VFD questions

[TorontoBuilder]

I don't know whether to say "what a great discussion"

or

"He opens the door, takes one look inside and closes the door again and backs away slowly..."

 

[BaitMaster]

I did the @Susquatch quip because I agreed with your take of the HP/torque/RPM relationship made in previous comments.

Your point about the SLV or closed loop controlled drives is of course true as far as rotor RPM is concerned. My only point about the Slip topic is that Slip as a physical principal (being expressed as rotor RPM < EMF RPM) is always present in an AC motor with a non externally excited rotor.

So when it’s said that the motor doesn’t slow down, I mean maybe the rotor RPM doesn’t change, but if the drive were not compensating for that, it would slow down, and the slip would increase, and it is often imperceptible.

Even if the drive compensates for the slip, it does increase with increased load. That was my only point.

I’ve had a VFD on a 1970s drill press (3/4 hp, 6 pole motor is not VFD rated) for ~10 years. The VFD is an Allen Bradley, vintage the late ‘90s. Per comments above it is a low duty cycle. I had one job of tapping hundreds of holes and set the drive up for rapid reversal on a foot controlled switch. The torque for actual job (3mm spiral point tap on a through hole) was minor - the endless fwd - rev - fwd - rev - fwd …….. (you get the point) at about 25Hz (on my lowest belt speed) should have been a workout for the drive/motor. All good, no problem.

Advise I was given was to keep the motor leads as short as practical (the longer the motor leads, the greater the potential for harmonic issues?).

Increased lead length can lead to cable / motor impedance mismatch and cause transient high voltages capable of damaging motor insulation. This typically would happen on cables longer then 5m and would only cause damage on motors where the insulation rating is close to the peak to peak value of the source.

So at 230v on a dual 230/460 motor, you are almost under no risk of this causing issues unless your cable is crazy long (100m+).

On 480v applications…. Ya. Definitely could be a legitimate problem for you. With P2P values of 1350 ish normal…. Insulation only rated at 1500v….. not very much does it take to degrade insulation.

1500v insulation and 650 ish volt P2P you have a lot of wiggle room.

 

[Susquatch]

So at 230v on a dual 230/460 motor, you are almost under no risk of this causing issues unless your cable is crazy long (100m+).

On 480v applications…. Ya. Definitely could be a legitimate problem for you. With P2P values of 1350 ish normal…. Insulation only rated at 1500v….. not very much does it take to degrade insulation.

1500v insulation and 650 ish volt P2P you have a lot of wiggle room.

Good comments all around. And we wonder why people are confused..... LOL.

Regarding the high voltage problems, there are really two issues. Beside the advantages of a VFD Rated motor that can handle them better, the output spikes and insulation issues can be dealt with through special insulated VFD cable and with an output filter near the VFD. I've been trying to source some of this cable for some time now without any success. The motor and VFD Suppliers specify it but don't sell it. They do sell output filters, but they are mega expensive. Sooner or later I will find the cable and I will make a filter. Normally, neither one is required for short runs (mine are all less than 10ft) but I don't like the spikes I see on my oscilloscope so I want to deal with them anyway. I'm just not willing to spend hundreds of dollars to do it. If I end up having to buy a whole roll of cable (which looks likely) I plan to make the remaining cable available to members at my cost per foot plus shipping. This is project 42g for me. It is still in the works. I'm not really terribly concerned about it. I just don't like the spikes I see on my output. Sometimes I wish I just never looked!

The other problem is feedback on the supply lines affecting other equipment in the shop. This is actually a bigger problem but it's a LOT easier to deal with. Line filters and ferrite cores are readily available at low cost.

 

[slow-poke]

but I don't like the spikes I see on my oscilloscope so I want to deal with them anyway.

Any chance you screen captured your oscilloscope spikes?

scope and or probe bandwidth?

I would be curious to see what they look like.

I'm guessing you have a high voltage differential probe(s), I would not want to risk my scope with a non isolated probe with > 600Vp-p

 

[Susquatch]

Any chance you screen captured your oscilloscope spikes?

scope and or probe bandwidth?

I would be curious to see what they look like.

I'm guessing you have a high voltage differential probe(s), I would not want to risk my scope with a non isolated probe with > 600Vp-p

No, I didn't. I'm not even sure that I can. All of my Scopes are CRT based Tectronix units. My digital storage scope might be a flat screen but I'm not sure of that. They are all BIG LONG boxes which suggests a CRT. If so, I don't think the screen will photograph well with a digital camera.

Yes, I used high voltage differential Probes.

When I get around to making filters, I'll try to take some photos.

I could prolly draw something for you from memory though. In the meantime, maybe a description would help. They look a lot like inductive spikes generated by a coil turning off. It's the kind of spikes you would kill with a diode in a relay. The VFD generates a DC voltage internally and then chops that to a duty cycle based high frequency (commonly called its carrier frequency). This chopping process generates the high voltage spikes on each edge as the electronic switches turn off. On the scope, it looks like duty cycle square waves at the carrier frequency with tall spikes on the edges. Spikes like that are common in almost all square waves because there is always a little reluctance in any circuit, but these are much taller than I expected.

I spoke to a number of experts at the motor and VFD manufacturers and they knew exactly what I was talking about but didn't feel it was really an issue without long cables which basically act like inductors and amplify the effect - think water hammer.

