Can typical (60 Hz input, 100 + horsepower ) VFD drives allow output at say, 1 Hz, 10 Hz, 30 Hz ? I understand that driven equipment may have problems at very low speeds, but the question here is if VFD systems usually have a "built in" minimum output frequency, or can you set the VFD at , say 2 Hz to 66 Hz on a 60 cycle system?
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
VFD Minimum Output Frequency 2
- Thread starter FredRosse
- Start date
Skogsgurra
Electrical
I have run at 0 Hz on several makes of VFD:s. So, there should not be any problem with that.
Motor cooling is very often dependent on speed. So may bearing friction/greasing be. But not a problem for the VFD as such.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Motor cooling is very often dependent on speed. So may bearing friction/greasing be. But not a problem for the VFD as such.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
If you are trying to do vector control at speeds below 5Hz you can also experience major stability problems with "sensorless" control.
Keith Cress
kcress -
Keith Cress
kcress -
As said, the problem is not in the VFD, it's in the motor. I can tell a motor to run at 0 Hz, 0.01Hz or 599Hz. That does not mean the motor is capable of any of those speeds.
Side not on the "599Hz"; that was not a random number. As of a couple of years ago, VFDs are not allowed to run at speeds over 599Hz without special permission, typically from the adept of Defence here in the U.S. and most other NATO countries. It has to do with centrifuges and weapons. So mfrs were required to gobble them with programming that cannot be changed by the user.
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
Side not on the "599Hz"; that was not a random number. As of a couple of years ago, VFDs are not allowed to run at speeds over 599Hz without special permission, typically from the adept of Defence here in the U.S. and most other NATO countries. It has to do with centrifuges and weapons. So mfrs were required to gobble them with programming that cannot be changed by the user.
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
![URL]](https://res.cloudinary.com/engtips/image/fetch/w_800,c_lfill,q_auto,f_auto,g_faces:center/[URL unfurl="true"]http://i11.tinypic.com/2pt4b3o.gif[/URL])
Skogsgurra
Electrical
That explains why we shall not! Been struggling with that until I gave up.
Yes, centrifuges for uranium. Just to make it a little bit more complicated to make THE BOMB.
But making THE BOMB takes a lot of effort. And reprogramming - or build from scratch - a VFD must be one of the lesser challenges.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Yes, centrifuges for uranium. Just to make it a little bit more complicated to make THE BOMB.
But making THE BOMB takes a lot of effort. And reprogramming - or build from scratch - a VFD must be one of the lesser challenges.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Caused a lot of grief at our end, particularly with the modular systems. A far more effective method of preventing uranium extraction would be to leave a back door in the plc software, tell mossad and the CIA how to exploit it, and get the centrifuges to overspeed until they destroy themselves. You could call it stuxnet, for example. oh, wait....
on topic, most VFD's have a programmable min. frequency, so the ramp up could begin at 2Hz, 4Hz, etc. But 0Hz is no issue for the drive, apart from localised IGBT heating, which can be significant. In fact, if you want todestroy a VFD, the easiest way is probably to run at high pulse frequency, slow speed, high current....
Cheers,
Mort
on topic, most VFD's have a programmable min. frequency, so the ramp up could begin at 2Hz, 4Hz, etc. But 0Hz is no issue for the drive, apart from localised IGBT heating, which can be significant. In fact, if you want todestroy a VFD, the easiest way is probably to run at high pulse frequency, slow speed, high current....
Cheers,
Mort
I do proofread, but sometimes when I post from an iOS device, I get an error when I try to go back and edit some of Apple's re-interpretations of what I was saying, such as it assuming I meant "adept" when I had typed "Dept." That's what happened that time, it kept giving me an "unexpected error, please try again later" message and eventually timed out on allowing me to edit it. I need to do more proofreading befor suiting, but my eyes are getting worse at seeing the tiny tiny test on my iPhone 5 screen. Time to upgrade to the big 6 I suppose, but I don't like the size iny pocket.
