joehpj

Does this amp need extra standalone input stage? And can this amp ran single ended input?

I believe SZ2 are 007A while SZ3 are 007mk2. (domestic / international version). All of my friends' 007A no matter manufacturing time are SZ2.

Any comparison with GG?

Both. I didn't see any extra amp for headphone jack. See pic 3. The rear output and headphone jack was controlled with relays by the switch on the front panel. One will automatically mute when the other is on. It will be really sloppy for amp but I don't think people buying this kind of DA will not have their own amps. Yes digital volume control. I assume it was done by FPGA. I didn't take the picture of the 48 lead QFP so I don't know if that was the volume control. It could be bypassed so not really an issue (if it's really done by FPGA).

This is not mine so I have to send it to the owner. I just got the chance of a brief listening. This DAC was really good for those have no skill/time to DIY. Compared with single 1021 w/ uber buffer fed into balanced CFA, uber has better sound stage and dynamics. 1541 has some more detail and sounded airy. The front panel is a convenient feature so changing digital filter and xfeed are much easier than DIY ones. The chassis however, was solid but compromised on finishing. It looked cheap. For a DAC now 1290 Euro, and made in Denmark, I can't expect it having a much better look though. The main issues here IMO, are 1) the performance of buck+LDO. I personally prefer some traditional way like transformer+ultra low noise linear PS. 2) the performance of the output buffer. It has only +-12V ps. But should be plenty for the output. It has different PS for digital and analog rail. The 1021 DIY module drive 5V digital parts form analog rail. Which was somehow compromised, too. So people have to choose form better buffer like Kevin's diamond buffer and some real ultra low noise LDO or isolated PS for digital/analog. PS. the OP based balanced buffer on 1021 is utterly shit. Use raw output and some better buffers.

I tried to take every detail. I still have some photos if someone need more details. Overall sound is good. This dac runs warm. Quiet curious about the buffer design. comments plz @kevin gilmore

Nice build! Quite interested about how will RK50 affects the sound. Never heard one before.

What's the current of -15V? I was thinking of using some ultra low noise LDO instead of GRLV to shrink the footprint. If I have to compromise one lead of the umbilical because of not enough lead for the TS. Should I use one 1085 for the front end tubes or tie F- to GND? I have 10 leads in total. +-400/+100/-15/GND/bias/6.3VAC are 8 already. For the tubes, I think 6922 will be a better choice because of higher max voltage for longer life. 6DJ8 is pretty limited. I was thinking using some 6H23p though.

What's the PS plan for front end tubes now? Separated GRLV for both positive and negative rail or use voltage dropper on negative rail?

A few noob questions. There are some more parts after voltage dropper to get +100V. Is that because the B+ in T8K is unregulated? It seems stax use regulated PS for the tubes. What's the function of Q19, Q21, Q27, Q30? These seems to first appear in T8K. Stax used some low-noise parts like KSA992 and KSC1845. And stacked KSA1156. Any chance to adopt them in Tube-carbon/Sandwich or even Carbon? Feedback is 200K in Carbon and 300K in T8K. So maybe there chance to lower the feedback in Carbon? Most important is how do you think about the front half of T8K comparing to your design? For the negative supply for the tubes, I like the way tube-carbon does which doesn't need -15V supply. But I don't know the noise level difference between GRHV and 7915/GRLV. Thanks.

Whoa! Looked much more complicated than previous Stax amps. Looked like regulated heater. Zener string for front end tubes?

sounded promising is tempting! detail please.

Did you use regulated supply for front end tube heater?

Connecting G3 to anode is theoretically better. But EL34 is a pentode so it has a real suppressor grid so tying G3 to anode should be fine. 6CA7 is a beam tetrode, so the G3 is actually a beam forming plate. By connecting G3 to plate will make it unable to form electron beams. I don't know if this will make the tube kinks or acts strangely as the plate was originally trying to avoid secondary emission. KT77 has G3 internally connected to cathode and pin 6 jumped. So far as I know EL34, 6ca7, KT77 worked fine in GG. I wonder if it will be the same under T2 configuration.

hey! wrong ID

I am thinking of building a Uber power amp. Can I just add a pot in front of the driver board to make a single ended amp? So I will just need 1*40V GRLV and 2 sets of driver/amp. Or should I use a dynalo as input buffer since I have one? What does the + reg drive on the uber2 stands for? And I plan to lower the bias to cool down the amp a little bit. Which resistor(s) should I change? Thanks!

Really? I don't even know RK50 will fit

Then 330V@400mA should be enough. Thanks for correction.

I use 325V/250mA secondary for 400V@20mA*4 for carbon. Megatron should be almost the same.

The voltage there was 10.xv as I recall. Change to 18v will avoid clipping but 12v should be plenty during normal usage. Congrats on the successful build!

Please call an ambulance. 1021's dead. BTW, the noise should be out of 20kHz so it shouldn't matter?

How would these compare with LDO+IXTP01N. I assume noise will improve a little bit by using using LT3042. I like the simplicity of the LDO and modern parts are easier to source.

Yeah... So, I will do it with separated winding with LT1083 regulator. All floated. I do it mainly because I plan to use separated ps box. Running AC with DC in a same umbilical cord seems somehow compromised for me.

Is there actual benefit to have both 6922 and EL34 filament regulated or only 6922 is required?

I was thinking to build one adjustable DC supply to use some P88CC or 6N23P. Those have much more reasonable prices. Filament will sit a little bit at negative. So negative LDO is preferred?