VFD worries...

Hi all,
So I’ve begun to buy and install some very basic little VFD’s, the KB Electronics KBMA-24D, very simple and easy to use. I’ve installed one on my J-head Bport, a-ok, never going back. The first one I bought, tho, well, I installed it on my old Walker-Turner 14” bandsaw; turned it on, got a rather distinct high-freq whine out of it, ran the motor a few 10s of seconds, and then...it croaked. Says it’s in short circuit sensing condition, even with nothing at all connected to it. Sigh. The eBay seller just sent me another (didn’t even want the dead one back, it appears) and that’s the one happily (and quietly) chugging away on the mill. So this THIRD unit, put it on the saw, again, and it’s whining loudly but doesn’t seem to be failing after a good few minutes running, no load yet. It’s just a plain-Jane 3ph motor on the saw, I find it hard to imagine it’s somehow blowing these things up (and indeed, the 1st unit may have simply been an “infant mortality”, ok; the seller certainly seemed to think so), but, ..what do you folks think? How come the 1960’s era 3/4 hp pancake motor on the Bport is quiet and this much newer 1 hp 3ph saw motor is whining up fairly close to my 60 yr old ears range? Is this indicative of it gonna fail soon, too? Thanks—C. Hale, Lafayette CO

—update; the KB VFD is installed on the saw now and seems a-ok, albeit whining compared to the one on my mill. But I’m seeing out there that the whine is characteristic of many VFD setups and isn’t indicative of the thing being stressed or failing etc. (so I really do believe that first VFD unit was an “infant mortality” case, now). The KB unit’s carrier freq can’t be easily directly adjusted (I get it now, that adjusting the freq higher gets rid of the whine), but there are a few comments out on the web about adjusting KB units’ “slip compensation” and “overload” (I believe it was) trimpots in these units can attack “whine”. So I may look into that. Of course, the saw makes enough noise while running, that it’s not really that big a deal—just mainly irks me that the saw unit whines, and the mill unit doesn’t...?

The whining sound is from what’s called “magnetostriction” of the laminated steel plates that make up the stator of your motor due to the high speed pulses of DC coming from the VFD that are “tricking” the motor into thinking it is seeing AC. The pulses are fired at what’s called a “carrier frequency”, the pulse rate if you will, and is typically set from the factory at between 1.4 and 4kHz.

If your drive allows it, you can increase that carrier frequency and if you get it above 10kHz it will be above the range of most human hearing (but your dog will not hang around your shop). The problem is that you are making the transistors switch on and off more often, and every on-off transition has what are called “switching losses” in the transistor, which is why VFDs have heat sinks. The heat sinks are typically optimized for the carrier frequency that the drive comes with from the factory, so it you increase that, you can overheat the drive and kill it. So typically a manufacturer will give you a “de-rate factor” with which you oversize the drive to compensate. So for example if you have a 3HP motor but want to run at 15kHz cf, you must buy a 5HP VFD. If you have already bought the VFD sized to the motor, you may be stuck with not being able to increase the cf by very much without putting the VFD at risk. Some low cost drives don’t even allow you to change the cf at all though fir this reason.

But increasing the cf ALSO can increase the potential for creating other problems with the use of VFDs such as motor insulation damage from standing wave generation and motor bearing damage from EDM. The best thing is to learn to live with the whine. If you are married, you probably already know how to do that...

Ha, yes, now that you mention it I have already been living with some so-far workable level of whine in my life...anyhow, great info, thanks. Yes, my 61 yr old ears, which evidently haven’t been completely burned out yet from a lifetime of rock and roll, do conclude that the whine is coming from the motor, not the VFD (they’re pretty close to each other but one can definitely tell all right). My KB manual indicates that the drive is “8/16 kHz” carrier freq, and, in this low-cost guy’s case ($100 new eBay, ‘make an offer’j, not user-selectable it seems (and yet it shows two values for cf...). I’ve got a spectrum analyzer app on my phone that makes it very clear that the drive is set to 8 kHz. I think I’ll try pulsing the factory and see what they say. And as you say, maybe just live with it since 8 kHz is probably pretty benign.

Jraef said:

The whining sound is from what’s called “magnetostriction” of the laminated steel plates that make up the stator of your motor due to the high speed pulses of DC coming from the VFD that are “tricking” the motor into thinking it is seeing AC. The pulses are fired at what’s called a “carrier frequency”, the pulse rate if you will, and is typically set from the factory at between 1.4 and 4kHz.

