I found the culprit. I have my DDDAC directly connected to my power amplfier and I know what's causing the problem. Both the "Neg" terminals are shorted by the signal ground in my power amplifier. Do any of you know any workarounds? Two monoblocks would work I think
Hi guys. I made some changes to my setup which I'm very happy with, so wanted to share.
I had a couple of issues I wanted to sort out:
I also wanted to employ a few things I've learned lately (mostly from Supersurfer and the experiments at Klaus's last weekend) I wanted to be sure the digital elements got as quiet a supply as possible and the analogue as low impedance as possible.
Well, I happened to have some leftovers from a previous project and a nice transformer I just got my hands on to try.
Ingerdients:
DIYinHK 2 x 5v 1A Ultra low noise power supply
4.17uV Ultralow noise DAC power supply regulator 3.3V/5V 1Ax2 - DIYINHK
Standard general use 15va 6v x 2 toroidal
A custom made Toroidal not available off the shelf and made to a secret recipe which I don't know, but what I do know is it's a potted 50va toroidal with a 2 x 16v centre tapped very low impedance secondaries and a interwinding shield which connects to IEC earth.
So the plan was to change the 160va toroidal in the 12v supply for the 50va low impedance shielded one and to see if it would be a good idea to exchange the 5v supply for the waveio and the switchmode supply for the wandboard with the dihinhk low noise 2 x 1A supply. This would also take a major heat source out of the case.
But is 5v 1A enough for the Waveio?
Yes, it seems so
And the Wandboard, which all the docs states it needs a 2A supply?
Yes. It peaks at 0.7A on bootup and settles to 0.3A on idle, but runs 0.4Aish when playing lovely music. The 2A recommendation must be enough headroom for when you plug a load of other devices into it.
Ok, let's go then.
Out with the 5V DDDAC supply and switchmode.
in with the low noise 2 x 1A
then just swap this for this (this 50va is almost as large and heavy as the 160va)
longer bolt or slight drilling of the potting needed, but it's good enough for now
and we're done
Now, cranking the volume to full with nothing playing reveals a massive reduction in horrid buzz and hum, so that's a good start.
What I wasn't quite ready for was the improvement in clarity, which really is quite significant. It's possibly due to the switchmode being so utterly rubbish it was polluting everything. When I find some time, I need to do a head to head and swap just the 12v supply transformers to see how much difference that has made alone, but anyway I'm VERY happy now
I now have ultra low noise 5v supplies for the wandboard and waveio, less heat and general crampedness in the case and a very quiet low impedance tx in the 12v supply. Not a bad day's work and none of it was particularly expensive.
The only thing left to change in the short term is the I/V resistors, but I've ordered some 120R GG102D Rhopoint precision wirewound today for that.
GG102 Series Precision Wire-wound Squaristor
After that, I think it sounds fantastic and I'm happy where I am for now and really enjoying the music
cheers,
James
I had a couple of issues I wanted to sort out:
- Because my case is wooden and doesn't dissapate heat well, my DDDAC has been getting really hot, so I have to run with the lid open
- I realised that the switching power supply I was using for the Wandboard was VERY noisy and even grounding the wandboard and waveio together was introducing a lot of noise at the RCAs
- I had fitted a large 160va Toroidal to my DDDAC 12v power supply, but it was just an off the shelf item and it seemed it was emitting a lot of noise. I could hear mains hum over the RCAs with nothing else powered up and with the 12v supply disconnected from the mainboard :s
I also wanted to employ a few things I've learned lately (mostly from Supersurfer and the experiments at Klaus's last weekend) I wanted to be sure the digital elements got as quiet a supply as possible and the analogue as low impedance as possible.
Well, I happened to have some leftovers from a previous project and a nice transformer I just got my hands on to try.
Ingerdients:
DIYinHK 2 x 5v 1A Ultra low noise power supply
4.17uV Ultralow noise DAC power supply regulator 3.3V/5V 1Ax2 - DIYINHK
Standard general use 15va 6v x 2 toroidal
A custom made Toroidal not available off the shelf and made to a secret recipe which I don't know, but what I do know is it's a potted 50va toroidal with a 2 x 16v centre tapped very low impedance secondaries and a interwinding shield which connects to IEC earth.
So the plan was to change the 160va toroidal in the 12v supply for the 50va low impedance shielded one and to see if it would be a good idea to exchange the 5v supply for the waveio and the switchmode supply for the wandboard with the dihinhk low noise 2 x 1A supply. This would also take a major heat source out of the case.
But is 5v 1A enough for the Waveio?
Yes, it seems so
And the Wandboard, which all the docs states it needs a 2A supply?
Yes. It peaks at 0.7A on bootup and settles to 0.3A on idle, but runs 0.4Aish when playing lovely music. The 2A recommendation must be enough headroom for when you plug a load of other devices into it.