One of the other things I thought of later and want to try, but didn't at the time, was to see how carrier frequency affected the spikes.

Edit Comments - somehow or another this post got all garbled up. I have fixed it as best I can.

 

[Susquatch]

scope and or probe bandwidth?

Sorry, I forgot to answer the scope question. To be honest, I don't remember which scope I used. I have a half dozen.

I'd guess I prolly used my
Tektronix 2465B 400MHz 4 channel scope because I was interested in looking at the subcomponents of a 16KHz waveform.

But it's also possible I used my Tektronix 2430A 150MHz Digital Storage Scope so I could freeze the display.

I just don't remember which one, or even if it was one of those or a different one.

 

[TorontoBuilder]

don't like the spikes I see on my oscilloscope...

Sometimes I wish I just never looked!

This type of regret may be why I don't own an oscilloscope

 

[slow-poke]

I would imagine that there would be some variability from one VFD to another especially newer vs. older, and better vs. cheaper. I read that the manufacturers are softening the edges on the newer and better quality VFDs

I'm most curious about the peak voltages.

I keep my VFD leads very short (2-3'), hopefully that helps.

 

[Susquatch]

This type of regret may be why I don't own an oscilloscope

Cats have 9 lives, but curiosity is what kills them.

One way for everyone to look at it is to keep in mind that there are tens of thousands of VFD's in use around the globe. Very very very few are filtered on the output side. Yet the motors do quite well anyway. In fact, I have yet to find even one user who has done it. I asked how many they sell and the answer was virtually none in small shop settings. They are mostly purchased by big shops with long cables.

The supply line filters are a whole nuther kettle of fish!

 

[TorontoBuilder]

Cats have 9 lives, but curiosity is what kills them.

One way for everyone to look at it is to keep in mind that there are tens of thousands of VFD's in use around the globe. Very very very few are filtered on the output side. Yet the motors do quite well anyway. In fact, I have yet to find even one user who has done it. I asked how many they sell and the answer was virtually none in small shop settings. They are mostly purchased by big shops with long cables.

The supply line filters are a whole nuther kettle of fish!

I had two different filters on the supply side of my old set up.

None on the Mill VDF and no issues related to it at all.

 

[Susquatch]

I'm most curious about the peak voltages.

I found my info.

The spikes were over 1000 Volts but extremely short duration. I didn't measure the time length.

I did use my good 400MHz 2465b scope.

The filters recommended by the VFD and motor companies are called reactors. If anyone knows a low cost supplier or has a low cost home design, I'm interested.

 

[Susquatch]

I had two different filters on the supply side of my old set up.

Ya, supply side is easy to deal with.

 

[Former Member]

Simple explanation.

One of the big things I learned when I sold the first VFD's was how high fast (how many steps) the DC was switch to generate the AC wave played a big role in interference. Generally lower quality (usually cheaper) units use a very low switch rate which as you decrease the frequency you get some feed back from the motors as they see it. The second is noise created by the switching that gets back into the mains which is why they recommend filters and isolation.

My unit even though is a high end unit does this and only effects the radio in close proximity. Cell and WiFi are not effected.

Remember all wires act as random length wire antennas and if you hit a harmonic right you'll have issues you need to address. Ie if it causes you issues address it otherwise leave well enough alone.

 

[slow-poke]

I realize that using a reactor is the normal "canned" solution. RC snubber would likely take the sting out of those spikes. I'm assuming your running at 240V, and I guess that's why the general consensus is that this not an issue with short leads and 240V. At higher voltages having VFD rated insulation seems like a necessity not a luxury.

 

[BaitMaster]

This one is basically right off of the drive terminals, before the cable.

 

[BaitMaster]

After the cable, at the motor terminals….. P2P value = 2.08 kV…..
you can see the spikes on top of the square waves. Most inverter duty 480v insulation is 1600v…

 

[BaitMaster]

This is motor terminals after applying a dV/dt style filter, at the drive, before the cable. To the same Motor that had the 2.08kV P2P.
Cable length approximately 60 feet.

 

[BaitMaster]

Motor terminals. 80’ of cable. 7.5hp 480v drive rigged up on the bench. Crappy picture but it was fluctuating from 2.05-2.3kV P2P.

 

[BaitMaster]

This is after the 80’ of cable…. After applying a sine-wave style filter. Yes. Satisfying.

After our researching of this and acquiring “drive rated” cable, no difference of the waveform was noticed by us when using special VFD cable. I’m sure it makes a difference in interference with neighbouring electronic signals, radios, encoders, etc. If you press the cable manufacturers and dig, 1000v rated Teck type cable is approved for use in 480v drive applications. No difference has been noticed by us at all between the special cable and 1000v teck properly used.

 

[BaitMaster]

3% impedance load reactors don’t make much of a difference in the readings. 5% impedance load reactors have very similar results to the dv/dt style filter image shown, which admittedly is a solid improvement.

Higher switching frequency marginally makes the spikes worse, but a decrease in audible motor noise is noticed. Lower switching frequency the spikes improve, but motor has a louder squeal.

The only “wow” factor solution we’ve seen in real life has been the sine wave filter. Completely perfect sine wave. No audible motor noise. Saves puppies from fires and nurses them back to health.

Different manufacturers of VFD’s have VERY similar waveforms. None have significantly outperformed any others, as far as I’ve seen.

Drives are cool, yes….. BUT…

There is no free lunch.

Hope this helps fellas.