Sweep Tube SEP

[Steve Cresswell]

Earlier this month I posted a few pics of a sweep tube SEP I designed.
There were one or two things about it's design that I felt were not quite right so it has been revised to great effect.

It was first built with Alexandar Kitic's philosophies about SEPs in mind but I found it lacking.
I redid it using more established practices and the improvement in sound quality was staggering.
It really ia a great sounding bit of kit now, fast, full blooded and powerful.

I've had a few email enquiries from people on here but don't have the time to deal with them so I decided to post the
circuit in the main forum where those interested can discuss it, point and laugh, talk about compatible alternative Russian tubes...whatever.
If the thread sinks without trace then so be it, but at least those interested will have the latest schematic to work from, and play with as they see fit
I'm very happy with it but as always there's bound to be room for improvement.

Here it is then:


Steve

[Doz]

Interesting Steve - I have a supply that's similar using a 6080, but for less current.
I think I have most of the tubes, except 6AU6, so I'll slip an EF94 in...
Where did you fall on the iron?
What kind of output power are you seeing?

[M Gregg]

Steve,

Your output Tx's were they a special from Majestic?
Can I ask how much they cost?

Aside from the grommet I think they look great. LOL
Whats the bass like from them?

Are they speaker sensitive..ie loading, I ask because you have no global FB?

Regards
M. Gregg

[artosalo]

My experience with PL519/-509 is that there is no reason the stabilize the whole supply voltage. Instead to stabilize the Ug2*and make it adjustable is a very good point.* I used 450 V anode voltage and 150 V at screen.
The amplifier was PP, not SE.

[the_manta]

Hmmmm
There is a major flaw on that design.
The Pentode driver stage has an unbypassed cathode resistor and thus strong feedback which increases its output resistance. So, the output resistance is completely determinded by the 47K resistor.

The power stage is directly feedbacked to its grid. This parallel derived and parallel applied feedback reduces the input resistance of the PL509.

Because of that overloading, lots of distortion is produced. I couldn't quite believe that myself, but switched to a parallel-derived-series-applied (here)and it sounded much better. Measurement was just with a simple soundcard, but distortion improvement was clearly visible.

Don't get me wrong, but you're using the feedback in exactly the wrong way
You're increasing the output resistance of the driver and decreasing the input resistance of the power stage.

regards, simon

[Steve Cresswell]

Originally Posted by the_manta

Hmmmm
There is a major flaw on that design.
The Pentode driver stage has an unbypassed cathode resistor and thus strong feedback which increases its output resistance. So, the output resistance is completely determinded by the 47K resistor.

The power stage is directly feedbacked to its grid. This parallel derived and parallel applied feedback reduces the input resistance of the PL509.

Because of that overloading, lots of distortion is produced. I couldn't quite believe that myself, but switched to a parallel-derived-series-applied (here)and it sounded much better. Measurement was just with a simple soundcard, but distortion improvement was clearly visible.

Don't get me wrong, but you're using the feedback in exactly the wrong way
You're increasing the output resistance of the driver and decreasing the input resistance of the power stage.

regards, simon

Yep That's how it's supposed to work I know what I'm doing
There is no "major flaw"

It's working exactly as intended

I don't intend to spend time arguing to justify the design philosophy.
Siffice it to say that it is the whole amp and how the two stages interact
with each other that determines the final sound quality
This can't be judged on how individual stages work in isolation.

Steve

[Steve Cresswell]

Originally Posted by Doz

Interesting Steve - I have a supply that's similar using a 6080, but for less current.
I think I have most of the tubes, except 6AU6, so I'll slip an EF94 in...
Where did you fall on the iron?
What kind of output power are you seeing?

Hi Doz

The iron was from Majestic Transformer Company in Poole Dorset.
£113 each.

I'm getting around 14WPC rms out of it.

Steve

[Steve Cresswell]

Originally Posted by M Gregg

Steve,

Your output Tx's were they a special from Majestic?
Can I ask how much they cost?

Aside from the grommet I think they look great. LOL
Whats the bass like from them?

Are they speaker sensitive..ie loading, I ask because you have no global FB?

Regards
M. Gregg

The bass is absolutely superb from this amplifier, powerful detailed, tight and clean throughout the range.
Huge soundstage with plenty of depth and width and image location.
It'll drive a wide range of speakers but as with most tube designs the more senitive the speaker the better.
Mine drives an open baffle setup.

Treble is sweet, clean and detailed... lovely cymbal sound.
Plenty of presence in the midrange without being too forward.
Hardness is not an issue with this amp.

It has two local feedback loops
The output stage has the venerable Schade plate-to-grid feedback scheme
and the input stage has an unbypassed cathode resistor. (on purpose )

[the_manta]

Originally Posted by Steve Cresswell

Yep That's how it's supposed to work I know what I'm doing
There is no "major flaw"

It's working exactly as intended

I don't intend to spend time arguing to justify the design philosophy.
Siffice it to say that it is the whole amp and how the two stages interact
with each other that determines the final sound quality
This can't be judged on how individual stages work in isolation.

Steve

But you don't see such a configuration on any good commercial amplifiers and that has a reason.
The 6AU6 sees two anode resistors in parallel. The 47k and the Rfb. Since the rfb is in the feedback loop, the driver sees this resistor reduced by the factor of µ of the PL509 ! This kind of feedback only works well when used in very small amounts. Your 6AU6 sees a very low dynamical anode resistance which causes more distortion than the feedback should compensate. Just read any book about feedback, it's the same whether tube or solid state.

Of course you can analyze each stage and with the laws of feedback the tube's parameters changes.

I'm sure this amp produces music and works, but I just wanted to point out that there are better ways.

[Steve Cresswell]

Originally Posted by the_manta

Your 6AU6 sees a very low dynamical anode resistance which causes more distortion than the feedback should compensate. Just read any book about feedback, it's the same whether tube or solid state.

Of course you can analyze each stage and with the laws of feedback the tube's parameters changes.

I'm sure this amp produces music and works, but I just wanted to point out that there are better ways.

Yes Simon and this argument has been done to death in this forum quite a few times, but the truth remains, Schade feedback or if you prefer, the anode follower configuration, when used with unbypassed pentodes (not triodes!) in the front end
produces arguably one of the finest sounds it is possible to get from a single ended pentode amplifier.

You're absolutely correct in what you say, it is well known that Schade feedback schemes if not implemented properly will cause a huge increase in second harmonic distortion from the driver stage. The value of the feedback resistor is critical to the distortion performance of these amps, but when done right, there is little to better them in the SEP universe.

If I seemed a little terse it was the grinning green smiley that was annoying about your post, easily taken the wrong way given the statement that preceded it.

There is a discussion going on right now about pentode drivers for the Kitic RH84 SEP using EL84s, you might want to look at.
You might not agree with what they are discussing, but plenty of people like the sound produced by pentode driven anode follower power amps
so it is a perfectly viable method that has been around since the Thirties. and incidentally I've read plenty of stuff on feedback, tried plenty of schemes and to my ears this one sounds best.

YMMV and only IMHO of course.

Steve