Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[MPaulHolmes]

Oh I think he was saying the absolute maximum rating on the part. 650v*800amp.

[Gigawatt]

Great information.......Thank you very much!!......

[thingstodo]

I am new to the thread - just read through all of it today. (ReVolt was the past two days ... my eyes are fuzzy)

I'm a Controls Engineer with about 25 years experience in three phase motors and industrial automation. I wanted to see if DIY is practical for a high reliability, high voltage project. The ReVolt tells me that it is. I'm hoping that lurking around this thread will get my butt off the couch and out in the garage to work on my project.

I have not seen much for specs on this thread. I've seen hundreds of amps a few times and some discussions about 150V - 300V versus higher.

I'm looking for a controller that will give me 600 VAC (or about 950 VDC peak to peak) at about 30 seconds for acceleration to highway speed, dropping to 30A or approx 30 HP for cruising at 65 mph or 105 kph.

I'll cut the first post right there.

[thingstodo]

I hope this is not too far off topic. I'm trying to establish some VFD requirements

The truck is a 91 Chevy S-10. It seems to me that the truck will take about 30 HP at 2700 rpm to keep me at 65 mph. See reference one below for where I got that number

I have tested (sort of) my 30 HP 575VAC three phase 1750 rpm motor (all 300 pounds of it) with a smaller 208V surplus industrial VFD. The motor is rated just under 30 amps and I fed is all 17 amps that the little VFD could produce. The off-the-line torque measures about 60 foot-lb. That's at about 30 VAC and 0.5 Hz, the starting frequency of the VFD. The motor is rated for about 90 foot-lb at 1750 rpm. See reference 2 for where that number came from.

The total drag from the first site includes a 30 mph head wind and a 2% grade on dry pavement. The 1994 S-10 on the site has an automatic, so there is a bit of loss from the auto transmission. I live in the prairies, so 2% is the highest grade this little truck will see.

A bit about motors. An industrial medium efficiency motor is designed to start 'across-the-line'. Full power is applied using a starter and the motor gets very high inrush current (6 to 8 times rated current) until it gets up to running speed. The big iron in these motors is there to absorb this large heat load on start. I'm hoping that I can get 60 HP out of this motor for about 30 sec to get up to highway speed. I plan to mount the motor beneath the bed and couple to the drive shaft. At that point, there is only the 3.43 rear end between the motor and the 205/75 r14 tires that meet the road. If it were a high efficiency motor, it would absorbe 10 times starting current, and the synchronous speed would be closer to 1800 rpm, like maybe 1790.

The numbers I get for over-driving the 30 HP up to 2X current are 180 foot-lb of torque * 3.43 gear box * 12 inches/foot /radius of tire which is 13.5 inches

So around 550 lb of force pushing the truck forward from a stand-still. With an estimated 4500 lb of truck and batteries, this gets me to 65 mph in about 30 or 35 seconds.

If I put a 75 HP VFD at 575 VAC on this motor, it would only be putting out 75 amps rms with a 950 VDC bus.

You guys are talking about hundreds of amps here, right?

Link 1 - evconvert. I used the 1994 S10, a Warp 9 motor, Optima 31T batteries, and a Zila 2K controller. I know it's all DC, but I'm looking for approximates here. 1 string of batteries at 144V, 80% DOD and a 2% incline with a 30 mph head wind. That gets me a 10 mile range. The weight and misc is where the weight comes from. The drag calculations is where the acceleration comes from. The tire sizes don't quite match and I think I have a different differential on my 91, so I need 2700 rpm at 65 mph not 2100 as the site shows. I'm using most of the rest of the information as is for a starting point. Prepend 3w's
evconvert.com/tools/evcalc/.

Link 2 - rotating horsepower. I entered 90 foot-lb at 1750 rpm and got 30 HP or close enough. Prepend 3w's, add htm
calculatoredge.com/new/horsepower

[thingstodo]

Some points about Vector control and VFDs in general.

I don't exactly understand vector control, but I use it. The parameters are measured by the VFD when you set it up during commissioning. It sends some DC pulses through the motor and measures a few items. Every VFD vendor is pretty specific about what they call the numbers and how they calculate them. Basically they get a resistance and an inductance for the windings. As the rotor starts to spin, the current taken by the motor changes. Another couple of inductances are measured.

The result is that the VFD puts out a 'magnetizing current' at the right voltage and phase to build a 'good' magnetic flux across the gap between the windings and the rotor. This generates current in the rotor. The magnetic flux 'rotates' as the voltage and current are changed and 'drags' the rotor with it. The current in the rotor cutting through magnetic flux is what generates torque. If the rotor was made out of magnets, then the rotor speed would be identical to the VFD output frequency. I think those are called BDLC motors. Induction motors turn slower than the VFD frequency.

