[mgalusha]

Over the weekend I finished adding a tube output stage to my already much modified Squeezebox 2. I had been using a pair of Burson Buffer modules but after doing some mods for a couple of friends on a Shanling tubed CDP and an Exemplar Denon 2900 with a tube output I found myself really missing the sound when I had to return them to their owners. This is a bit odd as I am using a much modified Behringer DCX2496 and normally use the digital in from the SB. I only use the analog inputs with my TT and for testing other gear.

With the Shanling or the Denon in the system using the analog inputs I found I enjoyed the system much more and it had a certain magic that was just not there using the digital in or using the analog inputs via the analog out of the SB2. Given that the analog inputs of the DCX are being converted to digital at 24/96 this kind of bothers me but I can't deny what I heard. It took me two weeks to get over the Shanling and after about of week of missing the Denon I decided I had to stop whining the build a tube output for the SB2.

I had inspiration from the Lampizator site as well, this gentleman has put tubes in as may CD players as he could get his hands on. The circuit he uses is a symmetrical SRPP. I played around in TubeCAD with this and decided it would work well for my purposes. Gain is fairly low, about 10 with the tube I'm using, a Russian 6H6P. This is still much more than I need so I am using a voltage divider between the SB's output and the input of the tube stage, netting an output of about 2.5V, just a little more than a standard CDP. I wanted a bit extra and this is just right. The downside of this circuit is a low PSRR but plenty of filtering in the PSU takes care of that. I have a tiny amount of hum if the volume control is pegged on the DCX but I will be doing a bit more work on the B+ PS to address that though it's inaudible at any sane listening level so it's not a high priority.

The SRPP inverts phase, which is a plus in a Squeezebox that has had it's opamp stripped as it normally does this. Net result is a non inverted output.

As for the sound, it's been running in for a day and a half of so and so far I'm quite happy with the results. The sound is very similar to using the Burson output in that it's clean and open, bass is tight and tuneful but more importantly it has that certain something that only vacuum tubes seem to provide. 

Total investment was less than $200 including the tubes from the Ukraine. Of course the biggest chunk of this was the transformer. I reused the output caps from the previous incarnation. The new parts are all on the PCB on the right. Schottky diodes for the filament supply and HexFRED's for the B+. 40,000uF on the filament supply in a CRC configuration. The resistors in the filament supply where chosen to get 6.3V with the Variac set at 120V. The B+ supply uses a RCRCRCRC filter with a 4 section JJ capacitor and a few extra caps after it. About 1mV of ripple with the B+ at 202V after the tubes are warm. Ignore the digital out cable with the masking tape on it, this will be going back in the chassis, it was just easier to work on with it out of the way.
 



mike

[JoshK]

Very nice.  I've been considering doing something similar myself, but then I have a Dac 60.

I have a few sketches for different tube output stages, but most have highish gain like yours.  12B4 would give you ~6 instead of 10, but that is a redesign.  Those 6N6p's are nice tubes.  One thing you could try that would lower gain, lower Zout and give a bit of drive current is a 4:1 (or 2:1) step down output transformer.  Most are expensive, but I've heard good things about the very inexpensive ones from Edcor.  They aren't gapped so you'd need to wire it in parafeed connection (sit it on a cap).

I am trying out the edcor's in another line stage project I am working on.  I know a couple of other gents trying them too.  For $15 they are worth a try.  It beats just shunting the voltage to ground and adding the extra noise of the resistor. 

By the way, what's the bias voltage you are using?  Is it bypassed or just resistor?

If you want to try to quite the supply, you can try out some of Gary Pimm's CCS boards or the Swenson/Pimm regulator.  I am going to place an order for some boards sometime soon, let me know if you want me to add a few more on, probably cheaper that way.    I was thinking of trying the Swenson regulator in the DAC 60.

Your hum might be heater related though.  Are you using AC or DC?  Did you reference the heaters up?

[JoshK]

4:1
http://www.edcorusa.com/products/transformers/xsm/xsm10k-600.html
2:1
http://www.edcorusa.com/products/transformers/xsm/xsm2_4k-600.html

[mgalusha]

Thanks Josh, I'm fairly happy with the results so far but like most things I'm sure it will evolve over time.

Bias voltage, according to TubeCAD, is -3.43V though I haven't measured it yet. To get to the pins I have to pull it out of the chassis. The cathode resistor is not bypassed, which in this circuit helps keeps the gain down to about .5 of mu.

I hadn't seen the Swenson/Pimm regulator, that looks very nice. Please let me know if you're going to order some boards, I'd definitely like to get in on that. I took part in the recent group buy on PinkFish Media for the TeddyReg boards. Let me know if you need a couple, maybe we could trade.

I'm fairly sure the hum is HV related. I am using DC on the heaters and originally had the heater ground separate from the signal ground and had a nasty buzz. Tying the heater ground to the main ground removed the buzz and all that is left is a very faint hum. At this point I'm not biasing the heaters. The circuit, again according to TubeCAD, only has a -6dB PSRR, so the supply needs to be really quiet. The regulator you show would certainly take care of that. I have a 30 Henry choke on order as well that should be very effective.