If your drive allows it, you can increase that carrier frequency and if you get it above 10kHz it will be above the range of most human hearing (but your dog will not hang around your shop). The problem is that you are making the transistors switch on and off more often, and every on-off transition has what are called “switching losses” in the transistor, which is why VFDs have heat sinks. The heat sinks are typically optimized for the carrier frequency that the drive comes with from the factory, so it you increase that, you can overheat the drive and kill it. So typically a manufacturer will give you a “de-rate factor” with which you oversize the drive to compensate. So for example if you have a 3HP motor but want to run at 15kHz cf, you must buy a 5HP VFD. If you have already bought the VFD sized to the motor, you may be stuck with not being able to increase the cf by very much without putting the VFD at risk. Some low cost drives don’t even allow you to change the cf at all though fir this reason.

But increasing the cf ALSO can increase the potential for creating other problems with the use of VFDs such as motor insulation damage from standing wave generation and motor bearing damage from EDM. The best thing is to learn to live with the whine. If you are married, you probably already know how to do that...

Click to expand...

The Bridgeport likely has fully potted windings, your standard dipped motor has half the winding saturated with resin, the upper half partly saturated. (Thermally cured polyurethane baked in an oven)

So i expect most of the noise to be from the windings, not the core, otherwise the bridgeport would be similar (though its possible the pancake core which has substantially different mechanical properties is much better dampened and thus makes far less noise)

Very interesting about the likely different construction of the two motors, thanks. The 1hp motor for the saw is a pretty modern, low-cost rebuilt GE motor that I got really cheap at a motor rebuilder’s place in Denver. So, I could see how this difference in whine could show up, for sure. (I did run my mill with its identical KB VFD, and my phone app analyzer could sure enough see a bit of 8 kHz coming out of it, too. I can’t hear it, but it could). Question, are there sensible filters that can be put in between the VFD and motor that I guess might smooth out the sharp frequencies (or discontinuities, maybe more correctly?) in the synthesized sine wave that would keep the VFD from exciting those windings in the motor?

chale4 said:

Very interesting about the likely different construction of the two motors, thanks. The 1hp motor for the saw is a pretty modern, low-cost rebuilt GE motor that I got really cheap at a motor rebuilder’s place in Denver. So, I could see how this difference in whine could show up, for sure. (I did run my mill with its identical KB VFD, and my phone app analyzer could sure enough see a bit of 8 kHz coming out of it, too. I can’t hear it, but it could). Question, are there sensible filters that can be put in between the VFD and motor that I guess might smooth out the sharp frequencies (or discontinuities, maybe more correctly?) in the synthesized sine wave that would keep the VFD from exciting those windings in the motor?

Click to expand...

Surely. Common problem.

Basic harf-assed "3%" to 5% "dV/dT" filter would do yah. "Sine Wave" grade filter would be overkill, but yah might find one cheaper, used. Either one takes away very little power but a tad more MONEY.

The most modern of "inverter duty" motors benefit. OLDER motors it can be more critical - good longevity or early-death critical.

Best advice is to use K-B's advice. Or whomever-other VFD maker's advice when using their units.

Even so, it isn't rocket insemination to get close enough.

chale4 said:

Very interesting about the likely different construction of the two motors, thanks. The 1hp motor for the saw is a pretty modern, low-cost rebuilt GE motor that I got really cheap at a motor rebuilder’s place in Denver. So, I could see how this difference in whine could show up, for sure. (I did run my mill with its identical KB VFD, and my phone app analyzer could sure enough see a bit of 8 kHz coming out of it, too. I can’t hear it, but it could). Question, are there sensible filters that can be put in between the VFD and motor that I guess might smooth out the sharp frequencies (or discontinuities, maybe more correctly?) in the synthesized sine wave that would keep the VFD from exciting those windings in the motor?

Click to expand...

