• 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.

Two stages EL34 power amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have build EL34 amp with differential phase splitter and tubes run in SRPP mode. It sounds very nice and has undistorted and very warm sound. It seems to be under power. I think that I have applied too much global feedback. I have no idea how to calculate it. Any help in solving this problem would be greatly appreciated. Also the topography of my design may be wrong. I have never seen anything like this.
Here goes the schematic.
:scratch:
 

Attachments

  • EL34-Amp01a.jpg
    EL34-Amp01a.jpg
    56.7 KB · Views: 585
... It seems to be under power. I think that I have applied too much global feedback. I have no idea how to calculate it... :scratch:

Do you mean that you need too high input level to get full output power ?

You can simply measure the amount of GNFB.
Disconnect the GNFB line from the OPT, measure the sensitivity and compare to the sensitivity when GNBF is connected.

There should be no reason to have more than 16...18 dB GNFB, especially when you have UL-connected output stage..
 
Thanks for reply.
When connected to scope 1Khz 0.2V sin wave produces 2V output. I don't know how many dB of GNFB I applied. How do I calculate it?
As to power, at full volume I hear no distortion and it seems to be long way from overdrive. I wander if less GNFB would improve sound quality.
 
Disconnect the GNFB line coming from the OPT to 4k7 / 360p and measure what input level is needed to produce the same 2V output voltage.
Then calculate the ratio of 0,2 V and the level just measured. And transform to dB's.
The amount of GNFB is the amount of the sensitivity is reduced when the feedback line is connected.
Less GNFB means higher distortion, lower S/N and worse frequency response - assuming the original GNFB was correctly accomplished.
 
Last edited:
Thanks for reply.
When connected to scope 1Khz 0.2V sin wave produces 2V output. I don't know how many dB of GNFB I applied. How do I calculate it?
As to power, at full volume I hear no distortion and it seems to be long way from overdrive. I wander if less GNFB would improve sound quality.


your closed loop gain of 10 is indeed small...at least twice that will give a perception of more power...

you play around with the 4.7k/47ohm feedback network, getting the ratio higher decreases feedback and increases your closed loop gain...
 
your closed loop gain of 10 is indeed small...at least twice that will give a perception of more power...

you play around with the 4.7k/47ohm feedback network, getting the ratio higher decreases feedback and increases your closed loop gain...

I agree. The gain is now small, but there is only phase splitter before output stage and it is therefore one reason for the situation.

Play around with 4k7 / 47 but do that after you know what the amount of GNFB of your amplifier is.
 
That design resembles a design by one finnish tube veteran, Mauri Pännäri. His design wasn't exactly as yours, but similar in that it also had a single stage up front - a differential amplifier as a phase splitter, just as yours does. It didn't have active loads, though. The power stage was pure class A push-pull, though. Yours seems to be class AB, right?

The first thing I'd do, is measure the gain and balance of the phase splitter. Although, if it does sound undistorted, as you state, then probably the balance at least should be great. By the way, in theory you should be able to ditch the balance pot from the input stage's cathodes. Check how the balance shifts with a scope when you turn the pot from end to end. Of course, disconnect the upper 4,7 kOhm resistor and the parallel cap before doing any measurements I proposed.

It still might be, that you need a preamplifier with more than 2Vrms output to drive this amp to it's limits. Although the amp by mr. Pännäri had an ECC81 up front, it was designed to be used behind a power amplifier. Then again, a gain of 10 should give you an output of 50 watts in to 8 ohms with full CD output (2 Vrms).

