power supply capacitor size in tube amps

[TomCarlos]

Can I chime in on an old post?

I too am looking for something physically smaller than the F&T Type A axial caps. I will open a thread in the next day or two on a Yamaha Tube Amp I am working on. The caps that were in there could hardly fit on the Power Supply PC board and the F&T Type A 450 volt caps were even worse!! So I am trying to find an alternative... either from F&T or any quality brand.

[xtian]

I'd advise you to use Nichicon or Panasonic radials. They're smaller than "made for tube amp" axials, but just as good, much less expensive, and they come in a variety of shapes, from short and stout to tall and skinny. Solder the +ve lead into the PCB, and lengthen the -ve lead if needed.

[Leo_Gnardo]

I'd advise you to use Nichicon or Panasonic radials. They're smaller than "made for tube amp" axials, but just as good, much less expensive, and they come in a variety of shapes, from short and stout to tall and skinny. Solder the +ve lead into the PCB, and lengthen the -ve lead if needed.

Good suggestion xtian. I recall Nichicon "pencil" caps rated 450V being mentioned at Audio Asylum as ideal for re-stuffing multisection can caps. If they're that skinny, should be an easy fit in your Yamaha. I also remember seeing them offered at Mouser. Good hunting, let us know how it goes.

[SoulFetish]

I dunno. I think I might be having a crisis of convictions. I famously held the opinion that with power film caps becoming increasingly smaller and cheaper, that the advantages in performance and life expectancy made them worth the increase in size wherever possible (okay, maybe not famously. Nobody really gives a sh*t how I feel about capacitors...).
Maybe I'm wrong. I mean, i'm having a tough time trying to make apples to apples comparisons looking at the datasheets.

Take the C4AQ line of film caps by Kemet – they have a max passive temp. rating of 125˚ with a life expectancy of 10k hours at a rated hot spot temperature (Thsp) of 105˚, and 100k hours at T_hsp of 85˚. Where as the two lines of electrolytic caps nsubulysses mentioned have life cycles rated from 10-12k at 105˚. But how do you factor in such a huge disparity between expected ripple current ratings between the film and electrolytics; where the electro's max out in the low hundreds of mA at 120Hz and the film caps provide ripple current figures in the 10s of amps at 10kHz. Obviously, the film caps were designed and marketed for switching supplies.

The other thing to consider is the low max voltage rating 450V of the aforementioned electrolytic caps. Not an issue for amps with lower idle HT supplies, but if you require two series connected caps, then the real estate ends up almost being a wash for the single higher voltage film cap.
What do you guys think?

[Leo_Gnardo]

The other thing to consider is the low max voltage rating 450V of the aforementioned electrolytic caps. Not an issue for amps with lower idle HT supplies, but if you require two series connected caps, then the real estate ends up almost being a wash for the single higher voltage film cap.
What do you guys think?

I noticed that those "pencil" caps topped out at 450V. Yes, it's a bit disappointing if you have to entertain higher voltages. Also, unless youi know you're going to be mounting filter caps in a hot place, like right atop the power transformer (I'm looking at you, Reinhold Bogner) you won't be approaching the 125C temperature mark. Operated at a lower temp, you'll be buying a longer lifetime for those caps, all to the better.

If it's a tossup between fiddle faddling series caps or choosing film caps, I'd opt for the films. Dare to be different, the future is here and it doesn't include flakey old fashioned electrolytics.

[TomCarlos]

MEF:

I wanted to circle back to the thread, post a couple photos, and add some closing comments. The photos come from two nearly identical amps I have been working on. They are both Yamaha / Soldano designs. One it a T50C and the other the T100C.

The photo showing the black caps comes from the T50C. I am not sure that these are original because as you can see, they are HUGE for the PC board. The PC board has two 80UF on the top side and another two 80uf on the bottom. This particular amp had a couple open caps. So I changed all the larger value/voltage caps. When trying to install the F&T type A, they were larger than what I started with!! So trying to push the PC board down so that I can attach the nut to the PC board took a little nudging. In the end, I got that amp to be very quiet and it sounds great.

