• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Designing a Single-Ended Triode EL34 Power Amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
WAXX:

You are welcome. I did none of the above with 'modeling' or 'simulation'. Back “BC” (before computers), one designed circuit after circuit on paper, with a pencil, and perhaps either a “scientific” calculator … or even older, just a slide-rule. Because ± 1% was just fine. For everything except the most technical things.

So … we “remembered formulæ”, so to speak. Simple stuff:

(1) Z = 1/(2πFC) … as the impedance calculation for capacitors vs. frequency.
(2) Z = ( 2πFL ) … for inductors
(3) P = IE … power
(4) E = IR … Ohm's law, resistance.

And a few things like when "Z" is equal to "R" (for pass capacitors), then one can expect a –3 dB response. So, for the instance of your 47 kΩ 2nd stage grid pull-down resistor and the 0.047 μF DC blocking signal-pass capacitor,

(10) Z = 1/(2πFC) which rearranges to
(11) F = 1/(2πZC) which substitutes in
(12) F = 1/(6.28 × 47,000 Ω × 0.047×10⁻⁶ farad)
(13) F = 72 Hz

That's where I got that recommendation number (6).

_______

The point is, to learn to “handle” equations in (1) thru (4) above. And to be confortable and confident in using basic algebra to rearrange 'em. As I did in (10 thru 13). Seriously: do try it on paper. With a pencil. And a handy erasor. Because "it is easy" and somehow writing down the numbers on the back of a junk-mail envelope is … satisfying.

GoatGuy
 
Last edited:
WAXX:

You are welcome. I did none of the above with 'modeling' or 'simulation'. Back “BC” (before computers), one designed circuit after circuit on paper, with a pencil, and perhaps either a “scientific” calculator … or even older, just a slide-rule. Because ± 1% was just fine. For everything except the most technical things.

So … we “remembered formulæ”, so to speak. Simple stuff:

(1) Z = 1/(2πFC) … as the impedance calculation for capacitors vs. frequency.
(2) Z = ( 2πFL ) … for inductors
(3) P = IE … power
(4) E = IR … Ohm's law, resistance.

And a few things like when "Z" is equal to "R" (for pass capacitors), then one can expect a –3 dB response. So, for the instance of your 47 kΩ 2nd stage grid pull-down resistor and the 0.047 μF DC blocking signal-pass capacitor,

(10) Z = 1/(2πFC) which rearranges to
(11) F = 1/(2πZC) which substitutes in
(12) F = 1/(6.28 × 47,000 Ω × 0.047×10⁻⁶ farad)
(13) F = 72 Hz

That's where I got that recommendation number (6).

_______

The point is, to learn to “handle” equations in (1) thru (4) above. And to be confortable and confident in using basic algebra to rearrange 'em. As I did in (10 thru 13). Seriously: do try it on paper. With a pencil. And a handy erasor. Because "it is easy" and somehow writing down the numbers on the back of a junk-mail envelope is … satisfying.

GoatGuy

I prefer Excel above an old enveloppe. But it's true that knowing the math is very important when designing something like a tube amp. That's why i'm studying it and make exercise designs like that above.

The amp is something like i want to build, but it is now mainly an exercise. The build won't be for the next months i think if i build it, and the design may (and has already due to your remarks) change(d) a bit. I'm full recalculating everything with your remarks in my mind...
 
I prefer Excel above an old enveloppe. But it's true that knowing the math is very important when designing something like a tube amp.

Excellent. Excel is. I've loved the power, the "immediate answer power" that spreadsheets deliver. I was a total fan of Lotus 123, of Quattro Pro, then eventually Excel. And still, Excel.

And most unfortunate is the move I made to Mac Excel. Unlike Windows Excel, it has no ability (or support for) writing code-as-callable-functions.

Which is really REALLY sad: I had a whole library of common formulæ that have no direct built-in Excel functions; I had (don't even know where they are) a whole library of hand-coded tube and transistor transfer functions. Gone. It is maddening. But The Big White Box company has decided that its level-of-customers simply are more gifted at getting themselves into support-hêll with callable code than its worth supporting.

GoatGuy
 
Been enjoying my two-week session of learning Python. Will never use Excel again for computations. Excel certainly has its place, and can do complex numbers with a little effort. But nothing beats being able to say a/b+c when they are all complex, and it 'just works'. To say nothing of arrays.
Learned matlab back in college, but not interested in paying the price.
 
Excellent. Excel is. I've loved the power, the "immediate answer power" that spreadsheets deliver. I was a total fan of Lotus 123, of Quattro Pro, then eventually Excel. And still, Excel.

And most unfortunate is the move I made to Mac Excel. Unlike Windows Excel, it has no ability (or support for) writing code-as-callable-functions.

Which is really REALLY sad: I had a whole library of common formulæ that have no direct built-in Excel functions; I had (don't even know where they are) a whole library of hand-coded tube and transistor transfer functions. Gone. It is maddening. But The Big White Box company has decided that its level-of-customers simply are more gifted at getting themselves into support-hêll with callable code than its worth supporting.

GoatGuy

That's one of the many reasons i avoid those computers. The Big White Box company makes good phones and tablets (but too expensive for what it is), but their pc's and laptops are way too limited for what i (as IT-er and nerd) want to do with it. Lenovo makes also very solid machines and windows is not that bad anymore today like it once was. And Linux is even better as OS but lacks the software i need. But i run both and keep OSX for my phone...

But like with most things, each to his own.;)
 

PRR

Member
Joined 2003
Paid Member
...Back “BC” (before computers)
… we “remembered formulæ”...........

Or just bench marks.

"Without a PC", I instantly spotted the C3 R5 dubious values. Self-bias EL34 can use 500K grid resistor; why be 10X smaller? Just makes the driver work harder. Which also means a big cap. No math: remember these happy combinations:

0.01uFd + 1Meg == 16Hz
10uFd + 1K == 16Hz

So 0.05ufd against 50K (always round-off) compared to 0.01u+1Meg. 20X lower R. 5X higher C. Both in the direction of less bass. 20/5 is 4. So the pole is near 16Hz*4 or 64hz.

Maybe 70 if you use the full numbers and proper pi. But already we do not care-- 64/70hz is way too high for "Hi-Fi". You can't even run a lot of 16Hz poles without them adding up to shaved bass.

Assuming that a 0.047u cap is in hand, try the max grid resistor. 0.05u+0.5Meg is 5X better and 2X worse than 0.01u+1meg. Counting on thumbs, 3.3Hz doubled is 7Hz, a fine value. We could even use a less than max Rg, 330K, and hit 10Hz. The very light loading of 330K on the 24K driver reduces driver distortion.
 
Here's an alternative view of cathode bypass and grid coupling cap values: with no (intentional) loop feedback, cap values don't need to be staggered, making cathode bypass cap values independent of each other. Both can be made very large on the grounds that the larger the value, the smaller the signal voltage that will appear across it. As long as the cap acts like a cap, brute force wins.

Choice of grid coupling cap value might also take into consideration overload recovery time. A properly designed amplifier's output stage clips first, and at point of grid drawing current. Grid current charges grid coupling cap, dynamically changing bias of output stage until bled back down. A bleed-down time contant of about a quarter second gives a good compromise.

Also, does your HV DC supply include another bypass cap after the 70R dropping resistor?

All good fortune, a Happy New Year to all,
Chris
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.