Off Topic?? - Serial coms.

Hi all,
This may be off topic, but since the project will be using PICs, perhaps it isn't?

What is the simplest way to initiate two way, ie point to point, comms with the remote end recovering power from the loop?
Protocal would be master interogate, slave answer, repeat ad infinitum.
I have tried RS485 over 500 metres and it works great, but am unable to recover DC from the loop to drive the circuitry at the remote end.
This comprises 2 X 12F675s ONLY to be powered. (fleapower really.)
Years ago I did have details for modified 4-20mA loop with DC recovery but have long lost all info.
Can anyone suggest a solution or point me at a site with one. Have checked a lot of circuit archives without success. BTW the method MUST be real cheap, or is cheap and good mutually exclusive?

Speed is not really important, even 300baud would be OK, anything PIC generatable is OK, up to about 4800.

You could try to modulate your serial data over the PSU line. At very slow baudrate it have to work. I know Melanie already gave an example of that somewhere here.

Just a thought...
Power lines sits at 12v normally, PIC and associated circuitry runs off 5v or less, so maybe it's got a regulator running to it, or it's running off a battery pack that's slow charged off the normally 12v line...whatever might work.
Modulate (send the data) the power line between 12v and 5v (i.e. 12v = 1, 5v = 0), use an op-amp at the RX end to 'subtract' 7v from the power line and feed that output into the RX pin of the PIC.
Got no idea how I'd do it off the top of my head. Could probably come up with something if I'd put my mind to it.
But it's just an idea that might get you going.

A quick idea. Have not tried.

Input from Serial charges cap. When voltage is high enough to start pic, P1 goes high switching signal to P2. P2 receives signal. P3 sends signal. Cap would be 4700 or larger.

Quote:

---

Originally Posted by mackrackit

A quick idea. Have not tried.

Input from Serial charges cap. When voltage is high enough to start pic, P1 goes high switching signal to P2. P2 receives signal. P3 sends signal. Cap would be 4700 or larger.

---

I'd use a small battery pack rather than a cap, trickle charge the pack off the serial line when you can, diode block it so the pack can't feed back into the data line. Cap's are too unreliable in cases like this.

Quote:

---

Originally Posted by skimask

I'd use a small battery pack rather than a cap, trickle charge the pack off the serial line when you can, diode block it so the pack can't feed back into the data line. Cap's are too unreliable in cases like this.

---

You might be right about the caps being unreliable, but the current is so low on the lines it might take some time to charge the battery pack.

Question - I thought RS485 voltage was negative without a signal on the line?

My original idea came from thinking about RFID chips.

Quote:

---

Originally Posted by mackrackit

You might be right about the caps being unreliable, but the current is so low on the lines it might take some time to charge the battery pack.

Question - I thought RS485 voltage was negative without a signal on the line?

My original idea came from thinking about RFID chips.

---

If it is, then hook the battery up backwards :)
Seriously...I'm thinking that even a small battery with an intial charge on it would probably maintain thru the '485 line. And a '675 should be able to run on sub-milliamp current.
How often will the RX be powered, what's it's average daily 'duty cycle'?

More fun to hook caps up backwards.:eek:

Quote:

---

Originally Posted by mackrackit

More fun to hook caps up backwards.:eek:

---

Now you're speaking my language...
I always say 'Everything comes prepackaged with a certain amount of smoke. You let that smoke out, it quits working.' (goes for cars too)

Thanks blokes, I have found a thing called Pbus which looks promising so I'm gunna have a further look at it.