Good vs. Bad Power Supply Differences Explained by PSU Expert Jonny Guru
video description
Date: 2022-05-03
Related videos
Comments and reviews: 9
44R0Ndin
I wonder if power supplies are ever going to go to active rectification (using MOSFETs instead of just a diode bridge rectifier) for the mains power rectification stage (as well as the already-common usage in the rectification stage after the high frequency step-down transformer) in further pursuit of efficiency, or are the gains to be had just not enough to justify the extra cost in that case because of the low frequency of AC mains power (50/60hz)?
I mean I'm sure there's some super-simple driver circuit out there that can control the gates of the MOSFETs in such an application (I'm just not sure if it's capable of being used at such a LOW frequency as AC mains power), but it would probably be the MOSFETs themselves that would be the expensive part (since they have to deal with at most 240V (with lower current) and at most 15A (with 120v input)), and to me that sounds like a MOSFET that costs about a dollar each (when prices are sane that is).
And then you'd probably also need a way to bring that mains voltage down to a level suitable to power the driver IC for those active rectification MOSFETs as well as tell the IC when to turn the MOSFETs on and off (so it would have to have both an AC output of low voltage and negligible current (few microamps, since it's signalling), as well as a DC output of low (24v) voltage and perhaps at most 50ma current to power the driver IC itself).
But all that sounds like a lot more complication for something that's already quite efficient since diodes are plenty efficient on a low frequency sine wave where the voltage drop (and therefore power loss) across the diode at maximum current is certainly under 5v if the diodes being used are any good.
In fact, I don't think they're even using Schottkey diodes for mains voltage rectification (and in the case of 120v input, voltage multiplication as well).
So they'd probably see if the Schottkey diodes work out first, before switching to active rectification.
An additional step that could be taken with this active bulk rectification approach is to further pulse-width-modulate the signal to the gate of the rectification MOSFETs, in order to actively compensate for any potential ill effects regarding power factor and/or EMI that might otherwise happen.
I think some LED driver power supplies do this in order to cut out the need to create a DC voltage rail between mains power and driving the LEDs, while still allowing them to regulate the LED's current enough to prevent failure, but I am not 100% sure on that.
You might even be able to reach a point where you could eliminate the X and Y capacitors, or at least significantly reduce their capacitance.
reply
I wonder if power supplies are ever going to go to active rectification (using MOSFETs instead of just a diode bridge rectifier) for the mains power rectification stage (as well as the already-common usage in the rectification stage after the high frequency step-down transformer) in further pursuit of efficiency, or are the gains to be had just not enough to justify the extra cost in that case because of the low frequency of AC mains power (50/60hz)?
I mean I'm sure there's some super-simple driver circuit out there that can control the gates of the MOSFETs in such an application (I'm just not sure if it's capable of being used at such a LOW frequency as AC mains power), but it would probably be the MOSFETs themselves that would be the expensive part (since they have to deal with at most 240V (with lower current) and at most 15A (with 120v input)), and to me that sounds like a MOSFET that costs about a dollar each (when prices are sane that is).
And then you'd probably also need a way to bring that mains voltage down to a level suitable to power the driver IC for those active rectification MOSFETs as well as tell the IC when to turn the MOSFETs on and off (so it would have to have both an AC output of low voltage and negligible current (few microamps, since it's signalling), as well as a DC output of low (24v) voltage and perhaps at most 50ma current to power the driver IC itself).
But all that sounds like a lot more complication for something that's already quite efficient since diodes are plenty efficient on a low frequency sine wave where the voltage drop (and therefore power loss) across the diode at maximum current is certainly under 5v if the diodes being used are any good.
In fact, I don't think they're even using Schottkey diodes for mains voltage rectification (and in the case of 120v input, voltage multiplication as well).
So they'd probably see if the Schottkey diodes work out first, before switching to active rectification.
An additional step that could be taken with this active bulk rectification approach is to further pulse-width-modulate the signal to the gate of the rectification MOSFETs, in order to actively compensate for any potential ill effects regarding power factor and/or EMI that might otherwise happen.
I think some LED driver power supplies do this in order to cut out the need to create a DC voltage rail between mains power and driving the LEDs, while still allowing them to regulate the LED's current enough to prevent failure, but I am not 100% sure on that.
You might even be able to reach a point where you could eliminate the X and Y capacitors, or at least significantly reduce their capacitance.
reply
kidman2505
I actually deal with EMI noise in my field of work. In Controlled Environment Agriculture (indoor gardening) lighting is obviously important. 1000 watt electronic ballasts were the mainstay in the industry for a long time. And as it got more popular and Amazon brands popped up, there was also a HUGE uptick in EMI issues. Usually Cable service would get terrible for anyone in the area, and it came down to cheap lighting ballasts just BROADCASTING massive interference.
Usually this was discovered by someone else and they would call their Cable company and they'd send someone out who would at some point end up at the address, because the interference was so great, that their meters led them directly there.
Crazy important part of an electronic device.
reply
I actually deal with EMI noise in my field of work. In Controlled Environment Agriculture (indoor gardening) lighting is obviously important. 1000 watt electronic ballasts were the mainstay in the industry for a long time. And as it got more popular and Amazon brands popped up, there was also a HUGE uptick in EMI issues. Usually Cable service would get terrible for anyone in the area, and it came down to cheap lighting ballasts just BROADCASTING massive interference.
Usually this was discovered by someone else and they would call their Cable company and they'd send someone out who would at some point end up at the address, because the interference was so great, that their meters led them directly there.
