Filtering the output of a VFD

[The Real Andy]

So after many months of assing about and pestering from the old man
(dad) I finally have a working prototype of my sinlge to three phase
variable speed drive up and running. The design uses an International
Rectifier IRAMX16UP60A IGBT module:
http://ec.irf.com/v6/en/US/adirect/ir?cmd=catProductDetailFrame&productID=IRAMX16UP60A

I have the output connected directly to the motor. PWM is around
20khz. Problem is, the motor feels to be getting quite hot, finger
test would put it at around 40-50degC. The module itself is supposed
to be able to cope with about junction temp of 150degC, i estimate it
to be at about 60-70degC running a 1/3hp motor at 60Hz.

The problem, i know nothing about how hot a 1/3hp 3phase motor should
be running at, this project is for my father, a retired engineer. I
can live with the module itself running at its current temp,
considering the relative lack of heatsinking. Nothing a bigger
heatsink and a fan cant handle.

My question is, should the output be filtered with a capacitor, or is
it ok to feed the raw PWM output straight into the motor? The output
looks very bad on the cro, but the noise induced into the ground lead
makes this hard to prove.

Please excuse my ignorance on this subject, but I really know very
little. Google has been very good to me so far, but I am struggling to
find an answer on this one.

Thanks.

Andy


BTW. I will do some proper temperature tests very soon!!

 

[[email protected]]

50deg is normal for a typical 3 phase motor, it may overheat if you run
it below its nominal speed, because its fan runs slower, if you dont
reduce the current. A RFI filter is often used on the output to reduce
radiated noise otherwise you dont need any unles the inverter is a long
distance from the motor typicaly over 100m.

 

[Terry Given]

So after many months of assing about and pestering from the old man
(dad) I finally have a working prototype of my sinlge to three phase
variable speed drive up and running. The design uses an International
Rectifier IRAMX16UP60A IGBT module:
http://ec.irf.com/v6/en/US/adirect/ir?cmd=catProductDetailFrame&productID=IRAMX16UP60A

what, no 9kV 6kA GTOs ?

I have the output connected directly to the motor. PWM is around
20khz. Problem is, the motor feels to be getting quite hot, finger
test would put it at around 40-50degC.

Thats fine. Small motors are amazingly lossy - they have a great Surface
Area to Volume ratio, so can easily dissipate lots of heat, thus its
cheaper to make them lossy. So the manufacturers do. Above a few kW this
stops being the case, and 100kW+ machines are very efficient, for the
converse reason.

As long as your motor case temperature sits below 80-100C you should be
OK. IIRC 150C insulation is the norm.

I presume you have fixed V/Hz control, perhaps with LF boost? (to
overcome stator IR drop at low Vout). If you just DOL the output, watch
for rotor heating after multiple starts - so-called "soft" starters are
notorious for cooking rotor bars this way.

The module itself is supposed
to be able to cope with about junction temp of 150degC, i estimate it
to be at about 60-70degC running a 1/3hp motor at 60Hz.

if Tj = 150C the module will *DIE* in a matter of weeks due to thermal
expansion & contraction. 60-70C is great though.

The problem, i know nothing about how hot a 1/3hp 3phase motor should
be running at, this project is for my father, a retired engineer. I
can live with the module itself running at its current temp,
considering the relative lack of heatsinking. Nothing a bigger
heatsink and a fan cant handle.

CBarn24050 makes a great point re. fan cooling at reduced speeds (cube
law IIRC).

Make sure your waveform doesnt have a high harmonic content, and the
motor should be OK. first-turn failure is a problem with inverter-fed
machines, usually only when fed from a long cable (peak voltage doubles,
and can even triple if your modulator is bad enough). inverter-rated AC
drives have heavy insulation on the first few turns for exactly this reason.

My question is, should the output be filtered with a capacitor, or is
it ok to feed the raw PWM output straight into the motor? The output
looks very bad on the cro, but the noise induced into the ground lead
makes this hard to prove.

stick filter caps on the output and watch your IGBTs shit themselves.
dV/dt is pretty high, many kV/us, so I=CdV/dt is nasty. Little dutch
time bomb, tick tock boom.

Chokes first....and watch them catch fire

Please excuse my ignorance on this subject, but I really know very
little. Google has been very good to me so far, but I am struggling to
find an answer on this one.

output filters for VFDs are very tricky. All the commercial ones I have
seen get *HOT* - 180C insulation on the chokes, that sort of thing.

I have set a number of large chokes on fire doing this

Thanks.

Andy

BTW. I will do some proper temperature tests very soon!!

Cheers
Terry

 

[The Real Andy]

what, no 9kV 6kA GTOs ?

Maybe in the future

Thats fine. Small motors are amazingly lossy - they have a great Surface
Area to Volume ratio, so can easily dissipate lots of heat, thus its
cheaper to make them lossy. So the manufacturers do. Above a few kW this
stops being the case, and 100kW+ machines are very efficient, for the
converse reason.

