Technical considerations for cable length between LED array and driver

Hi,

My first post to CPF. I am working on an idea for an indoor LED luminaire and would like to run a low voltage cable between the LED driver from the luminaire at a distance of 50 to 150 ft. The LED array is constant current...approximately 1050 mA at 35 volts. I've seen very little information about remote mounted drivers...the assumption seems to be that the LED driver is always co-located at the luminaire. Beyond choosing properly sized AWG cable (I am thinking about using CAT5 cable using all 4 pairs), I am concerned that EMI noise from the LED driver (especially when using dimming) may require a shielded cable. Does anyone have any experience with this?

Thanks.

Just look up "Wire Loss Table" on Google, and you'll get the accepted engineering chart for lengths of various gauge wires at various currents, and their related losses.

Regarding the EMI noise, if you are concerned about interference, enclose the driver circuit in a grounded metal box, and use ferrite beads on every wire that comes in or goes out of it.

Merrilite said:

Beyond choosing properly sized AWG cable (I am thinking about using CAT5 cable using all 4 pairs), I am concerned that EMI noise from the LED driver (especially when using dimming) may require a shielded cable. Does anyone have any experience with this?

Click to expand...

The goal would be to get rid of the high frequency content of the pwm voltage waveform, instead of using shielding to contain the high frequencies. I'd suggest some ferrites on each of the wires as they exit the driver, with a small cap wired across the wires just beyond the ferrites. This forms a low pass filter. The values should be chosen to round off the edges of the pwm waveform, or perhaps just get rid of any problems caused by the EMI. The downside is that this is easiest to do if you have an oscilloscope, and pretty hit and miss if you don't have one.

The fact that you are using 50 ft (or more) of wire means that there is a fair bit of capacitance between the wires, assuming that they are kept close to each other. It may be good enough to just keep adding ferrite beads to the wires until the problem goes away. The driver would probably be happier not having to drive directly into the wiring capacitance too.

Shielding can certainly help reduce EMI. It's not a cure-all, though, and it does add cost.

Steve K said:

The goal would be to get rid of the high frequency content of the pwm voltage waveform, instead of using shielding to contain the high frequencies. I'd suggest some ferrites on each of the wires as they exit the driver, with a small cap wired across the wires just beyond the ferrites. This forms a low pass filter. The values should be chosen to round off the edges of the pwm waveform, or perhaps just get rid of any problems caused by the EMI. The downside is that this is easiest to do if you have an oscilloscope, and pretty hit and miss if you don't have one.

The fact that you are using 50 ft (or more) of wire means that there is a fair bit of capacitance between the wires, assuming that they are kept close to each other. It may be good enough to just keep adding ferrite beads to the wires until the problem goes away. The driver would probably be happier not having to drive directly into the wiring capacitance too.

Shielding can certainly help reduce EMI. It's not a cure-all, though, and it does add cost.

Click to expand...

Thanks for the feedback. I agree that dealing with EMI issues is a black art, and an area I don't have much direct experience with. The ideal situation would be to keep it simple...use regular (not shielded) CAT5, with possibly some sort of off-the-shelf low pass filter module. For this application, the prospect of requesting that ferrites be threaded onto individual wires is not likely to go over very well.

I wouldn't say that EMI/EMC is a black art, but to do the job well requires some very expensive test equipment and facilities!

If you don't want to thread ferrite beads over wires, there are some snap-on ferrites on the market. I've used ones from Fair-Rite:
http://www.fair-rite.com/cgibin/catalog.pgm#select:freq1
These are usually considered to be used for troubleshooting EMI problems, but there are instances where they are the best solution for production too. If they will be used for a long term solution, I'd put a zip tie around the ferrite, as the built-in snaps don't age well.

cat 5 is 24awg

it can carry enough current for lights
but only about 3 amps max is advisable

power over ethernet standard is

44V - 57V
Max current 550mA
Typical current 10mA-350mA
Max power 15.4W

Merrilite said:

Hi,

My first post to CPF. I am working on an idea for an indoor LED luminaire and would like to run a low voltage cable between the LED driver from the luminaire at a distance of 50 to 150 ft. The LED array is constant current...approximately 1050 mA at 35 volts. I've seen very little information about remote mounted drivers...the assumption seems to be that the LED driver is always co-located at the luminaire. Beyond choosing properly sized AWG cable (I am thinking about using CAT5 cable using all 4 pairs), I am concerned that EMI noise from the LED driver (especially when using dimming) may require a shielded cable. Does anyone have any experience with this?

Thanks.

Click to expand...

Is this a one off or a product idea?

If a one off, unless its causing a buzz on an AM radio nearby I would not worry too much, though some filtering does help.

Are you talking PWM dimming or the inherent PWM in a switching power supply? You can get rid of the majority of the inherent switching noise and LC filter, that and if you know electronics by smoothing out the edge of the switching waveform to reduce EMI. The latter idea requires a customer driver, not off the shelf.

If you mean the PWM used for dimming, then that is a different beast. PWM dimming frequencies are pretty low, usually 100's of Hz up to 1KHz. If a packaged driver is meant to be PWM dimmed, then normally it will have some smoothing of the edges to reduce EMI. Ferrite beads are not going to be of any use to smooth this out. Ferrite beads are great in the 10's of megahertz (unless it is those enormous ones). For lower frequencies, you need larger inductance and capacitance.

Not sure I confused you more, but it may give you an idea of where to look.

Semiman

If you get lucky, the wire capacitance will help you. Most of the drivers I have worked with needed additional capacitance added near the LEDs if the wire run was long. (long, as in 10 ft vs 1 ft) Some of this depends on if your driver is really "constant current" or some kind of pulsing approach. Do you have a specific driver in mind ?