Author Topic: Primary wires to output transformer - which to twist together?  (Read 328 times)

Offline Qualityten

A couple of years ago I built a clone of the Mars EL84 P-P amplifier. I admire the look and neatness of his chassis layout, so I studied and followed the pictures given here: https://www.avforums.co.za/index.php/topic,66441.0.html

I later removed the tranformers from that build and to use on a Dynaco SCA-35 restoration.

Returning to this build with different UL output transformers, I was a bit surprised to see that I had twisted the two screen grid wires together and the two anode connections together for each side of the amp. Rechecking the pics linked to above, I see that's what Karel has done.

I'm very much still a learner, and would appreciate understanding what the thinking is behind this. Does it reduce hum more than if the wires from pins 7 and 9 of each valve are twisted together themselves?

Greetings from the UK and good wishes to all for these tough times. I had the privilege of visiting Karel in June last year. How different this year has been.

Offline fredeb

Re: Primary wires to output transformer - which to twist together?
« Reply #1 on: June 25, 2020, 03:55:01 PM »
If you left both Ultra Linear taps (the 2x screen grid taps )  disconnected and joined pin7 and pin9 , and connected the outer (anode) taps your amp will be connected in triode mode ( less power , more musical sound ) .

If you connect both Ultra Linear taps to pin9 (screen grid) of each EL84 , and  both outer (anode) taps to each EL84 your amp will be connected in Ultra-linear mode . Pin7 and pin9 not connected .

Take note that each Anode tap has Ultra linear tap closest to it going to the same tube . Check resistance between one anode tap and both UL taps . The lower resistance UL tap is the partner to that anode tap .




Below is a reference to different modes from https://dynacotubeaudio.forumotion.com/t3236-advantages-of-the-ultralinear-output-stage

Not shown on example is that in Pentode mode there will be DC connected to screen grid , probably about 2/3 the anode voltage . Notice that in triode mode the 3rd and 5th harmonics are much lower in distortion measurements .





...evolution is the gradual development and stratification of progressive series of wholes, stretching from the inorganic beginnings to the highest level of spiritual creation.
Jan Smuts

Offline handsome

Re: Primary wires to output transformer - which to twist together?
« Reply #2 on: June 25, 2020, 05:06:59 PM »
Twisting wires together aims to make their radiated magnetic fields cancel. Current flowing in a wire generates a field around the wire. Running two wires parallel to one another makes their fields add. Running them at 90 degrees to one another makes the fields cancel. A tightly twisted pair of wires criss-cross one another at almost 90 degrees with each twist. The current in one wire flows opposite to the other. This plus the twists reduces the radiation of hum - think wires carrying mains of 240V at 50Hz - or high voltages delivered to a transformer via a valve.

Twisting also helps reduce interference received by the two wires carrying a signal. Basically by twisting, the two wires now occupy the same space. So any interference should then be received equally by both wires. The signal voltage is the difference between the two wires (think one wire signal other wire earth). The interference is the same between the two wires. The input of your device amplifies the difference between the two wires; it subtracts the voltage on one wire from the other so interference is cancelled but the signal remains. This is the principle of balanced operation but it still holds for single-ended operation. You twist signal wires to try make them look like one wire to interference.

There is always a high DC voltage on all the taps on the valve side of an output transformer so careful with the fingers....

Offline Mars

  • Commercial Member
  • Trade Count: (+1)
  • *****
  • Posts: 1,250
  • Total likes: 77
  • “The path is the goal.” Mahatma Ghandi
    • Mars Tube Audio
Re: Primary wires to output transformer - which to twist together?
« Reply #3 on: June 25, 2020, 10:09:48 PM »
Excellent reply handsome. Thanks.

fredeb: I presume those UL graphs are for 43% UL taps? I cannot see any reference that clears that up.
« Last Edit: June 25, 2020, 10:19:17 PM by Mars »
OEM manufacture of Tube Audio
Power and Integrated Amplifiers
Phono Pre-amplifiers
DIY Kits
DIY Workshops

Offline fredeb

Re: Primary wires to output transformer - which to twist together?
« Reply #4 on: June 25, 2020, 11:53:13 PM »
Excellent reply handsome. Thanks.

fredeb: I presume those UL graphs are for 43% UL taps? I cannot see any reference that clears that up.

