Options to replace odd cap value

[Dr Tony Balls]

I'm recapping an old Ampeg Jupiter (B-42X) that is spec'd to have two 30uF 600V capacitors in the power supply. Schematic is below. The problem is that size is not exactly commonplace. Sprague makes a 600V cap but its 20uF, which is a bit off. So what would you do?

a) String two caps in series (say 100uF/300V and 47uF/300V) for an equivalent 32uF/600V cap

b) Not worry about the voltage and use a 500V cap (given that the schematic spec's 490V on these caps)

c) Not worry about the exact capacitance and use the aforementioned 20uF/600V caps.

https://ampeg.com/support/files/Schemat ... ematic.pdf

[Reeltarded]

490 is a little close for me to sub 500. If wall voltage creep is +10%, we're in danger!

I just googled and found a boatload of 600v axials and canned pairs in 33u.

Here is a mouser page. https://www.mouser.com/c/passive-compon ... 20VDC&pg=9

oh mouser i hate you

[Phil_S]

I think I'd opt for a pair of 68uF 350V to give you 34uF @ 700V. These are common enough in the $2-$3 range. Remember to use balancing resistors.

[martin manning]

I'm with Phil, go with stacked 350V, equal values, 68u. here are some nice ones: https://www.mouser.com/ProductDetail/Ru ... ltXQ%3D%3D Might depend on what kind of real estate you have to work with.

[Phil_S]

LOL, postage will cost more than the caps. Remember to add any other small parts you might need as the postage isn't likely to change very much.

[Reeltarded]

I'm with Phil, go with stacked 350V, equal values, 68u. here are some nice ones: https://www.mouser.com/ProductDetail/Ru ... ltXQ%3D%3D Might depend on what kind of real estate you have to work with.

or what conforms to original on a 50 year old amp you soul killing engineers.. jees

lol

[Cartman]

I have used Miec 600V electrolytics in several amps over the past five years or so and they seem to perform fine. Scroll down near the bottom of the price list below. The 30uF value is about $8 plus shipping. I think there is a seller on Amazon now too.

https://www.justradios.com/cart.html

[Dr Tony Balls]

Thanks for the input everyone!

[Stevem]

Those Sprague 600 volt filters are just two 300s in the same case, no?

[R.G.]

Before going too crazy running down exact filter caps for a power supply, it may help to remember what power filter caps do, and how electrolytics work.

Electrolytics have big tolerances compared to other parts. It was common for electros to have a new tolerance of -20% +100% up into the 1980s. Even with today's much more advanced manufacturing, they're commonly +/-20%. An amp from the "golden age" might have had filter caps with an actual value twice what the label said. A nominal 30uF cap at today's common +/- 20% might be 24uF to 36uF. Paying extra for tight tolerance filter caps might get you selected caps (which can then drift, another thing electrolytics do).

Back at what they do, power filter caps... filter power. The entire reason an amp designer puts in electrolytic filter caps is (1) to reduce ripple from the rectification and (2) to act as a low impedance AC shunt to keep stages from interacting. There have been and will continue to be arguments over "sag", but that's not why electros are there.

If I were trying to replace 30uF filter caps, I'd use 33uF modern caps and not look back. Or any series/parallel combination that was close.

I would dearly love to watch an honest blind test of picky guitarists trying to identify an amp which differed only in the power filter cap value. And the guitarists' faces upon seeing the results.

[LOUDthud]

I would dearly love to watch an honest blind test of picky guitarists trying to identify an amp which differed only in the power filter cap value. And the guitarists' faces upon seeing the results.

I know a guy who swears he can FEEL the difference between 16uF and 22uF in a tweed Deluxe. And it's not Neal Young.

[R.G.]

I know a guy who swears he can FEEL the difference between 16uF and 22uF in a tweed Deluxe. And it's not Neal Young.

Yeah. He's one of the ones I'd love to watch.

One of my friends is the amp-tech who does all of Eric Johnson's amps. I have watched him and Eric interview amp sounds A-B listening to the direction that the same exact resistor is soldered into position, swapping end for end a few times.

It is easy to concoct a situation where nearly anyone can hear a difference in component values. The further away from the intended value a component is, the easier to hear. This is especially true with things like filter caps being too small - the ripple rises a lot. Too big - that's harder. Once you suppress ripple enough, you get dramatically diminishing returns on ripple audibility. I fully believe that "sag" is ... something... , but I suspect that it is hard to quantify, and even harder to design for.

[roberto]

Some other options:
- if you don't want your nephews will ever need to substitute it: 6€ Mouser No:871-B32320I0306K000
- alternative with lower capacitance: 7 € Mouser No:505-DCP4I052507GD4KS
- standard electrolitic alternative: 8 € Mouser No:871-B43541B8476M000

[roberto]

Once you suppress ripple enough, you get dramatically diminishing returns on ripple audibility. I fully believe that "sag" is ... something... , but I suspect that it is hard to quantify, and even harder to design for.

What about a gyrator just after the rectification to eliminate the ripple (cheaper than a choke for screens only), then a small resistor in series to emulate PT secondary resistance, then small(er) capacitors to focus on sag only?

[LOUDthud]

Wouldn't you need an even bigger power supply to run the Gyrator ? As you can see from the simulation, most of the sag at the screen node and downstream is really caused by the ripple being filtered out by the choke and by PT losses at the reservoir cap.

I don't know if the resistances of the PT and choke are reasonable. Stepped currents are an approximation for a 50W amp.

50W_sim_1.GIF