5f2a filter cap variations

[nickrundall]

Hi everyone, what a great forum!

I'm just getting into my first amp build - I've gone for a 5f2a with a few switchable options (switchable 22k / 2k NFB resistor values , 4/8 Ohm speaker switch, 25uf / 100uf cathode cap switch). It's been suggested to me that it's a good idea to bump up the filter caps to 22uf 500v to reduce noise. My questions are: Should all four be the same value or is there an ideal configuration of values? Are there different benefits at each stage? Would you expect any tonal changes as well as reduced noise?

Thanks!

[Chuck H]

Posting a schematic here so that we can all speak from a referenced perspective.

[Enzo]

Hi
Read that note in the upper right corner of the schematic. And remember, this is just a guitar amp, not precision lab gear. Nothing is critical.

FInding old retro 8uf caps is expensive and pointless. 10uf is a standard value, but really 22uf is good advice. COmmon and effective. There is no magic formula. The first filter, right off the rectifier has the most smoothing effect. Note the schematic puts two 16uf in parallel there for a 32uf total. A 40uf there would be great.

This is a single ended amp, so there is no inherent hum cancelling like in a push pull. SO having the smoothest B+ possible helps hum. The second nore - the screen node - will have a lot less ripple (hum), and the remaining node should have none. This is a class A circuit, so there won't be much sag. SO I don't worry about filter cap value much.

[Chuck H]

switchable 22k / 2k NFB resistor values

Not sure what the goal is here. Reducing that resistor to 2k would INCREASE negative feedback enormously. That's the opposite of what most mods do and IMHE doesn't sound very good. Is the goal of this mod to increase the NFB for some reason?

4/8 Ohm speaker switch

Seems like a good idea.

25uf / 100uf cathode cap switch

First, which cathode? There are two bypassed cathodes in that amp. Second, for either a move from 25uf to 100uf isn't going to be an "instant gratification" difference at all. You may not even notice. If you're talking about the power tube then the higher cap value may give a very slightly more dynamic attack to the bottom end. It would be a small difference and I think it's probable you would find a switch position you like for the amp in general and probably never flip it again. If you're talking about the preamp tube then you almost surely won't hear any difference at all since 25uf already constitutes full AC bypass for all audible frequencies plus some. So 100uf could only serve to increase gain at subsonic frequencies. It might even tend to make the amp less stable. The usual goal with switchable preamp cathode bypass caps in later Fender designs is to reduce farty bass during overdrive. That is usually done by decreasing the value of the bypass cap.

It's been suggested to me that it's a good idea to bump up the filter caps to 22uf 500v to reduce noise

The real advantage would be ripple hum reduction. This seems like a fine idea. Though the difference would be within spec for capacitors of the time, so, not really a difference. You might use a 47uf for the main (first) filter. This is the capacitor that does all the heavy lifting WRT noise. I know that 47uf is beyond the max spec in the rectifier tube data, but capacitors of the time were typically rated at +80/-20%! I've used 47uf in a few amps with 5y3 rectifiers and never had a problem. Notice that on the schematic the first two filter capacitors are parallel in the circuit. For a stock uf value of 32. So even a 47uf cap is still within spec for caps of the period. If you find the over spec value distasteful you can do what's called a pi filter for extra ripple reduction. A capacitor followed by a low resistance and then another capacitor. This circuit has excellent ripple reduction and wasn't used on the early amps for the simple reasons that high voltage electrolytic caps were relatively more expensive and higher values even more so. Something like a (from the 5y3) 22uf to a 100r resistor and then another 22uf (or higher, like 47uf) to the plate supply should perform a lot better than a the parallel 16uf caps and doesn't threaten the rectifier spec on paper.

Should all four be the same value or is there an ideal configuration of values? Are there different benefits at each stage? Would you expect any tonal changes as well as reduced noise?

"I" would use (from the rectifier moving toward the preamp) 22uf-100r resistor-47uf (to plate supply, then)-10uf-10uf. The preamp filters aren't so much for noise reduction or ripple rejection. They are for "decoupling" They route any AC tracking on the low impedance of the power supply to 0V (ground) so that mixed signal in small quantities can't interact with the signal chain. In this regard you could use the stock 8uf value (or 10uf) and be fine. 22uf is a common value and is also fine. Too high a value here is reported to sound very hard and stiff. This is probably because there's simply more energy in the charge of a larger capacitor value for transients like pick attack. So perhaps larger values don't allow as much power supply reaction that is part of the characteristic compression in tube amplifiers (though in class A amps this is already minimized). In this light I think a little more is good, but a lot more isn't necessarily better. 8uf is an odd value today so I'd just go with 10uf. So my logic is to stiffen the supply at the plate for ripple rejection and leave it (basically) stock in the screen supply and preamp where ripple is less of an issue.

