VFD cable on line side?

Status
Not open for further replies.
A

adamscb

Senior Member
Location
USA
Occupation
EE
Hi Forum,

I've been doing some research on VFD cable and I had a question. I understand that VFD cable isn't as necessary when the wires are in their own metallic conduit. It's when the wires share a non-metallic conduit with other circuits that problems arise. Do you ever need VFD cable for the line side of the VFD? In detail, the power wires going from either an MCC or panelboard to the incoming of the drive. I would understand on the load side in which PWM is taking place, but I would think the sine wave would be unaffected on the line side. I have heard of line side filters however, so that makes me think that line side VFD cable is necessary. So I guess really my question is if VFDs affect the incoming power quality, to the point in which line side VFD cabling is necessary? I know there are instances where load side VFD cable going to the motor is necessary, but wasn't sure about line side.
 
The input to a VFD is a sine wave for the voltage but the current waveform is a series of pulses. Line reactors or filters take the edge off of this. I don't think you would gain anything from using VFD cable on the line side though.
 
Filters on the line side have to do with harmonics, not the same issues as what happens on the load side via PWM. No need for special wiring on the line side.
 
Jraef said:
Filters on the line side have to do with harmonics, not the same issues as what happens on the load side via PWM. No need for special wiring on the line side.
Click to expand...

The major problem on the line side is impedance. Most VFDs are designed for around 1-3% impedance. DC drives in particular are really bad without it. Most transformers are close to that ideal. The issue is that if the transformer is much larger than the VFD the apparent impedance is too low. The biggest issue this causes is that short circuits can destroy the drive. Most have an SCCR around 5 kA which is very low with some exceptions. Generally regenerative drives tend to be a little more hardy than diode front ends. A good rule of thumb is if the transformer kva is 10 times the VFD kva or larger, you need a line reactor. Or semiconductor fuses.

The other issue is harmonics but here’s the big lie. Which matters, current harmonics or voltage harmonics? And, does current harmonics cause voltage harmonics or vice versa? The answer is that current harmonics forced through a big impedance cause voltage harmonics. So if the drive is close to the same size as the transformer then as the current harmonics cause the current to have peaks exceeding the transformers capabilities it causes voltage dips...voltage harmonics. This overheats the transformer, causes issues with other devices on the same line, and so forth. It is common sense why this is a bad thing. IEEE 519 is the harmonic standard that is often quoted and it recommends keeping voltage harmonics, %THD, under 5%. As far as current goes as long as we are not reaching the limits of the transformer current harmonics are completely harmless and not an issue. In fact IEEE 519 puts limits in current harmonics but it uses %TDD, not %THD. TDD is total harmonic DEMAND, for harmonics in terms of the transformers capacity. Realistically if you are sizing everything that close you will have problems with a soft bus and voltage sags every time a simple contactor for a cooling fan kicks in. So it’s easy to scare someone with the threats of harmonics by showing them a nasty looking chart that happens because of the classic “bunny ears” current draw of a six pulse VFD but claiming that this is a huge issue is a big lie. The biggest problem is piss poor grounding and binding leading to electrical noise leaking out everywhere.

I’ve also heard wild stories about adding fancy filters to address it. The fact is that a line reactor and an isolation transformer provide EXACTLY the same performance in terms of knocking down harmonics but the isolation transformer is more expensive. You can do better with phase shifting transformers arranged in pairs with multiple drives but this is tricky to set up. The only time EMC filters are needed is in the EU. They have a bunch of anti-EMC directives that don’t make any sense in an industrial environment so in Europe you buy the drive with an EMC filter ant everywh
 
Hit the button accidentally...everywhere outside EU you don’t need the EMC filter. Power harmonics are weighted by the harmonic anyways...the higher the harmonic, the less it affects %THD. The EU EMC filters really only affect harmonics in the Kilohertz.to Megahertz range. They are ineffective against power harmonics.

Beyond a line reactor it is possible to build more complicated filters with multiple reactors and capacitors to knock down harmonics further or filter specific ones or buy a super expensive clean power drive but the fundamental issue of too small of a transformer remains.

I forgot to mention DC chokes. For harmonics and short circuit performance they have exactly the same performance as a line reactor but at about 25% of the size and 25% of the price. If the drive can use one, this is your best option. The only thing a line reactor does that a DC choke doesn’t is that line reactors provide a little surge protection to the drive.

Finally another lie involves triplen harmonics. These are odd harmonics divisible by 3 which are 3, 9, 15, 21, etc. A nice feature of delta wye isolation transformers is they cancel all triplen harmonics. But six pulse drives don’t cause any triplen harmonics in the first place so the transformer has no effect. That’s why I said they are identical in performance to line reactors. Computer power supplies and HID lighting does have triplens, just not drives.

Finally harmonics on the output DO NOT MATTER when using a drive to run motors. The output is rectangular waves, not sine waves. The output is designed to run a motor though.

There are two issues with it though. The first is when cables get electrically long rectangular pulses don’t couple to the motor well so sometimes we have to add filters to knock off the corners and make it more rounded pulses. The most effective filter for the vast majority of cases is the dv/dt filter. Load reactors work and are slightly cheaper but they kill output power and so do sine wave filters. Sine wave filters do the best but they are expensive, huge, and only necessary for truly excessive line lengths (over 1000 feet).

The second issue is that in some cases you can get bearing damage from unbalanced currents. Bearing damage has been around since electric motors existed. It is nothing new. Technical papers on it exist back to 1910. In the past though it was more of a large motor thing. It was a problem above 1000 HP. With VFDs it’s more of a small motor issue. The solutions are either to use a common mode filter or a grounding brush at the motor. A load reactor, VFD cable, and dv/dt filter do nothing, zip, zero, zilch because they don’t affect common mode voltages. Sine wave filters help marginally because they filter everything almost into true sine waves but you take a performance hit, make lots of waste heat, they need their own cabinet, and they are more expensive than the drive itself.
 
Status
Not open for further replies.
Top