Conduit as shielding

[megloff11x]

How good a job does conduit do in shielding PWM power to a motor? I have to run the wires through liquid tight anyway, and with the price of copper, the shielded cable preferred may as well be made of gold.

Noise leaks out of any seam, which is why we seal metal cases with metal mesh gaskets, and I'm sure that the conduit and its couplings will have the micro-gaps to leak EMI.

I was wondering if anyone had set up such a system and either had issues, or cured them this way. Liquid tight is still under $2.50 a foot and the THHN needed is still under $3 a foot for the 3 conductors plus ground. This vs. $12-$22 per foot for fancy braid & foil shielded cable. I'd melt my pennies if they weren't made of zinc.

Let me know your experiences in this area.

Matt

 

[peteo]

Matt,
You don't want to know about my experiences in this area. The issue with VFDs is that it is high frequency switching, in other words, impulses. This can easily be transmitted into other systems, obviously you know know this as you're taking shielding seriously.

You can, of course, run a motor on the floor with regular THWN wire to it. This is routine. What you're describing, if I understand correctly, is the need to avoid transmissions into other signal wires nearby. Take a 1990-1998 vintage inverter. Well... if you have a 40 foot run to a 202 Hz four pole, cable tied with a 5000 line differential encoder, you will have noise problems. The same installation with a 2000 line encoder can be done reliably as long as the shield drains are connected properly.

Now take a modern vector drive. They generate a lot less garbage than the old units; just listen to them. Just put it there and run it. Yes, the metallic flex is fine to filter out 'bad' signals. If there's a doubt, take the most sensitive instrument you'll be using down to the distributor's and run that setup right next to the proposed drive, using THWN. I think you'll be surprised at what things can take.

 

[RayS]

I'm thinking it would make great shielding. If I remember my electronics right (long time ago, tho), the small gaps from pipe threads, etc. are insignificant because they are so small compared to the wavelength.

You may want to investigate the effect of ferrous conduit on circuit impedance, or consider aluminum conduit, but there goes the cost savings...

 

[W6SJK]

Some reference material for you... http://steelconduit.org/free_downloads.htm

 

[robbietan]

sir matt,
if you can limit the length of the wires running from the supply to the motor, half the prblem is solved. some VFDs have a max cable length limit so that they could not affect any other load.

if you're forced to do long wiring, why not install some reactors at the line side (between supply and VFD input) to protect your system and at the load side (between motor and VFD output) to protect your motor?

you can even do a cost comparison between shielding and reactor cost.

 

[megloff11x]

regrettably the length is about 25ft and fixed by the way it's set up. I've gotten away with it in the past because we had no persnickety sensors and these were also 208V systems. This one's 480V, though I've gotten away with it at 480V in the past too.

I'm trying conduit next. But EMI will leak through the tiniest spots. We had to gasket very tight seams on gold plated (very thin) metal boxes to pass EMI tests.

This doesn't have to pass an EMI test. It just has to leave my %^%& sensors alone! These are newer drives, but the "factory endorsed" power cable costs more than the drive! Another lesson in read the fine print.

 

[petersonra]

This doesn't have to pass an EMI test. It just has to leave my %^%& sensors alone! These are newer drives, but the "factory endorsed" power cable costs more than the drive! Another lesson in read the fine print.

I don't believe you are forced to use the factory recommended cable. You might want to try taking your four wires, twisting them together using a drill, and make your own cable assembly. Its an old but effective way to reduce noise from VFDs.

 

[kc8dxx]

We have had a lot of success using modern VFD cabling. It is essentially a shielded cable. The stuff isn't cheap. One must very carefully ensure that the shield is properly grounded at both the motor and drive ends. Another issue we have seen is the motor and drive grounding. The motor ground needs to follow the drive manufacturer's instructions. Some drives require that the motor ground to come right back to the drive. Both of these techniques have helped in the past when a VFD motor/drive is causing problems with the sensors in the system.