A rebuilt motor, if that means it was rewound may not be dipped at all and the windings are merely laced. you could take your chances and dribble polyurethane varnish into the windings (takes a while to cure fully) or epoxy. i would mix up some epoxy that has a set time of an hour or two so you have plenty of time for it to wick into the windings before it sets.

if you want to get rid of the noise completely an off the shelf 3 or 5% reactor after the vfd, and put delta connected capacitors (5uF per hp is about all you need) at the motor. the motor will be silent and the reactor will make noise.. perhaps more than the motor did, but you can bury the reactor in a bucket of sand, or put it inside a stiff, box inside a box, etc to isolate the noise.

Common and common.
Buying a qaulity vfd in the first place and you usually will not have these issues at all, and most quality VFDs have built in filters too, VFD's without seems to be a CHinese/american thing generally.
Quite frequent work a place where there is 500+ VFD in anything from 1 to 2500 amps, never heard this issue even once.
Neither have i anywhere else there is used quality drives.
And the only VFD i've encountered in need of external filters is shabby ones from Allen Bradley

johansen said:

i would mix up some epoxy that has a set time of an hour or two so you have plenty of time for it to wick into the windings before it sets.

Click to expand...

Yah. sure. You would.

WTH ...That could lead to some seriously amusing side-effects...

More better to do what General Electric had its laboratories do. Solve the problem.

The solution has been a separate company for "a while" now. A long while.

Glyptal. Classical "red" 1201 // 1201B are alkyds. Instructions on the tin, and/or the 'net, Basically many thin treatments rather than fewer thick ones so as to penetrate, Xylol thinner. Not hard.

Vacuum chamber recommended, then bake. That's harder to DIY, best hired done as it is a major part of what motor shops DO:

Glyptal Inc.

Nicest part is that unlike general-purpose epoxy? Glyptal won't melt and run when the motor heats-up and glue the rotor to the stator as amateur-hour "solutions" are wont to do.

Guess that spoils the joke?

where is the evidence your

Alkyd - Wikipedia (s)

will stick to nylon 6 better than epoxy would?

Polyamide-imide - Wikipedia
Nylon 6 - Wikipedia

most but not all high temperature magnet wire has a coating of nylon over the top of the polyaimide-imide insulation, for the purpose of reducing friction.

if you really want to get into the details of it.. the alkyd compounds (might as well include polyurethane and polyester compounds) are used for dipping motor windings because they have a lower stiffness than epoxy does, and are far less brittle, and can be thermally cured rather than two part mixtures so you can run a continuous operation. (or in the case of your glorified glyptal, oxygen cured through polymerization of various oils)

Allegedly alliance scroll compressors are made in the same factory copeland compressors are.. the difference is alliance compressors have epoxy dipped windings. copeland compressors are laced and not dipped at all. why epoxy and not a thermosetting polymer i don't know, but they do run at ~200F and have to deal with several different types of refrigerant, both mineral and poe oil. so perhaps epoxy is the only thing that will hold up.

To be fair to thermite, it does not have to "stick", it just has to hold and prevent from vibrating, which it can do without necessarily "sticking". It might do that OK. usually it is the "end turns" that are the loosest, the part in the slots is not generally the issue.

But, the problem is not too much of one, a bit of noise. ignore it, filter it out, getting into the motor is probably not worth it. And even with glyptal, the penetration without doing vacuum is not likely to be that great, although thinned it should be a good deal better than most epoxy, which tends to be thicker.

main thing is to be sure that the wire insulation will not be damaged by whatever solvent / thinner is used. That might be a reason for epoxy, as it does not usually use a solvent as thinner, and may be less likely to bother insulation if it happens not to have nylon coating..

JST said:

To be fair to thermite, it does not have to "stick", it just has to hold and prevent from vibrating, which it can do without necessarily "sticking". It might do that OK. usually it is the "end turns" that are the loosest, the part in the slots is not generally the issue.

But, the problem is not too much of one, a bit of noise. ignore it, filter it out, getting into the motor is probably not worth it. And even with glyptal, the penetration without doing vacuum is not likely to be that great, although thinned it should be a good deal better than most epoxy, which tends to be thicker.

main thing is to be sure that the wire insulation will not be damaged by whatever solvent / thinner is used. That might be a reason for epoxy, as it does not usually use a solvent as thinner, and may be less likely to bother insulation if it happens not to have nylon coating..

Click to expand...

Thinned Glyptal actually "wicks" like a hungry bandit. The Vacuum chamber is more for CERTAINTY in more predictable time - and less of it - than futzing with many, coats over what can become several days.

