• 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.

EL34 Push Pull Amp Design

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello all! There was some previous discussion of this topic in a previous thread, but it was pretty off topic from the original title and question, and I figured it was time to start a new one. In the last thread, the conclusion I came to was that EL34s would be a good and fairly cheap tube to design something around.


I would like to build a monoblock, as I both don't have the budget for an integrated stereo amplifier, and it also has other power supply complications that I would prefer to not deal with, as I can always build a second monoblock.



In terms of what I would like for output power, something around 40-50 watts would be nice, as that should be plenty for driving my not-very-sensitive Koss bookshelf speakers, but I may also use it to drive a Bose speaker, which is currently plenty loud with a 50 watt solid state amp. I generally listen to a lot of rock music, and often want it quite loud, and I know that 50 watts is plenty. In the previous thread, I got recommended Dynaco A470s from Triode Electronics, and with it's 4300 ohm load, with a B+ of 400 volts, it's load line looks quite nice, provided the screen tap percentage is the 33 percent I read it is. I think UL is probably the way to go, as it means I don't have to deal with regulation for full pentode mode, but also don't sacrifice much power like I would with triode mode. However, the Dynacos are a bit expensive, as I don't have a huge budget. I'll probably build the chassis myself with help from my dad, but I'm a high school student, and so my income and budget are quite limited. Would there be any disadvantage to going with something like Edcor iron? (aside from the long lead time)


I also looked at some 6600 ohm load lines, and with a B+ of 450 volts, they also looked quite nice.


Mostly unrelated to my questions, but I was talking to my Dad, and he was saying how lucky I am to have this forum as a resource, compared to when he and his father would have been learning electronics. It really is like standing on the shoulders of giants.


Thanks, and let me know if I need to clarify anything!
 
50W from a pair of EL34s is definitely pushing it. Even 40 is probably a little on the high side but is certainly doable. If you really want 50W, I'd go with the 6550 / KT88, but you will quickly find a diminishing return as you push for more power. IMO, the Eico HF89 is a good schematic to start from. I really like the so-called "Mullard Topology".

The main advantage that the A470s have is superior HF performance, particularly above 20 kHz. Better performance at higher frequencies typically makes it easier to stabilize an amplifier with a lot of feedback.
Edcor makes excellent products, and they are hard to beat for the price. I really like the appearance of them as well- I think the blue color looks really nice.

I definitely recommend AnTek for the power transformer. Their power transformers are great quality and very economical. I'm about to order two of their 500VA units to use in a dual 0-70V, 0-7A linear power supply which I do not expect to treat gently. For a monoblock I would think that a single 100VA unit is probably adequate, though the 200VA unit is only about $4 more, so if you have the chassis space I would likely choose it since you will have more flexibility in the future.

AnTek's output transformers are new, unproven and need independent testing to verify if their atrocious response curve is accurate.




As you mentioned, this forum is a great resource with some incredibly brilliant analog minds and IMO is one of the better forums for analog design on the internet. Another great resource for beginners (and non-beginners) when it comes to tubes are the books by Morgan Jones. I also like the Radiotron Designers Handbook (PDF is available if you look) and some of the old ARRL handbooks, though they tend to be more focused on RF work. If you're interested in solid state, then the works by Douglas Self, Bob Cordell and of course AoE are all excellent resources.


Edit: If you haven't already, go download LTspice. It's a free and incredibly powerful circuit simulation program. I highly recommend checking out Mooly's guide for the basics: Installing and using LTspice IV (now including LTXVII). From beginner to advanced.
 
Last edited:
It's not too hard to build an amp on a budget and EL34's are a reasonably priced OP valve compared to KT66's/6L6GC's etc. I'd recommend using the old style valve bases rather than the Chinese ceramic jobbies which most in my experience are crap; no grip, hard to mount.

Cheap chassis can be easily constructed out of copper clad board, it's easy to solder bits together and easy to drill. A wooden mockup of you chassis layout can be made to get an idea where everything goes, saves having to through away a metal part and redo.

If you can find cheap used toroidal tfmrs of about the right VA it's pretty easy to rewind them for HT etc. I used Hammond OPT's in my EL34 monoblocks but mine were 3 paralleled to give 120w. As mentioned above 2 EL34's @ 50w is really pushing it, 4 EL34 in PP would give you an easy 50w,with lower HT/B+ just a thought.

Good luck, Andy.
 
50W is pushing it ? Look at the Mullard data, up to 55W no problem.
You want pushing, look at what Telefunken propose :cool:
Mona
 

Attachments

  • EL34-Mullard.png
    EL34-Mullard.png
    131.6 KB · Views: 1,093
  • EL34-Telefunken.png
    EL34-Telefunken.png
    82.7 KB · Views: 1,070
Okay, so the ham radio operator would say "the tube is rated for 25W dissipation, that means that if I accept a little bit of cherry on the plates it can take 30".

This isn't necessarily good practice for audio amps (not that it's good practice for RF amps either). I hate pushing tubes hard in audio service, because when they fail there's always a chance of them taking out something expensive. Especially in UL operation, a short involving the screen grid can take out an output transformer. It's rare, but less so on amps that push tubes hard in UL service. A 100R screen resistor and a 10R 1/4W cathode resistor can help, but I still don't like pushing my luck. Regardless, it's a massive transient that will get your attention.
 
