First-Time Tube Amp Design

[EtherealWidow]

It's a modded High Octane. Can't really even call it a High Octane anymore, so I just call it a "Higher Octane." I started a thread just now to give you and any others who want a look a little more info. There are so many aspects to amp building. One thing you'll have to find out almost entirely on your own is the physical execution of the amp. This "Higher Octane" is butt ugly, honestly. You'll see when I put pictures up. I used 14ga speaker wire for some of it because I ran out of my smaller, more appropriate stuff. The chassis is big enough for the board, sure, but once you start adding fuse holders, knobs, jacks, etc it becomes a very tight squeeze. I also found out I despise these boards.

[img:800:172]https://taweber.powweb.com/store/fb-v.jpg[/img]

I'm on my second build now. Finding out I hate forked turrets.

[img:657:204]https://taweber.powweb.com/store/turrets.jpg[/img]

[Theashe]

As a brief bit of practical advice: ask yourself how many of the bells and whistles (like reverb, an effects loop) do you actually need. If it's your first build, ignoring that it's to be an original design, these add extra layers of complexity are, well, extra layers of complexity. If you can find a way to build, and test, in stages and add on things as you go-- all the better.

I'm patient, and I'm willing to learn how to do it right since I'm putting in the effort to do it at all. The added complexity of the reverb and effects loops doesn't bother me much. My concern is that I think I have way too much gain (three gain stages before the PI). I could substitute V2 or V3 (or both) with a 12AU7, but the 12AU7 has more headroom than a 12AX7, so I might return to the original problem of only ever overdriving the PI. I've read a good overdrive sound comes from a couple of tubes being overdriven and that a single tube doesn't sound very good, so I'm avoiding that design. Perhaps that's heresay. Another option would be to include trimmers between each stage to dial down the signal, and play with it until it sounds right. The extra resistors might add noise though. Hmm.

Edit: Or use some local (variable?) NFB on V2 to control the gain.

[Theashe]

I also found out I despise these boards.

I'll check out the post. Any opinion on terminal strips instead of turrets?

[EtherealWidow]

Don't worry too much about resistor noise. For your first build, make sure your format is one that can be modded easy. Be prepared to mod a lot to get it to where you want.

[Theashe]

I've heard that - no problems. Once I finalize the design, I'm going to CAD it to get a rough idea what kind of space I need, and then buy a chassis with more space than that. In one of my past projects I crammed the PCB, wiring, pots, and switches for a portable piezo + JFET buffer/preamplifier with EQ into a stupidly small project box (so friends with acoustics with no pickups could plug in at backyard shows). Had to take the damn thing apart so many times - and every time I had to find a new way to get it back together because it wouldn't go back cleanly. Size is definitely not going to be a compromise on this one.

[EtherealWidow]

I would say they might be good for prototyping. Never worked with them. Seems like they'd be a tremendous pain. Someone more experienced should chime in.

[matt h]

(deleted)

[Theashe]

Gents - heater supply question.

The transformer I'm getting for my amplifier is the Hammond 273BX.

I'm wiring my 12AX7/12AU7 tubes with 12.6V and my 6L6GC's with 6.3V. I read from Merlin's book an artificial center tap referenced to a DC source (for DC elevation) will help reduce heater hum.

In my diagram, I've put a humdinger on the 12.6V terminals and the 6.3V terminals, with the idea that it should help reduce the noise for the 6L6GCs on the 6.3V. However, I suspect that there won't be enough heater noise on those tubes to need it, so I may remove it.

As for the 12.6V humdinger - since there is already a center tap on the transformer, is this even necessary? If I attach the center tap directly to a DC source such as the cathode of the 6L6GC (I'm cathode biased), would this accomplish the same DC elevation as putting in a humdinger and connecting the center tap of that to a DC source (see attached file, sticky note on the right)? However, I'm also running heaters at 6.3V, and connecting it this way would leave a DC voltage across the whole filament - probably bad.

Thoughts?

Edit: a follow up question. In my attached schematic I've placed a bleeder resistor on every smoothing capacitor as well as the reservoir. I've also seen designs out there where only the reservoir has a bleeder resistor. I assume that the more resistors, the faster the power supply capacitors can discharge because more parallel paths reduce the circuit's time constant. However, these resistors are effectively in parallel with the plate resistor/tube resistance/cathode resistance, so they should be made large so that most of the current does not go through the bleeder resistor while the amp is on. Does that logic check out?

[sluckey]

The transformer I'm getting for my amplifier is the Hammond 273BX.

That transformer does not provide 12.6V so just wire all filaments for 6.3v and ground the CT or elevate it. It will also produce more like 450V to 500V B+, rather than 350V.

Also, you should not use a bridge rectifier with that transformer. As drawn, you have the negative side of the bridge grounded and you also have the HT center tap grounded. Doing that will pop your fuses, or pop the two diodes that are connected to ground (hopefully open), or worse. You can safely use a bridge if you don't ground the negative terminal.

