• 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.

Will differential share the same THD characteristics as SE, if same triode is used?

I've embarked on a pre-amp project recently. Originally I wanted to use one SE stage plus one SRPP stage, but later I changed my mind to adopt topology as one stage of differential and one stage of SRPP, for differential offers better noice control performance, according to theory.

But I have one puzzle: SRPP, at certain load range, will exhibit higher odd distortions than even. To compromise this, I decide to use a triode that may produce higher even distortions as the first stage. If it is the original SE plus SRPP topology, there is no problem to achieve the compromising effect. But if I use differential, will the distortion characteristisc be the same as using SE? For instance, a 12AU7.

I've checked the school book and found differential seems to have more odd distortions than even. Am I right?
 

7N7

Member
2003-01-18 10:43 pm
England
Differential pairs (assuming perfect matching of both the valves and the anode load resistors together with an infinite tail resistance) will halve even order distortion and double odd-order.

The key, as Morgan Jones writes in Valve Amplifiers, is to select valves that produce low odd-order distortion.

Therefore do not use a 12AU7 in your differential pair!

7N7
 

7N7

Member
2003-01-18 10:43 pm
England
Majestic said:
I realised they only reduced even-order distortions; but they double odd-orders?! whoa! :bigeyes:

Yes, but it can still be very good.

For example, I have records of tests I did with 955, which I was planning to use for a pre-amplifier.

Configured as a diff pair, with 50k anode loads and just 148V on the anodes (6.5mA per valve) at +30dB output driving a 30k pad, I got THD of 0.07%. 2nd harmonic was -63dB, 3rd, -55dB. I could live with that!

As a matter of interest, when I prototyped the complete thing: the 955s driving 6AW8 DC coupled using the triode as a cathode follower and the pentode as its sink, I got -3dB at 180KHz.

7N7
 

7N7

Member
2003-01-18 10:43 pm
England
Re: Mr. 7N7, may you suggest triodes suitable for differential pair applications?

Kenneth Zhu said:
Thanks in advance.

Yes: the most reliable choice is 6SN7/7N712SN7/14N7. Jones demonstrated in his book that these have consistently low THD and in particular 3rd harmonic. If you need a bit more µ then I can recommend 13D3 (µ = 32) which gave good results when I tested it.

I have no information for higher-mu valves. It is vital to remember that operating point is very important; for example, it is well known that the 6SN7 family is best at around 8mA with Va at about 140V.

7N7
 

7N7

Member
2003-01-18 10:43 pm
England
ray_moth said:

I didn't know that. I thought 170v was at the lower end of the good performance range.

If you want a high-mu tube with low distortion, the 6SL7 does pretty well.


Yes, in fact anywhere from about 130V to say 220V is good; but of course since one should use a decent value of load resistance, say 47k, running at 170V and 8mA would require HT of 537V which is generally above typical values. Of course even my selected point with a 47k resistor would necessitate 507V!

Active loads is the way to go.

7N7
 

rdf

Member
2004-06-21 8:04 am
big smoke
Hey, that's my normal outlook on perfection. Nice fiction. ;) I also agree that anyone who thinks they automatically get complete cancellation doesn't meter. It's interesting to note that some designers (Allen Wright if I recall) 'tweak' differential stages for degree of even cancellation to create a still-low distortion but monotonic harmonic sequence.
 

7N7

Member
2003-01-18 10:43 pm
England
rdf said:
Hey, that's my normal outlook on perfection. Nice fiction. ;) I also agree that anyone who thinks they automatically get complete cancellation doesn't meter. It's interesting to note that some designers (Allen Wright if I recall) 'tweak' differential stages for degree of even cancellation to create a still-low distortion but monotonic harmonic sequence.

Well, in my rare moments of confronting reality, it was my practice to place a 50-ohm trimmer between the cathodes, so that I could at least get the anode voltages matched.

And if I was really trying, I would have a look at the outputs of the stage on a dual-trace 'scope and see if they remotely resembled each other!

7N7
 
rdf said:
{...} It's interesting to note that some designers (Allen Wright if I recall) 'tweak' differential stages for degree of even cancellation to create a still-low distortion but monotonic harmonic sequence.
I've heard of some 'philes using tubes of same type and matching electrical characteristics (eg balance, current draw, etc. -- as one should) but of differing brands in each half off a differential pair to achieve some kind of desired sonic effect. (some physical circuit constructions obviously won't allow such tomphoolery).

I've never had a chance to try it myself yet, but am intrigued given the effects of such tube-rolling combinations might be somewhat unpredictable in the differential pair.
 

7N7

Member
2003-01-18 10:43 pm
England
Majestic said:
I've heard of some 'philes using tubes of same type and matching electrical characteristics (eg balance, current draw, etc. -- as one should) but of differing brands in each half off a differential pair to achieve some kind of desired sonic effect. (some physical circuit constructions obviously won't allow such tomphoolery).

I've never had a chance to try it myself yet, but am intrigued given the effects of such tube-rolling combinations might be somewhat unpredictable in the differential pair.

Rules is Rules.

Alan Dower Blumlein invented the differential pair; I think he understood that each half must be as close as possible to a perfect match. He also understood the importance of a very large tail resistance. He did not anticipate the pentode CCS (at least I don't think so) but for the man who invented stereo and died at 42 with around 140 patents to his name I think we can allow a little latitude don't you think?

7N7
 
7N7 said:


Rules is Rules.

Alan Dower Blumlein invented the differential pair; I think he understood that each half must be as close as possible to a perfect match. He also understood the importance of a very large tail resistance. He did not anticipate the pentode CCS (at least I don't think so) but for the man who invented stereo and died at 42 with around 140 patents to his name I think we can allow a little latitude don't you think?

7N7

A differential par is excellent thing for DC amplification; also it is good to increase symmetry and a voltage gain when used after a phase splitter, but it's drawback is when an input signal and feedback are provided to different grids it produces odd order harmonics that is not good sonically. However, if somebody wants to recreate "Transistor Sound" using vacuum tubes, it is the way to go...
 

7N7

Member
2003-01-18 10:43 pm
England
Wavebourn said:


A differential par is excellent thing for DC amplification; also it is good to increase symmetry and a voltage gain when used after a phase splitter, but it's drawback is when an input signal and feedback are provided to different grids it produces odd order harmonics that is not good sonically. However, if somebody wants to recreate "Transistor Sound" using vacuum tubes, it is the way to go...


If you believe in push-pull as I do, then there are always compromises. A transformer is far from perfect, so are the concertina and the differential pair. Other phase-splitters are really rather unpleasant as far as i can see.

As I wrote above, life is not perfect, find the topology that appeals and live with it: the symmetry of differentials appeals to me - and I don't agree about "transistor sound"! I should add that I never used feedback in my own designs (apart from some cathode feedback in my 813 p-p amplifier) because the mathematics are far beyond me.

Edit: now, no longer able to build my own amplifiers, I am listening to a pair of Quad IIs: beam tetrodes, feedback, pentode inputs; they sound rather nice!

7N7