Single Ended EL84 tube amp plate voltage thoughts...

[eschertron]

All good points. Just speaking for myself, on a scope I can see cutoff a lot more distinctly than grid limiting, which appears softer (depending on grid stop value, another topic!). The softer transition affects the nature of the harmonics generated, and has informed my choice of Rk. For a more 'tubey' sound in a low/med gain amp, I'll definitely go with a smaller Rk. Sounds 'warmer' to me.

Now center-bias is warmer than cold-bias, so it may be all relative.

[Enzo]

Two points.

The tube only ever operates in the circuit. Whatever voltages are present when the tube and circuits are removed is not the operating conditions.

100k and 1.5k are so common because they are directly out of the RCA book in the recommended values section. PAge 647 in the RC30. 300v supply section.

[eschertron]

Two points.

The tube only ever operates in the circuit. Whatever voltages are present when the tube and circuits are removed is not the operating conditions.

100k and 1.5k are so common because they are directly out of the RCA book in the recommended values section. PAge 647 in the RC30. 300v supply section.

Surely. We got off track when I asserted that 100k/1.5k aught to work generically. What Kev wants for the stage is entirely dependent on what its desired effect on the signal is. Clean amplification? Maximum gain? I don't know, exactly. But throwing generic components into the circuit and then taking voltage readings gets us to a point where some additional calculations can be made.

[Malcolm Irving]

Like a lot of things with tubes, the more you look into it, the more complicated it gets.

The horizontal axis on the plate characteristics is Vak, so really we ought to add Vk to that to get Va.

The load line being straight is based on the assumption that B+ (or Vpp) stays constant while Ia varies. For audio frequency signals, that is a reasonable assumption (as the power supply filter cap tries to keep B+ constant).

To find the quiescent point, we get the correct values if we base the load line on the (sagged) B+ that corresponds to whatever Ia is at the quiescent point.

A long time ago, I experimented with warm bias (on a 12ax7 V1) by using a pre-set for Rk. I expected to get extreme distortion (with something like half-wave clipping) as I reduced the pot to zero ohms. It never happened – the signal stayed clean all the way. I figured out that at some point grid-leak bias takes over and holds the bias up after the cathode bias has gone too low. From memory the warmest bias I could get was with Rk at about 700 ohms – any lower value of Rk had no further effect - the bias had reached a ‘warm limit’. (I should have reduced the value of the grid-leak resistor from 1m to 470k, 220k etc. to take the experiment further – but I didn’t pursue that at the time.)

Edit: By the way, the sound I was getting at the 'warm limit' was quite acceptable - it seemed a bit 'vintage' to me - like the sound of an old 'worn out' amp. The grid-leak bias that you get does vary a lot between tube samples - which could be seen as a problem (or a virtue if you enjoy tube-rolling).

[kevtronic]

edit:...Duncan Amps' PSUD II is a very handy tool to have, but even without you can guesstimate the Vpp at the preamp node by figuring 2v drop from the node before for each 1k of dropping resistor, for each 12AX7 on that node...

Thanks for the info and sharing The PSUD II software. I've been to his site a few times and never saw it; if I did it didn't click.

Edit 1 Starts:
I ended up using a 1.5K cathode resistor and a 100K plate resistor. The load line showed a good match with about .9mA/-1.5v. I also ended up bypassing 12ax7(a) with a 25uf cap. 12ax7(b) isn't bypassed and I'm using a 22K resistor for the negative feed back. I tried 33K and 47K and didn't care for it. I am still using a screen resistor, but dropped it from 5.1K to 3K which balances the screen to cathode real close with plate to cathode (within a volt or less). If I use any less the screen voltage raises above the screen voltage. I thought maybe I reversed the plate and screen taps of the OT, thinking if I reversed them I'd have less on the screen and more on the plat, but they are how the original was. Not to mention the TR is based on the RED/Blue wire pair. So, I did some searching around for SE Ultra Linear amps and found others that, for what ever reason, have the screen voltage above the plate voltage. If anyone has a clue about why, I'd like to know.

Edit 2 Starts:
Thanks for all the input, etc, on this thread. Great stuff. I have waded deeper into the electrons. Learned a couple new things and cleared up (reinforced) other things. The amp is all soldered together. I am redoing the faceplate but it should be back together soon. I'm now going back through the thread to review and see if I can glean a few more pointers. Thanks again...

Edit 3 Starts:
Ok, I've gone back thru, took some notes, and I will come back to it later. I'm going to the brain electrons chill a bit while I create a new faceplate for the amp. It's Sci-Fi Saturday night on MeTV - don't miss out

[eschertron]

Thanks for the info and sharing The PSUD II software. I've been to his site a few times and never saw it; if I did it didn't click.

Yup. Another tool in the kit that will be handy again and again. I'm sure I'm not alone in using it from the point I start spec'ing PTs to deep into the design revisions. One trick I've learned is to ditch the resistive load elements and use the constant current load elements. The idle dissipation current can be read off a load line chart, tube data pages, or 'thumbruled' from experience. For example, a node that has a 12AX7 (2 stages) on it can be simulated with 2ma constant current. PSUD will take care of all the voltages as long as your loads are reasonable at each stage. You'll see what changing a resistor value (again, for example) does to voltages all through the rail. When I started using it, I modeled some classic power supplies from simple to not-so-simple; Champ, 5E3, Bandmaster, etc. just to get a feel for power supply behavior. There's a lot that can be done with it, but the good thing is you don't have to do "a lot" to start working with it.

[kevtronic]

Yup. Another tool in the kit that will be handy again and again. I'm sure I'm not alone in using it from the point I start spec'ing PTs to deep into the design revisions. One trick I've learned is to ditch the resistive load elements and use the constant current load elements. The idle dissipation current can be read off a load line chart, tube data pages, or 'thumbruled' from experience. For example, a node that has a 12AX7 (2 stages) on it can be simulated with 2ma constant current. PSUD will take care of all the voltages as long as your loads are reasonable at each stage. You'll see what changing a resistor value (again, for example) does to voltages all through the rail. When I started using it, I modeled some classic power supplies from simple to not-so-simple; Champ, 5E3, Bandmaster, etc. just to get a feel for power supply behavior. There's a lot that can be done with it, but the good thing is you don't have to do "a lot" to start working with it.

I am going to follow your lead on with the PSUD II Software. Thanks, for you help, time, and energy...