Ok, let's go then.
Out with the 5V DDDAC supply and switchmode.
in with the low noise 2 x 1A
then just swap this for this (this 50va is almost as large and heavy as the 160va)
longer bolt or slight drilling of the potting needed, but it's good enough for now
and we're done
Now, cranking the volume to full with nothing playing reveals a massive reduction in horrid buzz and hum, so that's a good start.
What I wasn't quite ready for was the improvement in clarity, which really is quite significant. It's possibly due to the switchmode being so utterly rubbish it was polluting everything. When I find some time, I need to do a head to head and swap just the 12v supply transformers to see how much difference that has made alone, but anyway I'm VERY happy now
I now have ultra low noise 5v supplies for the wandboard and waveio, less heat and general crampedness in the case and a very quiet low impedance tx in the 12v supply. Not a bad day's work and none of it was particularly expensive.
The only thing left to change in the short term is the I/V resistors, but I've ordered some 120R GG102D Rhopoint precision wirewound today for that.
GG102 Series Precision Wire-wound Squaristor
After that, I think it sounds fantastic and I'm happy where I am for now and really enjoying the music
cheers,
James
cheers, I thought there might be interest in that, but didn't want to be more specific until I know the score myself. I'll find out more now I know it's good and ask if there's more where it came from and report back in the next day or so
Well yes, but not necessarily...
I changed all 3 power supplies for the wandboard, waveio and dac in 1 go and altogether there was an improvement. If I had to guess, I would say that changing the waveio one will have made the least difference if any out of the 3, and the wandboard one, possibly the most.
Really I need to find some time to swap them back in and out individually and identify the differences, but I'm really short on time after this week.
Is that an aerial lead next to your RCA connectors? I suspect moving this would be a good idea and also I would move the system/wand board as far away from the DAC as possible to reduce high frequency interference.
I'm about to install the same 5Vdc PSU's this weekend. Your results sound very promising.
Last edited:
I plan to wire up a 150va to the ddac supplying the WAVIO first, to see what improvements it makes.
Then I'll be wiring up a 300va to the DAC boards. I shouldn't get any hum as my whole hifi is running off a 3Kva balanced power supply.
Yes, that's a Wifi aerial. Good call. I'll get them both moved over away from the DAC and RCAs.
something else that occurred to me this morning, maybe a question for Doede. If the new mainboard has a good quality clock on-board for SPDIF input, is there some way that it's possible to use direct I2S from something like a Raspberry Pi, but using this external clock? Or is there some basic reason why this isn't possible?
I think that there may already be a driver for the wandboard to use I2S with external clock as part of the CSOS project. I'll investigate...
James, the DDAC is perfectly fine run direct from 3-wire I2S out of the WB.
It's actually the one case where there isn't a sonic advantage, (to these ears at least), from re-clocking. (The use of the on-board pll4 audio clock is fine. Ultra-high quality MCLK doesn't come into play in this case anyway. No MCLK needed and BCLK doubles as MCLK/SCLK in non-oversample Doede's 1794 design. Although, I would recommend isolating the I2S, between the WB and the DDAC.) If I manage to clear the decks of everything else I'm currently working on, I'll document the what and how, next week.James - Direct output seems to be the holy grail of output coupling with a single DAC board. It appears to me that the reason your direct balanced outputs are working so well is the wooden box. It prevents the WaveIO power supply (-) from grounding to a chassis through the WaveIO metal mounting spacers. Your power supply floats to allow the balanced outputs to not have a voltage offset with the preamp power. Does this explanation make sense?
I have Salas shunt power supplies which have an output impedance of less than 0.01 ohms. I wouldn't expect that a huge power transformer would make any difference with the Salas. Does anyone have any experience with the Salas and over-sized transformers? If James ever decides to put a Salas PS in his wooden box, he will be able to bake bread in it.
I have Salas shunt power supplies which have an output impedance of less than 0.01 ohms. I wouldn't expect that a huge power transformer would make any difference with the Salas. Does anyone have any experience with the Salas and over-sized transformers? If James ever decides to put a Salas PS in his wooden box, he will be able to bake bread in it.
Hey Clive, thanks for the answerJames, the DDAC is perfectly fine run direct from 3-wire I2S out of the WB.
I really wanted to ask you, but I've been trying hard not to bug you as it seems a busy time in the world of CSOS at the moment.
That's brilliant
That also means I can do it just fine without changing my older 1794 mainboard. Excellent news After experiencing last weekend how well a Raspberry Pi works with direct I2S into this DACusing Volumio, it seems that any benefit to be gained from the posh clock and good circuit design of the WaveIO board is negated by the additional process of converting to USB and back again. So this way would be much simpler and cheaper with potentially better results
If there's anything I can do to help you sort the documentation for this, please just ask.
thanks again,
James