I misunderstood the OP's question Karel . Sorry .  :-X
...evolution is the gradual development and stratification of progressive series of wholes, stretching from the inorganic beginnings to the highest level of spiritual creation.
Jan Smuts

Offline Qualityten

Re: Primary wires to output transformer - which to twist together?
« Reply #5 on: June 26, 2020, 12:51:54 AM »
Thanks Handsome.  I understand the value of twisting AC wires, e.g. wires to to heaters, to reduce the magnetic field.  but I'm wondering whether it matters which wires are paired together.

So, thinking of just one channel of an amp, is there a difference between:
(a) Karel's practice of twisting the two screen grid connections together and the two anode connections together?
(b) And the practice I've seen elsewhere, e.g. in Dynaco assembly manuals, where the screen grid and anode connections for one valve are twisted together and the screen grid and anode connections of the other valve are twisted together?

In (a) the two screen grid wires are twisted together (one twisted pair) and the two anode wires are twisted together (another twisted pair).  I've not seen this done elsewhere.  Is there any difference in the cancellation effect between (a) and (b), given the difference in phase of arrangement (a)?

I hope that makes my question clearer.

Offline handsome

Re: Primary wires to output transformer - which to twist together?
« Reply #6 on: June 26, 2020, 11:29:08 AM »
Whoops sorry, i was rambling without addressing your concern!

Currents should flow in opposition in a twisted pair so you would want twisting anodes as a pair and screens as a pair for cancellation. For the maximum cancellation you would want to (3-wire braid?) twist the HT feed to transformer center tap in there too. The problem is that the valves are (or should be for heat reasons) physically far apart so you can twist screens anodes etc but they eventually have to part. Dynaco probably twisted anode and screen for neat wire dressing.

One place where twisting is very important is the loop from power transformer output to rectifier to first capacitor. This path has very high current spikes within it relative to the rest of the amplifier's current draw. The currents are high and the spikes introduce high frequencies into the mix. So you want this loop as short as possible as it is potentially the noisiest part of the circuit. If you are using a center-tapped power transformer you should must twist the center tap as well. You also want to be sure your circuit's earth connection goes to the first capacitor and not somewhere within that loop

Offline Qualityten

Re: Primary wires to output transformer - which to twist together?
« Reply #7 on: June 26, 2020, 01:48:47 PM »
Thanks very much Handsome.  So Karel's wiring is an example of best practice--not that I doubted that!

I appreciate your explanation, as good wiring and earthing practice is not always easily available to amateur builders, and yet these are the areas that cause hum. E.g. I have just been rereading the relevant sections of Morgan Jones' Building Valve Amplifiers and he doesn't cover this question of which primaries to twist together, nor the relationship between signal earth connections to the ground bus and the power supply earth connections.

Incidentally, Jones recommends decoupling heaters to the chassis at the valve base with 10nF capacitors. Would this be at each valve? Do others usually do this?

Offline handsome

Re: Primary wires to output transformer - which to twist together?
« Reply #8 on: June 26, 2020, 02:11:56 PM »
Well I am not sure how much of a radiation problem you would have at the output transformer...OTOH wire dressing is crucial in the PSU.

Jones is an excellent resource - although he tends to a 'stream of conciousness' type of presentation. So you have to read and think, to get the full value of his knowledge

Sent from my SM-A500F using Tapatalk


Offline handsome

Re: Primary wires to output transformer - which to twist together?
« Reply #9 on: June 26, 2020, 02:49:52 PM »
10nF at each heater pin per valve, to chassis would be bypassing noise right where you would want to - at the "input" which would in this case be the heater.

Sent from my SM-A500F using Tapatalk


Offline Qualityten

Re: Primary wires to output transformer - which to twist together?
« Reply #10 on: June 26, 2020, 09:45:10 PM »
 Thanks again for your advice Handsome, including how to get best value from Reading Jones!