SuperSurfer, if you have lots of hirez stuff please test it and post the results, I only have a few 96 and 192 to test with. But doing a long session with hirez stuff will test how well can the wireless stream can be maintained. I had nothing on the usb/ethernet on the rpi for the test, except the wireless dongle.
Your previous test was using usb for audio and usb wireless ethernet for receiving the stream, that alone may have been enough to overwhelm the rpi? Doede, please tell us when we've reached the limit of OT posting...
Your previous test was using usb for audio and usb wireless ethernet for receiving the stream, that alone may have been enough to overwhelm the rpi? Doede, please tell us when we've reached the limit of OT posting...
Clean power source is critical in audio. I have a separate 240V 20A power line into my audio room and running another 10A for lighting and others. All my audio gears including this DDDAC are connecting to my PSAudio P10 power plant which is connecting to the 20A line. May be I can post a photo later?
Appreciate if you could pls direct me to Salas Shunts for each analog and digital for DAc boards? Where to buy and how should the be soldering to 8decks DAC board? Please explain?When you would want to shunt regulate an 8 dac stack the shunts will be quite costly but I had an idea for this:
Just take off all regulators from the dac boards and make two separate Salas shunt regulators and feed it directly to the dac boards. One for analog and one for digital stages. You can than use the digital salas to also feed the motherboard. This can pretty easily be done by not soldering the power supply wire from the motherboard to the dacs, if you leave a cap position open near all dac chip feed points, than you can use that hole to feed a power supply wire from top to bottom dac board. Connect them from top to a Salas and you are in business!
As local decoupling just use some small and fast wima capacitors, you can leave out the bank of electrolythics. This will bring you a very speedy and responsive power supply.
The Salas regulates with kelvin sensing wires so the voltage will be regulated directly at the dac boards.
Another advantage of this appoach is that you can keep the digital and analog power supplies separated.
I havent tried this myself but I believe this would be a very nice improvement.
Anyone dare to try this?
Cheers,
Last edited:
Hey Sovereign,
If the psu load is no different even with the larger trans the heatsink will not require changing, another part of the change to the larger transformer is how the secondary of the trans reacts with the remainder of the psu in regards to the RLC component, this forms a tank circuit and will cause a resonance and ringing from the diodes and generate noise on the psu ground which will also be amplified as it is the reference for the other voltages in the system. I use tank curcuits for tuning RF Welders in my work, these run on 27 or 40 MHz at 4-20kW output at a plate vilrage of 7.5-10kV. Makes plenty of noise around the place.
Hope that makes sense.....
Chuz,
Drew
Hi Drew, kind of makes sense, as my knowledge is limited.
I currently have two PSU's, one to the DDDAC and the other to the wavIO I plan to replace the tranny of the DAC psu to 300va, 15v,1A and the tranny of the wavIO to 150v, 9v, 1A
I guess what your saying is leave the rest as it is, fit the new transformers and if you have problems regarding resonance then there is a solution? Is that about right? Thanks!
I currently have two PSU's, one to the DDDAC and the other to the wavIO I plan to replace the tranny of the DAC psu to 300va, 15v,1A and the tranny of the wavIO to 150v, 9v, 1A
I guess what your saying is leave the rest as it is, fit the new transformers and if you have problems regarding resonance then there is a solution? Is that about right? Thanks!
interesting topic on the tanks. goes on my short list for the "lab". this is easy to test, like with small and large transformer ...if no difference is measured, still does not mean we cannot hear it though. but as always, I am looking for clues by listening to other people, listen my self and do measurements... mostly a new direction or at least a deviation is found.
On a general note, this thread is developing very nicely I must say, many view points, also based on listening to the DAC... keep going team !!
