EMI from long PWM cable run

[HPsauce]

I'm working on a project that's going to involve running PWM at somewhere between 2.5 and 25kHz, down a 30m long pair of wires (hanging in space). The max current will be 6A and EMF 48V.
Can anyone tell me how much of a problem EMI is going to be? I'm looking for a rough answer, along the lines of:
1. no problem
2. might upset sensitive audio equipment
3. will cause serious interference

Thanks

[Christe4nM]

Others may have a more readily available answer, but still it comes down to "it depends."

Some factors to consider:
- What kind of cable do you use?
- Is it shielded?
- How is the cable connected to the system. For example, shield connection matters if it is to be effective
- Signal frequency alone does not say a lot I'm afraid. So what are the rise/fall times of your PWM; In other words; what spectral content does it produce and at what levels?
- Is there any other signal that can couple into the cable and turn into EMI emissions?

Note that a EMI emissions and immunity are two sides of the same medal. If a systems has high emissions in a certain frequency band you can count on it being susceptible to disturbances in that band as well. It goes both ways.

[HPsauce]

It's a pair of 2mm stranded wires, running parallel 20mm apart, hanging free.
No screening
PWM details are TBC. I will make it pretty slow to reduce emissions as much as possible.
The cable is powering 50 lighting fixtures (evenly spaced on the cable), each consisting of PWM driven LEDs, so yes, these will contribute too.

[nickm]

When you say the pair is running 20m apart, is it one single conductor and then another single conductor 20m away? Or does the wire have two conductors and you are running two of them 20m apart?  Ideally the send and return paths are right next to each other because blasting 6A edges around a 50m loop is just about as bad as you can get.

I'll assume the wire has two conductors.

Any filtering?  It will go a long way.  Upsetting audio equipment will probably more of an issue of what kind of current it is drawing from the line side.  If emissions upsets the audio equipment then the equipment isn't that good.  What you'd most likely have EMI issues with is interfering with the am band 500kHz - 1.6MHz. 

At 25KHz squarewaves, 500kHz is 26dB down so you'd see about 3Vpp of 500kHz on your lines which you could probably pick up on nearby radios.  When you squarewaves turn into low duty cycle PWM it only gets worse.  If you put a second order LC at 25kHz on the output, the 500kHz would drop another 52dB and would not cause much trouble. 

Now if you are actually trying to pass FCC emissions you would be lucky if you got away with only a single LC.

[CaptnYellowShirt]

It's a pair of 2mm stranded wires, running parallel 20m apart, hanging free.

Looking at this in a plan-view you'll see a 20 x 30m loop of wire?

[HPsauce]

Doh! 20mm apart. Not 20m. I've updated my post. Sorry

[HackedFridgeMagnet]

The pwm can be filtered out to give what you really want for driving LEDs, a PWM controlled current source.

So put a filter on your output. But make sure you are sensing the current that is going through the LEDs.

[SeanB]

Unless you are using a twisted pair wiring scheme you have a pretty effective long wire antenna there, it will definitely radiate a lot. You definitely will need a decent LC filter there to provide close to DC to the LED's, and preferably a LC filter on each LED module as well to reduce EMI, otherwise you will be making a wideband noise source.