Picaxe to Picaxe communication over 48v AC line

Hi Forum

Can anyone help with a solution for sending serial data between two picaxes over a low voltage (48v) AC powerline.

I am sending HiDef Serial Digital video over 100m of fibre optic to a receiver unit. The fibre incorporates 4 copper cores that we are pairing to supply 48v AC to a tranciever at the distant end.

I would like to use the Picaxes to pass serial data and monitor activity lights on the distant unit.

Cheers
Al

Cheers, bit pricey though. I think i will try to steal two of the copper lines for the serial

100M is a long way for serial .......

And if the PICAXES are ones without resonators, the RX and TX clocks could be out quite a bit, which wont help.

DMX-512 will work over 1km of mains cable at 250000 baud. That uses a differential line but considerable distances can be achieved with standard single line serial.

As serial has a combined tolerance of +/-5% or so, and most PICAXE with internal oscillators have +/-2% accuracy or better, there shouldn't be major problems there under normal operating conditions but those with external resonators will have greater stability.

Best thing to do, as often the case, is to prototype things before settling on a solution.

Cheers Guys

Idea:

At the TX end. Use a current transformer in line with your 48v cable. Use the Picaxe INFRAOUT command with some current amplification to drive the transformer. INFRAOUT generates up to 127 unique codes modulated at 38KHz.

At the RX end. Tap the 48v cable, straight voltage tap. Use a simple 38KHz bandpass filter and the Picaxe INFRAIN2 command.

Not full ASCII, and much slower than serial, but better error rejection and 127 codes should be good enough for jazz.

My own experiments with serial using multiple 10M serial extension leads to see how far I could push serial from the PICAXE resulted in dropouts in the data at the other end(garbled characters) once I got to about 60 meters.

I was using 2400 baud.

You might be able to get further then that using a 232 converter at each end to up the transmitt voltage into the cable run, but I have not tried that. My experiments were at TTL directly out of the PICAXE pin and into the cable, then out the other end into a laptop terminal program.

I would expect problems trying to push the data over 100 meters...

I've been tempted to suggest radio modules instead but I believe there shouldn't be much problem doing a slow serial connection over 100 meters, especially after remembering one of my earliest threads on this forum about serial comms and wire distances:

http://www.picaxeforum.co.uk/showpost.php?p=111171&postcount=9

Also the thread is an interesting read because I wanted to do 2-way comms over 2 wires, hippy came up with a resistor wiring diagram to achieve this without damaging the Picaxes in case both tried to send data at the same time, the idea works fine.

http://www.picaxeforum.co.uk/showthread.php?t=13251

XLSERVICE said:

Cheers, bit pricey though. I think i will try to steal two of the copper lines for the serial

Click to expand...

If the 48V AC power is running parallel with your data lines, it will induce a substantial 50/60 Hz voltage over that length.

You will have to modulate the data with a higher frequency carrier, and then demodulate it back at the receiver end.
As such, Boriz's solution is the one I like. I would add some coupling cap between the powerline itself and the 38Khz bandpass filter to substantially attenuate the powerline frequencies.

May be worth giving it a try.

Thanks Fernando. I was beginning to think it was a stupid idea.

Another approach cold be switching from power to data for some small duty cycle with a battery at the far end being charged to hold up the electronics whilst the data is sent - the change could be sensed by sensing the AC current in the lines - when the current goes down it's time to switch to data for a designated short period... practicality of this approach would be determined by how much power is being sent by the AC and data speed etc...

What is the load?

What is the 'tranciever at the distant end.'

I'd go along with Boriz's idea too. Aftrer all it's the method used by several data transmission manufs (with a little ZX sensing chucked in).
If you have 2 spare conductors you may consider differential drivers-receivers to negate any induced signal issue. This would be faster and possibly easier.

Bugger. There's me trying to be original again. What an ass. Will I never learn.

Well there's nothing wrong with applying the grey matter - even if you come up with an idea already used from years ago.
There has been data transmission on power cables for donkey's ears ; maybe you read it as a boy and had forgotten?
Maybe you saw a spec on X10 or similar?

Will you ever learn? Who knows .... I doubt if any of us will ever learn