How Is This Poweramp and Power Supply Looking?

[lkrasner]

Hey all, just looking for some approval on my design before I go design the preamp. This schematic has 2 parallel tubes with a separate MV for each one. The idea is that you can then blend between them. I'm sure there are a million things wrong, so don't be afraid to rip me apart.

Here's the schematic: http://s3.postimg.org/3mep8ener/amp.png

A few big questions:

How are my supply resistor values, assuming a normal 2 or 4 stage preamp?

Will my bias circuit work? I added the switch to avoid putting 4 test points.

[HotBluePlates]

What output transformer will you be using?

[lkrasner]

What output transformer will you be using?

Probably one of the Hammond 125SE series depending on what size I need (here: http://www.hammondmfg.com/125SE.htm) . I've also looked at the mercury magnets ones, but there site makes it somewhat difficult to browse around for generic transformers.

Really I am open to suggestions, but I would like something high quality for a reasonable cost. Expensive is fine, but not ridiculous.

[kagliostro]

Between Hammond 125 and MM I see a very long way

something in the middle will probably better suit your need

you put together an el84 and an el34, both tubes are near in sensibility (~11mA/V)

but the el84 at 250v requires a 5.2k load and the el34, for the same B+, a 2k load

also, many people, referring to tone, refers to the el34 as a bigger el84 tube

why don't mix a pentode with a tetrode ? an el34 with a 6l6
(at 250v the first requires a load of 2k the second a load of 2.5k, very close each other)

or el84 with 6v6
(at 250v the first requires a load of 5.2k the second 5k, also this very close)

K

[lkrasner]

Between Hammond 125 and MM I see a very long way

something in the middle will probably better suit your need

you put together an el84 and an el34, both tubes are near in sensibility (~11mA/V)

but the el84 at 250v requires a 5.2k load and the el34, for the same B+, a 2k load

also, many people, referring to tone, refers to the el34 as a bigger el84 tube

why don't mix a pentode with a tetrode ? an el34 with a 6l6
(at 250v the first requires a load of 2k the second a load of 2.5k, very close each other)

or el84 with 6v6
(at 250v the first requires a load of 5.2k the second 5k, also this very close)

K

The plan was to use an el84 and 6v6 when used in parallel mode. Or a single other tube if desired.

[HotBluePlates]

What output transformer will you be using?

Probably one of the Hammond 125SE series depending on ...

Previous posts aside, let's assume you'll use the EL34 & EL84 (or it could be a 6V6 if you use an octal socket instead). Your 270FX is 250-0-250vac output, which will rectify to ~318vdc at the 1st filter cap. Ignoring the 100Ω dropping resistor for a moment, let's pencil "300vdc" for B+1.

There's little sense idling single-ended tubes much below 100% dissipation, so figure 25w (EL34) + 12w (EL84/6V6) = 37w, which gives a total idle current of 37w/300v = ~123mA. The 125SE series tops out @ 100mA of idle current, so you'll need something bigger. Pencil in the Hammond 1627SEA (30w, 2.5kΩ primary Z, 160mA unbalanced d.c.). The EL34 will idle about 25w/300v = ~83mA and the EL84 at about 12w/300v = 40mA. The 30w 1627SEA is really too-big from a core-size standpoint, but the overriding factor is the allowable unbalanced d.c. We really don't know if we can push the 125SE series 25% above its rating in this regard. The 1627SEA is ~50% more expensive than the 125GSE; you may wish to try to contact Hammond to see how rigid that unbalanced d.c. spec is. The 1627SEA is also 11lbs, and will need a thick chassis and/or some angle pieces to support the weight.

A first-approximation of an ideal Class A load is plate volts / idle current; ignoring bias for a moment, that's 300v / 0.123mA = ~2400Ω, so 2.5kΩ is the closest standard value. Don't worry about "what the tube wants to see" because you're really defining the power output with the PT and OT, then going back & forth to confirm the output tubes can drive the chosen setup within their ratings. Assuming the tubes can manage, they will have a peak current of double-idle current, or another 123mA (for the total output stage) above the idle value. Power output will be 123mA² * 2.5kΩ / 2 = 18.9w RMS, which is probably a little optimistic being over 1/2 of the plate dissipation rating. The output stage will manage this much output after some distortion is occurring.

