Gungnir with tube output stage

Testing my Gungnir with a Decware tube output stage in balanced mode.

 

Can you describe what we're seeing here?

 

Nice! I assume you are familiar with Gumbfreya?

 

This is what I'm using. Since it's setup as balanced, I had to use two complete power supplies. I used the bottom PS on the page with the OC2 regulator tubes, which are on top of the PS black box.

ZKIT4 (decware.com)

The top board in the Gungnir Multibit has through-hole pads that connect to the analog outputs of the four DAC chips, plus ground pads making wiring a snap. Since the output of the tubes is around four volts, I installed two Chinese stepped attenuators to be able to adjust the output. At present, I'm going to set up the wiring after the attenuators so that I can use both the tube output as well as the stock, from the same XLR jacks.

As you can see, the tube PCB's just fit in the space ahead of the USB board. They will be mounted using the standoffs to an aluminum plate that I designed through Front Panel Express that will itself mount on top of the gaping hole that I will have to make in the case top.

 

Yes, I've read about what Marv did to create the Gumbreya. That's what got me thinking about doing this. It's all Marv's fault.

 

Very cool. What does it sound like?

 

No noise at all. Still evaluating the sound. So far, bass is solid, very good transients, treble may be a bit rolled off, but my old ears aren't the best judges.

 

This is a cool project. The "better" power supply seems like it could be a problem though...



The glow tube is basically a zener, i.e. it has a lowish impedance, generates some noise, and drops a fixed voltage. The claim that it is better than a 33k resistor at filtering B+ is surprising. Typically this sort of tube would be used in a shunt configuration, as in the below circuit.



Flanking the glow tube with big caps could cause it to oscillate as well. The datasheet specifies 0.1uF max. If nothing else, it will definitely cause a current spike that will shorten tube life when the supply is turned on and the caps are not charged.

It's also unusual to put a resistor in series between the final cap of the power supply and the load, as this increases the supply's impedance. Could be voicing, but be aware that it does degrade the common mode rejection significantly in a balanced circuit.

I would suggest trying something like the schematic below if you feel like experimenting. the combined resistors set the supply for a total 15mA (which you may need to adjust). Note that it brings B+ down from 95V to 75V.

 

Someone here did do a tube rectified power supply and a tube buffered digital output on a CD player.

https://www.superbestaudiofriends.o...nsport-player-thread.3028/page-12#post-353913

 

Thanks. I had already considered redoing the PS since I'm not particularly happy with my layout. I'll try your schematic: it does make more sense. Decware's schematic lists the output as 95VDC, but the OC2 datasheet has the output at 85V, and that's what I'm getting.