As long as your motor case temperature sits below 80-100C you should be
OK. IIRC 150C insulation is the norm.

I tried looking on the manufacurers website to no avail. However they
did mention some ratings of the bearing at 100degC.

I presume you have fixed V/Hz control, perhaps with LF boost? (to
overcome stator IR drop at low Vout).

Yup.

If you just DOL the output, watch

for rotor heating after multiple starts - so-called "soft" starters are
notorious for cooking rotor bars this way.


if Tj = 150C the module will *DIE* in a matter of weeks due to thermal
expansion & contraction. 60-70C is great though.

This is all under no load btw. My plan is to mount the heatsing in the
case with a fan blowing over it. This should keep things cool.

CBarn24050 makes a great point re. fan cooling at reduced speeds (cube
law IIRC).

Make sure your waveform doesnt have a high harmonic content, and the
motor should be OK. first-turn failure is a problem with inverter-fed
machines, usually only when fed from a long cable (peak voltage doubles,
and can even triple if your modulator is bad enough). inverter-rated AC
drives have heavy insulation on the first few turns for exactly this reason.

There is a qustion. How can I check the output? I have a high voltage
probe for the cro, but it is terrible for noise. The probe is 1000:1.

stick filter caps on the output and watch your IGBTs shit themselves.
dV/dt is pretty high, many kV/us, so I=CdV/dt is nasty. Little dutch
time bomb, tick tock boom.

Chokes first....and watch them catch fire


output filters for VFDs are very tricky. All the commercial ones I have
seen get *HOT* - 180C insulation on the chokes, that sort of thing.

I have set a number of large chokes on fire doing this

I will skip output filtering then! Many thanks.

Andy

 

[Michael A. Terrell]

what, no 9kV 6kA GTOs ?

I prefered my 389 CID GTO! ;-)

--
Link to my "Computers for disabled Veterans" project website deleted
after threats were telephoned to my church.

Michael A. Terrell
Central Florida

 

[Terry Given]

Maybe in the future




I tried looking on the manufacurers website to no avail. However they
did mention some ratings of the bearing at 100degC.

small motors run stinking hot, which is probably why they dont mention it.

Yup.

jolly good jelly.

If you just DOL the output, watch



This is all under no load btw. My plan is to mount the heatsing in the
case with a fan blowing over it. This should keep things cool.

Depending on the IGBTs you use, conduction losses may or may not
dominate. I suspect at your Fsw then Psw dominates, so an unloaded motor
is a not-bad test. But it will get hotter under load.

make sure you have a heatsink over-temp. 80C is good, and in conjunction
with a decent current limit will keep the IGBTs safe in the event of an
overload.

There is a qustion. How can I check the output? I have a high voltage
probe for the cro, but it is terrible for noise. The probe is 1000:1.

I have a bunch of 4th order (IIRC) bessel filters, 50R in & out that I
use in conjunction with a P5200 diff probe (Zout = 50R) and a coax cable
+ 50R terminator to the scope. A little bit of gain peaking (thanks
HP3577), but monotonic rolloff from 6kHz to 50MHz.

Lin = 2.5mH
Cshunt = 1.44uF
L = 6mH
Cout = 1uf

I forget what cores I used (blue, 25mm OD, 14mm ID, 10mm thick) but my
tech sectional-wound them with wire-wrap wire . IIRC they came from
Jaycar or DSE. Its built on a piece of Cu-clad PCB, mounted in a 92mm x
38mm x 30mm diecast box with a BNC poking out either end.

I will skip output filtering then! Many thanks.

You shouldnt need it - unless the drive makes all adjacent electronics
go nuts. In which case, use a screened cable from drive to motor, screen
co-axially bonded to drive & motor (ignore instrumentation guys who say
this is wrong)

Andy

You're welcome. Post some pics of the toy.

Cheers
Terry

 

[Terry Given]

I prefered my 389 CID GTO! ;-)

the specs, the specs, my kingdom for the specs.

Cheers
Terry

 

[The Real Andy]

small motors run stinking hot, which is probably why they dont mention it.


jolly good jelly.


Depending on the IGBTs you use, conduction losses may or may not
dominate. I suspect at your Fsw then Psw dominates, so an unloaded motor
is a not-bad test. But it will get hotter under load.

make sure you have a heatsink over-temp. 80C is good, and in conjunction
with a decent current limit will keep the IGBTs safe in the event of an
overload.


I have a bunch of 4th order (IIRC) bessel filters, 50R in & out that I
use in conjunction with a P5200 diff probe (Zout = 50R) and a coax cable
+ 50R terminator to the scope. A little bit of gain peaking (thanks
HP3577), but monotonic rolloff from 6kHz to 50MHz.

