Rating ecaps.

[Mike_J]

In the never ending quest to make the growler I started out with 22uF capacitors in a dog house the way Fender Bassman were made many years ago. I used F&T axial lead caps which are fine caps. Prior to that time someone read that his ecap blew up and of course it didn't register with me but later it did after I installed these caps the same way Fender did in 1959.


Thinking that the higher AC line voltage contributed to the exploding ecap I went on a quest to rectify this risk factor. Not hard just stole the idea the blackfaced amps had. Couldn't use the F&Ts for this because they were too big to fit in the dog house. So I procure some caps that would fit, two 100uF @ 350VDC for the first stage which gave me 700VDC and just to be super safe went with two 47uF @ 350VDC for the screen supply another 700VDC. We are protected now.


Well then what do I read by Merlin but "The smoothing capacitor should therefore be positioned close to the valve stage, and the connections to it should be kept as short as possible (radial capacitors are sometimes preferred for this reason)." Of course the dog house is in another galaxy from where the valve stages are making it impossible to make short connections. So I reason although I don't know how reasoned it is, let's get some radial caps because we can get them close enough to make short connections.


Purchased radial caps from an online company some time ago but they still haven't come in. Figured out a way to fit them on the board for short connections. Plan on using the F&T axial ecaps for the preamp and LTP. Cement the preamp ecap in so I have no idea how I will ever get it out.


Then I look at the chassis for the reissue Bassman. They have two 22uF in parallel for the  preamp. Remember back to when I first made this amp and recall using an 8uF @ 450VDC for the preamp. Figured Fender buys so many 22uF @ 500VDC ecaps that they probably get a price break that makes it cheaper to install two of the 22s than one of the 8s because they wouldn't buy anywhere near as many of those.


So I have a few choices. Stick with the F&T and go with 22uF for the preamp filtering but oh no that would be far too easy. Remove the F&T at considerable effort and replace it with an 8uF which of course I have on hand. Oh no, again a little too easy. Purchase some 22uF radials and install them right on the board in parallel. Only will have to wait weeks for the supplier to provide them. Paid for two day shipping.


Now I am trying to justify all this effort by saying I will have better caps. Have no idea what the rating on the F&Ts are but do know the ratings on the axials and radials I purchased.


                                                                          Radial          Axial
100uF
- Ripple Current                                              .800mA       .460mA
- Maximum Temp                                             105C             85C
- Life                                                                  10K Hrs         2K Hrs

47uF                                                                 
- Ripple Current                                              .970mA        .450mA
- Maximum Temp                                             105C               80C
- Life                                                                   5K Hrs            5K Hrs

Seems like the radials were an upgrade in this case don't you think? Sure it doesn't apply to all the caps but at least in this case seems obvious to me.

[Mike_J]

Have attached a picture of the schematic showing where the two caps in parallel are. The same schematic has a .047uF cap across the entrance to the standby switch to the ground side of the first stage cap. What is this?

[PRR]

> what do I read by Merlin but "The smoothing capacitor should therefore be positioned close to the valve stage, and the connections to it should be kept as short as possible

How big is your amp? Feet? yards?

This isn't transistors. You do not have to be obsessive about length. Merlin is right of course but you seem to be trying to take it too far (too short). Millions of Fender dog-houses have played just fine.

[Mike_J]

> what do I read by Merlin but "The smoothing capacitor should therefore be positioned close to the valve stage, and the connections to it should be kept as short as possible

How big is your amp? Feet? yards?

This isn't transistors. You do not have to be obsessive about length. Merlin is right of course but you seem to be trying to take it too far (too short). Millions of Fender dog-houses have played just fine.

What can I say I like to go overboard and I am experimenting. I am retired. What else am I going to do? We will see how it works. If it doesn't the dog house is still full of caps and all the holes are still in the board. It looks like swiss cheese.

[Mike_J]

> what do I read by Merlin but "The smoothing capacitor should therefore be positioned close to the valve stage, and the connections to it should be kept as short as possible

How big is your amp? Feet? yards?

This isn't transistors. You do not have to be obsessive about length. Merlin is right of course but you seem to be trying to take it too far (too short). Millions of Fender dog-houses have played just fine.

PRR, you are very knowledgeable. Two obvious things are a cap with higher maximum operating temperature and life expectancy is good. Not so sure about the ripple current. Is bigger better. Have heard some people say it is a good thing and others say it doesn't make much difference. Do you have an opinion concerning ripple current?


Thanks
Mike

[PRR]

> opinion concerning ripple current?

Not an issue in speech/music amplifiers. We design for low power ripple Voltage, which makes low ripple current in large cans, not a problem. At Full Power Output an Output Cap (some transistor amplifiers) may exceed rated ripple current in bench-tests, but never in normal operation.

And how long do you expect your amp to live? Bad old parts in 1960s amplifiers often play OK today. E-caps are service parts anyway. (One reason to keep them in a remote doghouse.) What kills amps are abused tubes and tinfoiled fuses melting power transformers. Like most manmade disasters, a combination of things.

[pdf64]

My thinking is that ripple current is important for reservoir caps, as, during high loading, that will be having to work hard to fill in the dips between the rectified pulses.
My perception is that a valve guitar amp, although an outwardly similar application to a speech / music, is actually rather different because it has to accommodate continuous heavy overdrive (at least when I use them anyway ). Hence the reservoir cap will be passing continuous heavy ripple current.
Dunno how to assess the magnitude of that ripple current though, simpler just to use caps with a good spec, where required.

[PRR]

Mocking-up a DeLuxe power supply, I figure 144mA RMS in the cap.

A random CDE 40u 450V cap is rated Ripple Current 653mA @ 120Hz.

A larger 400u cap takes nearly the same ripple current, but at about 1/5th the ESR and about 4X the surface area to throw-off the heat.

Ripple current rating is for HIGH ripple applications. Welders, not audio amps.