I'm currently working on a PhD about a DC-DC multiphase 3 level flying capacitor converter. It's always refreshing to see a clear and detailed presentation, it helps not to lose sight of the main goals whenever I'm making a big deal of small but tedious technicalities. In fact, I checked you channel and saw (by the titles) many videos that might help me in my investigations. I'll be sure to watch them. Thank you for your work, professor.
Congratulations on an excellent series of lectures. Another advantage of multiphase converter presumably is that the power dissipation will be shared between output transistors. So in high reliability/safety critical applications the semiconductor junction temperature rise (above the heat sink temperature) of each transistor will be reduced. Lower semiconductor temperature rise increases reliability (Arrhenius law).
This is another great video. I know this is a very old video but I would like to see how the number of stages affects the control loop and how to deal with compensation in designs that do phase shading. You have taught me so much thank you for your time.
Thank you so much for the great presentation. Do you think lossless current sensing is practical considering tolerance of inductors, resistors and capacitors? Thanks
Sam, have you written hard copy/print books on power electronics?? I respect your material sir and would love to read your book on this subject. RUclips is great but printed text is easier on the eyes for long hours. My opinion
@@sambenyaakov no problem. Even printed lecture notes could suffice. Slides that could be printed would work as well. Just I love how you analyze this topic from multiple angles. It's refreshing to see someone elaborate on power electronics discussing not just power stage but also multiphase converters and closed loop compensation, and practical topics such as EMC, PCB of power supplies etc.
Mirror current sensing has an issue. It will not follow the current ramp in main MOSFET, which is dictated by inductor in series to main MOSFET. But mirror MOSFET has a resistor which means no ramp and flat current once the MOSFET turns on completely.
Mr. professor, I can not figure out, why I have different currents in my 2-phase BUCK MPPT, at 20 kHz, currents are 7:5 AMPS, at 25 kHz curents are 17:17, but if I go higher 23:21 AMPS, so I need to change freg. for IR2104 to 28 kHz, than 22:22 AMPS is OK, but at 27:23 AMPS and 28 kHz I have again uneven currents. Where is the problem ? ESP32 to IR2104 connections ? IR2104 - resistors + diodes - IRFP4110/4227 FETs ? Connection FETs to 45 AMPS / 3.9 mOhms toroids ? Can problems arise from not properly designed snubbers atr FETs ? Is problem Trr and Qrr of the FETs ? Or problem is deadtime of IR2104 ? Or problem is low drive current at gates from IR2104 ? I need to know, what is with high probability the cause of the problem.
Dear Professor Yaakov Great and informative video as usual. How does one generate phase shifted pwm's ? And is there a risk of shoot through, if top gate of phase B is on at the same time as bottom gate of phase A ?
At 14:10 you say that load sharing leads to a higher switching frequency, beacause we have less switching losses. Why is that the case? If we share the current between many inductors we can make the switches smaller, but we also have more switches to turn on. Shouldn't the switching loss stay the same?
Hi, you got a point there. What I had in mind when saying this was the practical ability to work in borderline mode and the fact that many transistors have better thermal conduction so you can allow higher dissipation.
Hi Professor, Under low loads, we are doing diode emulation(DCM) to reduce conduction losses due to reverse current. However, isn't the diode more lossy than FET in terms of conduction losses?
Hi professor, recently multiphase buck converter using coupled inductor get many attention by claiming to have large inductance in steady state while small inductance during transient and therefore achieved small current ripple and fast transient, can you make a video to give a brief overview of it?
Professor Sam, do you have any other video where you go into more details on the bandwidth of the control mentioned at 18:26 ? I didn't understand this concept.
Also, you have to think about your mechanical structure. You will have one busbar collecting the output current on all buck phases. Some of those bucks will be closer to the load, some further, so the phases will have slightly different impedance between them and the load. Your cooling system can also be an issue : if you don't balance your cooling system (for example, you put a fan in front of your phases, but some of them are a bit far away from the airflow), then some phases will heat more than others for the same current passing through. Since higher temperature often means higher RdsON, it will lead to lower currents in those undercooled phases, and so higher currents in the other phases. In the end, this phenomenon tends to balance itself naturally (as you can logically guess), but putting some thoughts into your cooling system may help reduce the overall conduction losses, minimize fault risks, etc.