Lin = 2.5mH
Cshunt = 1.44uF
L = 6mH
Cout = 1uf

I forget what cores I used (blue, 25mm OD, 14mm ID, 10mm thick) but my
tech sectional-wound them with wire-wrap wire . IIRC they came from
Jaycar or DSE. Its built on a piece of Cu-clad PCB, mounted in a 92mm x
38mm x 30mm diecast box with a BNC poking out either end.



You shouldnt need it - unless the drive makes all adjacent electronics
go nuts. In which case, use a screened cable from drive to motor, screen
co-axially bonded to drive & motor (ignore instrumentation guys who say
this is wrong)


You're welcome. Post some pics of the toy.

No no! not yet! I am just getting a second cut of the board done. Once
i build up version 2, i will post some pics and schematics on my
website.

 

[Michael A. Terrell]

the specs, the specs, my kingdom for the specs.

Cheers
Terry

Specs?

1966 Pontiac GTO.
Red exterior with a black & wood grain interior.
389 CID V8 engine W/ Turbohydromatic 400 heavy duty three speed
automatic transmission.
Soft shell 2 door body.

The best part was that I did the entire restoration myself, back when
I could handle the heavy labor. It ran like a dream! ;-) It could lay
rubber on hard packed sand, and idled at 300 RPM. It had a pair of
seven and a half foot 3 inch straight exhaust piles that could rattle
buildings, at the right engine speed. Ah, I miss the good old days.

--
Link to my "Computers for disabled Veterans" project website deleted
after threats were telephoned to my church.

Michael A. Terrell
Central Florida

 

[Terry Given]

Specs?

1966 Pontiac GTO.
Red exterior with a black & wood grain interior.
389 CID V8 engine W/ Turbohydromatic 400 heavy duty three speed
automatic transmission.
Soft shell 2 door body.
nice.


The best part was that I did the entire restoration myself, back when
I could handle the heavy labor. It ran like a dream! ;-) It could lay
rubber on hard packed sand, and idled at 300 RPM. It had a pair of
seven and a half foot 3 inch straight exhaust piles that could rattle
buildings, at the right engine speed. Ah, I miss the good old days.

that would have sounded great.

The father of a friend of mine recently bought a nice new merc. He's a
retired mechanic, so was poking around trying to figure out where to put
oil etc. but was unsuccessful. He asked the dealer, who replied "the car
will tell you when it needs oil, bring it to us and we'll do it". Sheesh.


Cheers
Terry

 

[Michael A. Terrell]

that would have sounded great.

The father of a friend of mine recently bought a nice new merc. He's a
retired mechanic, so was poking around trying to figure out where to put
oil etc. but was unsuccessful. He asked the dealer, who replied "the car
will tell you when it needs oil, bring it to us and we'll do it". Sheesh.

Cheers
Terry

That's not a car, its a moneypit. ;(

--
Link to my "Computers for disabled Veterans" project website deleted
after threats were telephoned to my church.

Michael A. Terrell
Central Florida

 

[Rob Votin]

So after many months of assing about and pestering from the old man
(dad) I finally have a working prototype of my sinlge to three phase
variable speed drive up and running. The design uses an International
Rectifier IRAMX16UP60A IGBT module:
http://ec.irf.com/v6/en/US/adirect/ir?cmd=catProductDetailFrame&productID=IRAMX16UP60A

I have the output connected directly to the motor. PWM is around
20khz. Problem is, the motor feels to be getting quite hot, finger
test would put it at around 40-50degC. The module itself is supposed
to be able to cope with about junction temp of 150degC, i estimate it
to be at about 60-70degC running a 1/3hp motor at 60Hz.

The problem, i know nothing about how hot a 1/3hp 3phase motor should
be running at, this project is for my father, a retired engineer. I
can live with the module itself running at its current temp,
considering the relative lack of heatsinking. Nothing a bigger
heatsink and a fan cant handle.

My question is, should the output be filtered with a capacitor, or is
it ok to feed the raw PWM output straight into the motor? The output
looks very bad on the cro, but the noise induced into the ground lead
makes this hard to prove.

Please excuse my ignorance on this subject, but I really know very
little. Google has been very good to me so far, but I am struggling to
find an answer on this one.

Thanks.

Andy


BTW. I will do some proper temperature tests very soon!!

Andy,
At the risk of asking an obvious question, the motor you are using is
"inverter rated", right? Otherwise a VFD would cook an ordinary AC
motor...

 

[Terry Given]

Andy,
At the risk of asking an obvious question, the motor you are using is
"inverter rated", right? Otherwise a VFD would cook an ordinary AC
motor...

nah, only if there is a long lead, enough that transmission-line
behaviour can double the peak voltage seen by the first turn (and, if
your modulator is as bad as those used by Allen Bradley, triple it).

Cheers
Terry