A Volts per Herz VFD puts out a pretty much linear voltage as you vary the frequency, up to the point when the voltage is at the maximum that the VFD has. Then the voltage stays the same and the frequency keeps going up.

A vector VFD puts out what the motor needs to turn at the specified frequency. It knows the magnetizing current. It increases the voltage with frequency to keep the magnetizing current the same as the speed changes. The current increases non-linearly with load on the motor. Pardon my personification of electronics.

I have 400 HP motors driving belt conveyors that are 1500 feet long started with tonnes (literally) of product sitting on the belt conveyor. And the VFDs we use don't put out thousands of amps. Our VFD vendors (Mitsubishi, Siemens, Toshiba, Allen Bradley) all use their own variants of IGBTs and they all seem to agree that a single IGBT should not have to deal with more than 100 - 150 amps at 600V. All of our VFDs 150 HP and below use 1 set (6) IGBTs, 200HP VFDs use two sets of IGBTs, the vendors disagree about the 300's, some use 2, some 3, the 400 and 500 HP VFDs have more variation in how many IGBT sets are used.

Driviing more current into a motor gives more torque for the same speed. If a operations person screws up and overloads a belt, we can put 200% current through the motor for a couple of minutes to get it cleared. But we keep an eye on that motor for an hour or so to make sure the temperature is not rising.

There is a limit. I have not tried 300% on a larger motor, but there is definitely not the same gain going from 200% to 300% that there is going from 100% to 200% current on a 25 HP motor.

The IGBT's, at least the ones I deal with, put out a lot of heat. Running a VFD at rated power will give you 1.5% to 3% heat. So a 10 kw VFD is also a 300W heater. Air flow is important.

[thingstodo]

Ok last post for the night.

The pulses sent out by the VFD need better insulation on the motor than a standard motor. The very square pulses, the ones that reduce the heat from the IGBT switcfhing, also stress the motor insulation since it was designed for the nice sine wave that the power company puts out. The dv/dt insulation rating is what gets you an inverter rated motor. The inverter rated motor insulation is normally about 1900V versus 1000V for a typical motor. That's because there is a weird physics issue with the impedance of the motor versus the impedance of the cable. It's hard to explain, but there is a reflected wave that makes its way back to the VFD for every pulse that is sent to the motor. The worst case magnitude is about 200%, so a 950 VDC pulse is reflected as 1900V. You can use a large filter to reduce this reflected wave, or a large inductor only to put a bit of slope on the pulse and reduce the stress.

The VFD vendors we use have their IGBT's rated at 2200V or so. So there are details somewhere that require the extra rating. These companies do not design in a safety factor if they can get away without it.

Sorry about the bad news on the motor insulation. If you stay below 500 VDC or 350 VAC a standard motor should take the voltage as long as you can round out the pulses a bit.

[thingstodo]

I just re-read my posts.

My apologies to the group. I do now know nearly as much as these posts make it sound like I do. I'm sure that several points that I have made will be clarified or corrected by various contributors.

If I came across like a know-it-all (Like I think I did) please chalk it up to me being new at posting.

[DawidvC]

If you have a look at AEVA forums, you will find that they have a lot of experience on AC motors. (It seems if all the ac-fanboys at least lurk there). The way to go forward is to rewind the motor for lower voltage, then run it at ICE speeds. That way, when choosing your motor, you only need to be concerned with the torque, as you calculated. Your rate of acceleration then becomes the main measure of what you need. As far as vector control is concerned, not all "vector control" schemes are the same, but all improve motor response, which is needed for ev-applications. Welcome to ecomodder forum.

Dawid

[Automcdonough]

2200V??? Do the IGBT's have to be rated for this or can it be snubbed without losing the information in the signal? Someone who knows more about vector control will have to chime in on that one. Or perhaps just dealing with position encoders suddenly seems like a great compromise!

The IGBT's I'm planning to use are 800A/650V. After 650V the Vf practically doubles, making things a lot less efficient. I was already planning on building some headroom into the voltage rating there by wiring the motor for 240V instead of 480V, the bus will be in the 320~380V range using active clamping to prevent reaching 400V.

Part of this decision was also rooted in arc-flash safety, I just don't want the battery voltage any higher than this from a liability perspective. (does everyone here know about arc-flash hazards? If not look it up, it's a very big deal!!)

[dcb]

Originally Posted by Automcdonough

... Or perhaps just dealing with position encoders suddenly seems like a great compromise!...

I don't think position encoders solve the whole problem on an induction motor, took me a few tries to start to get it myself, but Paul spells it out pretty well here:

http://ecomodder.com/forum/showthrea...tml#post142677