Multiple +/- channel pulse output transmissions

[montyg11]

Hello All,

I have six channels each with a +/- output. The negative channel is a referenced ground, and the + would be a 5-12V pulse output which would output a pulse of 5-100ms upon random detection.

I basically need to transmit these pulse outputs over a distance of about 2000 feet and assign them to the same respective channels.

I looked at a Null Modem to 8-port serial device server to ethernet/fibre, but I don't believe the device servers are capable of converting back to 8-port fiber.

The Null Modem was considered because I require 1-way passive transmission without handshaking.

Any guidance please?

Thank you in advance.

 

[VEBill]

For a one-off solution (not a production-ready solution), perhaps those discreet signals could be interfaced to a local PC, transmitted to a remote PC, and then converted back to discreets. It would require software at both ends. Suitable discreet signal interface cards or boxes are avaiable. Beware of latency. And ensure the complete system doesn't miss short pulses. And if pulse width fidelity is a consideration, then the software will have to pay attention to that.

If not PC, then use one of those embedded CPU modules.

Once the data is in the PC or similar world, there are many options to transmit the near real time status over 2000 feet. A radio Ethernet link might be the best option. Fibre optic would also be good. Copper cable over long distance will "attract" (not really) lightning.

 

[VEBill]

Spelling correction: discrete signals.

 

[BrianG]

The Null modem suggestion might be a possible simple transmission medium, but like most engineering solutions "it all depends..."

You have not told us how "random" is the pulses occuring on the different channels or whether there is some regular frequency or synchronisation of pulse event, where only the pulse width varies. Also, how important is it to not miss a pulse? These sort of questions will determine if the channel pulses can be grouped and sent as some sort or composite data, or whether they must be deal with individually. The more random the channels the higher the 'data sampling' rate and transmission rate must be in order not to miss an event in real time at the far end. As always, the more information the better.

 

[montyg11]

Thanks gents...

@VE1BLL
The software/PC approach is not feasible at this stage in the design and I was hoping to avoid it. But I will definitely have to consider it if I can't figure out anything else. Also, given terrain, fibre optic will be the way to go.

@BrianG
A pulse would be output from each channel every second on average with the pulse width varying. Each channel output is interdependent and of the next and should be independent at the receiver end. As such the data should be dealt with individually and in real-time. Missing about 5% of pulses would be acceptable. I never really considered the sampling rate as I hope to transmit exactly what is output at the detection end.

Sorry if I am still unclear. I will continue to look into this myself as well and post any solution.

Thanks.

 

[VEBill]

... hope to transmit exactly what is output at the detection end. ...

When you have (for example) six (parallel) discrete signals and you wish to send them over a single (serial) link, this necessarily demands (*) a parallel-to-serial conversion step. And that implies a sampling time for the conversion. And that implies that you have to be very careful to define what is "exact" and what is 'close enough'.

(* Unless you used six different optical wavelengths in parallel within a single fibre optic link. ;-) )

 

[BrianG]

Quote: "Each channel output is interdependent and of the next.."

Please clarify if this should mean INDEPENDENT at the transmitting end. "Interdependent" means to me that there is some sort of specific relationship or sequence, for example, channel 5 does not pulse until channel 1, 2, 3 and 4 have occured.

If there IS a sequence, and each output varies only in pulse width between a minumum and a maximum width, it may be possible to devise some simple hardware for generating a sequential self-clocking serial data stream. This data could be sent over a single fibre optic link and then de-serialised at the other end to recover the individual pulse widths.

One simple idea:
Depending on the importance of transmitting an exact pulse width to the receiver:
Is the minimum pulse width adjustable at the transmitter or is the exact width at the receiver critical? It may be possible to have a similar scheme to that used by multi-channel remote control radio systems for models. In that scheme all channels are always transmitted with at least a minimum pulse width in sequence (as a timed channel position marker). This minmum width is decoded by the receiver as the "null" width. Greater pulse widths than this minimum are then expanded as valid proportional pulse outputs to individual outputs to servos etc. as required.

 

[itsmoked]

Do you have to do this via radio?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[montyg11]

@itsmoked

No, I require a wired setup over fibre.

 

[BrianG]

Hi Montyg11, picking up on what Itsmoked asked, I may have confused you by mentioning 'radio control' for models. I was not suggesting that you use radio, just using the same sort of modulation scheme on the serial bit-stream sent over a fibre of cable connection.