And that's why I design my circuits always to run at 3.3V. 😂 This was an eye opening experience. A circuit that kills itself is definitely undesirable. Never thought a diode capacitance can act to destroy itself. Brilliant!!!
Thank you professor. I was recently dealing with this problem as diode turn-off ringing due to stray/leakage inductance and junction capacitance was causing EMI issues(it was an output rectifier diode of a flyback converter). I have then used RC snubber to solve EMI issue. But with RC snubber input current THD got deteriorated. I see whenever we use dv/dt capacitor or snubber across switches, iTHD do get deteriorated. Can you please shed some light on it?
Hi. Thank you for this nice presentation. I'm curious how to apply it to flyback secondary diode. Should I use leakage inductance of transformer as Ld in this boost example? And I assume I should use delta Vd as not voltage above Vo, but above Vo+n.Vin_max. Correct?
Hello, Professor. I use the boost diode absorber in the APFC BOOST circuit, the R value is 51R, and the capacitance value is 150PF. This can improve EMI very well, but it will bring about MOSFET and diode heating problems. I debugged and found that the heating problem is much better when the capacitor value is 100PF, but the EMI is worse.
What a timely resource, I was just doing similar calculations with one of my team on a diode snubber for a two swicht forward circuit that we learned the hardway is really important, but we came up with a different calculation for the snubber losses at 7:29. As the resistor has to discipate the energy of the current flow in both directions, isn't it 2 X the Prs value you calculated. Thanks again for the great videos and if I am missing something I appreciate being corrected :-)
Hi Dave, Thanks for input. Yes indeed, there is a power loss during the charging time. I did not dive into it to keep the video short😊But.. it is not X2. This is because, during the cap charging, the resistor is dissipating both the charging energy AND the energy originally stored in the stray inductance LI^2/2. However... these losses are not to be taken in their average value because the duty cycle of the charge/discharge processes is small so the thermal impedance of the resistor is much smaller than the steady state thermal resistance and consequently the temp rise will be much lower than estimated by the average power loss. Not trivial.
Actually currently we have similar problem with our product - overvoltages on diodes during switching. The converter is of another topology (hard-switching half bridge dcdc), but still this lecture was truly helpful, thanks Professor.
Thanks Sir for excellent Video. But on 13:12, I am not getting similar V(BUS, LS) waveform in simulation on Ltspice. My second query is that in my flyback project, I am getting 620V reverse voltage from 5nS on MUR1560 diode. Though the diode has not failed, but is it acceptable. Thanks for your help. Regards
Sir, Thanks for your reply. I have put RC snubber with 22ohms and 470pF values. It resulted in 620V peaks across the didoe. If I increase 470pF to 680pF, I am getting 580V across diode , but the resistor of 22 ohms is getting very hot ( 145C against 115C with 470Pf) hence I decided to use 470pF to avoid heat in the resistor. Thanks again for your kind help. Kindly advise if there is scope for further improvement wrt to voltage across diode and heating in the resistor. Sanjay @@sambenyaakov
Hello, i'm currently woking on a led project where 10 LEDs are placed in series as a load, but during testing, I've noticed the the first diodes that are in touch with the output are the first to fail, would a snubber Network like this in parallel with each LED balance the peak current through each LED??, i was thinking of this soultion until i saw this video, my guess here is that first input LEDs are having to put up with the highest peaks and absorb extra power all the time, thanks for sharing
@Tauquir Iqbal oh i'm driving the led strip with a sepic constant current converter so, 10 leds are connected in series, and the maximum output is 130Vdc@1Amp, so I don't want series resistors as to keep high efficiency, but usually the 2 LEDs that get the energy input are the ones that fail, inner ones will be fine.
Is there difference in cooling of LEDs? Perhaps the first ones are getting hotter? Also , look for spikes and oscillations. The first ones will channel these to ground and may get damage by that.
That was sobering! I never thought such tiny inductance would cause such trouble!
Thanks for note.
Thanks prof! You speak with great clarity.
Thanks.
And that's why I design my circuits always to run at 3.3V. 😂 This was an eye opening experience. A circuit that kills itself is definitely undesirable. Never thought a diode capacitance can act to destroy itself. Brilliant!!!
Thanks for input.
