Another RTZ SDM (DSD) DAC

[Thermionic Idler]

It's high time I spawned my own RTZ DAC thread so that Ray gets some peace and quiet in his.

Current sit-rep. Ray has very kindly lent us the breadboard he recently completed, which has his power supplies and Beaglebone Black, together with a DSD'it PCM to DSD upsampler 'cape' that I have purchased from him, plus the RTZ DAC and filter boards that he has very kindly put together for me. I have already described the sound which I'm entirely happy with, and it works faultlessly as a Roon endpoint, so very easy to use. Plug in a network cable at one end, point Roon at it and get sound out the other. Simples!

At some point Ray will want the bits that aren't mine back, so I've started the process of designing the RTZ into a proper case. The power supplies I will be using will be:

AckoDAC 'Super Audio' PSU - 2 x 5V. This will provide independent 5V supplies to the Beaglebone and the DSD'it cape.
Salas Reflektor-D - 1 x 5V. This will supply power to the digital side of the RTZ DAC board.
AMB sigma 22 - 2 x 15V. To supply the analogue circuitry on the RTZ DAC board.

The first two (AckoDAC and Salas) were purchased from Ray some time ago, originally for a Twisted Pear DAC project that ended up on the back burner for ages whilst I faffed around with phono and line stages. The AckoDaC supply is built up and ready to go, but I do need to populate / configure the Salas supply.

The AMB board formerly saw service in my last linestage (the one with the VU meters in it) where I naively thought that replacing it with a battery powered switch-mode supply would somehow improve things. So that will be brought back into use, I just need to change a few components to convert it from ±10V to ±15V. Everything I've powered with the AMB supply has turned out sounding superb - I use it in the Millett phono and there are two of them in my new linestage, so bringing it back into service for the DAC was a no-brainer.

I did a very rough CAD layout this morning having cloned the design file for the linestage, as the dimensions will be exactly the same. This confirms there's ample space for three power transformer PCB's on the left, the three power supply boards and the RTZ board on the right hand side.

[Ray P]

Looking good Dave.

There's no immediate hurry for the BBB and breadboard power suppiies Dave, although I'll probably head down your way in early June I'm content for you to hang onto them for a while longer until you get your project progressed. As it happens I'm currently working on the power supplies for my own RTZ project and I'm hoping to have them u and running in a week or so. I can test my RTZ build with the JLSounds USB module I have too.

Looks like there could be several DSD DACs at owston this year!

[Cressy Snr]

Look fwd to seeing/hearing some of this stuff. Things have certainly changed over the years at Owston. Scary, humming, 1000V breadboards seem to have gone the way of the Dodo.

[Thermionic Idler]

Look fwd to seeing/hearing some of this stuff. Things have certainly changed over the years at Owston. Scary, humming, 1000V breadboards seem to have gone the way of the Dodo.

Until I turn my attention to my next phono stage...

[Cressy Snr]

Yikes!

[Thermionic Idler]

Yikes!

TBH that won't be for a while yet (and 12AX7's probably don't need 1000V B+). Once this DAC is done, I am going to have to freeze the budget for a time on new projects as we have some matters on the horizon that need to take priority first. So there'll a bit of designing and modelling but no actual building. But the phono stage will be LCR (as I've already bought the Sowter modules), and I am determined to make it valve-based - just using commonly available valves, not exotica.

[Thermionic Idler]

Orders went off to Mouser and Digikey tonight for the Beaglebone board and various bits and bobs like the three power transformers and remaining parts for the Salas power supply.

Next I need to design a PCB for soldering the three power transformers onto. Once those get here I should have everything I need.

The Salas power supply needs to be assembled and the AMB power supply needs to be modified for a different output voltage, but that's about it. After that it'll just be a matter of plugging PCB's together.

[Thermionic Idler]

PCB design done for the power transformers to mount to. All PCB dimensions and hole locations are now known, meaning that I can go ahead and finalise the chassis layout this weekend.

