Passive I/V + step-up transformer + tube

[regal]

After building a pass D1 DAC and also owning a PS Audio Digital Link III DAC, I have come to the conclusion that I don't like SS in my DAC analog stages. No surprise since I came to this same conclusion with regard to amplifiers a few years back. I have a highly modified DAC-60 (tube)that just sounds great.

The Dac-60 is using a rather large 50 ohm I/V presently. I find a real improvement if take this down to 6 ohms (V seen on the DAC output to under 25 mV's). Problem is I have to turn up my amp to the point where I am hearing some hiss and microphonics.

Does anyone have experience using step-up transformers like the Sowter 8347 or the LL1636 before the tube output?

I know there is an issue of high impedance at high frequencies but with an I/V resistor on the primary this should not affect the DAC since it would be in parallel with 2-5 ohms.

Any other issues? These things are expensive, if anyone has a preferred model please let me know.

 

[Mazuki]

I just ordered parts for a Pass D1 and just recently completed a 6N6P Lampizator circuit. I was going to do a comparison.

Is the Pass D1 really that bad?

 

[regal]

Its very good for solid state, but I feel there is a certain lack of life that the tube DAC has that the D1 doesn't.

I don't recommend the Lampizator, it an SRPP which never makes sense in a DAC. Any of the single ended stage sound better. There is no need for push-pull in a DAC. My DAC-60 came stock with a lampizator (SRPP.) I tried various configurations and setting it like the first stage of the Aikido (totem pole) sounds best .

 

[Mazuki]

But SRPP has lower output impedance no?

What kind of DAC are you using? I have a parallel TDA1541.

 

[regal]

An SRPP is very load dependant, read Broskie's article. For a 6DJ8 it works out to only 7 kOhms. A typical amp has an input impedance of 50k ohms. You don't need push-pull for this. I got much better sound taking the output from the plate of the bottom tube which is single ended with an output impedance of 3500 kOhm (6DJ8.)

 

[nikongod]

After a moment tripple checking SRPP and white cathode followers, I no longer feel the need to build an amp with an SRPP output stage. YUCK!

Anyways, to the credit of a "lower output impedance through complication" you can drive capacitive loads better. Driving a cable&amp that combine to around 100pf or more could be a problem to a source with a high output impedance.

 

[Mazuki]

Quote:

An SRPP is very load dependant, read Broskie's article. For a 6DJ8 it works out to only 7 kOhms. A typical amp has an input impedance of 50k ohms. You don't need push-pull for this. I got much better sound taking the output from the plate of the bottom tube which is single ended with an output impedance of 3500 kOhm (6DJ8.)

3500kohms? That's a bit high wouldn't you say? Fikus provides measurements for his Lampizator, unless you have similar measurements for comparison, I wouldn't be so quick to dismiss his design.

Plus I'm using 6N6Ps, which really are an entirely different animal than 6DJ8s.

At the very least we're all in agreement on the benefits of a circuit with greatly reduced simplicity (over opamps).

 

[dsavitsk]

Quote:

Originally Posted by regal /img/forum/go_quote.gif These things are expensive, if anyone has a preferred model please let me know.

Take a look at these: EDCOR Electronics Corporation - Audio Transformer & Equipment Manufacture

They are cheap, made with high quality stuff, and a very good way to experiment. Raw DAC output is not a lot different from a microphone output, so this might be a good way to experiment for not much cash. When you hit on something you like, you can "upgrade" to the Lundahls or whatnot. Only issue is that they are not shielded so you need to be very careful with where they are, and how they are oriented, or else they will pick up hum.

Specs look pretty good: http://edcorusa.com/images/graphs/MX1_0720.pdf

 

[dsavitsk]

Quote:

Originally Posted by Mazuki /img/forum/go_quote.gif 3500kohms? That's a bit high wouldn't you say?