Fantastic explanation and very helpful, thanks professor 🙏
Thanks
Thx for truly educating your viewrs, is there a way I can contact you, i need to ask you about a schematic
Thx in advance
sby@bgu.ac.il
@@sambenyaakov thank you sir, i have send you an email
Thanks professor. Very helpful presentation
Thanks
Thank you professor. I was recently dealing with this problem as diode turn-off ringing due to stray/leakage inductance and junction capacitance was causing EMI issues(it was an output rectifier diode of a flyback converter). I have then used RC snubber to solve EMI issue. But with RC snubber input current THD got deteriorated. I see whenever we use dv/dt capacitor or snubber across switches, iTHD do get deteriorated. Can you please shed some light on it?
THD. Is this a PFC converter?
@@sambenyaakov it's a single stage cc flyback design based on PFC flyback controller.
Hi. Thank you for this nice presentation. I'm curious how to apply it to flyback secondary diode. Should I use leakage inductance of transformer as Ld in this boost example? And I assume I should use delta Vd as not voltage above Vo, but above Vo+n.Vin_max. Correct?
Yes on both
Thanks for the video.🙏
How to obtain Ld?
Not trivial. In practice trial and error is used to select the cap.
@@sambenyaakov OK. thank you.
Hello, Professor. I use the boost diode absorber in the APFC BOOST circuit, the R value is 51R, and the capacitance value is 150PF. This can improve EMI very well, but it will bring about MOSFET and diode heating problems. I debugged and found that the heating problem is much better when the capacitor value is 100PF, but the EMI is worse.
Have you tried increasing the resistor?
What a timely resource, I was just doing similar calculations with one of my team on a diode snubber for a two swicht forward circuit that we learned the hardway is really important, but we came up with a different calculation for the snubber losses at 7:29. As the resistor has to discipate the energy of the current flow in both directions, isn't it 2 X the Prs value you calculated. Thanks again for the great videos and if I am missing something I appreciate being corrected :-)
Hi Dave, Thanks for input. Yes indeed, there is a power loss during the charging time. I did not dive into it to keep the video short😊But.. it is not X2. This is because, during the cap charging, the resistor is dissipating both the charging energy AND the energy originally stored in the stray inductance LI^2/2. However... these losses are not to be taken in their average value because the duty cycle of the charge/discharge processes is small so the thermal impedance of the resistor is much smaller than the steady state thermal resistance and consequently the temp rise will be much lower than estimated by the average power loss. Not trivial.
Actually currently we have similar problem with our product - overvoltages on diodes during switching. The converter is of another topology (hard-switching half bridge dcdc), but still this lecture was truly helpful, thanks Professor.
Thanks for sharing
Thanks Sir for excellent Video. But on 13:12, I am not getting similar V(BUS, LS) waveform in simulation on Ltspice.
My second query is that in my flyback project, I am getting 620V reverse voltage from 5nS on MUR1560 diode. Though the diode has not failed, but is it acceptable.
Thanks for your help.
Regards
5ns is probably OK, but what not pure a simple RC snubber?
Sir,
Thanks for your reply. I have put RC snubber with 22ohms and 470pF values. It resulted in 620V peaks across the didoe. If I increase 470pF to 680pF, I am getting 580V across diode , but the resistor of 22 ohms is getting very hot ( 145C against 115C with 470Pf) hence I decided to use 470pF to avoid heat in the resistor.
Thanks again for your kind help.
Kindly advise if there is scope for further improvement wrt to voltage across diode and heating in the resistor.
Sanjay @@sambenyaakov
@@sanjayagrawal6143620V for 5n is I think OK
Nomally, P(Rs)=E*f*2 because of the power of 2 of the voltage right ?
Pease indicate to which minute in video or slide number you are referring to.
What about the discharging of the ground capacitors ?
The snubber caps are in the 100s pF range.
Hello, i'm currently woking on a led project where 10 LEDs are placed in series as a load, but during testing, I've noticed the the first diodes that are in touch with the output are the first to fail, would a snubber Network like this in parallel with each LED balance the peak current through each LED??, i was thinking of this soultion until i saw this video, my guess here is that first input LEDs are having to put up with the highest peaks and absorb extra power all the time, thanks for sharing
@Tauquir Iqbal oh i'm driving the led strip with a sepic constant current converter so, 10 leds are connected in series, and the maximum output is 130Vdc@1Amp, so I don't want series resistors as to keep high efficiency, but usually the 2 LEDs that get the energy input are the ones that fail, inner ones will be fine.
Is there difference in cooling of LEDs? Perhaps the first ones are getting hotter? Also , look for spikes and oscillations. The first ones will channel these to ground and may get damage by that.
Thanks
🙏
👍🙏💖
🙏😊👍
Power Electronic Gems
😊🙏