Thank you so much ! I have learned a lot from you sir . I had a question though, I have seen some multiphase dc-dc systems where they use multiple converter ICs controlling their individual phases but are connected through a clock synchronization input pins. Why it is used ? And how does it help ? Thanks in advance
Thanks for the video, im gonna look at it in more detail, but i wanna ask a question, what would you suggest if i wanna use that topology for high voltage conversion lets say converting 160 VDC to 28vdc @300a? İ couldnt find a suitable controller.. :(
Do you have experience with a multiphase buck with coupled inductors? It would be interesting to learn about the performance as compared to the non coupled case.
![Tee Grizzley - Robbery 8 [Official Video]](http://i.ytimg.com/vi/34_vkNV6wrU/mqdefault.jpg)
Pure gold, never seen anything explained so clear and you gave it for free.
Thank you sir!
Thanks
Yes excellent teacher
Thank you so much
I'm currently working on a PhD about a DC-DC multiphase 3 level flying capacitor converter. It's always refreshing to see a clear and detailed presentation, it helps not to lose sight of the main goals whenever I'm making a big deal of small but tedious technicalities.
In fact, I checked you channel and saw (by the titles) many videos that might help me in my investigations. I'll be sure to watch them.
Thank you for your work, professor.
Thanks. Comments like yours keep me going.
I did enjoy the presentation. It's great to have good lectures on more advanced topics out of school without pursuing a master's or PhD.
Thanks
Very well explained. Thanks for sharing your information and time. You are great teacher.
🙏😊
Congratulations on an excellent series of lectures. Another advantage of multiphase converter presumably is that the power dissipation will be shared between output transistors. So in high reliability/safety critical applications the semiconductor junction temperature rise (above the heat sink temperature) of each transistor will be reduced. Lower semiconductor temperature rise increases reliability (Arrhenius law).
You are absolutely right. I thought this is clear, but perhaps I should have emphasized the lower power dissipation (assuming low Rds(on) transistors)
This is another great video. I know this is a very old video but I would like to see how the number of stages affects the control loop and how to deal with compensation in designs that do phase shading. You have taught me so much thank you for your time.
Thanks for comment. Good points. Will try to prepare a follow up.
Excellent lecture. Even touched upon the magnetics design aspects briefly. Adding this to my bookmarked list.
🙏🙂
very great，professor，you explain so clearly，i have learned a lot from you，think you！
Thanks for the comment, I appreciate it,.
Thank you Professor for sharing your knowledge! It is a pleasure for me to hear you, because you deeply understand what you are talking about.
Thanks
Learn a lot from this video, thank you! Please take care, prof
Thanks.
Thanks a lot. It is right what i need for my new project. Really helpful.
👍😊
Thank you professor! Interleaved and multiphase is the same thing?
Yes
Sir, can you please make a similar video for coupled multiphase buck converter
Good subject, will try. Thanks
This was a very helpful video!
Thanks
Great video, do you have any video of design of the coupled inductor for multi phase Buck converter?
On my todo list
Thank you so much for the great presentation. Do you think lossless current sensing is practical considering tolerance of inductors, resistors and capacitors? Thanks
I guess you mean by RC? Is, it is practical and is being used in interleaved Buck converters. There are some reference designs of that.
Sam, have you written hard copy/print books on power electronics?? I respect your material sir and would love to read your book on this subject. RUclips is great but printed text is easier on the eyes for long hours. My opinion
Never got around to do that. Just too busy. Thanks for comment.
@@sambenyaakov no problem. Even printed lecture notes could suffice. Slides that could be printed would work as well. Just I love how you analyze this topic from multiple angles. It's refreshing to see someone elaborate on power electronics discussing not just power stage but also multiphase converters and closed loop compensation, and practical topics such as EMC, PCB of power supplies etc.
There are audio amps who use this, seems to work very nicely, but control is complex as you say. Delay differences can give problems?.
Good for this application.
