• 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.

MULLARD 5-20 build with mods

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm building two amps using the Mullard 5-20 circuit with mods. One is using negative bias on the op valves. Another is replacing V1 the EF86 with an ECC83.
Any thoughts?diyAudio

Note the use of the negative bias rail and driving both grids of V2. What gain and distortion can I expect compared to the original EF86 circuit.
 

Attachments

  • Scan0037.jpg
    Scan0037.jpg
    49.3 KB · Views: 939
Last edited:
I'd stick to the original topology, a single-ended input stage direct coupled to a long-tail pair. An ECC83 with a very high plate load and minimal cathode degeneration should work great for that. You might think about biasing it with an LED in series with a 47R (47R connected to ground), then return the feedback to the junction of the 47R and the LED. A CCS plate load would be icing on the cake and further improve linearity.

Because the ECC83 current is low, you'll want to run a resistor from the B+ rail to the top of the LED to increase the current through it, if you don't use a CCS at the plate. If you size it so that 6mA of current run through the LED, the 47R will contribute another 0.3V or so, so you might want to use an IR LED for that position to get the right bias voltage on the ECC83 cathode.
 
I am thinking of using the Mullard circuit too;
On another thread I found the following.
Best to keep the EF86 and get rid of the 12AX7 and use a 6CG7/6SN7 from what i have read.
Phil
 

Attachments

  • harmon_kardon_citation_5_schematic_diagram.gif
    harmon_kardon_citation_5_schematic_diagram.gif
    32.8 KB · Views: 864
  • amplifier_30wpp.pdf
    260 KB · Views: 135
What gain and distortion can I expect compared to the original EF86 circuit.

It is essential to know that the distortion is mainly generated by the output tubes, not by voltage amplifying stages. And the THD is mainly reduced by GNFB.
Therefore there is no big difference on THD level if anything is modified at the front end of the amplifier assuming the amount of GNFB is not changed.
 
I'm building two amps using the Mullard 5-20 circuit with mods. One is using negative bias on the op valves. Another is replacing V1 the EF86 with an ECC83.
Any thoughts?diyAudio

Note the use of the negative bias rail and driving both grids of V2. What gain and distortion can I expect compared to the original EF86 circuit.

what output transformer will you use?
 
I am thinking of using the Mullard circuit too;
On another thread I found the following.
Best to keep the EF86 and get rid of the 12AX7 and use a 6CG7/6SN7 from what i have read.
Phil

I'd agree with this. The 12AX7 in the position of Long Tail Pair phase splitter is not the best tube to use for that purpose. It's got quite high plate impedance and not a lot of current drive. The 6GC7 or 6SN7 works extremely well in this position.
 
It's not really a 5-20 since you have the input stage set up as a diff amp. Your issue is likely to be in the second stage, since the first stage plates are unlikely to have the same DC voltage on them.

I don't understand. I've read in a number of places that cascading diff pairs will improve the balance at the output of the second diff pair's plates, as compared to the balance achievable by a single long-tailed pair fed from a single ended source.

I know that's in "Valve Amplifiers" but I can't remember exactly where. Probably somewhere in the discussion of the Crystal Palace amp, which uses cascaded diff amp pairs.

Also, if the first stage LTP has a constant current sink in the tail and matched plate load resistors, that will enforce balance in the first stage, correct? Then that pretty good balance should be made a little better by its connection to the second diff (push-pull driver) stage, no?

That all seems to contradict the idea that imbalance in the first stage will "upset the apple cart" and make good balance difficult to achieve from the 2nd stage's plates. However, I'm sure I'm missing something.

--
 
I don't understand. I've read in a number of places that cascading diff pairs will improve the balance at the output of the second diff pair's plates, as compared to the balance achievable by a single long-tailed pair fed from a single ended source.

You're confusing AC and DC balance. Look at the first stage- the cathode voltages and grid voltages are forced to be equal for the two sections. So if the sections aren't perfectly (and I mean PERFECTLY) matched, the DC plate voltages will have to be different, even if the long tail is a CCS. You can verify this on the bench with a quick test circuit. But all the long tail (or CCS) does is insure that the sum of the two sections' plate currents is constant- the two currents do not have to be equal. So even with a DC imbalance, the AC balance is maintained, since a change in one side must have an equal and opposite change on the other side.

I posted a demo a while back showing the results of a real mismatch, with two different tubes used in the LTP (I think it was a 12AT7 and ECC88). The AC balance was, as advertised, perfect. But the DC plate voltages were severely different, sort of a reductio ad absurdum of this point.
 
You're confusing AC and DC balance. Look at the first stage- the cathode voltages and grid voltages are forced to be equal for the two sections. So if the sections aren't perfectly (and I mean PERFECTLY) matched, the DC plate voltages will have to be different, even if the long tail is a CCS. You can verify this on the bench with a quick test circuit. But all the long tail (or CCS) does is insure that the sum of the two sections' plate currents is constant- the two currents do not have to be equal. So even with a DC imbalance, the AC balance is maintained, since a change in one side must have an equal and opposite change on the other side.

I posted a demo a while back showing the results of a real mismatch, with two different tubes used in the LTP (I think it was a 12AT7 and ECC88). The AC balance was, as advertised, perfect. But the DC plate voltages were severely different, sort of a reductio ad absurdum of this point.

I understand that.
The Michaelson and Austin TVA1 power amp has extra resistors and capacitors between V1 & V2. Also, V2 has been changed too. I am trying to keep it simple. Maybe fitting a 50k pot between the 330k resistors and the HT - and the same for the 150k resistors. I don't mind re-adjusting these from time to time, likewise the bias pots.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.