Author Topic: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?  (Read 1843 times)

0 Members and 1 Guest are viewing this topic.

Offline theleakydiodeTopic starter

  • Regular Contributor
  • *
  • Posts: 118
  • Country: gb
  • I've got no idea what I'm doing.
I've found lots of papers about them online but can't find practical examples or people discussing actual circuits using them, has anyone here ever used this scheme? And if so are they generally a good substitute for lossy RCD snubbers or more complex active clamps?

I assume the diodes themselves will need snubbers too.
« Last Edit: November 22, 2022, 11:17:31 pm by theleakydiode »
 

Offline boB

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: us
    • my work www
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #1 on: November 22, 2022, 11:26:02 pm »
Rudy Severns may explain this one in his free snubber design book.

https://rudys.typepad.com/files/snubber-e-book-complete.pdf

I have used that particular snubber on a flyback and it worked well.

It returns the snubbed energy back to the input source rather than dissipating it in the discharging of the capacitor.

boB
K7IQ
 
The following users thanked this post: theleakydiode, spostma

Offline theleakydiodeTopic starter

  • Regular Contributor
  • *
  • Posts: 118
  • Country: gb
  • I've got no idea what I'm doing.
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #2 on: November 22, 2022, 11:32:08 pm »
Rudy Severns may explain this one in his free snubber design book.

https://rudys.typepad.com/files/snubber-e-book-complete.pdf

I have used that particular snubber on a flyback and it worked well.

It returns the snubbed energy back to the input source rather than dissipating it in the discharging of the capacitor.

boB
Thanks for the link. What were some real world limitations you found whilst using it and did those diodes need any additional RC snubbers? I've found them to be quite ringy.

But it's odd how they're not more widely used.
« Last Edit: November 22, 2022, 11:37:00 pm by theleakydiode »
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6686
  • Country: nl
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #3 on: November 24, 2022, 01:24:05 am »
Can't you use Schottky's?
 

Online coppercone2

  • Super Contributor
  • ***
  • Posts: 9167
  • Country: us
  • $
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #4 on: November 24, 2022, 01:27:57 am »
I accidentally saw a circuit like that kind of work in a simulation before
 

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3285
  • Country: fr
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #5 on: November 24, 2022, 07:06:33 am »
Rydy Severens was an old friend, very specialized in SMPS topology.

Low dissipating snubber, by return of power to bus, a very old idea.

See Vicor line of  SMPS, patented by Vincerelli decades ago, basis of very efficient and compact DC-DC modules.

Jon
Jean-Paul  the Internet Dinosaur
 
The following users thanked this post: boB

Offline boB

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: us
    • my work www
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #6 on: November 24, 2022, 07:10:09 pm »
The circuit shows Schottky's I think.

Have to choose an approximately correct inductor as well I believe.

boB
K7IQ
 

Offline theleakydiodeTopic starter

  • Regular Contributor
  • *
  • Posts: 118
  • Country: gb
  • I've got no idea what I'm doing.
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #7 on: November 26, 2022, 08:45:05 pm »
Can't you use Schottky's?
Doubt it in my case, there's a few hundred volts coming from the primary coil at turn off.
Rydy Severens was an old friend, very specialized in SMPS topology.

Low dissipating snubber, by return of power to bus, a very old idea.

See Vicor line of  SMPS, patented by Vincerelli decades ago, basis of very efficient and compact DC-DC modules.

Jon
Cool! Got any schematics you could share? And has he ever chatted with you about these snubbers? I've had a look in that pdf but this style of snubber isn't really covered in depth.
I accidentally saw a circuit like that kind of work in a simulation before
I've ran some basic LTspice simulations and built a flyback back converter using them, but I'm wondering if there's any real world pitfalls since they seem to be uncommon.
The circuit shows Schottky's I think.

Have to choose an approximately correct inductor as well I believe.

boB
I'm using GUR460 diodes and an aircored inductor of a few uH.
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #8 on: November 27, 2022, 02:14:11 am »
I have done some simulation in qucs, and so far I give this snubber a fail.

I don't remember this compound snubber, but it might have come from the old days " My Generation" the Who, 1965 when waveforms were more trapezoidal.

Capacitor has charge(+) while switch_Q is On. At Off, capacitor takes a pulse of current and ends
with charge(-). That is good if capacitor and diode can take the di/dt in a circuit with no defined R.

At next On transition, switch_Q has to take the load current, plus a pulse re-charging the capacitor via the low side diode and (hopefully lossy) inductor.

On EMI side, in total, there are 6 second derivative peaks of I_snubber, the highest one being  in Off time when the dv/dt abruptly changes from (+) to (-)

Perhaps I have used components with low side excessive di/dt in my simulation, but I think the  expectation of high side diode and capacitor reliability  is in doubt.
In a high power converter, the high side diode will have to be rated for the falling di/dt into recovery.
« Last Edit: November 27, 2022, 02:26:04 am by mag_therm »
 
The following users thanked this post: theleakydiode

Offline theleakydiodeTopic starter

  • Regular Contributor
  • *
  • Posts: 118
  • Country: gb
  • I've got no idea what I'm doing.
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #9 on: November 27, 2022, 06:14:56 pm »
I have done some simulation in qucs, and so far I give this snubber a fail.