Mirror current sensing has an issue. It will not follow the current ramp in main MOSFET, which is dictated by inductor in series to main MOSFET. But mirror MOSFET has a resistor which means no ramp and flat current once the MOSFET turns on completely.
www.ti.com/lit/an/slva859b/slva859b.pdf
Mr. professor, I can not figure out, why I have different currents in my 2-phase BUCK MPPT, at 20 kHz, currents are 7:5 AMPS, at 25 kHz curents are 17:17, but if I go higher 23:21 AMPS, so I need to change freg. for IR2104 to 28 kHz, than 22:22 AMPS is OK, but at 27:23 AMPS and 28 kHz I have again uneven currents. Where is the problem ? ESP32 to IR2104 connections ? IR2104 - resistors + diodes - IRFP4110/4227 FETs ? Connection FETs to 45 AMPS / 3.9 mOhms toroids ? Can problems arise from not properly designed snubbers atr FETs ? Is problem Trr and Qrr of the FETs ? Or problem is deadtime of IR2104 ? Or problem is low drive current at gates from IR2104 ? I need to know, what is with high probability the cause of the problem.
Do you have current sharing control?
Dear Professor Yaakov
Great and informative video as usual. How does one generate phase shifted pwm's ? And is there a risk of shoot through, if top gate of phase B is on at the same time as bottom gate of phase A ?
There are dedicated controllers, too complex and costly to build from discrete. Sure one should wtch it not to cause shoot through..
Thank you sir, excellent explanation.
Thanks. You are welcome to www.linkedin.com/groups/13606756/
At 14:10 you say that load sharing leads to a higher switching frequency, beacause we have less switching losses. Why is that the case? If we share the current between many inductors we can make the switches smaller, but we also have more switches to turn on. Shouldn't the switching loss stay the same?
Hi, you got a point there. What I had in mind when saying this was the practical ability to work in borderline mode and the fact that many transistors have better thermal conduction so you can allow higher dissipation.
Hi Professor,
Under low loads, we are doing diode emulation(DCM) to reduce conduction losses due to reverse current. However, isn't the diode more lossy than FET in terms of conduction losses?
Yes it is but since the current is small the absolute loss is small. If the transistor will be kept on the losses might be higher.
excellent video. very informative
Thanks
Hi professor, recently multiphase buck converter using coupled inductor get many attention by claiming to have large inductance in steady state while small inductance during transient and therefore achieved small current ripple and fast transient, can you make a video to give a brief overview of it?
Thanks for comment. I may add the issue of coupled iductors in interleaved.
@@sambenyaakov Isn't interleaved same as multi-phasing?
Professor Sam, do you have any other video where you go into more details on the bandwidth of the control mentioned at 18:26 ? I didn't understand this concept.
Sorry I don't. Try to look up application notes of multi phase controllers.
Hi Professor,
Why does a current imbalance occur?
Could you explain?
The channels can never be equal. It takes only a small change in duty cycle, for example, to make a large change in current.
@@sambenyaakov Thanks for answer!
Also, you have to think about your mechanical structure. You will have one busbar collecting the output current on all buck phases. Some of those bucks will be closer to the load, some further, so the phases will have slightly different impedance between them and the load. Your cooling system can also be an issue : if you don't balance your cooling system (for example, you put a fan in front of your phases, but some of them are a bit far away from the airflow), then some phases will heat more than others for the same current passing through. Since higher temperature often means higher RdsON, it will lead to lower currents in those undercooled phases, and so higher currents in the other phases. In the end, this phenomenon tends to balance itself naturally (as you can logically guess), but putting some thoughts into your cooling system may help reduce the overall conduction losses, minimize fault risks, etc.
@@mythicdawn9574 Thank you so much for your answer.
Thank you so much ! I have learned a lot from you sir .
I had a question though, I have seen some multiphase dc-dc systems where they use multiple converter ICs controlling their individual phases but are connected through a clock synchronization input pins.
Why it is used ? And how does it help ? Thanks in advance
You obviously need to synchronize the multiple converters.
Sir when different dc source is connected during off state off switch in buck....does system work as buck overall with common inductor for both states
Not sure I follow your question.
@@sambenyaakov can u please give ur mail..i will upload question i have
Thanks for the video, im gonna look at it in more detail, but i wanna ask a question, what would you suggest if i wanna use that topology for high voltage conversion lets say converting 160 VDC to 28vdc @300a? İ couldnt find a suitable controller.. :(
Multi phase will be OK but you may have to build your own controller preferably around a MCU.
Hi prof, If my input is 24V,+-2V and output is 12V,20A max ..would you suggest two level synch buck converter or single level synch buck converter.
Eff should be greater than 90%
Two level or interleaved?
This is un-coupled multiphase
Do you have experience with a multiphase buck with coupled inductors? It would be interesting to learn about the performance as compared to the non coupled case.
E X C E L L E N T ^n
Thanks😊