VFD Noise Making My Head Spin

[calexanian]

I got a call from a customer today stating that they are now using VFD's in many of their applications. This is for agricultural irrigation and we manufacture what can be viewed as a glorified sprinkler timer. From time to time these VFD's make enough noise that will interfere with our unit, and other equipment at the installation. I realize this is a common problem, but our unit has only basic protection and I was curious if somebody has come across a good quality filter for this sort of combo EMI RFI situation. The primary of our unit is 120 or 240, and at idle is less than a watt of power, and at full output is about 20 watts.

Thanks

[NiHaoMike]

Just add some EMI filtering on all lines going in and out of the unit.

[Rerouter]

RFI injected into your unit can be fixed by either using a metal box, a foil lining or a metal impregnated paint on the inside of the box

external emi is more fun as it pays to know what its actually upsetting, i would guess you cant share any documentation of the product, so if you could perhaps better explain how it is interfered with,

[NiHaoMike]

At the (relatively) low frequencies VFDs operate at, it's much more likely for interference to occur as a result of capacitive coupling to a wire connected to the unit.

[SeanB]

You probably want to have some RF chokes and capacitive filtering on all inputs and outputs. Depends on how it comes in, most likely from long cable runs capacitively coupling to the output or sensor lines or via the power supply. For sensors you will have to make LC filters on the inputs, and similarly for the outputs, along with adding VDR's to handle overvoltage events on the lines. Power supply wise a common mode choke, and capacitors to ground, like in any common mains filter, possibly with a set of inline chokes to increase rejection to supply side transients, along with more VDR's.

IIRC you supply sprinkler systems, so likely have both long wiring to valving and power, along with sensors. Sensor wiring I would recommend having both separation by at least 300mm from other cabling and using a cheap shielded cable like Cat5 shielded cable, as they likely are low voltage and you can get a direct buriable outdoor rated CAt5e shielded cable, to not increase cost too much. Valving I suggest the use of a shielded mains PVC cabling, and ground the shield to the case of the valve and to the controller.

[calexanian]

At the (relatively) low frequencies VFDs operate at, it's much more likely for interference to occur as a result of capacitive coupling to a wire connected to the unit.

I have yet to go out and actually experience it in this particular situation. I have used VFD's in the shop to see what happens and typically our products handle it well. I believe from their description the units are actually going into reset. This is an older design based on the PIC16C62, or 73. (Microchip CMOS chips not world known as being bulletproof by any means. I am currently migrating designs to flash chips but this product will not be changed due to its age)That in itself is not an issue as an occasional reset is not cause for alarm. Our units just pick right back up where they were when they come out of reset. There is nothing proprietary or special about our designs.  Generally our line introduced faults are of the over voltage, transient, or switch noise spike kind. We have a rather beefy TVS and are supplied by a split bobbin 24VCT transformer. Also we have gobs of MOV'S on the field wiring, however they are realistically ineffective against radiated noise, and MOV's basically can act as small value capacitors and our field output wires as antennas so there are many modes of conduction I can think of. First thing first though, I think a really kickass line filter is the first step. I started doing a literature print spree of all the various brands before I left work and Monday I will pick it back up.

[calexanian]

BTW we use powder coated mild steel enclosures.