PWM over a distance

[chico]

I have a PWM signal which I would like to send 50 ft.
Right now I have it at 20 kHz I had not taken into account cable inductance, is that an issue?
Its just a signal, so its little current and only 5 V.

I can lower the frequency if its needed.

 

[crutschow]

Cable inductance is not a problem. Cable capacitance could be depending upon how you drive the signal. Cable capacitance is generally several tens of pF per foot.

I would suggest you use a pair of chips for RS422/485 transmission, such as Maxim sells, to send the signal down a twisted wire pair. That will give a good signal and minimize any induced or common-mode noise problems.

 

[ronv]

There are a lot of variables to your question, like what are you driving it with, is the power for the receiving end also going down the cable or just the ground?

 

[chico]

I am driving it with a 555-timer and the power for the receiving end is elsewhere.
There would be other signals with it but, they are all just 5V or 0V permanently.

If I use the twisted pair wires but no RS422/485 transmission ICs would this deal with the majority of the problem?
I routed the boards already before I realized this, a different cable is easy, newer boards is harder... I guess I can just probe it at the end of the cable with a scope.

 

[crutschow]

If it's already designed then just use some twisted pair wires and see what the signal looks like at the cable end. Twisted pair doesn't help a lot without a differential transmitter and receiver, but it's probably the best you can do without changing the circuit.

 

[ronv]

When I tried to model it as 33Uh inductors and 750Pf capacitor it looked pretty ugly until I clamped it at the receiving end with some small signal diodes (one to keep the signal from going above +5 and one to keep it from going below ground. So if you see a lot of ringing try it.

 

[unclejed613]

did you terminate it with a 470 ohm resistor? that will definitely be a step in the right direction. put a 470 ohm resistor across the receiver end of the twisted pair and see if it cleans it up a bit. you may want to use a schmitt trigger as the input to the receiver which will clean it up a lot (in addition to the 470 ohm resistor. that would be a 7414 or equivalent

 

[RCinFLA]

Cat5 and Cat6 twisted pairs have a 100 ohm characteristic impedance. Cat6 is more consistant because it has a plastic 'X' separator to keep the four twisted pairs in their own little space avoiding possible inter-mingling of the pairs in cable bends that occurs with Cat5.

Drive them differentially with 100 ohm source and termination impedance and you will get a good renditions of input at the end. Propagation delay is about 5.47 nsecs. per meter.

 

[unclejed613]

transmission line characteristics don't really come into play here as we're talking 20khz, and 50ft is far less than 0.1 wavelength. if this was 2Mhz or so, then we begin talking transmission line characteristics for a 50ft run.

 

[chico]

OK what do the RS422/485 transmission ICs do? are they like high speed buffers? do they go at both ends? I could wrangle another board onto my original board but there isnt space for anything at the receiver end. A 470ohm through-hole resistor on the wires maybe. The resistor for the PWM signal I assume goes from PWM signal to Vcc (or ground if it wanted it to be a pull low which i dont).

This is feeding to a PWM controlled fan, im not sure what the fan circuit does but it is the type of fan that has a PWM input as well as Vcc+ and Vcc-.

 

[crutschow]

The RS422/485 ICs are high speed differential drivers and receivers. They convert a single-ended signal (referenced to ground) to a differential signal on two lines (one line goes high while the other goes low and vice-versa). The advantage is that the differential receiver (which converts the signal back to single-ended) only looks at the differential signal between lines and ignores any common-mode noise (noise common to both lines such as between the transmitter and receiver grounds or noise coupled externally into the wires). You thus can transmit the signal over long lengths of twisted pair unshielded wire without problems from noise.

If you don't have room for these devices, then I would suggest you use a shielded wire to carry the signals. Shielded CAT5 will give you several twisted pairs inside a shield.

 

[Mr RB]

20kHz is a really low frequency for 50 feet of Cat5 cable.

Just put the 555 at one end of the cable and the fan circuit at the other end and see if it works...