News: Trouble logging in? Log in from the login page.
September 19, 2020, 02:04:49 AM

Author Topic: Thoughts/advice on using thermistors with voltage doublers for current inrush?  (Read 109 times)

0 Members and 1 Guest are viewing this topic.

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
My PSUD model of the power supply of my Kaiju rebuild shows that the voltage doubler caps briefly draw a ton of current at turn on. I've doubled the capacitance of the doublers to reduce ripple even lower, but the increased capacitance causes even greater inrush current.  I'm a bit concerned by this and want to explore using a current limiter. Thermistors seem to be the go-to device. 

But before I explore thermistors further, I'd like to pick the brains of the BH experts as to possible pros and cons. Are there any reasons not to use thermistors with a voltage doubler?  Are some thermistors "noiser" than others and to be avoided?  This is new territory for me and I'd appreciate any pointers that people may have.

many thanks, Derek
« Last Edit: August 31, 2020, 02:10:18 AM by Deke609 »
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13536
How much is the inrush current?
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
The initial spike at turn on is around 18A.

many thanks, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13536
A UF4007 can handle a peak current of 30A briefly.

The power transformer is also going to have a ton of voltage drop if you try to pull 18A through it, though PSUD doesn't take that into account super well. 

A CL-90 between hot and the power transformer primary will slow this down, but it wouldn't seem to be explicitly necessary based on the information you're providing.

-PB
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
Thanks PB. I'll be using some beefier SiC Schottky rectifier diodes rated for 42A non-repetitive surge current, and for 21A of repetitive surge current: FFSP05120A.  I'm more concerned about the life of the doubler caps. I'll be using some long life Rubycon BXW 220u 400V caps. Datasheet attached. The datasheet doesn't mention max instantaneous ripple current -- maybe that doesn't exist? (But I thought I read about that somewhere when I was selecting caps a few months ago). According to PSUD, peak current drops to below 2A after 1500 ms, so maybe I don't need to worry about the doubler caps heating up.  Dunno. I'm totally ignorant here.

Good point about voltage sag @ 18A draw. PSUD shows a modest sag, but nowhere near as low as I would expect based on my limited experience.

Thanks for the suggestion of the CL90.  Is there a particular brand you would recommend? I may install it just out of an abudance of caution.

cheers and many thanks, Derek



Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13536
IMO you're miles away from needing a precharge circuit.  You are using 4x220u/400V caps for the doubler filter node?  (since you said more capacitance than stock)
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
Yeah, 4 caps: 2 pairs of paralleled 220u's.

many thanks, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
A follow-up on this. The caps are rated for 1.2A rms of ripple current at 120Hz.  Since I'll have two in parallel, I assume that they will split the current - perhaps not equally, but hopefully not entirely lopsidedly.  PSUD shows peak current at the doubler caps after 1500 ms of approx. 1.85A - so approx 1.3A rms.  So as long as one caps takes at least 100mA of that, I should be good after the turn-on surge. What I'm concerned about (out of ignorance) is that first 1500 ms when current bounces pretty high.  Is there no cause for concern?

many thanks again, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13536
What is the actual RMS current displayed on the data table for each of the doubler caps in PSUD?  I would doubt that it's 1.3A RMS.  That's a lot of ripple current for a high voltage supply.

If you believe the current bounces that high, multiply that AC current value times the DCR of the transformer.  That's instantaneous loss, and I bet it is several times as high as the transformer voltage rating. 
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
Huh. PSUD shows about 465mA rms.  I just took the peak value and multiplied by 0.707.  Guessing it's not a sine waveform?  But whatever the reason, that's a relief. I was a little surprised by the the high current peaks -- but I'm surprised by so much of this stuff, so didn't think to question my RMS calculation.

cheers and thanks, derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13536
It is most definitely not a sine waveform!  There's not really a negative peak ;)

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
Right! It's recitfied. 

So to cap this off: no reason to worry about inrush current?

cheers and thanks, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13536
You should always worry about everything, but the next step should be to quantify what you're worried about, then reexamine your paranoia.
Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man

Offline Deke609

  • Hero Member
  • *****
  • Posts: 1188
... then reexamine your paranoia.

You're freaking me out, man! How did you know I was paranoid?!!!   ;D

Many thanks again - I can get back to putting the amp together now. Finally finished the metal work (more than 300 holes!), so now on to the fun stuff.

cheers and thanks, Derek
Derek
______

Roon Intel NUC ->  Yggdrasil DAC -> BeePre (w/ BeeQuiet and EML 300B)  -> Kaiju (w/ DCFil and EML 300B) or Stereomour II (2A3 [EML Mesh] and 45 Conversion [EML 45B])  -> Audeze LCD4

Offline oguinn

  • Sr. Member
  • ****
  • Posts: 663
    • DIY Audio Gallery
You should always worry about everything, but the next step should be to quantify what you're worried about, then reexamine your paranoia.
Somebody ought to put that on a t-shirt
Jameson O’Guinn

-

Main System: Schiit Bifrost MB, Schiit Sol, Eros 2, BeePre, Kaiju/Stereomour II, Jägers, Mainline

Desktop System: Crack with Speedball