EMC/EMI filter question

[brandon_costa]

Does an EMC/EMI filter like https://www.valinonline.com/products/508-120 go before the VFD or after the VFD on the cable that goes to the spindle? Also, if I use a filter on the output side of the VFD, would this protect both the VFD and the motor?

[magicniner]

On the mains feed into the VFD.

[Neale]

Be aware that using an EMI filter can cause problems if your supply comes via a RCD, as it can cause unbalanced live/neutral currents. I have an inverter with built-in EMI filter and there is a warning in the manual about this. I personally don't use one with either of my inverters and don't have problems but it's a very situation-dependent thing. The filter is nothing to do with protecting motor or VFD - it is to prevent interference generated by the VFD going back up the line and interfering with other equipment.

[routercnc]

Whilst we are on that general topic - I have a vague recollection that inductors used as chokes can suppress high frequency and reduce interference. Is that something that can also be added on the mains side either separately or together with the filter, or is the filter based on a choke anyway.
My reason for asking is that I got very occasional false estops (even though I have a filter fitted) until I used an extension reel on the VFD and plugged it in the other side of the workshop and that sorted it out. Did I create an inductor of sorts ?(cable was mostly still wound on the reel) and if so can a commercial choke be added inside the control box before the filter to dispense with the extension reel?

[Neale]

Yes, I think that typically EMI filters do include some kind of inductor/choke - at least, the one I have with a circuit diagram on the outside is like that! I would expect to see a couple of chokes and a few strategically placed capacitors.

Couple of reasons why moving the connection to the other side of the workshop might have helped. One is that possibly the coil of cable but possibly also the increased length of cable between VFD and control box was the reason - the high-frequency noise would be reduced just by having to travel further along the cables. The other reason (depending on the kind of workshop you are in) is that the sockets on the other side are either on a separate ring main (and hence separated by a couple of MCBs, probably, which have coils in them) or even, if it's a non-domestic workshop, on different phases. My own control box and VFD plug into the same multi-way socket and I have never had noise problems, but I deliberately went for a 24V system (limit switches, e-stop, etc) rather than the more common 5V purely because it is much more intrinsically noise-resistant. Maybe 5V would have worked for me, but I wanted to frighten away the gremlins with a big stick even before they appeared! Don't actually know if my HY inverter has a built-in EMI filter or not, though.

[Voicecoil]

Anything that has lots of power switching going on in it like a VFD ought to have a mains filter as part of the design, it likely won't meet the relevant EMI regs without it. Note that a filter to stop stuff creeping OUT of the device may have a slightly different design to one aimed purely at stopping crap coming in. And yes, it's pretty certain it will contain an inductor or two along with some capacitors and if it's an input filter ideally a surge protector (VDR).

[john swift]

the EMI filter could be like this



the choke has a low impedance at the mains frequency but
a high impedance to the high frequency noise your trying to block

the capacitors are just the opposite
a high impedance at mains frequency and a low impedance to high frequency noise

so the mains supply passes through uneffected but any high frequency noise between live & neutral are short circuited by capacitors C1 & C2

any noise between the supply lines ( live & neutral) and earth will pass through capacitors C3 & C4 to be grounded

the filter may be potted in a metal enclosure

if VDR's are added to provide transient voltage protection like this



you need a fuse in the live connection as VDR's very often go short circuit when they fail


John