How much capacitance?

[Mike Hanson]

I put the smoothing module together on some strip board (a.k.a busboard). I was able to get both channels on the single board. I started with a basic plan for the layout, then marked it out for one complete channel, then just enough to keep my bearings on the second. To make the capacitors fit nicely, I used two strips for the ground running down the center:

I installed the resistors first. The 1R is on the upper area, oriented vertically. The 10R is horizontal, enabling the series connection with the inductors.

Then I added the capacitors and inductors:

Due to the size of the inductor, it sits at an odd angle, reaching over its series resistor:

Then I decided the reinforce the tracks by adding a bit of solder. Given the low current, they probably didn't need it, but I also figured it would visually clarify the circuit, for me to double check that everything was fine. I also created more breaks in the tracks, to ensure that all that solder didn't cause a short circuit:

It's not the tidiest soldering job, but I probed about and everything seemed fine. I soldered a bunch of wires to the inputs and outputs to finish the job.

Finally, I hooked up a little transformer with 12VAC secondaries plus an Avondale FWB rectifier module. Using my Variac for safety, I confirmed that no magic smoke came out, and both sides are acting exactly as expected.

Thanks for your guidance, everyone!

 

[martin clark]

@Mike Hanson You have slightly more caps and inductors than I think I expected : D

Mind scribbling a simple schematic of one rail for quick sense-check (though I'm sure it will be fine)

 

[Mike Hanson]

@Mike Hanson You have slightly more caps and inductors than I think I expected : D

Mind scribbling a simple schematic of one rail for quick sense-check (though I'm sure it will be fine)

I'm using your exact schematic, which was for a single rail. I have two rails for each of two channels, so 4 inductors and 4x3=12 caps.

 

[martin clark]

Right - gotcha. I missed that this is two , +/-v supply rails, at output. : )

 

[Mike Hanson]

Back to my thread's original theme of "How much capacitance", I will probably use a pair of big 20,000 µF caps for each channel's output stage (40,000 µF total), as I don't think I have enough room for my HackerCap modules. The HackerCaps have 6 caps each with a total of 60,000 µF, but because it's 6 in parallel, the ESR would be lower. Should I try harder to make them fit?

 

[Gervais Cote]

Depends……..

If you can use a pair of these below, you won’t need to worry about the ESR :



And they will last forever.

 

[Mike Hanson]

Depends……..

If you can use a pair of these below, you won’t need to worry about the ESR :



And they will last forever.

Impressive, albeit very expensive. I notice they also have a 47,000 µF version, so I could really shoot for the moon.

Fortunately, I've figured out a way to fit those HackerCaps into the build, so I think that will do fine.

 

[Gervais Cote]

Impressive, albeit very expensive. I notice they also have a 47,000 µF version, so I could really shoot for the moon.

Fortunately, I've figured out a way to fit those HackerCaps into the build, so I think that will do fine.

You know you owe us a few pictures of the final assembly now !

 

[Mike Hanson]

I've been playing with using this new capacitor bank board, and was somewhat surprised.

Between tests I'll drain capacitor banks by applying a 10Ω 50W resistor to the output. With this particular supply, after zapping it I would see the voltage climb up again over the span of minutes (quickly at first, then slowing). If I zapped it again, it would do it again. I start with about 150VDC from positive to negative. After zapping it, it will rise to about 40VDC. If I zap it a second time, it rises to 20VDC.

It took me a bit of pondering to suspect I was zapping only the last pair of capacitors. The others were "hidden" behind the resistor. However, I'm surprised that they would retain that much power. Would those intermediate components cause it to drain that slowly? How long should I apply the zapper to fully drain the supply?

 

[martin clark]

..welcome to 'dialectric absorption' ...!

This is why a final resistor across the supply rails, is always used. Veh esp for HV rails like valve supplies - but the trick is, not to load-down so hard, the resistor runs hot forever.

10k/1w is, at a guess / very roughly/ about the ballpark you want for this application, +/- 20%, but do: run you own numbers

 

[Mike Hanson]

Thanks, Martin. The 10K resistor would draw only 7.5mA from 75V (x2). The input stage takes only 30mA, so a total of just 45mA. The transformer can manage 330mA, so lots of headroom.

I think I may need to go that direction. It's going to be a tight squeeze to get those in there, but I should be able to make it work.