I have one point for supersurfer... the 8 volt shunt, do you feed the DAC before or after the inductor, which originally comes after the LDO? by pure listening tests, the inductor was preferred, but introduces like 100mOhm at DC and very high impedance for AC. This was done to give the DAC a very local decoupling for its internal circuitry, also this will totally overrule the output impedance from the shunt I would think. looking at your pictures the inductor is still in the power circuit? might want to try with and without
I have received a few shunts now as well from Guido and I like to test some variations. just build a one deck DAC and make it A-B able...
original situation
3.3 volt digital only
8 volt only
both voltages
with and without capacitors
with and without inductors.
this absolutely need to be done. we need to understand the individual response to theses changes, not pay the before an after in a multiple item change
work to do..... would appreciate anyone to do the same !!!
On a general note, this thread is developing very nicely I must say, many view points, also based on listening to the DAC... keep going team !!
I have one point for supersurfer... the 8 volt shunt, do you feed the DAC before or after the inductor, which originally comes after the LDO? by pure listening tests, the inductor was preferred, but introduces like 100mOhm at DC and very high impedance for AC. This was done to give the DAC a very local decoupling for its internal circuitry, also this will totally overrule the output impedance from the shunt I would think. looking at your pictures the inductor is still in the power circuit? might want to try with and without
I have received a few shunts now as well from Guido and I like to test some variations. just build a one deck DAC and make it A-B able...
original situation
3.3 volt digital only
8 volt only
both voltages
with and without capacitors
with and without inductors.
this absolutely need to be done. we need to understand the individual response to theses changes, not pay the before an after in a multiple item change
work to do..... would appreciate anyone to do the same !!!
Hi Doede,
That's a very good point you've got there. I left the inductors in but you are absolutely right that it makes the impedance higher. I need to try without it.
I am also very interested in your own findings with the shunts.
@ Chanh,
In addition to the PM: Salas boards can be found here:http://http://www.diyaudio.com/forums/group-buys/188974-gb-thread-salas-sslv1-1-bib-shunt-reg.html
@ Palmito,
The limited bus speed is indeed the problem when streaming through wifi and out to wave io.
With i2s out we've eliminated one datastream on the usb bus. Wifi however is a very inefficient way in comparison to ethernet. Wifi has a lot of package loss and therefore needs a much bigger data throughput than strictly necessary. This may also work counterproductive with high res files.
That's a very good point you've got there. I left the inductors in but you are absolutely right that it makes the impedance higher. I need to try without it.
I am also very interested in your own findings with the shunts.
@ Chanh,
In addition to the PM: Salas boards can be found here:http://http://www.diyaudio.com/forums/group-buys/188974-gb-thread-salas-sslv1-1-bib-shunt-reg.html
@ Palmito,
The limited bus speed is indeed the problem when streaming through wifi and out to wave io.
With i2s out we've eliminated one datastream on the usb bus. Wifi however is a very inefficient way in comparison to ethernet. Wifi has a lot of package loss and therefore needs a much bigger data throughput than strictly necessary. This may also work counterproductive with high res files.
Hi All,
Hope one of you can help with this:
I'm underways with building a DDDAC, however, will use the amanero USB-to-I2S board. And in one of the latest versions the amanero allows for outputting the data as right justified. Which is what the DDDAC - as I understand it - should be using.
However, in the DDDAC circuitry description on the webpage it says:
"Delay line to convert I2S Data into Right justified Data and align Left and Right Channel".
So my question is if outputting the data from the amanero board as right justified would allow me to do without the shift registers in the DDDAC mainboard - or if I for some reason I still need these?
Also, except if the answer to the above question makes this one irrelevant: anyone know of a source for the 74VHC164MX shift registers that doesn't require me to buy many (and also doesn't have a high handling cost)?
Thanks for reading and maybe replying ;-)
Jesper
Hope one of you can help with this:
I'm underways with building a DDDAC, however, will use the amanero USB-to-I2S board. And in one of the latest versions the amanero allows for outputting the data as right justified. Which is what the DDDAC - as I understand it - should be using.
However, in the DDDAC circuitry description on the webpage it says:
"Delay line to convert I2S Data into Right justified Data and align Left and Right Channel".
So my question is if outputting the data from the amanero board as right justified would allow me to do without the shift registers in the DDDAC mainboard - or if I for some reason I still need these?