I don't remember this compound snubber, but it might have come from the old days " My Generation" the Who, 1965 when waveforms were more trapezoidal.

Capacitor has charge(+) while switch_Q is On. At Off, capacitor takes a pulse of current and ends
with charge(-). That is good if capacitor and diode can take the di/dt in a circuit with no defined R.

At next On transition, switch_Q has to take the load current, plus a pulse re-charging the capacitor via the low side diode and (hopefully lossy) inductor.

On EMI side, in total, there are 6 second derivative peaks of I_snubber, the highest one being  in Off time when the dv/dt abruptly changes from (+) to (-)

Perhaps I have used components with low side excessive di/dt in my simulation, but I think the  expectation of high side diode and capacitor reliability  is in doubt.
In a high power converter, the high side diode will have to be rated for the falling di/dt into recovery.
So it's very harsh on the snubbing components and requires expensive capacitors to take the current pulses?
 

Offline boB

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: us
    • my work www
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #10 on: November 27, 2022, 06:51:26 pm »

I used this circuit in a flyback converter on years ago and it worked pretty well as I remember
K7IQ
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21566
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #11 on: November 27, 2022, 08:17:31 pm »
Note that the current pulse is given by Vpp / Zo, where Zo = sqrt(L/C) in the loop.  So, if it's "excessive", you've probably used too much C and not enough L.

It is a downside, that to get complete reset in a minimum duration on-pulse, the peak current may need to be comparable to the nominal load current.  The trouble with linear networks is, they're linear, everything's proportional.  So, it's hard to have operation remain useful over a wide operating range.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #12 on: November 28, 2022, 04:38:45 pm »
The snubber of this thread title was described by:
Domb, Redl and Sokal,
"Nondissipative turn-off snubber alleviates switching power dissipation, second-breakdown stress and VCE overshoot: Analysis, design procedure and experimental verification"
Published 1982

Paywalled by IEEE, not a member, I might try to get a copy through Engineers Australia


The latest converter project here is mostly completed and running. it is a converter from 32V DC link to 14V 20Amp part of the 85 A.h. DC Ups for  ham radio gear.
The converter presently has what I call "floating" snubber.
Here is qucs showing the "floating" snubbers. They do return part of the energy to the DC link, and they only dissipate about 2 Watt in this converter.
https://app.box.com/s/2orjo7boyqa74c0siq88yy8ydzspus4i
"Floating" Snubbers only conduct when the collector voltage is > 2 * 36V
So they do not reduce the off dv/dt which is set by the IGBTs.

Here is the spectrum today of the converter with "floating" snubbers running at 7.9 Amp output.
 Measuring on the DC(+12V) output bus via a pad: Level is about -88dBm to the pad,from 2 to 20 MHz
marker at 18 MHz. (Note this spectrum is using snubbbers different from subject of this thread !)
https://app.box.com/s/vtljdvw7wqwh1odq3h1zqvxo0m695j9g

----------------------------

Thanks Tim for your post #11,
I will continue with that in qucs with R being of the transformer and the ESR of the DC Link capacitor.
Here is qucs circuit for the "Domb" snubbers that are subject of this thread:
https://app.box.com/s/vw2jfnoy2ejb40ch5be32pc9va1fc8nb
I am presently having difficulty (long solving time, instability) with the transient solver due to snubber di/dt .
Time steps are down to 5 ns but using the whole inverter circuit, it take 600 us for the start transients to die out sufficiently to zoom in on snubber.
Also note that I have added Miller capacitors to the IGBTs to slow down the voltage rise time.
That is for qucs solution and I might try it on next real converter too.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21566
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #13 on: November 28, 2022, 04:48:12 pm »
If it's freezing up (so to speak) on the falling edge, maybe some loss/damping (R+C?) on the R+D is needed?  I don't know QUCS quirks personally but that can give regular SPICE some trouble.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #14 on: November 28, 2022, 05:12:30 pm »
I just got the first complete run with the zoom at the end of 600 usec.
The circuit is :
https://app.box.com/s/vw2jfnoy2ejb40ch5be32pc9va1fc8nb
Note that C_Miller is used to slow IGBT transitions.

The graphs:
https://app.box.com/s/6ix8or7x1cpmamp9qpj9jeqxxm3nwyzf

Looks like smaller steps needed. Note(bottom graph) that the body diodes are now conducting
 

Offline Neilm

  • Super Contributor
  • ***
  • Posts: 1545
  • Country: gb
Re: Anyone used this non-dissipative snubber design for SMPS DC-DC etc?
« Reply #15 on: November 28, 2022, 07:15:08 pm »
I have used this circuit in flyback transformers. It does need some careful work to ensure you don't overstress stuff and you do need to discharge on a cycle by cycle basis.
It is more expensive and you may find you need a weird value or high pulse current for the inductor which makes it expensive.

In the application I used it for, it replaced a 20 W resistor which improved the efficiency of the converter and saved the battery life considerably
Two things are infinite: the universe and human stupidity; and I'm not sure about the the universe. - Albert Einstein
Tesla referral code https://ts.la/neil53539
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf