain't it..its very crude and the actual data rate is about 8 bits per servo 20 times a second..and maybe 8 servos.. 1280 baud.

To get ore would require a pretty sophisticated setup..maybe a modem chip set to 4800 baud using te FM channel as an audio channel would work..but unlike a phone line, where phase delays and frequency response is fairly fixed..an FM channel in the sky is going to have awful variations as you fly through multipath zones etc.

A 4800 baud serial data signal is already beyond the receiver bandwith anyway more or less. 2400 is about tops for simple FSK.

If you put that into a tx modulator, don't expect to find it on any given channel either..you need to look at the decoder INPUT and you MIGHT have something a bit like it.

[vintage1]

Unless you get into special modulation techniques (ie, non-standard RC gear), you are going to be limited to the delta of the frequencies for the FM modulation. If you are using a 27MHz system where channels are spaced 50kHz apart, you can get up to about 50kHz data through (38400 bps would be highest standard rate). If you are using modern 72/75MHz stuff where they are spaced at 20kHz, the max would be 19200. But you would be stepping on somebody's toes doing that, since there are other non-RC channels spaced in between ours and would really need to consider 9600 as the top speed.

http://webpages.charter.net/rcfu/HelpsHints/RFC.html

Andy


This is wrong on two counts. What counts is the reciever bandwidth and that is typically 3-5Khz only.

And then you have to assess the S/N ratio. For theoretical best complexc modulation anyway. Using simple 'crude serial' to FSK. its going to come out at less than half the receiver bandwidth. Less than 3Khz for sure.

[vintage1]

OK,
I just tried sending at 4800 and I don't receive the same character that I send.
I first thought the signal is reversed but the scope showed that on the receiving end the data is not nearly the same as the transmitter.

I am sampling the output of the comparator in the receiver just before the shift register.

could this be because of bandwidth problem?

Hell-e

Yes. The pluses are really crudely shaped to avoid frequency smearing in the transmitter, and the IF filters will do their worst at the far end as well.

1200 baud should look pretty decent, and with some pulse shaping you may be able to get to 2400.

[simingx]

If you're looking at sending data out from the plane, a dedicated 433MHz (or whatever) module would be smaller and faster...

[MatC]

The bit about signal to noise ratios is correct, but my understanding is that in normal noisy environments you get about the bandwidth. So 5KHz wide gives you about 5k/s.

The nice (or bad...) thing about PPM is that a small amount of noise will not affect the signal too badly, but it will change where the thresholds are and this will affect the pulse widths.

So although PPM may be around 1300 baud when noise-free, you won't get out exactly what you put in when there is noise or at the limits of range: the pulses will stretch or contract a little. Happily however for PPM this has only a soft performance degradation, not the all-or-nothing you tend to get with PCM systems.

If you want to put non-control data down it, you want something that sends more information and recovers from more errors. Look up turbo codes if you want the best known method to date :)

[Hell-e-Guy]

Thanks for all the replies !
Peter,
I am feeding the ppm input pin directly with the data and not through a channel.

V1,
If the receiver's bandwidth is around 3-5k (sound's about right as the IC is made for old cordless phones), the maximum baud rate should be around 1.5k to 2.5k (about half). This is because the data I am sending needs 2 changes for every bit ??!!??

Please correct me if I'm wrong...

Hell-e