PFC
#26 material is 100% the wrong material. It does bot work well at 100KHz (where most PFCs work).
I have worked closely with Micrometals (they are 2 hours away from our factory) and their engineers are quite clear that #8 and #18 are the only suitable cores for PFC. Yes making a core simply huge may get rid of some problems but all of us have space issues.
Interwinding capacitance if too high is the death of PFCs, and too high an inducatnce allows the PFC coil to saturate easier on a given core
Stephen Mantz
Zed Audio Corp.
#26 material is 100% the wrong material. It does bot work well at 100KHz (where most PFCs work).
I have worked closely with Micrometals (they are 2 hours away from our factory) and their engineers are quite clear that #8 and #18 are the only suitable cores for PFC. Yes making a core simply huge may get rid of some problems but all of us have space issues.
Interwinding capacitance if too high is the death of PFCs, and too high an inducatnce allows the PFC coil to saturate easier on a given core
Stephen Mantz
Zed Audio Corp.
I'd rather spend a few extra bucks than let the magic smoke out of my MOSFET because I got cheap! I will reserve the T-200-26 cores for the cross-coupled output inductor for the main PWM section.
Re: PFC
It all seemed pretty straightforward if one pays attention and does his homework so none of this really surprises me. I'm just deathly afraid (err...respectful!) of plasma-emitting semiconductors - likely due to some traumatic experiences as a younger boy. I built a Lexan box into which my "I'm not sure what might happen" designs go during initial power-up. It's less shocking and probably much safer.
I used IGBTs years ago - completely unsucessfully. Using an IR part with bootstrapping - the IGBT would instantaneously C-E short. A few back-and-forth emails to IR yielded no fruit. I gave up and went to a FET.
I'm a little older and a lot wiser now and willing to give them another go. At least now I have the proper test equipment to know exactly what's going on.
You've been very helpful - now I need to build it!
MOER said:
It all seemed pretty straightforward if one pays attention and does his homework so none of this really surprises me. I'm just deathly afraid (err...respectful!) of plasma-emitting semiconductors - likely due to some traumatic experiences as a younger boy. I built a Lexan box into which my "I'm not sure what might happen" designs go during initial power-up. It's less shocking and probably much safer.
I used IGBTs years ago - completely unsucessfully. Using an IR part with bootstrapping - the IGBT would instantaneously C-E short. A few back-and-forth emails to IR yielded no fruit. I gave up and went to a FET.
I'm a little older and a lot wiser now and willing to give them another go. At least now I have the proper test equipment to know exactly what's going on.You've been very helpful - now I need to build it!
PFC
I don't follow you here, dead time between what? Do you mean time before the PFC circuit starts chopping? Or max duty cycle? Did you use an extra diode which lets the startup surge bypass the inductor?
Dead time of the main switching device. Yes decrease the maximum duty cycle on start up.
No need for the bypass diode as the PFC inuctor and PFC diode on start up are passive parts in my design. Since the PFC controller is off, the bulk capacitors are charged via the PFC coil ( a zero ohm part now) and the PFC diode. Once the PFC controller is activated the precharging of the bulk capacitors has been done (voltage dependent on the incoming AC mains)
Stephen Mantz
Zed Audio Corp.
I don't follow you here, dead time between what? Do you mean time before the PFC circuit starts chopping? Or max duty cycle? Did you use an extra diode which lets the startup surge bypass the inductor?
Dead time of the main switching device. Yes decrease the maximum duty cycle on start up.
No need for the bypass diode as the PFC inuctor and PFC diode on start up are passive parts in my design. Since the PFC controller is off, the bulk capacitors are charged via the PFC coil ( a zero ohm part now) and the PFC diode. Once the PFC controller is activated the precharging of the bulk capacitors has been done (voltage dependent on the incoming AC mains)
Stephen Mantz
Zed Audio Corp.
Re: PFC
Most of the PFC I've researched have that "bypass" diode installed. I could be completely wrong, but wouldn't the diode reduce the surge current from passing through the inductor and also reduce the chance of a Miller component turning on the switching transistor? It seems like some brief resonance would occur between the switching component and the inductor causing the device to behave destructively...
MOER said:
Most of the PFC I've researched have that "bypass" diode installed. I could be completely wrong, but wouldn't the diode reduce the surge current from passing through the inductor and also reduce the chance of a Miller component turning on the switching transistor? It seems like some brief resonance would occur between the switching component and the inductor causing the device to behave destructively...
PFC
Simple test, take a PFC coil, put a diode in series and then put your bulk capacitors after the diode.
Monitor the current across the coil. My PFC operates this way on turn on, the main IGBT is off and the IR1150S is off. So the PFC coil is just some wire on a core. The inductance I use is always less than 200uH due to the size of the power supply.
Stephen Mantz
Zed Audio Corp.
Simple test, take a PFC coil, put a diode in series and then put your bulk capacitors after the diode.
Monitor the current across the coil. My PFC operates this way on turn on, the main IGBT is off and the IR1150S is off. So the PFC coil is just some wire on a core. The inductance I use is always less than 200uH due to the size of the power supply.
Stephen Mantz
Zed Audio Corp.