[Thermionic Idler]

With every build, there is a tedious job that we dislike, for me this is one of them. Painstakingly modelling each PCB board in TurboCAD to map out where all the holes are drilled, so that there is an accurate 'map' to eventually send off to the machinists, and all the board pin headers / sockets will line up properly when the thing is eventually screwed together.

Modelling the 'stack' with the RTZ boards and Beaglebone/DSDit cape was the trickiest bit. Very glad to have the breadboard. Been using a very old ruler we found recently that actually has tenths of an inch marked instead of sixteenths / thirty-seconds.

Having to constantly switch between imperial measures (which PCB's are generally sized in) and metric (which is what my chassis is sized in) just makes things even more fun.

I do believe I now have a PCB layout that will work. Still got to do cutouts for front panel LED, rear mains entry socket and LAN ethernet connector.

[Thermionic Idler]

Final layout design pretty much done - when I added the RFI shielding mains inlet I realised there wasn't enough clearance between it and the power transformer boards so they all had to be moved back slightly. This is why we lay out everything in advance kiddies (at least I do, the money that Schaeffer charges I always, always have to make sure I get it right first time).

Also my 30cm Ethernet male to female chassis mount connector lead arrived from Amazon today so I was able to cut out the aperture required for that.

[Thermionic Idler]

Sit-rep, not a whole lot to report. The order went off to Schaeffer early last week for the panels, and I am still waiting for them to ship. I have to populate some PCB's - the ones for the mains transformers arrived from JLC this week and I also have to put together the Salas power supply board, as well as desolder and change out some components on the AMB sigma-22 board to convert it to the right voltage. So this weekend will be devoted to that. Hopefully by the following weekend the chassis panels will have showed up, and I can get on with completing the thing.

[Thermionic Idler]

Power supply boards completed and ready for installation.

Had some fun and games with the sigma22 board (bottom right) where R10 had to be desoldered and replaced with a different value. Turns out getting solder out of a through-hole with a simple solder sucker is nigh on impossible if it's connected to a ground plane on both sides, which acts as a nice little heatsink.

I must have spent over two hours popping the solder sucker, applying more solder to try and transfer the heat, trying different shaped solder tips and solder removal braid (which only stuck to the hole), swearing profusely and getting very cross, before close examination of the board offered me a workaround in the form of two unused holes that were also connected to each end of R10, just the correct distance apart to accomodate it. Spot the slightly off-kilter resistor next to the bottom right electrolytic and 'unpopulated' R10.

Yes, I probably need a proper rework station but it's difficult to justify the cost given how little rework I do.

[simon]

I feel your pain! Sometimes I've got lucky by putting the point of the iron into the hole (iron perpendicular to pcb) to get the solder and adjacent copper pour hot then quickly place the suckered over the hole as the iron's removed and with luck it comes out quite cleanly. But if the sucker isn't placed close and quickly it doesn't work. Then I start to worry about damaging the pcb.

If I can get the worst of the solder out then I'll heat the component leg and slide it in when the remnant solder on the board melts.

[Thermionic Idler]

This screenshot was taken about 5 minutes ago. I love the way that the Fedex website still seems to think I'll receive the parcel between 9:20 and 13:20 today when the thing is evidently still at Stansted and not even on the fecking van.

With the exception of JLCPCB shipments from China, I don't think I've had a single Fedex package show up on time in the last two years. I don't know what happened to Fedex, they used to be faultless.

So I'm evidently not going to be finishing the RTZ this weekend.

[Ray P]

Had some fun and games with the sigma22 board (bottom right) where R10 had to be desoldered and replaced with a different value. Turns out getting solder out of a through-hole with a simple solder sucker is nigh on impossible if it's connected to a ground plane on both sides, which acts as a nice little heatsink.

I sometimes use a micro drill bit and a pin vice to avoid delaminating the copper foil.