With a 100K input and short cables, I've used tubes in a grounded cathode configuration with rp as high as 8K or so. Not ideal, but sounded fine. I don't like the 6DJ8 here as they always sound edgy to me. 6N1p is a decent choice, though c3g, d3a, etc. are better. In fact, a very high mu tube with a OPT can improve your ability to drive a load. Lundahl LL1930 for a cheapish option, or LL1674 for a better one. As for the I/V, I'd split the difference -- 25 to 30R or so, though obviously it depends on the chip. AD1865 is happy with a 200R, 1541 w/ 33R, PCM1794 w/ a pair of 50's.

 

[regal]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif Take a look at these: EDCOR Electronics Corporation - Audio Transformer & Equipment Manufacture . Specs look pretty good: http://edcorusa.com/images/graphs/MX1_0720.pdf

Problem is their frequency response goes up over 50khz, we need the opposite on a DAC?

How about these:

http://edcorusa.com/products/transfo...x/mxl10cs.html


As far as push pull in a line stage, trust me I have done extensive testing as long as you can keep the Zout under 4 grand (50k Zi) you are much better off single ended grouded cathode or totem pole. I agree about the 6DJ8's, this new DAC will probably use 6N6P's.

 

[regal]

Here are a couple other posibilities:

http://cinemag.biz/mic_input/CMMI-10PCA.pdf

http://www.jensen-transformers.com/datashts/115ke60.pdf


.

 

[dsavitsk]

Quote:

Originally Posted by regal /img/forum/go_quote.gif Problem is their frequency response goes up over 50khz, we need the opposite on a DAC?

I guess I never worry about filtering a DAC. I have transformers on the DAC output, transformers in the preamp (TVC now, formerly input and output and a tube stage) and OPTs on the power amp. If anything is doing any HF filtering, it is probably on the power amp. Or rather, it is probably the speakers/headphones, which seems fine to me.

Quote:

Originally Posted by regal /img/forum/go_quote.gif How about these:

I think they are fine, although experience suggests that that particular shield is not going to do all that much, and the M6 core won't sound as good. The others are cheaper which makes them better for experimenting in my eyes.

Back to your noise concern, a tube like the triode strapped D3a has a noise equivalent to a 65R resistor and a mu of 77 (rp~2K) which should solve your hiss issue. I'd just use the I/V as a grid leak and use the transformers on the output. My $0.02 anyway.

 

[regal]

dsavitsk I greatly appreciate you sharing your experiences as it will save me and others much time/expense.


I guess I ruled out OPT's because I read where a typical SE 10k:600 output tranny would not load the tube correctly with a typical 50kohm load. It seems analogous to hooking up 300 ohm Sennheisers to an 8 ohm OPT tap. Is this correct?

 

[dsavitsk]

Quote:

Originally Posted by regal /img/forum/go_quote.gif I guess I ruled out OPT's because I read where a typical SE 10k:600 output tranny would not load the tube correctly with a typical 50kohm load. It seems analogous to hooking up 300 ohm Sennheisers to an 8 ohm OPT tap. Is this correct?

I think the issue is that the transformer load on the tube is inductive, so the impedance load doesn't matter so much. You can figure out the effective Z from an inductive load from

Z = 2 * pi * Hz * L

Also, you can always put a 600R resistor on the secondary, but I've never found it to be necessary.

fwiw, I like a CCS loaded parafeed setup in line level apps because you can shorten the signal path, keep the big caps out, and there is no need for the big swings that a choke load or a single feed transformer provides.

The other thing I have been playing with lately is using either a TVC or an AVC as a parafeed output. You get to eliminate the resistive volume control for free, sort of (TVCs and AVCs are aften a little more expensive than other transformers), and the performance improvement is significant. And, if you have low Z phones, you can drive them directly from your DAC. Here are the beginnings of my work on this: ecp.cc Take a look at the AVCs from Intact Audio.

 

[regal]

These step-up mic input transformers are made of core material of Nickel alloy lamination square stack.

The first one has very similar specs to a $110 Sowter 8347.


Microphone




.