A quick look at the data sheets for both tubes shows they can each easily reach double-idle current at 250v on the screen, so they won't have any problem in the circuit as-proposed, even if the screen-to-cathode voltage is reduced due to bias and dropping resistances. The EL84 needs -8v of bias at 300v plate, 250v screen to idle at 40mA; screen current at idle will be ~4mA. The EL34 will need about -14v of bias at 300v plate, 250v screen to idle at ~83mA; screen current will be ~10mA.

Ignoring the preamp current draw, the 100Ω dropping resistor ahead of the B+1 node will drop (40mA + 4mA + 83mA + 10mA) * 100Ω = ~14v, so the guess of 300v for B+1 was good. A 150mA choke like the Hammond 158Q or 159Q will filter very much better than the 100Ω resistor and result in very similar B+ voltage, but adds $24-26 to the parts tally plus shipping. The 100Ω resistor would dissipate a bit under 2w, so your 5w rating would be ample if you go that route.

[HotBluePlates]

I'd give each tube its own 1kΩ screen resistor rather than the shared resistor you have, but that's mostly personal preference. Penciling ~4mA for the preamp (which may be high depending on the circuit you choose), current through R3 is ~14mA, so B+2 will be ~290vdc. So we're again safe with the currents guessed by using the 250v G2 curves in the data sheet (they may be a bit higher, but will result in a little more dropping of B+1 and B+2, dropping the current a bit closer to the estimates so far).

I'm not inclined to keep the variable cathode bias arrangement you have drawn, mainly because when Tubenit tried varying cathode resistors to tame idle current he needed to make fairly large changes to have an impact. The EL34 will likely need -14v/(83mA+10mA) = 150Ω, or up to as much as the ~200Ω shown on the data sheet for 300v plate and screen. The EL84 will need -8v/(40mA+4mA) = 181Ω, which you'd round up to 200Ω or so to get a standard value. In the end, you're really gonna find a value that lands on where you want the tubes to idle and leave that value alone. All tubes plugged into the socket will settle on a slightly different idle current and bias voltage as a result of the tube's individual characteristics. The upside of cathode bias is you don't need to adjust tube current, especially with SE operation.

I think KOC of London Power drew up a circuit for variable cathode bias which involved returning a portion of the voltage developed across the cathode resistance to the grid as a way to set plate current. I'll wait to see if someone else has a drawing handy to verify, because I have that book in storage in another state & not handy.

All that said, you probably can't go wrong guessing the cathode resistance downward 20-50Ω and adding a 100Ω pot as you've drawn.

The value for R4 can't really be known until you have an idea of what your preamp will be, how much current it will draw and what supply voltages you'll want in the preamp. But it's probably close enough as a guess to start.

[lkrasner]

I have this same basic bias circuit in another amp I build. The 100R fixed resistor is to ensure enough resistance to prevent immediate damage if the pot fails. the pot is a high power 1K linear model, which gives plenty of range.

The reason I am keeping this is to allow for different tubes when not using the amp in parallel mode (in which case only an el84 and 6v6 will be used). pulling the noval tube will increase b+ significantly (depending on the octal tube used). Then the pot gives enough range to tune for nearly every tube.

[HotBluePlates]

I have this same basic bias circuit in another amp I build. The 100R fixed resistor is to ensure enough resistance to prevent immediate damage if the pot fails. the pot is a high power 1K linear model, which gives plenty of range.

Fair enough. If you use the outer 2 lugs and tie the wiper to one, an open wiper will default to the full pot resistance. Then you shouldn't need a high power rating.

Referring to the plate curves, ~1kΩ of cathode resistance will cause the bias to settle around -21 and plate current at ~21mA; really it will be a little higher bias and less plate current because I've ignored screen current. -21v implies 21²/1kΩ = 0.44w dissipated in the pot. In that case a 1-2w pot is perfect.

Since you noted you've already tested this in another build, I'd say everything is good to go. I would think 1kΩ against 100Ω of fixed resistance might make the adjustment jumpy or hard to dial in a fine adjustment. However, you made a good point about what you'll need if you pull one of the output tubes.