Best capacitors for power supplies update

Jump to Latest
#22
popchops said:
........

So... I'm currently looking at the EPCOS Sikorel B41570 which I would mount horizontally off the heatsink in a 2U case. 5mOhm ESR and ripple rating off the scale. And lifetime is about a million years which is good because I plan to do this once.🙄

..........
Click to expand...

I actualy run the 550 and 560 series from EPCOS TDK , they cost alf the price of the 570 with great result 😉


.
 
#25
my heatsinks will be the case walls... not internal. Not likely to be more that 10 deg above ambient and the caps are rated for 105deg so prob ok.

Just mounted off the walls for convenience not for cooling - it means you can get the 80mm or even 105mm can into a 2U case if you bolt it in horizontally.
 
#32
carlmart,
About shunt regulator:
Initially I used regular Erno Borbely regulator (EB208/418 design), and I had a decent sound. I read few articles about "Super Shunt" regulator, and decided to build it. Unfortunately, Erno Borbely closed his business and I was not able to purchase original kit, so I bought clone PCBs on ebay (Jim Audio). After I built it (all parts genuine) I had self-generation, so I was able to remove it only increasing capacitors between gate and drain of 2SJ148 and 2SK982 from 47pf to 470pf (really, about
400pf helped). And, I was really disappointing with the sound - although bass became more "filled", something was wrong, the sound was much worse than with original regulator. I had such feeling, that the problem is in PCB layout, but I had no idea how to fix it. Once I saw the picture of original EB208/418 layout, and figured out, that voltage reference and voltage divider has separate "signal" ground, that connected to main ground like "star" at the output of regulator. I draw similar topology for shunt,
build it - again self-generation, but after few tests I was able to remove it with capacitors 75 pf(instead of 470pf). I connected and listened - and, or miracle, I heard decent sound, with "filled" bass better than regular voltage regulator.
Finally, I replaced 2SJ313 to 2SJ201 and 2SK2013 to 2SK1530, as I wanted to use high shunt current (resistor 2.2R in current source). The sound did not change and no self-generation. So, conclusion is that Erno Borbely "Super Shunt" regulator is very
sensitive to PCB layout, Jims Audio clone does not work as it should (although I can not see self-generation on my scope).
And, yes, to buld Borbely preamplifier is very hard now, the main issue is input matched octet of 2SK170BL/2SJ74BL,
I was able to select few, so I have some materials from my games🙂 But the amplifier that you build requires matched
quad of 2SK147/2SJ72 - to obtain them matched is even harder than 2SJ74/2SK170, do you have them? I would like to build
John Curl Vendetta to compare with Borbely EB199/320, but still can not find 2SK147 normally (less than 0.05 mA) matched.
In Borbely preamplifier I use octet matched better than 0.02 mA - it took a lot of time to select and retest JFET to get it.
 
#34
alweit said:
carlmart,
About shunt regulator:
Initially I used regular Erno Borbely regulator (EB208/418 design), and I had a decent sound. I read few articles about "Super Shunt" regulator, and decided to build it. Unfortunately, Erno Borbely closed his business and I was not able to purchase original kit, so I bought clone PCBs on ebay (Jim Audio). After I built it (all parts genuine) I had self-generation, so I was able to remove it only increasing capacitors between gate and drain of 2SJ148 and 2SK982 from 47pf to 470pf (really, about
400pf helped). And, I was really disappointing with the sound - although bass became more "filled", something was wrong, the sound was much worse than with original regulator. I had such feeling, that the problem is in PCB layout, but I had no idea how to fix it. Once I saw the picture of original EB208/418 layout, and figured out, that voltage reference and voltage divider has separate "signal" ground, that connected to main ground like "star" at the output of regulator. I draw similar topology for shunt,
build it - again self-generation, but after few tests I was able to remove it with capacitors 75 pf(instead of 470pf). I connected and listened - and, or miracle, I heard decent sound, with "filled" bass better than regular voltage regulator.
Finally, I replaced 2SJ313 to 2SJ201 and 2SK2013 to 2SK1530, as I wanted to use high shunt current (resistor 2.2R in current source). The sound did not change and no self-generation. So, conclusion is that Erno Borbely "Super Shunt" regulator is very
sensitive to PCB layout, Jims Audio clone does not work as it should (although I can not see self-generation on my scope).
And, yes, to buld Borbely preamplifier is very hard now, the main issue is input matched octet of 2SK170BL/2SJ74BL,
I was able to select few, so I have some materials from my games🙂 But the amplifier that you build requires matched
quad of 2SK147/2SJ72 - to obtain them matched is even harder than 2SJ74/2SK170, do you have them? I would like to build
John Curl Vendetta to compare with Borbely EB199/320, but still can not find 2SK147 normally (less than 0.05 mA) matched.
In Borbely preamplifier I use octet matched better than 0.02 mA - it took a lot of time to select and retest JFET to get it.
Click to expand...

I am not sure what you mean by "self-generation". Is that distortion or oscillation?

It seems you only tried shunt regulators on your preamp. The original Borbely preamp had a Sulzer regulator, modified by EB.

The ones I thought might be interesting to try are the Jung Super Regulators. But you are right one one thing about their layout: grounds have to be separated and joined as Jung shows on the original articles.

In fact that ground separation should be used in all regulators.

Separating signal and supply grounds is also essential in the preamp PCB, but I'm sure EB provided for that.

Jim's audio has clones for EB's all-fet preamp and for shunt regulator. Maybe the grounds can be cut and arranged as they should with separate wires.
 
#35
Fabrice - at 40 deg C (or thereabouts) the B41560/80 caps do seem to last. ESR is as good as the 550/570. Much cheaper too. Since mine will be mounted off a sidewall heatsink I cant really see them getting much above 40 deg.

Also slightly narrower and available in single units from Mouser. Good choice Patrice.
 
Last edited:
#36
Simple Snubbing

The DA Universal PS PCB is a bit big for my application. I'm considering standalone integrated bridge rectifier and 2x 22mF filter caps for a +/-55V DC linear power supply.

I'm not sure if I'll need to add CRC snubbing to attenuate high frequency ringing so I want to lay out my single bridge and 2x caps to make this possible later if needed.

Hagermans 'optimum snubbing' calculation shows both the RC and C crossing from the AC side of the bridge to the ground return. Is this right? Seems like the DA Universal PS does not facilitate this arrangement. Not sure if anybody can help me figure out a minimalist snubbing circuit.

Has anybody done this without a PCB? Patrice?

Thx. Pops.
 
Last edited:
#37
A single secondary transformer gets a snubber across the transformer leads, if it needs one.

A dual secondary gets two snubbers, one across each secondary leads.

A centre tapped transformer needs two snubbers. one from each output lead back to the centre tap.

I have used a few PSU PCBs. But the majority of my builds hardwire the PSU from transformer through the rectifier and caps to the load.
 
#40
What I have read, and it makes sense, is that a snubber should also have a resistor, which is the really important part to absorb the ringing.

The question is that to find the ringing point you need to measure that.

And yes, if you use just a capacitor it is best to solder it directly on the diode pins.
 
You must log in or register to reply here.
Top Bottom