Tube theory discussion

[Wolfeman]

If you haven't discovered this forum yet, start reading- http://music-electronics-forum.com/f43/

It's where all the tube snobs hang out. Not worth posting and asking questions there, it always turns into holier than thou debates. But there is some great reading in that place. Tons of topics on dual rails/split rails.

 

[Omega_Void]

...what would be the best way to drop that screen voltage without messing the with the HT of the preamp.

You could use a BJT or MOSFET regulator/filter. You're probably familiar with the circuit:

Connect the drain/collector to the B+ where the screen was connected, and two resistors from there to ground to make a voltage divider giving whatever voltage you want on the screen. Connect that divided voltage to the gate/base of the transistor through a stopper resistor, and connect the screen to the source/emitter.

Stopper resistor value depends a bit on what transistor you use. It's the gate/base capacitance which is important here. Usually needs 100R to 1k, try 470R to start or SPICE it for a better idea.

Transistor will dissipate voltage dropped times screen current pulled, so will need to be bolted to the case at least, and may need a heatsink if you're dropping a lot.

Come to think of it, you could maybe put another resistor between the B+ and the transistor to share the dissipation, if you're dropping a lot of voltage. That's more math, though.

Bypass the output of the reg and the lower resistor in the voltage divider with capacitors to reduce noise.

Use a zener diode instead of the lower resistor in the divider to regulate to a fixed voltage -- the upper resistor then should be sized to give about 25% of the zener's maximum rated current for best results.

You still want resistors on the screen pins, as stoppers to protect against oscillation.

Screen will pull less current than before of course, but if you dimension the voltage divider carefully you can get it so the current pulled from the B+ is pretty close to what the directly-connected screen pulled. That way the voltage dropped in the choke will remain about the same as it was before, and the preamp will get about the same voltage no the plates, although it will vary less than it did of course.

Or you could fit an extra transformer, as Wolfeman said.

I'd go the regulator way 'cos it's smaller, lighter, and doesn't add a new hum field to the amp which might get into the signal.

 

[Omega_Void]

...if there is idle current on both sides of the push pull amplifier, doesn't that mean that both sides will conduct when the signal is present until the signal reaches the bias voltage, making the amplifier not a class B, but a class AB?

Sorry -- you posted that like two weeks ago, you've probably figured this out by now.

You can stop a Class B output from coring out the middle of the waveform by biasing the bases to a voltage equal to the forward voltage of the base-emitter junctions (~0.65V for silicon substrate). At quiescent, the bias current flows in the circuit which elevates the base voltage, but not in the output transistors.

The base-emitter junction is ON and ready to start swinging, but draws no current because the voltage elevation on the base is eaten by the base-emitter forward voltage -- so the emitter is at 0V, same as the other end of the load.

It's usually done by putting a pair of diodes in parallel with the base-emitter junction. Obviously, the forward voltages of the diodes need to be matched to the BJTs. For a high-powered output any matching error will be comparatively negligible -- you wouldn't call it class AB just 'cos the BJTs are swinging a few micro-volts the wrong side of 0V, would you?

 

[philby82]

What program do you guys use when designing your preamps? I've just started playin with ltspice and it seems ok, but is there a better one for tube amps? One where i can play with biasing, and cathode capacitor values etc and see exactly what it does to the waveform?

One where i can feed in audio into the designed circuit and actually hear the output would be sweet, but im guessing not available?

Cheers

Phil

 

[Wolfeman]

LTspice does all that, you need to read the manual.

 

[TheWinterSnow]

The base-emitter junction is ON and ready to start swinging, but draws no current because the voltage elevation on the base is eaten by the base-emitter forward voltage -- so the emitter is at 0V, same as the other end of the load.

I got that part for a BJT. There will be Emitter-Bass current but there will be no Emitter-Collector current which means VCC is being dropped on the transistor. But for a Pentode if the bias voltage on the control grid is low enough (negative) then it will cut off all current from the cathode to plate, and there cannot be control to cathode current when the tube is in cutoff as that would mean (with electron flow) that current would go from the control to cathode which is not possible.

Another thing that is really bothering me, is I have been trying to do load lines for poweramps but for whatever reason I am getting extremely off numbers and I think it has something to do with the anode to anode primary impedance, which I am not really understanding because there is very little on the internet and the terminology is all over the place including on datasheets.

