Cathode Bypass Capacitor Options

[jammybstard]

I've quite new to valve amps and I've been playing around with some of the components in my ampmaker SE-5a, so far mainly the bypass capacitors:
[img768]http://ampmaker.com/images/ak01kit/ak01sc2.jpg[/img]
I've replaced the two cheap electrolytic pre-amp stage bypass caps (C1 and C3) with poly's and reduced the value of the C3 to 470nf to reduce the flabby bass sound a bit.

I'm now thinking about the output stage Coupling cap C7 currently a cheap 47uf Electrolytic; I dont thing there's any point in adjusting the value much, but could a change of cap type effect the voicing much? I'm thinking a better electrolytic perhaps in combination with a decent poly film cap. I was reading that Messa even use Tants for bypass caps! (Tants in audio!!!) so I might give a wet tant a try.


Any thoughts on what could be achieved here? your experience would be welcome.

[Structo]

What did you notice when you changed the bypass caps to plastic?

The cathode bypass cap on the power tube could be changed to a higher value.
The 18 watt Marshall guys frequently install 100-500uF bias caps.

Fun ain't it?

[jammybstard]

What did you notice when you changed the bypass caps to plastic?

Classic Audiofool trap!
When I changed the first stage cap C1 I thought perhaps it sounded a little rounder but I wouldn't swear to it; C3 was a value change as-well as a type change and that reduced the bassy distortion.
There's an element of Nine out of Ten Cats would choose Polyproperlene involved in all of this!

The cathode bypass cap on the power tube could be changed to a higher value.
The 18 watt Marshall guys frequently install 100-500uF bias caps.

What would that achive? more gain I suppose? more Bass?

[jammybstard]

anyone?

[jhaas]

The cathode bypass cap on the power tube could be changed to a higher value.
The 18 watt Marshall guys frequently install 100-500uF bias caps.

What would that achive? more gain I suppose? more Bass?

"Tighter" bass, not really more, closer to the vibe of fixed bias. Actually, the 18Watt guys have been putting 1000uF - 2200uF in that position. Myself included. In my case I've used a 1000uF in both a LiteIIB and a Superlite TMB. I haven't tried it in any of my SE builds.

[jammybstard]

Fixed bias would be no cap at all??

So it would be worth just trying a few different values in that position then; Is the value of the bias resistor important if your changing the caps?

[tubeswell]

Fixed bias would be no cap at all??

No - fixed bias is no cathode resistor and (therefore also) no bypass cap (so that the cathode then sits at 0V potential) - and the grid then has to have a negative DC voltage applied to it to drop it about 10V or so below the cathode (i.e.; -10V potential). To do this, you have to have a bias voltage supply to the EL84 grid (which isn't in your amp, but you could make one) - but that's another story

So it would be worth just trying a few different values in that position then; Is the value of the bias resistor important if your changing the caps?

Anything below about 47uF will start to roll-off bass frequencies. But reducing the value of the bypass cap in the output stage won't change the tone noticeably. You'd have more luck removing bass by reducing the coupling cap after V1a to something like .01uF (10nF) or .047uF (Note that at least a 400V rating is required on this coupling cap).

Completely removing the cathode bypass resistor on the EL84 will reduce output power by reducing the overall gain of the stage, because without a bypass cap there is nothing to keep the cathode voltage constant, and the cathode voltage will then be pulled up and down slightly, in synch with the rise and fall of the plate voltage, thereby reducing the overall voltage differential between the cathode and the plate at all parts of the signal swing. (AKA 'cathode current feedback')

[Structo]

Fixed bias would be no cap at all??

No - fixed bias is no cathode resistor and (therefore also) no bypass cap (so that the cathode then sits at 0V potential) - and the grid then has to have a negative DC voltage applied to it to drop it about 10V or so below the cathode (i.e.; -10V potential). To do this, you have to have a bias voltage supply to the EL84 grid (which isn't in your amp, but you could make one) - but that's another story

So it would be worth just trying a few different values in that position then; Is the value of the bias resistor important if your changing the caps?

Anything below about 47uF will start to roll-off bass frequencies. But reducing the value of the bypass cap in the output stage won't change the tone noticeably. You'd have more luck removing bass by reducing the coupling cap after V1a to something like .01uF (10nF) or .047uF (Note that at least a 400V rating is required on this coupling cap).

