First-Time Tube Amp Design

[Theashe]

Glimpsing your PSUD faux-schematic, off the top of my head I'd be worried about adequate de-coupling between nodes. (ignoring not being able to figure out why there are so many or drawing so much current). I'm also not sure why there's a 260k resistive load on top of all of that (left in for a bleed resistor?)

PSUD barked at me for trying to change the load resistor into a current tap, so I just left it in as an equivalent resistance. Although now that I'm looking at it, I only treated the current draw as if it were for half of the tube. The resistance should actually be half of that (290V/(1.1mA+1.1mA).

Edit for clarification: What I've just bolded was incorrect - for some reason I thought it was my PI. It was an equivalent resistance instead of a current tap for V1, which should be seen as 260V/(0.55m + 0.55m) = 236k, not 260k.

[jazbo8]

Hi jazbo.
I changed the output stage that I posted earlier.

I see, so Ep=350V, and Eg2=300V (both adjusted for Vg1=-23V)? Also a comment on the loadlines, to get a better matching and higher output power, the class B loadline should pass through the knee instead of the class A loadline (which I believe is based on Couch's method).

[Theashe]

I see, so Ep=350V, and Eg2=300V (both adjusted for Vg1=-23V)? Also a comment on the loadlines, to get a better matching and higher output power, the class B loadline should pass through the knee instead of the class A loadline (which I believe is based on Couch's method).

Okay, I'll take a look at drawing the transfer characteristics for a higher screen voltage (320ish). That should bring the class B loadline into the knee without having to look for a different output transformer. I'll see where to re-bias it to get back to around 30 watts from there. However, I want to stay away from triode-strapping my power tubes, since I want to try a morph control on it.

Edit 1:
Another idea I looked at was to switch to a 6.6k OT and run the screens at 250, plate at 350. Rebiasing for 70% would put me at -18V and about 25W of output power. I wanted a little more headroom and output power than that, unfortunately (see attachment 2)

To get the class B loadline to run through the knee with my current plate/screen voltages, I'd need a 5.5k OT.

I should mention that Hammond transformers are the easiest and most affordable for me to get. I don't live terribly far from Guelph and a local supply store stocks their transformers (no shipping).

Edit 2:
Using this transformer (meant for Marshall JTM30), I could run my B+ a little closer to 375 and get the class B loadline through the knee. This again puts my power closer to 26W, but with more clean headroom (this is a mostly clean amp). See attachment 1.

[jazbo8]

There are many ways to get the desired 30W output with good headroom using the Hammond OPT. I have not seen a morph control used on the output tubes, let us know how it turns out.

[matt h]

(deleted)

[Theashe]

Back-peddling on a design idea here. From the picture I uploaded before, I showed that I want to have a switch between a 5U4GB and SS rect. I was designing the circuit with the 5U4GB in mind as the primary rectifier and worrying about the voltage increase when switching to the SS rect. Reversing the idea and instead designing with the SS rect as the primary makes this easier - then the voltage drops instead.

Since I figured the circuit might need re-biasing to maintain a similar volume because of the drop in B+, I had an idea to use a 4P2T switch to simultaneously switch the rectifier and switch between two different Rks. Here's the schematic I came up with.


Anywhoo, I had a question about selecting transformers. On the Hammond site I see several different size transformers for each terminal voltage that I might be interested in, and they're all rated for a specific current. Should the transformer I select be rated based on my idle current draw, or should it be rated based on the maximum current I'll push through under full signal?

[jazbo8]

Should the transformer I select be rated based on my idle current draw, or should it be rated based on the maximum current I'll push through under full signal?

For guitar amps, it depends on how stiff you want the supply to be, i.e., use the maximum Ia for a stiff supply and around the average Ia for a saggier supply. Since you said, the amp is going to be mostly clean with good headroom, I guess you would prefer a supply that is on the stiff side.

[Theashe]

Should the transformer I select be rated based on my idle current draw, or should it be rated based on the maximum current I'll push through under full signal?

For guitar amps, it depends on how stiff you want the supply to be, i.e., use the maximum Ia for a stiff supply and around the average Ia for a saggier supply. Since you said, the amp is going to be mostly clean with good headroom, I guess you would prefer a supply that is on the stiff side.

I'm still having trouble reconciling what Matt said earlier about DCV = 0.707 * Vpri with results from PSUII.

When I use a well-rated 300-0-300 transformer with an impedance of 80 ohms per winding and a FW 1N4007 rect, which should have a DCV of 424V, PSUDII gives me a voltage of about 355V under load.

This makes me step back up to 350-0-350 and use a dropping resistor to get my desired B+ of 400V (375 + 24 for cathode bias). I looked at two alternative Hammond transformers to try to get a compromise in the middle, but the only 320-0-320 and 325-0-325 transformers have much higher impedances, which drop my voltage back down to the same as the 300-0-300 transformer I was looking at.

Matt had some trouble understanding my PSUDII schematic from before, so I've attached one that's been marked up to hopefully make it easier to understand. The 1n capacitors with the red X's across the terminals are ones I don't plan on putting in the power supply but had to include in order to get the dropping resistors, so I left them really small so they would be insignificant in the simulation. Any thoughts? I could try simulating the voltage I'd get using the 115V primary tap instead of the 125V tap in my simulations to get more voltage from the 300-0-300.

Also, huge ripple?

[Theashe]

Glimpsing your PSUD faux-schematic, off the top of my head I'd be worried about adequate de-coupling between nodes.

Hey Matt, can we come back to this point that you made? If I have a resistor to drop voltage and a capacitor to supply current between each stage, why would the decoupling be inadequate?

[matt h]

(deleted)

[jazbo8]

I'm still having trouble reconciling what Matt said earlier about DCV = 0.707 * Vpri with results from PSUII.

I think you meant Vdc = 0.707 * Vsec (rms). In PSUDII, you need to enter the rms voltage not the peak voltage for the secondary, so it seems you would need a PT with higher HT winding to get B+ of 400V. The large ripple isn't too much a problem for class AB amps, so it should be fine. Also, you should let the supply settle down a bit in PSUDII, e.g., run the simulation for 100ms after a delay of 10s.

[matt h]

(deleted)

[Theashe]

I think you meant Vdc = 0.707 * Vsec (rms). In PSUDII, you need to enter the rms voltage not the peak voltage for the secondary, so it seems you would need a PT with higher HT winding to get B+ of 400V. The large ripple isn't too much a problem for class AB amps, so it should be fine. Also, you should let the supply settle down a bit in PSUDII, e.g., run the simulation for 100ms after a delay of 10s.

Yes, yes I did. I was concerned that PSUDII was returning lower voltages than I was expecting, but I suppose when loaded that drop is to be expected.

Also use the "stepped load" fucntion to see what changes occur between idle and full load. That'll let you know how saggy things are.

This is a fantastic suggestion - I hadn't even considered it. Thanks for this one.


I have another question about the number of filter capacitors I need. I have six preamp tubes in my design (for various features), but I don't like the idea of having that many large filter caps. All of the designs I've looked at out there fed each individual tube from a separate capacitor - but then, those designs also had fewer tubes. Can I connect more than one preamp tube to the same supply to cut down on parts? What would this do to performance?

[matt h]

(deleted)