JM(very biased)2C

And welcome to the forum

[nickrundall]

Thanks for such amazing answers - I'm glad I asked instead of just jumping in!

NFB switch - that totally makes sense - I just want to switch it on and off so I can get Champ tones with it off and better cleans with it on. Sounds like the best thing is to just switch the 22k in and out of the circuit.

Preamp Cathode cap switch - also makes sense, I might leave this for a future mod once I know how the amp sounds - see if I even need it. I was interested in the bass cut mod.

Filter caps - thanks for the great explanation. I think I'll start with 22uf/22uf/8uf/8uf and see how the amp turns out, and then maybe mod it to the pi filter suggestion with 22uf-100r resistor-47uf-10uf-10uf to see how it changes things.

Thanks again!

[mozz]

I've read more than once people say to not go higher than 8uf in the preamps, changes the sound. You can get 16uf F&T caps and Sprague still makes 8uf, i think you can get both from Tubedepot.

[Chuck H]

Well alrighty. It's pretty important to include some sort of referencing for switches in some circuits to avoid annoying POP's. Here's a modified schematic that includes the power supply as I described it, a bass cut circuit on the preamp cathode and a NFB cut switch.

Notice (please) that I also removed the "death cap", added an earth ground and moved the fuse.

[Enzo]

To those people telling you never to go over 8uf in the preamp, I again point out that important note upper right. But beyond that, electrolytic caps of that era had - seriously, not exaggerating - tolerances like +80/-20%. That means your 8uf cap can be over 14uf right out of the box, and still be within specs...Or for that matter well under 7uf.

Look at the earlier versions: 5C2, 5D2, 5E2, 5F2. The earliest has a pi filter stage before the plate node. By 5E2 they had a choke in the pi filter. Finally 5F2A, the schematic we have been looking at, eliminates the pi before use node. Then Chuck added it back in for his version of 5F2A. So there were various versions of teh factory circuit being produced in that era. Hard to say one is the good one and all the others suck.

I am not generally a conspiracy nut, but... No one needs 8uf caps except guys trying to build old guitar amp circuits. it is a tiny niche market electronic part. I would not be at all surprised to find they made batches of 10uf caps and the ones that checked low for value at QC became "8uf" caps.

[Justin Thomas]

Chuck,

Not to be an arse, but shouldn't the fuse go before the switch? I thought the theory is that the fuse should be the first thing juice sees...

Justin

[Chuck H]

Yep. You're right. I was moving too fast trying to cut and paste on a schedule. It's not even convenient to physically wire it with the fuse after the switch Will fix.

[g1]

My thinking on the 'don't go over 8uF' thing is they mean 8 or modern equiv. (10) but don't step up to 16 or 22. Many will be able to tell the difference from 8 to 16, I don't think any claim to be able to tell 8 from 10.

Chuck,
Not to be an arse, but shouldn't the fuse go before the switch? I thought the theory is that the fuse should be the first thing juice sees...

Justin

I've seen good arguments for both methods, but modern manufacture seems to favor switch first.

[Justin Thomas]

Rilly? I think I'll just settle the debate & include neither on my amps. Just unplug it from the wall!

I wish the codewriters could make up their minds...

Justin

[Chuck H]

I think the only reason to favor the switch first would be to keep AC off the fuse body when the amp is switched off. Then the manufacturer can say in the lit: "Never change the fuse with the amplifier switched on." and their liability is reduced.?. More of an issue with some older fuse holders but even with new design fuse holders a mistake could happen. Then again, in the event of an AC input short having the fuse after the switch adds one more possible component fault that could cause a hot chassis with the amp switched off. I think this must be more rare than guitarists poking their fingers into the fuse hole though.

Since the fuse and AC mains input is typically on the back of the amp and the power switch on the front it's certainly easier/tidier to come in with your AC and hit the fuse first. Rather than go to the switch and then back to the fuse.

[Enzo]

Instructions are no defense. The unit must be made so no one can contact the mains. The "no user serviceable parts inside" is the only real warning.

Just my opinion: I think the don't go over 8uf comes from people conflating arguments. Larger amps tend to sag when you crank at them loud. and larger filter caps hold up better, so the amp loses some compression. Amp gets "stiffer". But on little one-lungers like this, we don't get huge sag when you pound out a chord. SSo I think the cap size has way less effect. So the cap size idea is just a pro-forma argument.

[nickrundall]

On a side note, I've populated my turret board and I'm using a pair of 0.022uF 500Vdc SoZo NexGen Blue Molded Vintage Capacitors as coupling caps. I want to connect the outer foil side to the incoming signal side of the circuit, but I'm not certain which side is the foil side. I know the other Sozo caps have a black line that indicates the foil side, but I haven't found any info on the Blue molded ones. I've emailed Sozo, but haven't had a response. Does anyone know?