The slower path is OK for an impoverished retiree/hobbyist. If impatient? Bill P built hisself a chamber - but he had more than just the one task for it.

Or third-word, where labour and time are cheap, options fewer, as with the several rebuilds done "all manual" to our goods in 'nam by expat Korean techs at the PA&E Fire Department. Rock solid results. Not their first rodeo.

DIY by hand application is a waste of fully-burdened shop time and a detriment to quick turnaround against deadlines for a pro rebuild shop, so they WILL have the vacuum chamber and bake oven.

Also alternative brands to Glyptal, which good as it is, does command a non-trivial price off the back of "legend".

Side note, but thanks for the reminder as to why I did not, and will not, use Copeland OR Alliance HVAC components.

And "Oh, BTW...' WHEN and WHERE an Epoxy IS the/a right answer .. Glyptal and competitors sell those, too. But spec'ed for particular motor/generator situations. Few of us have time to learn what they already know about that, nor what the research cost in money, ruint goods, and mess to be cleaned-up, either.

JST said:

main thing is to be sure that the wire insulation will not be damaged by whatever solvent / thinner is used. That might be a reason for epoxy, as it does not usually use a solvent as thinner, and may be less likely to bother insulation if it happens not to have nylon coating..

Click to expand...

that is one argument for something other than epoxy because you want it to stick. not that i want the potting compound to dissolve the "enamel" in minutes but if it can at least partly attack the surface perhaps it will do better than a few hundred psi worth of Van der Waals attractions.

i know from experience that polyurethane varnish which is about as thick as motor oil, in the space of hardly a minute will soak its way into a motor winding completely. i have no reason to believe epoxy which is say, 100 times thicker, wouldn't do the same thing if it had at least an hour before it set.


i also know from experience that most of the time i can, with judicious care, pull the motor windings out of the thermosetting compound they were baked in without ripping the insulation off the windings. this is across a half dozen different motors. which means the thermosetting compound didn't stick nearly as well to the insulation as the insulation did to the copper. big deal, but there is a tremendous difference between 5000 psi adhesion between copper to polyaimide, and 500 psi from nylon to polyester. not that the 500 psi isn't good enough, but if 5 minute epoxy is just as good then there is no reason to spend 200$ for a gallon of glyptol that i'm never going to use. i sill think there is a good argument to use something other than epoxy on the basis of stiffness alone. you want something that is less stiff than the nylon and the polyaimide, and that is a good argument for what i suggested already which is polyurethane.

the older the motor the more likely its a 50/50 shot of the motor insulation approaching the tensile strength of the potting compound. older still and the potting chemicals are stronger then the wire insulation, you simply look at the shellac insulated wires and it crumbles. But then there is the modern epoxy dipped windings where no, you cannot pull the wire out of it without breaking the wire.

Well my point is that if there is rigid "stuff" all around the wire, even if it does not "stick", there is not going to be much if any vibration, the wire has nowhere to go, it can't move. if epoxy will do it, fine, all it has to do is fill up all the space, and not soften up the existing varnish etc insulation enough to allow a "cut through" in a tight spot, which would ruin the motor. That's if wire vibration is the issue, it may not be.

if the wire can pull out of the potting that may not be an issue with good fill. but if you are depending on just a coating to stick the wires together, then yes, you need t to "stick", because it is not going to surround the wire and make the winding into a "block" of solid material.

The heat cure varnishes are sort of "in between" as to sticking, and do tend to stay in the innermost parts, although even they are not always a totally solid block after curing.. they drain out to some degree at least, depending on wire size in the winding, and how good penetration is..

As for glyptal... if thinned enough, it wicks. But no guarantee of no "holidays" down in the windings. Once it fills, if there are any gaps, they can never be fixed, because the rest of the winding is filled up and blocks access. Epoxy etc would be the same unless vacuum is used, which should work with basically any filler material..

And watch out for the solvent used to thin it, with regard to softening of the wire insulating varnish and potential creation of internal shorts.

it may not need to be very good to cut noise.... a lot of the noise does not come from the windings anyhow, but from the core... some vibration and buzzing between laminations, which varnish fixes, but some you may never get rid of due to "magnetostriction". The filter works by preventing the HF "signal" from getting to the winding and core in the first place.... a "sine filter". Then it does not matter what the exact cause is.