Also, except if the answer to the above question makes this one irrelevant: anyone know of a source for the 74VHC164MX shift registers that doesn't require me to buy many (and also doesn't have a high handling cost)?
Thanks for reading and maybe replying ;-)
Jesper
Hey Jesper,
Not yet owning one but reading alot I think what Doede has done by using the shift registers is that the data for the right channel and data for the left channel arrive at their respective chip at the same time, I think there is another board about here that will do the same thing via an FPGA kind of setup. Methinks the motherboard from Doede will be the easier option.
Not to put the fire out but that seems easiest, no idea where to get the 74 chips from, the usual suspects I guess, but google is your friend.
Sorry if that doesnt help, but that is my understanding. Also the 1k resistor at one of the chips delays that data the last half bit or something too.
Much to ponder.....
Chuz,
Drew.
Not yet owning one but reading alot I think what Doede has done by using the shift registers is that the data for the right channel and data for the left channel arrive at their respective chip at the same time, I think there is another board about here that will do the same thing via an FPGA kind of setup. Methinks the motherboard from Doede will be the easier option.
Not to put the fire out but that seems easiest, no idea where to get the 74 chips from, the usual suspects I guess, but google is your friend.
Sorry if that doesnt help, but that is my understanding. Also the 1k resistor at one of the chips delays that data the last half bit or something too.
Much to ponder.....
Chuz,
Drew.
Hi Drew,
Thanks for replying ... your understanding also is close to mine, however, what could be interesting is if the shift registers could be omitted altogether by setting up the amanero to output right justified data instead ... Then a resistor could still be added if it was necessary (as I understand it it is there due to the shift registers' delay ...(?)) ...
But maybe one of the thread readers know about this?
Cheers,
Jesper
Thanks for replying ... your understanding also is close to mine, however, what could be interesting is if the shift registers could be omitted altogether by setting up the amanero to output right justified data instead ... Then a resistor could still be added if it was necessary (as I understand it it is there due to the shift registers' delay ...(?)) ...
But maybe one of the thread readers know about this?
Cheers,
Jesper
Hi All,
After reading the thread for days I cannot help but placed an order for a 4-deck kit.
I recall reading in one of the posts that each DAC board with the Tentlab shunt regulator consumes ~250mA current. Now the DDDac 12V supply can supply up to 1A so it seems it is very marginal? Plus the mainboard needs some current of its own so does it mean that one single 12V supply won't suffice?
If not then should I order a second 12V supply and how to connect it? Just parallel the two?
Thanks.
Jack
After reading the thread for days I cannot help but placed an order for a 4-deck kit.
I recall reading in one of the posts that each DAC board with the Tentlab shunt regulator consumes ~250mA current. Now the DDDac 12V supply can supply up to 1A so it seems it is very marginal? Plus the mainboard needs some current of its own so does it mean that one single 12V supply won't suffice?
If not then should I order a second 12V supply and how to connect it? Just parallel the two?
Thanks.
Jack
The power supply can handle the extra current but will need a larger heat sink for the TIP122. I used 2 stacked up and runs at 55ºC now. It will do even more with bigger transformer and heat sinks.
I run a two board shunt reg dac. The heat sink on the PSU is almost too hot to touch and after a while I need to let go as it's too hot, (hope you like my scientific approach to temperature measurement ;-)
I think I need to upgrade the heatsink aswell
Last edited:
Able to hang on to it is around the 50-55 mark. About every second over 55 is a degree. i.e. hang onto for 2 secs around 57.
Not super scientific but an approximate. Also depends what one does for a living and how much tough skin lives on the finger tip. A bloke I work with seems to have asbestos lining his fingers......
Im not liable if you get burnt calibrating your fingers for more accurate results.....
Chuz,
Drew.
Not super scientific but an approximate. Also depends what one does for a living and how much tough skin lives on the finger tip. A bloke I work with seems to have asbestos lining his fingers......
Im not liable if you get burnt calibrating your fingers for more accurate results.....
Chuz,
Drew.