Losses in an AC motor: some hint as to differences between inverter and non-inverter motors

[rabler]

I’ve let myself get dragged into a few discussions on VFDs, and I see a bit of confusion on why inverter rated motors matter.

To start with, motors all generate some amount of heat. That heat is wasted energy, electricity that could have been turned into motion, but instead just generated heat. If you get nothing else out of this, just remember heat is wasted energy as far as motors are concerned. (If you’re trying to heat you room, that’s another story). Heat is also what eventually kills motors. Bearings wear faster when hot. Lubricants break down faster. Insulation fails faster. Copper wire corrodes faster. Most of these aren’t instantly catastrophic if not taken too far. Like running your car hard, or pulling a heavy trailer, your car’s life will suffer, but unless you get wild, it won’t die immediately. How badly a motor will suffer is almost like predicting how long your car will last. Manufacture, make, model, road conditions, and luck of the draw all play some part in that.


Of course temperature is the actual killer above. Sure, more heat increases the temperature, but fans, and lower ambient temperature will reduce the temperature rise. Motors are designed to not overheat by adding enough fan air movement to keep them within reasonable temperature. Just like a car’s radiator. But if the motor is operated outside it’s designed parameters it risks overheating.

Since we’re often using general purpose motors in weird ways, we have to have some understanding of what causes overheating to make a good guess at whether our motor will work well for a given application. Of course, if you’re in the business of using motors to make high efficiency systems like HVAC systems, you also want to get the most motion for your dollar of electricity. The cost of the motor may be significantly less than the lifetime power costs. ( This was the real motivation for early VFDs, saving electricity). As home, or even professional machinists, electrical costs are rarely a big factor in how we use or set up our equipment.

Many motors actually give you an efficiency rating. That rating is at optimum conditions, sort of like MPG in your car. Usually its at 60Hz, and about 75% of rated power. Generally anything else is going to be less efficient for that motor. But it is not easy to determine how much less efficient. Efficiency is simply the power creating motion divided by the total power used. Heat is pretty close to total power used minus power of motion. (air movement from the fan and noise created also use a small amount of power.). What this means is if the motor is generating less power of motion, than for the same efficiency it is generating less heat. Or,
heat = (1-efficiency) * actual horsepower output. Unfortunately efficiency does change a bit with horsepower output, but not drastically, so you can get dome idea from this of whats going. And it’s pretty intuitive, that motor is going to get hotter if it’s cranking out more power.

With all that said, here are some things that waste energy in a motor:
The fan cooling the motor.
Friction in the bearing.
Resistance of all that long thin wire.
Eddy currents in the material of the electromagnets
Magnetic hysteresis and saturation
Air gap between stator and rotor

Yikes, what a list! I’m sure I’ve missed a few too.


The first two are actually fairly trivial. But the fan plays an important role in getting rid of heat. Fans are noisy so generally an oversized fan is avoided for that reason. And increasing airflow also means a bigger motor. If a motor is always going to run at good efficiency we can optimize that fan and airflow. And of course fan speed matters if we start using a VFD. Bearing friction depends on rpm and load on the bearings. But they are pretty good, which is why we use them.

One way we can shortcut all of this loss stuff is just make some measurement.
Power in is easy to measure electrically. P = Volts * Current. But that only works for DC.
For three phase, P = Volts * current * sqrt(3), if the current and power are in phase. If not in phase, which means the sine waves are shifted left/right of each other, then
P=V*I*sqrt(3)*cos(c). C is the amount of shift between them, and is usually written as the greek letter phi. I just don’t know how to get that greek letter while typing this on my ipad . Phi is also related to what is called power factor.

Usually we know our voltage. 230V, 120V, etc. We match our motor voltage and our utility voltage, hopefully.
An ammeter will measure current. But that damn phi gets in the way. You need an oscilloscope or other fancy equipment to measure that. One way this shows up: most AC motors run a fairly high idle current. You can measure it with a clamp ammeter. Often 30 - 40% of full load current, which is called FLA (full load amps) in some motors and VFDs. So we’re using 1/3 of the electricity and getting zero useful motion? Well yes. But the motor at idle is basically borrowing that current and then sending it back to the power company with every 1/2 cycle of your 60Hz. It is acting mostly as a straight inductor, which doesn’t use real power. It uses reactive power, or as the engineers sometimes call it, imaginary power. That current still generates loss in the power distribution system though, so the power company isn’t keen on that borrowed power. Industrial customers generally pay a fee for that, homeowners typically get away “interest free” on that very short term borrowing. As the motor goes from just spinning with no load, the current will increase, AND the phase angle will reduce. As that angle gets less, cos(c) gets closer to 1. Borrowing less power and actually using it to generate useful motion.

Ok, enough for one night of single finger ipad typing. I will take a crack at more tomorrow.