A bleeder resistor across the first filter cap is all you need.

[Theashe]

The transformer I'm getting for my amplifier is the Hammond 273BX.

That transformer does not provide 12.6V so just wire all filaments for 6.3v and ground the CT or elevate it.

You're right - I misread the sheet.

It will also produce more like 450V to 500V B+, rather than 350V.

I simulated the current demands of my amp in Duncan's PSU designer and got a B+ near 350V. I also did some calculations backwards, starting with my current demands and rectifier choice and found I needed a transformer with an RMS voltage of near 350V per section (see attached if interested). I also thought the 350V sections would be quite high. Maybe I'm still missing something. My total current demand is 160mA. (60 x 2 for 6L6GC biasing @ 350V, and the rest is made up of four 12AX7s and two 12AU7s).

Also, you should not use a bridge rectifier with that transformer. As drawn, you have the negative side of the bridge grounded and you also have the HT center tap grounded. Doing that will pop your fuses, or pop the two diodes that are connected to ground (hopefully open), or worse. You can safely use a bridge if you don't ground the negative terminal.

Thanks for this. If I'm going to leave the negative terminal ungrounded, then I might as well just use a couple of 1N4007s instead.

A bleeder resistor across the first filter cap is all you need.

Check.

Edit: if you do look at the attachment, don't worry about the smoothing filter section. It's left over from some experimenting I did a few weeks ago, without any meaningful numbers.

[matt h]

(deleted)

[Theashe]

Thanks again for your insights Matt. The idle current in my amp is 160mA (60mA per power tube + 40mA for all of the preamp tubes). The power tube current when signal is at a peak should be around 215mA if I've drawn the load line correctly (see load line attachment), and higher still if you include the preamp tubes.

I've seen that sheet, but chose instead to see where this tutorial would get me:
http://diyaudioprojects.com/Technical/T ... -Supplies/

It roughly agreed with the output of Duncan's PSU Designer II when using a 5U4GB FW rectifier. The 5U4 has a fairly large drop (~50V at peak), which helps account for the discrepancy. If I'm misleading myself, please point me in the right direction.

Edit:
As I look at this more, I'm wondering if the output voltage in Duncan's tool is given as RMS. If it is, then my voltage is actually 350*1.414 = 495V, which is very close to the 700*0.71 = 497V given in the Hammond sheet.

[matt h]

(deleted)

[jazbo8]

The idle current in my amp is 160mA (60mA per power tube + 40mA for all of the preamp tubes). The power tube current when signal is at a peak should be around 215mA if I've drawn the load line correctly (see load line attachment), and higher still if you include the preamp tubes.

The datasheet you used does not correspond to the output stage schematic that you posted earlier, has it been changed? Also, the 40mA draw for the preamp tubes seems rather high, but without seeing the preamp schematic, it's hard to tell. Also, it seems somewhat unusual to use 12AU7's for a pair of 6L6 operating in class AB1.

[Theashe]

The idle current in my amp is 160mA (60mA per power tube + 40mA for all of the preamp tubes). The power tube current when signal is at a peak should be around 215mA if I've drawn the load line correctly (see load line attachment), and higher still if you include the preamp tubes.

The datasheet you used does not correspond to the output stage schematic that you posted earlier, has it been changed? Also, the 40mA draw for the preamp tubes seems rather high, but without seeing the preamp schematic, it's hard to tell. Also, it seems somewhat unusual to use 12AU7's for a pair of 6L6 operating in class AB1.

Hi jazbo.

You're right, I changed the output stage that I posted earlier. I made an edit to the bottom of the original post to say that, but I didn't upload the new schematic since I wanted to preserve the original question. I'll upload it tonight when I get home.

As for the 40mA, I'm running a parallelled 12AU7 for a reverb driver and another for V2 as a recovery stage plus cathode follower. Each section of each tube uses 4-8mA at idle depending on what its used for. The rest of the 12AX7s make up the difference. I chose to swap in a 12AU7 for V2 because I'm running a parallel effects loop with a 12AX7 common anode mixer in V3 and I get gain from that. I found swapping in a 12AU7 got the gain to a more appropriate level.

The general layout is:
V1A - 12AX7 - Gain
V1B - 12AX7 - DC Coupled Cathode Follower
Tonestack
V2A - 12AU7 - Gain
V2B - 12AU7 - AC Couple Cathode Follower
FX Out/In in parallel with dry signal
V3A - 12AX7 - Dry signal volume + gain
V3B - 12AX7 - FX signal volume + gain
V4A/B - 12AU7 - Reverb driver
V5A/B - 12AX7 - Reverb recovery
V6 - 12AX7 - PI
V7/V8 - 6L6GC