Re: IR1150S
Using a high value gate resistor might fix EMI but the slowed transitions are going to make the switching device get way hotter than necessary. I worked in a SMPS design lab for 10 years until ~2 years ago and the biggest I worked on was 3.3kW. The last PFC I worked on was a 300W unit and had a Fairchild IRFP460A that would turn off in 30nS and this fed a Cree 600V silicon carbide schottky diode. The results were unbelievably better than so many of the units I had worked on before. The waveforms were like in a textbook. 240VAC mains to DC bus (400v) efficiency was 98% despite the higher forward voltage drop of the SiC diode. You really owe it to yourself to try one of these things. Diode reverse recovery and all the difficulties it brings are a thing of the past. It's very much like going from bipolar transistors to mosfets for switching use.
We never found IGBTs to be very good; maybe they are better now?
MOER said:
Using a high value gate resistor might fix EMI but the slowed transitions are going to make the switching device get way hotter than necessary. I worked in a SMPS design lab for 10 years until ~2 years ago and the biggest I worked on was 3.3kW. The last PFC I worked on was a 300W unit and had a Fairchild IRFP460A that would turn off in 30nS and this fed a Cree 600V silicon carbide schottky diode. The results were unbelievably better than so many of the units I had worked on before. The waveforms were like in a textbook. 240VAC mains to DC bus (400v) efficiency was 98% despite the higher forward voltage drop of the SiC diode. You really owe it to yourself to try one of these things. Diode reverse recovery and all the difficulties it brings are a thing of the past. It's very much like going from bipolar transistors to mosfets for switching use.
We never found IGBTs to be very good; maybe they are better now?
Re: Re: IR1150S
*sigh*
I began writing a couple things, but gave up. I have no experience with PFC, but it doesn't seem all that complex with today's integrated drivers. I have great test equipment, current probes, spectrum analyser, EMI testing chamber...and oh, yeah - good PCB layout skills (really my only saving grace, I guess). The worst I could do is overbuild it.
Circlotron said:
*sigh*
I began writing a couple things, but gave up. I have no experience with PFC, but it doesn't seem all that complex with today's integrated drivers. I have great test equipment, current probes, spectrum analyser, EMI testing chamber...and oh, yeah - good PCB layout skills (really my only saving grace, I guess). The worst I could do is overbuild it.
So far so good. I started a thyristor inrush reduction circuit, and quickly found ST has a component - the STIL-08 - with everything inbuilt for a smidge more in cost (but savings in parts placement). The PENTAWATT package is ridiculous, but I just staggered the leads making it all work just fine.
I've had no problems with surges using this switched thyristor circuit and thermistor, the 1150s behaves just fine and I've TRIED to blow it up by switching ON/OFF quickly at 220VAC input under near full load. The STIL doesn't trigger (to bypass the thermistor) until the PFC has started - and the PFC doesn't start until the bulk caps are charged.
This particular design is commercial (and my first PFC), so I may not be posting pics. Use your imagination until I do. Thanks to those suggesting the Micrometals cores - they're great.
I've had no problems with surges using this switched thyristor circuit and thermistor, the 1150s behaves just fine and I've TRIED to blow it up by switching ON/OFF quickly at 220VAC input under near full load. The STIL doesn't trigger (to bypass the thermistor) until the PFC has started - and the PFC doesn't start until the bulk caps are charged.
This particular design is commercial (and my first PFC), so I may not be posting pics. Use your imagination until I do.
Thanks to those suggesting the Micrometals cores - they're great.
PFC
We do not have a problem of the PFC inductor saturating on turn on. Two reasons. The PFC chip is not powered on turn on and I have a series resistor with my bulk capacitors so my turn on surge is almost zero. Once the bulk caps are charged to the peak of the incoming AC, I start the PFC chip with maximum dead time which again prevents any surge currents. Once the PFC voltage has settled at 385v, I release the dead time circuit, then release the surge resistor and of course the PWM switcher is up and running at this time.
Stephen Mantz
Zed Audio Corp.
CA USA
We do not have a problem of the PFC inductor saturating on turn on. Two reasons. The PFC chip is not powered on turn on and I have a series resistor with my bulk capacitors so my turn on surge is almost zero. Once the bulk caps are charged to the peak of the incoming AC, I start the PFC chip with maximum dead time which again prevents any surge currents. Once the PFC voltage has settled at 385v, I release the dead time circuit, then release the surge resistor and of course the PWM switcher is up and running at this time.
Stephen Mantz
Zed Audio Corp.
CA USA
Attachments
PFC
Re: PFC Post #37
Trader Joe's Granola/Cereal? Just leave the bag open on the table...neat idea, but maybe messy.
A bit off topic, but: Do you still sell off components from Zed that you're no longer using in production? Shoot me an email @ postmaster_at_envisionelec.net. Thanks, Steve.
Well not messy for us as it works well and the issues with the IRS1150 are all bypassed
The email address you posted does not work. Email me at ZEDAUDIO@AOL.COM and I shall then be able to reply
Stephen Mantz
Zed Audio Corp.
CA
Re: PFC Post #37
Trader Joe's Granola/Cereal? Just leave the bag open on the table...neat idea, but maybe messy.
A bit off topic, but: Do you still sell off components from Zed that you're no longer using in production? Shoot me an email @ postmaster_at_envisionelec.net. Thanks, Steve.
Well not messy for us as it works well and the issues with the IRS1150 are all bypassed
The email address you posted does not work. Email me at ZEDAUDIO@AOL.COM and I shall then be able to reply
Stephen Mantz
Zed Audio Corp.
CA
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.