Completely removing the cathode bypass resistor on the EL84 will reduce output power by reducing the overall gain of the stage, because without a bypass cap there is nothing to keep the cathode voltage constant, and the cathode voltage will then be pulled up and down slightly, in synch with the rise and fall of the plate voltage, thereby reducing the overall voltage differential between the cathode and the plate at all parts of the signal swing. (AKA 'cathode current feedback')

Just trying to help here.
Tubeswell meant to say capacitor where he used the word resistor.

[jammybstard]

Anything below about 47uF will start to roll-off bass frequencies. But reducing the value of the bypass cap in the output stage won't change the tone noticeably.

So how do the larger caps mentioned above tighten the bass?

You'd have more luck removing bass by reducing the coupling cap after V1a to something like .01uF (10nF) or .047uF (Note that at least a 400V rating is required on this coupling cap)

400V sugar Really? I have a 250V in there at the moment; I only measured 150V accross that cap without any signal? is the Peak to Peak of the signal that high?

[tubeswell]

Anything below about 47uF will start to roll-off bass frequencies. But reducing the value of the bypass cap in the output stage won't change the tone noticeably.

So how do the larger caps mentioned above tighten the bass?

By 'below 47uF' I mean lower capacitance (i.e. smaller caps)

You'd have more luck removing bass by reducing the coupling cap after V1a to something like .01uF (10nF) or .047uF (Note that at least a 400V rating is required on this coupling cap)

400V sugar Really? I have a 250V in there at the moment; I only measured 150V accross that cap without any signal? is the Peak to Peak of the signal that high?

Okay I retract that bit - 250V should work fine in V1 if the HT voltage at that point in the power rail (for that stage) is less than 250V, which it probably is given that the B+ for the entire amp is only about 270.

[jammybstard]

Anything below about 47uF will start to roll-off bass frequencies. But reducing the value of the bypass cap in the output stage won't change the tone noticeably.

So how do the larger caps mentioned above tighten the bass?

By 'below 47uF' I mean lower capacitance (i.e. smaller caps)
?

I get that bit, but above people are talking about higher valour caps 100uf to 400uf etc and that this can tighten the bass; i was wondering how?

[brewdude]

I don't pretend to really understand it, but the explanation I recall was that the large bypass cap, say 1000uF, stabilized the cathode voltage swing and enabled it to behave similar to a fixed bias amp.

[jammybstard]

I don't pretend to really understand it, but the explanation I recall was that the large bypass cap, say 1000uF, stabilized the cathode voltage swing and enabled it to behave similar to a fixed bias amp.

That would make sense

[tubeswell]

I don't pretend to really understand it, but the explanation I recall was that the large bypass cap, say 1000uF, stabilized the cathode voltage swing and enabled it to behave similar to a fixed bias amp.

47uF is enough to stabilise the cathode for all the audible frequencies that the human ear can detect in an EL84. You won't notice any difference really on an EL84 geetar amp between 47uF and 1000uF bypass cap. But in reducing (say halving or quartering) the value of the 47uF bypass cap, you will start to notice reduction in the bottom end.

On the other hand, what fixed bias does, is increase the overall gain of the tube by increasing the voltage differential between the plate and the cathode - by putting the cathode at ground potential instead of at 10V above ground potential. But to do that you have to have the grid now set at -10V with a negative bias voltage supply. That is how fixed bias give the stage more power.

[tubeswell]

Fixed bias would be no cap at all??

No - fixed bias is no cathode resistor and (therefore also) no bypass cap (so that the cathode then sits at 0V potential) - and the grid then has to have a negative DC voltage applied to it to drop it about 10V or so below the cathode (i.e.; -10V potential). To do this, you have to have a bias voltage supply to the EL84 grid (which isn't in your amp, but you could make one) - but that's another story

So it would be worth just trying a few different values in that position then; Is the value of the bias resistor important if your changing the caps?

Anything below about 47uF will start to roll-off bass frequencies. But reducing the value of the bypass cap in the output stage won't change the tone noticeably. You'd have more luck removing bass by reducing the coupling cap after V1a to something like .01uF (10nF) or .047uF (Note that at least a 400V rating is required on this coupling cap).

Completely removing the cathode bypass resistor on the EL84 will reduce output power by reducing the overall gain of the stage, because without a bypass cap there is nothing to keep the cathode voltage constant, and the cathode voltage will then be pulled up and down slightly, in synch with the rise and fall of the plate voltage, thereby reducing the overall voltage differential between the cathode and the plate at all parts of the signal swing. (AKA 'cathode current feedback')

Just trying to help here.
Tubeswell meant to say capacitor where he used the word resistor.

Indeed I did