So would it be better to shoot for something like 35 watts? That will almost certainly still be sufficient. I'll do some looking at those schematics later! For tubes, I was going to buy a matched pair of JJ EL34s, and the datasheet does spec them at the same level as the Mullard one does, but I have no plans on running them at 800 volts on the plate.


In terms of amp design, I've read over all of Morgan Jones' website and it's been immensely helpful. Maybe I should pick up his books. I also do have a copy of the Radiotron Designers handbook as a PDF.



Thanks everyone for your suggestions. I have a lot of reading and research to do!
 
I've been looking over schematics, and quite like that Eico schematic. It's easy to understand how it works and is quite simple overall.



I think I'll probably wind up going with an Antek toroid for the power supply. However, I'll probably use PSUD2 to come up with a choke filtered supply instead of the one without that the Eico has.



I was looking at Hammond output iron, and it's both significantly more expensive and people don't seem to like it that much. The Edcor iron looks really good, but the lead times are very long according to their site. I may wind up using the originally recommended Dynaco iron if the Edcor + shipping isn't much less.


Thanks for all the input!
 
I see from the measurements you are using the 8 Ohm tap from the secondary.

You could wire the secondaries for 16 Ohm, ground the central (it would be 4 Ohm tap) and do some cathode feedback as well. Then the 8 Ohm speaker will be connected across three secondary windings in series.
 
The ear's response to sound levels is logarithmic.
The ear responds to average sound level, and then the bones change the
mechanical advantage (or disadvantage) from the eardrum to the cochlea, accordingly.
For the ear to perceive a doubling of sound level, you need 10 times the power.

For the bodies perception to vibration levels, perhaps it is more linear.
(there is no changing of the mechanical advantage).
So double the vibration power may be perceived as double the vibration.

Want more vibration?
Then someday later, add a real powered subwoofer.

30 actual clean watts will almost certainly be enough for your loudspeakers.
You are not deaf, and you do not want to become deaf.
The federal safety sound levels versus hours of exposure, are not very high (but they are probably not conservative enough, and some get deaf at those levels).

Some research into all the above may prove worthwhile for you when you get the time.
Might be another interesting science project.

I purchase my JJ tubes from Eurotubes.com (they do lots of re-testing). No, I am not paid by eurotubes (But I can drive over and get the tubes, no shipping required).

My only Encore output transformer works great. But it is only a 10 Watt model. I use it in Ultra Linear mode.

A local group I am in has used many of the Hammond push pull output transformers, from 10 Watt to 60 Watt models. They work very well.

One of the more difficult parts of building a push pull amplifier is designing and adjusting the negative feedback.

For my latest single ended amplifier, just 2 stages, I am using output tube plate to driver tube cathode negative feedback. It has been easy to adjust that feedback.
It eliminates the larger frequency response and phase response problems when taking the negative feedback from the output transformer secondary (higher group delay from the secondary than from the primary).

I am considering using that negative feedback technique for push pull, but with 3 stages:
Current sink cathode coupled Phase Inverter, Balanced Driver, push pull output stage, with the feedback from the output tube plates to the driver tube cathodes. The idea is to make the negative feedback more easy to adjust and stabilize.

All the above are just my opinion.

"All Generalizations Have Exceptions" - Me

But you will find that there are many push pull EL34 circuits that can conservatively put out 30 to 40 Watts.
Pick one or two, and these who already posted can give some hints and kinks.
Some want power.
Some want finesse.
Some want both.
 
Last edited:
Thanks for the informative read! I'll absolutely consider getting a sub in the future. I agree that 30 watts should be plenty, so I'll shoot for 35ish at low distortion.


I was looking at loadlines and the attached one looks pretty good. Not sure what people think of the site though. Let me know your thoughts!


It's 400v of B+ into a 5k P-P transformer with 40 percent screen taps, which is a spec available from Edcor for a decent amount of money. I plan on using fixed bias, so the loadline is biased at around 70 percent of max dissipation, which at 400v is 62.5 ma.



Thanks!



View attachment Loadline 5k 400v B+.pdf
 
400V x 0.0625 Amps = 25 Watts.
The EL34 plate can take up to 25 Watts,
The EL34 screen can take up to 8 Watts.
That is 33 Watts total.
But you might not be able to easily divide the power between them at the ratio of 25:8.
And 25/33 = 75.8% of total plate plus screen dissipation (Hot!).

And Ultra linear screens are at a higher voltage than the plate, because the DCR times plate current = voltage drop in the output transformer winding to the plate (DCR from the screen tap to the plate tap).

Lots to look at here.

For fixed bias, I recommend a separate potentiometer for each output tube (set the current individually for each tube).
Push Pull output transformers like it when the plate currents are very well matched (it prevents early core saturation).
Of course the screen currents need to be closely matched, but the current is smaller, and the turns are 40% of the total plate turns (so screen Amp X 40% Turns = magnetic flux).
The normal / often practice with fixed bias is to put an individual 10 Ohm resistor from each cathode to ground, and then measure the voltage across the resistor.
62.5 mA across 10 Ohms = 0.625 Volts.

The software printout has some disagreements.
The screen says 200V; but Ultra linear puts it at 400V.
The load line is broken, one slope from 400V up, and a much higher slope from 300V down.
It also crosses the maximum plate dissipation dotted curve quite a bit.

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