Ultra low baud rate.

[dr pepper]

I have in the past developed data over power circuits, I have another project for a friend that also needs this but in a very electrically noisy environment, so I'd like a baud rate of maybe 30 or 40 baud.
Looking at the datasheet for a pic16f628 the lowest rate on the list for the generator is 300 baud, can I get lower than this without lowering the clock freq.
I have a software routine I use for lower end pics that handles rs232, but its a little heavy on processor time.

 

[3v0]

If you can't do it in hardware rewrite the bit bang stuff to use timer generated interrupts.

 

[dr pepper]

Yes I díd that before chips with rs232 became popular.
I'd like the ease and power of using the uart, but it sounds like it'll have to be bit banged.

 

[3v0]

I ran into much the same thing with NIST radio.

 

[dr pepper]

One benefit from using the pic's uart would be noise immunity, as far as I know the pics usart takes multiple samples of the incomming data stream and calculates a 1 or 0 depending on the majority of the samples.

Isnt nist radio the Wwvb time signal?, I wrote some code to decode the english version msf, it operates at only 1 baud!, the slowest I've ever seen.

 

[3v0]

Yes on NIST.

You are right about the multiple samples. But you can duplicate that behavior.

 

[Nigel Goodwin]

Looking at the datasheet for a pic16f628 the lowest rate on the list for the generator is 300 baud, can I get lower than this without lowering the clock freq.

Why not just lower the clock frequency? - is the actual program processor intensive?, PIC's commonly spend almost all their time sat in loops waiting for the real world to catch up.

 

[dr pepper]

Yes sort of, its a datalogger and has I2c and a few other things happening, at present it just uses the built in osc.
The 'f628 can switch to a lower freq, 32kc, maybe I could switch to that to transmit then switch back, it'd still kill the processor though I want to send 20 bytes, at these sort of speeds that'd take over 2 seconds.
Bit banging with interrupts might be a less intesnive way, and I spose I could implement multiple sample averiging in software at these low data rates.

 

[NorthGuy]

If you only need a 30-40 baud rate and don't use built-in UART module anyway, have you thought of creating your own protocol synchronized with AC 50/60Hz signal?

 

[dr pepper]

I want to communicate with a laptop so I need something that will work with the serial port, doesnt have to be exact rs232 protocol though, just compatible with the ports hardware.
The logger will be logging traction motor data on board a tram, ac is available, I could use an input port on the pic to detect zero cross and sync the centre of every bit at the power line zero cross point, the laptop will automatically synchronize, noise immunity will also increased if I get the centre of every bits timeframe at the power lines zero cross creating the max no. of samples where noise will be lowest.
Good idea NG.

 

[nsaspook]

There is an X10 app note for some circuit ideas. https://ww1.microchip.com/downloads/en/AppNotes/00236a.pdf

 

[dr pepper]

Interesting, however I want to sync comms to the power line not use the power line itself.
Noise on the line is tremendous 1930's electric motors didnt bother too much about noise, and power to the vehicle comes from a roller on an overhead 600v cable which isnt exactly clean by any stretch of the imagination.
I could use screened cabling however we are trying to do this using an existing multicore, which is screened but does carry other high voltage loads.

 

[nsaspook]

Interesting, however I want to sync comms to the power line not use the power line itself.
Noise on the line is tremendous 1930's electric motors didnt bother too much about noise, and power to the vehicle comes from a roller on an overhead 600v cable which isnt exactly clean by any stretch of the imagination.
I could use screened cabling however we are trying to do this using an existing multicore, which is screened but does carry other high voltage loads.

I see. We used 60ma current loop connections to send serial data in conditions like that.
http://en.wikipedia.org/wiki/Digital_current_loop_interface
http://ckp.made-it.com/currentloop.html

We also used it to send data over miles and miles of crappy wiring for teletype networks.
http://www.baudot.net/teletype/M28.htm
I did field repairs to these monsters. The current loop operated a high speed solenoid that at the start bit started encoding the received data on a rotating cam that during the stop bit engaged the typing gearing to the cam code.

 

[nsaspook]

dupe

 

[dr pepper]

Yes of course current loop, hadnt thought of that, should have as they use that for some of the instruments where I work, 4-20mA is the industrial standard.
What I have set up is I spose to an extent a current loop, transmission is done by dumping current through one resistor for a 1 and another for a 0, the receiver discriminator detects current my measuring voltage across a resistor inline, thinking about it the only thing I need to do to improve noise immunity is lower the value of the resistors to make current mod higher, so long as I dont create too much ringing, its a 2.5mm straight run cable.
I have some experience of telex and rtty, what I'm doing is similar only 8 bit instead of 5, an digital instead of tone.
I have a test jig on breadboard held together with elastic bands and luck, I'll have another trial session with greater current mod this weekend.

 

[nsaspook]

Good luck, we ran 120vdc 60ma loops for rtty circuits. Completely bullet-proof for noise immunity as long as the receiver detector loop is isolated from ground.
The modern opto based circuits should work nicely. https://www.avagotech.com/docs/5953-9359E

 

[dr pepper]

Yes ground might be an issue.
I have 2 wires to play with and was hoping to power the aquisition pics from the host laptops usb, however as per usual the project has expanded way past initial specifications.
Because the power comes from the laptop host I dont need to isolate the aquisition devices from the laptop, however I thought of using an opto to isolate the mains sync signal, I could use a nasty cap/zener circuit but I think this time I'll go with a transformer then I wont even need an opto, just stuff the trans o/p to the pic through a high enough value resistor not to overloads the pics clamp diodes.

 

[dr pepper]

433 mc's RF module worked well, from anywhere in the car to the motormans seat, never even bothered with wired connection.
My current transformer clamp worked well, on its limit at 100a a couple of times.
Now to get on and write the software, I have mmtty to help prove the sending end is working, and at the laptop end I'll have a black box with a rf module, lm567 and a pic to convert the rtty into serial for the laptop.
Been an interesting project so far.

 

[dr pepper]

Well I got the system to work well.
I can decode rtty from a silly distance off a little 100mW chinese 433 mc's rf module.
Current and volatge probes simple enough.
Now its time for the visual basic software to add the measurements and the gps data then work out how much energy therefore £'s is being used by the motors and dumped in the braking resistors over the course of the days up and down the track.

 

[NorthGuy]

I can decode rtty from a silly distance off a little 100mW chinese 433 mc's rf module.

Are they legal? Linx has TRM-433-LT modeles, which are 10mW. In their docs they say that an adjustment resistor is necessary to lower the power (to around 5-7mW I guess) to comply with regulations.