I sort of thought of a transformer output after looking at Jim Hagerman's Castanet but the transformers he uses would cut my output voltage too much as the turns ratios are 8 and 12 something to 1. The 4:1 Edcor looks perfect and as you say I can nuke the voltage divider and get the Zout down to about 60 ohms. I called Edcor and ordered a pair, the price is certainly right. Thanks very much for the link and info. I'll let you know how they work out.

I did take some basic measurements so I have something to compare to with the transformer output's. As is the FR is flat to 10Hz, which is the lower limit of my signal generator with a -3dB point at 250kHz. THD is about .3% at 10V output at both 1kHz and 20kHz, nearly all 2nd harmonic. S/N is only about 60dB referenced to 1V. A bit more work on the PS and the addition of the output transformer should get that down pretty low. It will swing just over 40V on the output before the waveform shows visible distortion, headroom is not a problem. This matches the simulation from TC perfectly as it predicts -42/+48 and the bottom of the wave form just starts to look bad at 42V. Gotta love it when the sim is close to reality.

Mike

edit: fixed typo's

[SET Man]

Hey!

  Wow! Very nice work there. I wish I know enough how to rip thing apart and make it in to tubed one.

Take care,
Buddy

[mgalusha]

Thanks Buddy. I admit to an almost uncontrollable desire to tinker with things.

[JoshK]

Cool Mike.  Looking forwards to seeing your results with the Edcors. 

P.S. There is a thread on AA somewhere where Swenson describes how he came to this regulator.  He has been playing with numerous applications of Pimm's CCS.  He mentions that he has built a plethora of regulators and likes this one by far the best.  The measured results look really promising.

[Occam]

Mike - Very Cool!

[poseidonsvoice]

Mike,

Fantastic work. I would echo Josh's comments about referencing the heater's voltage, it might take care of your hum problem. So will a stepdown transformer of course.

The 6N6P is a great tube, I'll be experimenting with it when I get to building the active output stage Kevin Carter has newly designed for his RAKK dac II. Great ideas there. He also sells some very well designed regulators. Don't know how it compares to Pimm's or Swenson's. I use Pimm's Pentode Mosfet Constant Current Source in my own preamp actually with the EL84 serving as the pentode. Rock solid voltage, no hum issues whatsoever, and it does double duty as a headphone amplifier.

Anand.

[jrebman]

Mike,

Sounds good, and I was just about to email you to see how it was coming along.  There's a 2900 sitting here waiting for something like this (and an SB as well .)

Josh, count me in on a few of the Swenson boards as well.

How did you guys get your 6n6ps?  I only saw one seller on ebay who was selling them in lots and I wasn't sure about sending money to the eastern block.

-- Jim

[mgalusha]

How did you guys get your 6n6ps?  I only saw one seller on ebay who was selling them in lots and I wasn't sure about sending money to the eastern block.

Jim,

I bought a lot of 8 from someone in the Ukraine (I believe) about a year ago. They were $19.00 plus shipping. I've purchased tubes from the Ukraine, Russia and Lithuania and have never had a problem. The last batch I picked up in may arrived in about a week from the Ukraine, no complaints there. I can spare you two of the 6N6P's if you would like a set. I had 8 but now it's 7 as one rolled off the bench and my shop floor is concrete. I'll trade you for a set of 6GM8's, I know you have just a few of those.

Mike

[kyrill]

Wow nice.

Doe it soundwise matter where the tube "magic" is inserted in the chain?
could it be between (SB2 dig. out-->Dac-->preamp-->amp)   Dac and preamp, so that the digital out could still be used?
 

[jrebman]

Mike,

Sure, sounds like a good trade.  Do you have a tube tester?  I have an unopened box of 25 6gm8s and they've all been tested, but they are not matched, but I'm sure we can find some nicely matched pairs in that lot.  Who knows, they may even have the test results marked on the boxes.

-- Jim

[mgalusha]

Kyrill,

Not sure if it matters where the "magic" is inserted. In my case I'm using a modified Behringer DCX2496 as DAC/Preamp and of course crossover. Even though it's re converted the analog output from the sb back to digital it still seems to preserve the qualities of the tube output stage. I've said before this bugs me from an engineering perspective as ultimately the DAC's in the DCX are the one's you are hearing but it seems the hot rodded DCX is transparent enough to allow the flavor from the source through. I am under no illusion that the new output stage is uncolored, it just so happens I like that particular palette.

Jim,

I do have a tube tester but it's a very basic good, bad, maybe type. On my list of someday projects is to build Steve Bench's tube tester which will measure mu and transconductance but who know's when that will happen. Hmm, I supposed I could reuse the chassis from the current old Heathkit for that.

Mike

[audiobasics]

I really like the wood you chose for the front panel inlay.

[mgalusha]

Quick update. I finally got around to installing the choke in the power supply so now the PS is LCRCRC and this has knocked the noise down considerably. I do have a pair of the Edcor transformers Josh mentioned and one of these days those will go in for a test drive. I am still very happy with the results and now even more so since it's very quiet.

mike