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
Skogsgurra
Electrical
It is an interesting point that you bring up, Mort. I never thought about the local heating that you get when running four out of six IGBT:s.
My first thinking was that the heat supervision would take care about that. My second thinking is that there's just one heat sink and one sensor. So, the heat generated by four semi-conductors before the alarm goes off will be 50% higher for each IGBT. And that is probably above the limit for destruction. Good point. Will remember.
So, how come that I never experienced a problem with that? Probably because all applications where I have used "zero frequency" wasn't actually 0 Hz because the induction motor that was standing still needed a low frequency rotating field to keep the load and that low frequency was enough to even out heat between the transistors.
When Synchronous motors become more widespread, we may very well see IGBT destruction in such applications. Very interesting.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
My first thinking was that the heat supervision would take care about that. My second thinking is that there's just one heat sink and one sensor. So, the heat generated by four semi-conductors before the alarm goes off will be 50% higher for each IGBT. And that is probably above the limit for destruction. Good point. Will remember.
So, how come that I never experienced a problem with that? Probably because all applications where I have used "zero frequency" wasn't actually 0 Hz because the induction motor that was standing still needed a low frequency rotating field to keep the load and that low frequency was enough to even out heat between the transistors.
When Synchronous motors become more widespread, we may very well see IGBT destruction in such applications. Very interesting.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Jeff:
Now, you're funning me right?
I use mine horizontally he uses his vertically. You can actually retrieve your phone while sitting which is impossible from a pocket. Also with an otter box you have zero issues with dropping it EVER. You can then go to a larger phone with no pocket involvement at all. Heck go with a phablet! LOL
Keith Cress
kcress -
Now, you're funning me right?
Pocket?!! I gave that up long ago. I switched to the Otter Box Defender Series with Holster and couldn't be happier. Showed my son and he switched and thanks me about once a month. He now packs two phones with Otter Box defenders in holsters.
I use mine horizontally he uses his vertically. You can actually retrieve your phone while sitting which is impossible from a pocket. Also with an otter box you have zero issues with dropping it EVER. You can then go to a larger phone with no pocket involvement at all. Heck go with a phablet! LOL
Keith Cress
kcress -
-
2
- #13
Hi, Gunnar,
Thanks for the reply. I can only speak from my personal experience with our particular products, but it is a topic I have been doing some work on in the past few years. The following refers to small (<30kW) drives.
You correctly point out that there is only 1 temperature measurement point, usually on the same substrate as the IGBTs (and the freewheel diodes, and the rectifiers). The path for heat to reach this component is long (IGBT > module plate > heatsink > along the heatsink > back to the plate > thermistor) and depends on the thermal impedances of the materials, esp. heatsink paste. For this reason, knowing the temperature of the heatsink does not tell you the temperature of the IGBT die.
The die temperature is modelled, based on power loss in the devices (from DC link volts, output current, pulse frequency, etc.) and a dynamic model of the thermal system (i.e. the system described above).
For low output frequencies, there is a multiplying factor applied to allow for local heating. at 0Hz this factor is 2.5 - i.e. the IGBT temp rise at 0Hz is 2.5x that at 50Hz, for example. This might mean that larger IGBTs are required than would be considered at first glance. The frequency at which localised heating begins to occur depends on the thermal system characteristics, and so varies from drive to drive.
Our thermal model (and hardware sizing) allows for all this, but I wouldn't like to speak for anyone else. To find out, try doing some heavy overload cycling at 0Hz, high supply voltage, and high pulse frequency. Only try if the product is under warranty!!!
Cheers,
Mort
Thanks for the reply. I can only speak from my personal experience with our particular products, but it is a topic I have been doing some work on in the past few years. The following refers to small (<30kW) drives.
You correctly point out that there is only 1 temperature measurement point, usually on the same substrate as the IGBTs (and the freewheel diodes, and the rectifiers). The path for heat to reach this component is long (IGBT > module plate > heatsink > along the heatsink > back to the plate > thermistor) and depends on the thermal impedances of the materials, esp. heatsink paste. For this reason, knowing the temperature of the heatsink does not tell you the temperature of the IGBT die.