So what the hell is the anode to anode impedance. I know is the impedance between one plate to another through the primary winding of the output transformer and appears as the load for the tube and dictates the maximum amount of current a tube can conduct under the best conditions. I am having a hard time finding out how to determine the Ra-a so I can actually do a load line. The only thing I could gather was that a 5150's Ra-a was 4235 ohms which under those conditions the class B load line is almost entirely above maximum dissipation (with 450mA peak current) and Class A gets close with to Max Dissipation (with 225mA peak current), where class AB is 337mA putting a 6L6 over the maximum dissipation of a KT88 around 90% of the time (with an average dissipation of 74 watts per 6L6). So either the Ra-a that I was getting was extremely off or I am horrible at deriving it, becuase the rest of the load line is easy, the only thing that is off is the Ra-a.

 

[Wolfeman]

Well you need to know the p-p impedance of the OT. Start the load line at 0 ma and the plate voltage. Take the OT's p-p primary impedance(lets pretend it's 4.3k) and divide it by 4. This gives you 1.1k. Pick a rise in current, lets say 200mA, so the voltage drops to 220. (200mA X 1.1k = 220). If the Pv was 400v, we now have 80v. So plot that at Vp= 80 and Ip=200mA.

That would represent all possible conditions at which the power amp operates when producing signal. Is that what you were looking for? Or are you talking plotting the net plate current through the OT primary?

 

[TheWinterSnow]

What I gathered from is that the 5150's OT impedance was 4235, which 4235/4=1058. 486v/1058=459mA for Ipeak which is just absurd, that placed the q point above the max dissipation of the tube.

I was reading Merlin Blencowe's page on push pull amplifiers because the current book I have on power amps makes no mention of this. He did an example, and it worked out really good, its just that his example had a much larger OT impedance and his HT was much lower. I really can't see a 5150's p-p primary being so low.

 

[Wolfeman]

Where did you get that number for the 5150 OT? For a 50watter that would be correct, but a 100w amp would need more like 2k

 

[TheWinterSnow]

wouldn't 2K (if your talking about p-p impedance) mean more current though? Because I am going back to how I learned load lines, where the cutoff is the supply voltage and the max current is the supply current divided by the load (plus cathode resistance if applicable) and for power tubes the max current they will put out is at the knee where Vg=0v.

Thats the whole thing though, I know the only thing I did wrong was use the incorrect p-p impedance, just getting the correct impedance would make everything work.

 

[Wolfeman]

Hmm I may be confused by what you are confused by, but this might help you out-

When you are drawing the load line, for class AB, for a 4 tube push pull amp, you just draw the load line for one tube. So you draw the load line from the knee of Vg=0 to the idle plate voltage. You would then use (Vpp -Vp)/Ip=Rp.

You then need to take that Rp and determine what it would be for push/pull with parallel tubes. So for 2 tubes in parallel, Ip would be double. This means the equivalent for the push or pull tubes in parallel is now half the value of Rp. The plate to plate OT primary impedance needs to be four times the final value of Rp. It's working out on my end.

BTW- now you can easily figure out the output power of the amp at this point, by ((Vpp-Vp)x(2x(Ip)))/2

Edit- You shoudl really check out this book for all things power amps, so far it's the only way I've been able to understand them- http://www.ampbooks.com/home/books/power-amps/

Haven't read Merlin's new power amp book yet, but it's probably just as good. He tends to reference Richard Kuehnel a lot anyway.

 

[TheWinterSnow]

This is what I was reading: http://www.freewebs.com/valvewizard1/pp.html

I just want to draw the load lines so that I can see that a specified tube won't run too hot but get full power given the plate and screen voltages. I was trying to go off Merlin's page because I couldn't understand anything else I was reading, and his page on the topic actually is closer to how I understand load lines from transistor/triode preamp stages.

I have Richard Kuehnel's book and I can't make sense out of it or find specific topics and he wants to go on tangents about single ended amps, including giving SE example in the push pull chapters. I can't for the life of me find him just going over a load line for a class AB push pull, with actual pictures showing the load lines. I have actually turned to Merlin's preamp book as he seems to have more organized information on the topic of pentodes, but its a preamp book so it doesn't have everything.

 

[Wolfeman]

Page 257-274 in Richards book should be all you need. It's strictly about a push pull power amp, and doesn't stray. He starts with selecting a tube(6L6) and goes on from there.

I understand what you are saying about him going on tangents, because he does throughout the first 17 chapters. It's not nearly as bad as Kevin O'Connor's books though, I have those damned books tabbed with reference pages to other volumes, that's how bad he jumps around.

 

[TheWinterSnow]

yeah I am actually on chapter 18 right nowit makes a bit more sense than the other chapters.