RTZ SDM (DSD) DAC

[Ray P]

I'm currently listening a nice DSD256 Vanessa Fernandez recording via a prototype RTZ (Return to Zero) SDM DAC - build on a piece of MDF, so bread-boarded to some extent The design is by Marcel van de Gevel, who is the man behind the Valve DAC. This one is all solid-state with an op-amp based output filter stage.

It was a real pain to get it playing to start with as some tall caps on the DAC board mean it's really hit and miss to get the headers to connect the filter board properly. At first it sounded a bit edgy so I just left it playing for an hour or so before returning for a proper listen - I'm impressed, it sounds clean and dynamic, well balanced, with lots of detail, not as sweet as the Valve DACs I have available but then it is somewhat compromised being a prototype.

For the input I'm using a JLSounds i2soverUSB board (version 2) interfaced to the RTZ DAC in the same way as my recent Valve DAC project. Because the output is balanced I've had to botch a single-ended connection to a 3.5mm headphone jack to feed my Noir HPA.

In my opinion Marcel is to be congratulated on another successful project and I believe the concept is worthy of further developement, I would definitely add one to my DAC stable if that happened.

Here's the DAC board:



and this is the prototype assemble and playing music;

[Ray P]

If anyone is interested in finding out more here's the link to the DIY Audio thread

https://www.diyaudio.com/community/thre ... ac.379406/

[Ray P]

I did some rework on the RTZ DAC - I did a bit of tidying up of the hook-up wiring and the whole thing is now screwed down to a piece of MDF - and I'm listienng again. It's sounding as good as I remember from last weekend - currently listening to Vivaldi's Four Seasons from a DSD256 recording via my Noir HPA.

Replacing the headers connecting the boards with longer ones was the main area of rework and it has ceraintly made the connections between the boards more reliable than before when the slightest bump could result in silence.

I'm hoping the DAC designer will update the boards with a few improvements as I would like to build an updated version.

[Ray P]

I started putting this together this afternoon to try a simpler output/filter stage with the RTZ DAC.

Capture.JPG (85.23 KiB) Viewed 2328 times

There's an article here;

http://itsonlyaudio.com/ultranalog/cden ... ncer2.html

[Nick]

Couple of things I would opinion about that schematic. First why the cathode resistor shared between channels? Second, that's very low current for a 8650, seems a bit of a waste, Third, the 8650 is very sensitive to DC, you need some way of balancing the two half's of the LTP.

[Nick]

OK, so its a shared cathode resistor because the filaments are connected. I would use two heater supply's. Well, TBH, I would (and did) use a 5687 with added balancing resistors.

[Ray P]

Couple of things I would opinion about that schematic. First why the cathode resistor shared between channels? Second, that's very low current for a 8650, seems a bit of a waste, Third, the 8650 is very sensitive to DC, you need some way of balancing the two half's of the LTP.

Funny you should say that Nick. I've started to build it as two mono circuits and I think the 3A5 would also benefit from decent filament supplies, such as Andrew's DHT boards.

I landed on this circuit simply because I have a pair of 8650s and a small stash of 3A5 (along with just about everything else needed) and I've heard positive things about the approach from an acquaintenance who has used it.

For the moment though I just want to get a feel for the RTZ DAC with an alternative output stage before I need to return it to its designer/builder. He's going to modify the PCB to incorporate the various tweaks he had to make in developing it and I'll probably build one afterwards so it'll be good to get a feel for the relative merits of the output stages.

Another possibility would be some 6SN7s in the 5687 circuit - perhaps i shouldn't be so tight and get a cheap pair of Octal sockets to try it as well (as I have 6SN7s but not 5687s).

5687 schematic.JPG (62.37 KiB) Viewed 2314 times

[Nick]

Well, 5687 are close enough to being b9a 6sn7's so either would be worth using and 6sn7 has the advantage of still being made.

[Ray P]

Well, 5687 are close enough to being b9a 6sn7's so either would be worth using and 6sn7 has the advantage of still being made.

Cheers Nick, I've had a look and I do have a pair of NOS 6SN7 GTB valves - 5687s are a bit pricey now.

Using 6SN7s would certainly be easier regarding filament and B+ supplies as I have some ready-built boards and a transformer i can utilise to breadboard the circuit. I'll order some suitable sockets for the 6SN7s.

Do you recommend any changes to the circuit shown on the 5687 schematic - you commented about the Sowter's DC sensitivity?

[Nick]

What is the DC output level of the dac?

[Ray P]

What is the DC output level of the dac?

I measured the DC (rel. to ground) at the inputs of the current op-amp based filter board as below;

Capture.JPG (182.69 KiB) Viewed 2255 times

[Nick]

The image copy went a bit wrong, but this is the general idea I used. The output stage is in the bottom right corner. Cathode resistor values may need changing for 6sn7 and your output DC levels.

[Ray P]

Thanks Nick, it's a bit difficult to read but I get the gist of the schematic, basically just an additional resistor between the cathode cap -ves and ground.

[Nick]

Not really. The important point is the cathodes each have their own resistor to help DC balance but the cathodes are tied together at AC to work as a LTP.

[Ray P]

After Owston yesterday I've been motivated to get on with a valve stage for the RTZ DAC. As it happens when I checked I found i don't have a transformer suitable for the 6SN7 power supply to hand so I've reverted to the 3A5 and have assembled and tested power supplies for the B+ and filaments.

The filament supply is an LT1085 based regulator and is set to 2.8VDC delivering 500mA to a dummy load.

The B+ supply is a Maida regulated supply delivering 125VDC delivering 50mA to a dummy load. 125V is a little higher than specified on the 3A5 schematic but is 10V under the max plate voltage so should be OK. I could change some components on the reguulator board to lower the voltage but i would prefer to avoid messing with it.

I've started putting the 3A5 circuit together on some vero type PCB and have ordered the 1800pF caps I need.

The 3A5 is a DHT so if the breadboard shows promise I'll look to swap in a pair of Andrew's DHT filament supplies.