The die temperature is modelled, based on power loss in the devices (from DC link volts, output current, pulse frequency, etc.) and a dynamic model of the thermal system (i.e. the system described above).
For low output frequencies, there is a multiplying factor applied to allow for local heating. at 0Hz this factor is 2.5 - i.e. the IGBT temp rise at 0Hz is 2.5x that at 50Hz, for example. This might mean that larger IGBTs are required than would be considered at first glance. The frequency at which localised heating begins to occur depends on the thermal system characteristics, and so varies from drive to drive.
Our thermal model (and hardware sizing) allows for all this, but I wouldn't like to speak for anyone else. To find out, try doing some heavy overload cycling at 0Hz, high supply voltage, and high pulse frequency. Only try if the product is under warranty!!!
Cheers,
Mort
Skogsgurra
Electrical
Mort,
even more interesting. You deserve an LPS. Sorry, there are no super-sized ones...
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
even more interesting. You deserve an LPS. Sorry, there are no super-sized ones...
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Gunnar; First, I'm confused why are you talking about running only 4 of 6 IGBTs??
Secondly, why would running the inverter output at high freq near zero speed with high current be any worse than high freq at whatever speed with high current?? The VFD doesn't cool any worse putting out zero speed than any other speed.
Keith Cress
kcress -
Secondly, why would running the inverter output at high freq near zero speed with high current be any worse than high freq at whatever speed with high current?? The VFD doesn't cool any worse putting out zero speed than any other speed.
Keith Cress
kcress -
Skogsgurra
Electrical
Answers:
1. I was back in older days (no PWM) and simplified thinking.
2. I hoped that no-one would notice. (Silly me)
If you run at zero frequency, let's say with peak current in phase 1, then the two other phases will carry 50% each. The IGBT with maximum current will then heat more than it would if current was distributed across all three phases, as would be the case at higher frequencies.
Losses in IGBT:s are more or less current "squarish" (the constant initial voltage drop, as seen in diodes and thyristors, is quite low in an IGBT) and that makes losses in the phase 1 IGBT substantially higher at 0 Hz than at higher frequencies where load is more periodic and spread out over all thyristors. The thermal resistance between die and heat sink plays a role. More than one thinks as an everyday user.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
1. I was back in older days (no PWM) and simplified thinking.
2. I hoped that no-one would notice. (Silly me)
If you run at zero frequency, let's say with peak current in phase 1, then the two other phases will carry 50% each. The IGBT with maximum current will then heat more than it would if current was distributed across all three phases, as would be the case at higher frequencies.
Losses in IGBT:s are more or less current "squarish" (the constant initial voltage drop, as seen in diodes and thyristors, is quite low in an IGBT) and that makes losses in the phase 1 IGBT substantially higher at 0 Hz than at higher frequencies where load is more periodic and spread out over all thyristors. The thermal resistance between die and heat sink plays a role. More than one thinks as an everyday user.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Ahhh I'm getting it. This would be the special case of zero speed so the VFD may never actually commutate to anywhere else but just sit thrashing one pair of output drivers. Thanks.
Keith Cress
kcress -
Keith Cress
kcress -
Skogsgurra
Electrical
Right, Smoked! You are not only good-looking. You got brains, too! 
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
controlsdude
Electrical
zero speed would be like a motor with a brake (example transfer car). You zero speed the motor to bring it to a stop, then set break, then disable the drive. Go to next move or command. Used this with motion using cheap vfd technology versus servo. This way whatever your moving does not slip out of position.
Also, most motors have a minimum speed where a external fan is used. I thought it was around 15-20 Hz and below. This is so that the motor will not over heat. I would ask the manu of motor for that speed or hertz. I got this same advise from this website, forgot who told me that issue.
Also, most motors have a minimum speed where a external fan is used. I thought it was around 15-20 Hz and below. This is so that the motor will not over heat. I would ask the manu of motor for that speed or hertz. I got this same advise from this website, forgot who told me that issue.
- Status
Similar threads
- Question
- Question
- Question