 

[Ulma Doctor]

another subject not touched upon is the voltage and frequency of the pulses supplied from the VFD
old style motors were designed to run on 60Hz and a given AC input, they do not get that at all times from a VFD
since the frequency and voltage output vary with operator input, high performance windings are installed in inverter rated motors to operate under varying conditions. the windings do not heat up as fast as in a normal induction motor as a result.
it is generally safe to operate normal induction motors from 30 to 90 Hz, for short periods of time
as well as from 50 to 70 Hz for even longer periods of time ( nearly indefinitely under very light loads for most quality non-inverter rated motors)

 

[rabler]

another subject not touched upon is the voltage and frequency of the pulses supplied from the VFD
old style motors were designed to run on 60Hz and a given AC input, they do not get that at all times from a VFD
since the frequency and voltage output vary with operator input, high performance windings are installed in inverter rated motors to operate under varying conditions. the windings do not heat up as fast as in a normal induction motor as a result.
it is generally safe to operate normal induction motors from 30 to 90 Hz, for short periods of time
as well as from 50 to 70 Hz for even longer periods of time ( nearly indefinitely under very light loads for most quality non-inverter rated motors)

I’m hoping to work my way up. Get the motor basics covered first before going into VFDs and the issues there.

 

[rabler]

Hopefully you've seen a simple electromagnet, wrap some wire around a nail and hook it up to a battery, and presto, a magnet. What makes that magnet stronger? From an electrical side, you want more current, and more wraps of wire. More wraps of wire is easy, just keep looping wire around the nail. Just make sure you don't reverse the direction of the loops. But as you make the wire longer, the resistance goes up. As the resistance goes up, more energy is lost as heat in the wire. And as the resistance goes up, it takes more voltage to push that current through the wire. You can overcome that by using bigger diameter (larger gauge) wire, but then your magnet gets awkward and bulky, and costs more (copper is getting pricey). You can get rid of some of that bulk by using thinner insulation, as long as you don't run the voltage up too high.

That iron nail is a magnetically permeable material that does two things. It amplifies the magnetic field, and it directs the magnetic field. Think horseshoe magnet vs bar magnet. But you can only squeeze so much magnetism into a nail before you need a bigger nail. If you change out that battery for an AC power source, an additional problem happens in that nail, each time the current switches direction, the iron in the nail needs to re-arrange it's electrons to align with that new field induced by the current. And that re-alignment causes some heat to be generated in the iron due to magnetic hysteresis. Higher frequency causes more frequent realignment, and therefore more heat.

To further complicate things, currents induce magnetism. But magnetism in turn causes currents. Gross simplification here to avoid calculus and vectors and such. With AC current, the magnetism in our nail will cause currents inside the nail. They're called eddy currents because they are little spinning circles like eddies in a river. Those currents experience resistance in the iron, which in turn causes ... heat. And higher AC frequency causes more currents, and thus more heat. (Sound familiar). Eddy currents are reduced by layering insulators throughout the iron that makes up the core of a motors magnets. Higher frequency AC power requires more layers (closer spaced) to reduce the eddy currents.

So motor heat goes up, which means efficiency goes down, as frequency increases. Again, design can overcome this. But more copper, more laminations translate to bigger, more expensive motors. And we'll all naively buy the cheapest motor that meets our HP and RPM needs, right?

 

[rabler]

Last big detail to talk about before moving into hooking up our motor to a VFD, is the actual mechanical load. A 1 HP motor turning at 1800 RPM CAN generate around 2.9 foot-lbs of torque. If it is spinning with nothing attached, it is essentially generating zero torque. The motor generates the amount of torque that it needs to maintain that it's RPM, up to it's rated horsepower. Probably even over that horsepower by a pretty good margin. But as the torque goes up, so does the amount of energy going into heat. Motors actually "slip" a bit, with more torque required, it'll drop off below our 1800RPM, but typically only by 40-60 RPM, depending on the motor.

So your motor can run at rated power and steadily generate a fair amount of heat. Or it could mostly run at only a little bit of needed power, and only rarely need full power output. In that case it can readily burst to quite a bit more than full power briefly as that brief burst of power, and the resulting heat won't last long enough to over-heat the motor. (I alluded to this above.)

 

[rabler]

Now, lets throw in a VFD. Even if your VFD is set to 60Hz, it actually generates a carrier wave that is used to approximate that 60Hz. That carrier wave has a frequency, called the carrier frequency. Typically carrier frequencies are around 3kHz - 15kHz, but different VFDs have different ranges. The carrier frequency is a bit like listening to music on the radio (actually sent over the airwaves radio. Does anyone still use them?) The carrier frequency is a bit like broadcast frequency. The music is "modulated" onto the carrier. For our VFD case, we're not broadcasting, we'll use want actual wires. But there are two components to what is sent over the wire from the VFD, the digital carrier and the actual operating frequency for the motor. When we change the speed of the motor, the carrier frequency doesn't change, just the signal modulated onto it.

Engineers use something called Fourier analysis, either Fourier series or Fourier transforms, to analyze this. We won't go into that, but you need to know that there are two frequencies being sent from the VFD to the motor. And that means that some of the energy from the VFD is put into the carrier, although hopefully most of it goes into the main signal.

Using that Fourier analysis, power at that carrier frequency is basically wasted energy. When it hits the motor, that energy gets converted to heat. Even inverter motors are not designed to do anything useful with energy at the carrier frequency. Obviously we want most of our power at the main frequency, but that carrier frequency requires some small amount of energy.