Crazy important part of an electronic device.
reply
Bjorn
- 22:18 No loss? How? Does the Vf of the diodes magically disappear? Also, rectified 120V is not 120V, it's SQRT(2) Vrms-2 Vf with a regular bridge rectifier.
- Where's the PFC circuit? Is it the MOSFET rectification that Jonny is talking about?
- Bridge rectifier with mosfets: does the controller get its power from the 5Vsb, via a capacitive dropper, a separate transformer,...?
- Are there any MOVs inside Corsair's PSU's?
- What happens when the AC sine voltage is lower than the Vf of the diodes in the bridge rectifier? Does this get thrown away, lowering the efficiency and power factor or does the PFC circuit fix this?
- Do manufacturers still make PSU's with a voltage selector switch? Why (not)?
reply
- 22:18 No loss? How? Does the Vf of the diodes magically disappear? Also, rectified 120V is not 120V, it's SQRT(2) Vrms-2 Vf with a regular bridge rectifier.
- Where's the PFC circuit? Is it the MOSFET rectification that Jonny is talking about?
- Bridge rectifier with mosfets: does the controller get its power from the 5Vsb, via a capacitive dropper, a separate transformer,...?
- Are there any MOVs inside Corsair's PSU's?
- What happens when the AC sine voltage is lower than the Vf of the diodes in the bridge rectifier? Does this get thrown away, lowering the efficiency and power factor or does the PFC circuit fix this?
- Do manufacturers still make PSU's with a voltage selector switch? Why (not)?
reply
Dan
I absolutely loved Corsair's HXi series power supplies. I still have, in active use, an HX750, and HX750i, and an HX1000i. I could presently sell any of those for more than I initially paid for them, the HX750i could sell for three times what I paid for it. That demonstrates the popularity of these power supplies. Anyone know if Corsair are planning any return of the HXi series, or producing something similar with USB monitoring and 80+ Platinum rating?
reply
I absolutely loved Corsair's HXi series power supplies. I still have, in active use, an HX750, and HX750i, and an HX1000i. I could presently sell any of those for more than I initially paid for them, the HX750i could sell for three times what I paid for it. That demonstrates the popularity of these power supplies. Anyone know if Corsair are planning any return of the HXi series, or producing something similar with USB monitoring and 80+ Platinum rating?
reply
SSTC
3:10 470v volts, yeah...;
4:12 it is regulated;
4:30 wrong mosfets shown;
4:45 ...output caps... points at standby smps ;
5:10 ...5volt standby... points at PFC inductor ;
13:02 a gate drive transformer.
I understand that it could be a bunch of silly mistakes on camera, but at least, fix it in b-rolls and captions. Do not release bad quality info.
reply
3:10 470v volts, yeah...;
4:12 it is regulated;
4:30 wrong mosfets shown;
4:45 ...output caps... points at standby smps ;
5:10 ...5volt standby... points at PFC inductor ;
13:02 a gate drive transformer.
I understand that it could be a bunch of silly mistakes on camera, but at least, fix it in b-rolls and captions. Do not release bad quality info.
reply
itech
I have a question about PSUs. When my Seasonic SSR-850TR is plugged in from cold (as in, caps are empty), my UPS briefly clicks into overload & bypass - presumably from inrush current. Presumably it's possible to limit inrush current in PSUs, but presumably it was a design decision not to do that. What was the likely reason(s)? Was it a tradeoff for some other benefit?
reply
I have a question about PSUs. When my Seasonic SSR-850TR is plugged in from cold (as in, caps are empty), my UPS briefly clicks into overload & bypass - presumably from inrush current. Presumably it's possible to limit inrush current in PSUs, but presumably it was a design decision not to do that. What was the likely reason(s)? Was it a tradeoff for some other benefit?
reply
Siegbert
Nice one. Very informative. Did not know about that there is dedicated 5v standby rail. Got myself a Corsair RM850 and it perfectly handles my 3 GPUs (mining optimized settings, 400w total). Also one of the few psu that offers 4 8pin pci-e power connectors in this wattage class. Thankfully I got it for under 100 back then. Now its 150 msrp.
reply
Nice one. Very informative. Did not know about that there is dedicated 5v standby rail. Got myself a Corsair RM850 and it perfectly handles my 3 GPUs (mining optimized settings, 400w total). Also one of the few psu that offers 4 8pin pci-e power connectors in this wattage class. Thankfully I got it for under 100 back then. Now its 150 msrp.
reply
NeoDaOne
Love how enthusiastic Stone was getting with each explanation. Its fun to see folks get excited about stuff majority of folks dont think to realize.
Jonny - Your bridge rectifier is not always your bridge diode, you also have a thing called a mosfet bridge on some of the higher end products
Stone - Oooohhhh!
Good stuff GN
reply
Love how enthusiastic Stone was getting with each explanation. Its fun to see folks get excited about stuff majority of folks dont think to realize.
Jonny - Your bridge rectifier is not always your bridge diode, you also have a thing called a mosfet bridge on some of the higher end products
Stone - Oooohhhh!
Good stuff GN
reply
Lan
Whoever is still insisting on masks is way out of touch. If it s to virtue signal in videos, the fact is nobody cares about that anymore and actually it makes you look and sound weird. If it s because you think you re protecting yourself, the facts and even Fauci himself say masks don t work. So either way it s just pointless.
reply
Whoever is still insisting on masks is way out of touch. If it s to virtue signal in videos, the fact is nobody cares about that anymore and actually it makes you look and sound weird. If it s because you think you re protecting yourself, the facts and even Fauci himself say masks don t work. So either way it s just pointless.
reply
Add a review, comment
Other channel videos