After that, there are several resistor values in the schematic I would like to see a bit smaller, but those have nothing to do with the perceived loudness of the thing. One thing I'd like to state, that with 330k grid leaks on the power tubes G1, you risk the power tubes running amok when they age. In the aforementioned class A -amp I had 330K grid leaks on EL34's, and I never got more than 4 months out of a pair of output tubes. After that at least one would go south by turning it's anode nicely cherry red. Switched the grid leaks from 330K to 220K, no more problems. What I observed when I dug into the problem, was that the grid was starting to go emissive with age and heat (it was class A, thus HOT), and the current caused the voltage at G1 to rise, which again worsened the situation, and the result would be a nicely glowing catastrophic power tube failure. I remember my findings being in line with what was written in Robert Tomer's 1960 book "Getting the most out of Vacuum Tubes".
 
if we assume that the LTP has voltage gain of about 50, the EL34 around 5, so that is x250 or around 48db....assuming 4.3k:8 opt, turns ratio is 23 or 27db so that overall gain is 48-27 or 21db....so that closed loop gain of 20db gives a feedback factor of only 1db.....you can try operating without the feedback loop connected and see what gives....

more knowledgeable members, please check my math, i could be wrong....
 
Thank You All for replies

I took some advice and I'm listening to the music now.
It sounds very good, nice bass and high frequencies. On the bench I run it without feedback resistor. I thought amplification was 48. Next I added 250k pot and run it down till amplification starts to drop. It started dropping at 20k, then I soldered in 20k resistor. Now amplification is 13.5.
I'll do it over in a few days. My scope is a cheap piece of junk from Radio Shack and I think I'm getting wrong readings. As to loudness, vol control needs to be at about 11 o'clock for comfortable listening and at 1 o'clock I don't hear my telephone ringing. Sound is good and beats my David Hafler amplifiers so I'll keep it as is for now, till I get better scope.
Once again, thank You All for help and cheers.:drink:
 
Last edited:
Hi again
I made mistake. I wired pentodes in triode mode. Stupid me. Now I corrected it and amplification is 20 without GNFB and 16 with it. Not much but I can hear the difference. Sound seems to be more dynamic. As to my scope it is still piece of junk. Switches are worn out and need tapping to keep the waves from jumping up and down the screen. One of these days I'll tap the whole thing with the sledge hammer and fix it for good.
On the picture You can see where I made mistake.
Cheers again
 

Attachments

  • _IGP1716.jpg
    _IGP1716.jpg
    117.6 KB · Views: 207
Hey voyciek,

Can you explain the necessity of the 1Meg/4,7u in the feedback loop?

750ohm at the Sg seems high, any reasons for that? In the ballpark of 100ohm normally works best.

NFB should be about 1db with the values you have given, 47ohm-4.7k. Some of you guys miscalculated with a factor of 10;). You need for an acceptable result with UL wrt dist and Zout. Triodestrapped you shouldn´t need any NFB at all.
 
Last edited:
There is a danger with your Bias circuit. If or when the wipers on the 100k pots go intermittent then you'll lose bias completely to the relevant EL34. I would swap the pot & the 100k fixed resistors positions. Ground the wiper with one end of the pot. The 100k resistor should then be split in two, choose values to give maximum bias and minimum bias dependent upon the pot wiper position. Other possible ideas are also available on here if you care to look around. :drink:
 
Don't know

That was not answered. I've never seen that before. What is the purpose of the 1M/4u7 in the feedback loop, on the grid of the 12AX7?

thanks
--
I see it this way in differential amplifiers but I don't why it is this way. I was hoping that someone can give me some idea what to do with them apart from sticking them where sun doesn't shine.
 

Attachments

  • Spliter.jpg
    Spliter.jpg
    21.5 KB · Views: 154
Good idea

There is a danger with your Bias circuit. If or when the wipers on the 100k pots go intermittent then you'll lose bias completely to the relevant EL34. I would swap the pot & the 100k fixed resistors positions. Ground the wiper with one end of the pot. The 100k resistor should then be split in two, choose values to give maximum bias and minimum bias dependent upon the pot wiper position. Other possible ideas are also available on here if you care to look around. :drink:
Thank You for suggestion. It makes sense to me. When I get some time I will rewire BIAS circuit his way. On diagram I wrote optional values in red.
 

Attachments

  • BIAS.jpg
    BIAS.jpg
    45.1 KB · Views: 154
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.