Why do VFD's use this "carrier frequency"? Well basically the only efficient way to deal with electricity is to turn it all that way on, or all the way off. You can shape it by using resistance, but resistance is purely converting electricity to heat. Each time the power gets turned on or off at the rate dictated by the carrier frequency, the IGBT transistors in the VFD also generate a bit of heat. Higher carrier frequency means more on/off means more heat. They don't care about the main frequency really. So ideally we set our carrier frequency as low as possible for efficiency. The problem with that is the carrier frequency is in the human hearing range, at least the low end is. Most people want to set their VFDs around 10kHz to 12kHz though, because lower frequencies are audible, usually as an annoying buzz or whine which can be quite loud.

So our motor, operating on a VFD, is already somewhat less efficient than just being plugged straight into utility 3phase. Running just a bit hotter than otherwise. How much hotter depends on how well the motor handles that carrier, and how much heat tolerance and extra cooling it has.

Next I'll talk about what happens when a VFD operates with the main frequency at other than 60Hz.

 

[Karl_T]

Say, i tried running my new 10EE with 10Hp 1800 RPM VFD rated motor WAY outside normal speed range - 3Hz. made several cuts threading a 2" pipe thread. Put hand on outside of motor, it was only just warm. I am sure efficiency was under 10% here, but do I care? or any chance this will hurt the motor?

Same thing at other end, ran it up to 200 Hz and made a skin cut on 1/4" rod. Motor was just warm. Again, efficiency has to be terrible, but do i care?

 

[markba633csi]

Probably more efficient than you think Karl. Cool temperature means happy motor, no matter what
-Mark

 

[rabler]

Say, i tried running my new 10EE with 10Hp 1800 RPM VFD rated motor WAY outside normal speed range - 3Hz. made several cuts threading a 2" pipe thread. Put hand on outside of motor, it was only just warm. I am sure efficiency was under 10% here, but do I care? or any chance this will hurt the motor?

Same thing at other end, ran it up to 200 Hz and made a skin cut on 1/4" rod. Motor was just warm. Again, efficiency has to be terrible, but do i care?

Exactly. You put in three times the original HP so power to spare. Inverter rated, which means plenty of extra cooling capacity. And made cuts that lasted a few seconds at a time. You don’t need to be efficient, you only briefly generated any load, with some pauses to cool between cuts as you backed up. Unlikely that you’ll be in a seriously long cut on a 10EE. The inverter rated motor also is designed to be efficient over a wider range of operation than a regular motor.

But the answer is that it works. And that is all most people care about. Which is fine. A few people do care about “why”. So I’m throwing out the list of major issues. Not as show-stoppers, but as answers to the ”why”, and hints on why determining exact limits is complicated.

An interesting comparison would be how heavy of a cut you could make, at different rpms, between this configuration and the original Monarch DC motor. Of course that would require an original configured lathe side by side with yours. As hobbyists most of us aren’t pushing machines at this level to their limits, but it is kinda fun to know where those limits are.

More to come when time permits.

 

[rabler]

I was babbling on about carrier frequency in the VFD. Carrier frequency has several effects. Inductors create voltage spikes when you try to change the current through them, especially when you try to do so quickly. Since the carrier frequency is pulse modulated, rather than a simple sine wave, this is compounded (Quickly gets back to that Fourier analysis). Motors are in part, big inductors. So that carrier frequency can cause some voltage spikes much higher than the line voltage, inside the motor. They are VERY brief. But these spikes can start chipping away at the insulation. Inverter motors use better insulation or more insulation to compensate. Higher frequencies also have a tendency to generate radio waves, and thus interference on other devices. Inside the motor, that radio wave issue can, in some bizarre circumstances, cause the bearings to arc and pit, thus ruining bearings. (Have you ever left an old metal wire twist tie on something and thrown it in the microwave?) Some VFD motors actually use a mechanism to short out the center shaft/rotor to the frame, so that there is no voltage across the bearings.

Electromagnetics also adds another complication with high frequencies. We have to start worrying above wave propagation. If the wires are long, those high frequencies can start bouncing back and forth between the motor and the VFD. You can think about this wave propagation issue a bit like a whip, long enough wires and the motor starts seeing more brief voltage spikes at the input terminals, up to twice the expected voltage. Adding a bit of inductance to the wire can damp that out, like tying something heavy to the middle of a whip. But keeping the wires short is best. 3-5 feet of wiring from your VFD to motor is fine. Don't put your VFD on the other side of the room! Sometimes industrial controls system want to organize VFDs away from the motors, but they have to design in ways to mitigate that. Not much reason for your wires to be that long. Obviously the better insulation in the inverter motors is much more likely to be able to handle this.

While we're on this, this is also related to why VFDs don't like a switch between them and the motor. The energy racing down the wire toward the motor suddenly sees a dead end, and "bounces" back to the VFD. This causes the output IGBTs on the VFD to get a jolt. Of course, more expensive components can handle that better. But it is hard to tell if a more expensive VFD is using better components or is just making more money for someone.