The other photo with the grey IC caps comes from the T100C amp. It looks like that amp has had some work done to it. I can see that the screen resistors have been replaced (originally they were carbon comps) and two of the grid resistors were also replaced. As you can see, all four 80uf caps also appear to be replaced. Even the grey IC caps are too big for this pc board. The amp had a noticeable hum. When I lifted the PC board, one of the caps fell loose - the solder connection did not hold. So I am hoping that when I reflow the solder, it will solve the problem - but that is for another post if needed (and oh how I hate to leave IC caps in this amp).

I have seen a photo of this same PC board where someone mounted all four caps on the top side - Sprague Atoms in fact. It was really tight but he must have encountered the same problem I did; not enough clearance on the bottom side to the chassis.

So that is why I chimed in on this old thread. I have no idea what type of caps would have come with this amp but they must have been physcially small! These are 80uf, 450 volts.

[TomCarlos]

G1 - You are correct. The schematic shows two 80uf in series for 1/2 of the PC board and then the other half. But there is no mention of the voltage rating. Again, two different amps and each one had 450v caps. Maybe the prior tech only had 450v caps? So perhaps 300v caps might be ok here?

Here are the images from the full schematic: yamaha_t50_t100_schematic.pdf

[TomCarlos]

Perhaps I am reading this wrong, but with the caps in series don't we get a total of 40uf with approx 250v across each cap? I believe the (safe) rule of thumb is to have your caps at twice the actual voltage? So two caps rated at 450 volts is the minimum you would want to see?

[Leo_Gnardo]

I believe the (safe) rule of thumb is to have your caps at twice the actual voltage?

Caps rated twice the actual voltage is a way big safety margin. Good, not a problem, but more than necessary. In this case I'd be happy to install 350V rated caps.

That series of Yamaha amps is in my opinion, excellent. I call them "the poor man's Soldano", sleepers that often trade at a bargain price. On tour 2004, I saw Michael McDonald's guitarist Bernie Chiaravalle used his T100 through a Marshall 4x12 loaded with Tone Tubby's for excellent results. Dont'cha know, the amp sprouted a problem, and I had to crack it open and replace a filter cap. Since then only one T series amp has showed up here for repairs, also needed some filter cap replacement.

[TomCarlos]

Thanks again G1 and Leo.

I would think that the original design would have all caps mounted on the top side of the PC board. The only reason for seeing caps on the bottom side is because of the cap dimensions.

[Leo_Gnardo]

The only reason for seeing caps on the bottom side is because of the cap dimensions.

A living illustration of a lyric from Neil Young's gem "Cripple Creek Ferry" - "it's a mighty tight . . . squeeze."

FWIW I've seen over/under technique in some Mesa amps. Squeeze 'em in, any way you can.

[g1]

Apologies Tom, I completely missed those caps on the underside of the board. I thought they had changed over from 4 cap to 2 cap arrangement.
I'll delete my earlier posts.

Perhaps I am reading this wrong, but with the caps in series don't we get a total of 40uf with approx 250v across each cap? I believe the (safe) rule of thumb is to have your caps at twice the actual voltage? So two caps rated at 450 volts is the minimum you would want to see?

Twice the actual voltage is overkill. The caps are designed to work full time at up to their rated voltage. Some leeway is advisable to account for power line fluctuation or the amp being tested with power tubes removed (which will raise B+).
I'd guess the originals were either 300V or 350V, so there would have been a 100 or 200V margin of safety.
The lower voltage caps being smaller size may have all fit on the topside of the board.

[TomCarlos]

No Problem G-1.... I was guessing you missed the underside.

Ok, no sense in beating this thread to death! I am now on the search to see if I can find some caps that will fit with room to spare - but that might be a tall order.

Perhaps next week I will open a thread on the Yamaha T100C. That amps needs some caps and tubes. More to come on that one.

Thanks again friends.