Very nicely put together! I do have few questions, if you do not mind?I've put the unregulated powersupply in the box..
Let me know if you want to know more or have any suggestions.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
1. Raw Unregulated PS, I see you have approx 6 x 2200uf, same size, and one seems a little bigger in size? Also one bleeder resister with heatsink, but what is other little resister does? Would you able to give us a run down on your ps configuration?
2. DAC boards, I see you use mixed bag of Sanyo OS-Con, a bigger skinless Electrolytic, and retained the stock Muse? What is that bigger electrolytic at the corner, and your drive for this caps layout/specs?
3. USB/SPIDF switch, very interesting! Where did you source it? How is it better than the standard stock toggle?
Sorry if I am being a pain in the neck?!?
Hello Stijn and the others,
I will use a similar set up.
But the common mode input choke will be a ll2733 ( 400mH so double the inductance but 2*1,7 ohm serie resistance)
Will use a 2*15 volts 300va R core. Smaller bleeder because of the bigger choke and probably 4 boards with shunts.
Second choke will be ll1694.
Will put the bleeder at the first cap to reduce voltage loss.
Will also try to reduce cable length even more!!! Stijn you can confirm that common mode sounds better?
So far i only did see positive comments about it.
I hope i will end up with enough voltage.
Sincere greetings, Edward
I will use a similar set up.
But the common mode input choke will be a ll2733 ( 400mH so double the inductance but 2*1,7 ohm serie resistance)
Will use a 2*15 volts 300va R core. Smaller bleeder because of the bigger choke and probably 4 boards with shunts.
Second choke will be ll1694.
Will put the bleeder at the first cap to reduce voltage loss.
Will also try to reduce cable length even more!!! Stijn you can confirm that common mode sounds better?
So far i only did see positive comments about it.
I hope i will end up with enough voltage.
Sincere greetings, Edward
.. I do have few questions, if you do not mind?
..
Hi Chanh, you have a very sharp eye! J
Ad 1. You are right, the caps in the PS are 6x 2200uf (25V Silmic) and the big one is a 3300uf (50v Silmic), I couldn’t find a lower voltage value for the 3300uf cap. The big cap is there to give a bit more extention and impacts around the lowest frequencies, and does this very well. The little resistor is there for a LED connection on the on/off switch of the case, nothing fancy..
Ad 2. My plan was to replace all the Muse caps with a mix of Oscon’s and Silmic’s of various sizes. However I lost heart when I started lifting traces, so I only did a few and left the muses around the dac chip in place. But it’s clear to me that the Muses can really do with augmentation with higher values.
This what I changed (much inspired by Stefan en Cees their findings)
Ad 3. The rotary selector switch, is a pretty standard run of the mill (http://www.banzaimusic.com/Rotary-Switch-2-level-2x3x3.html) . It allows me to eventually select between more than two inputs when I fit the BBB and an additional I2S switch in the case as well.
I can't wait to see what you'll make of it edward. Putting the bleeder before the caps and keeping the leads short are great idea's.
I would say serial common input choke sounds a bit cleaner, but there is very little in it as far as I can tell (but I've been changing a few to many things lately to be absolutely sure about what's what). A main concideration for me too, is the fact that the first set of caps are given an easier life, hopefully prolonging their life.
Hi Doede, I'm intrigued by this comment. I don't have the PCM1794 spec sheet, are the 47uf's at the VCOM's the recommended factory default values?
Why did you chose to have a low frequency roll-off or was this a compromise to retain sufficient bias coupling?
yes they are the recommended values.... low freq roll off for bias, means basically, that for the audio band there is power supply filtering and short return paths for the bias electronic circuit inside the chip. unfortunately there is little know what goes on exactly inside, but it looks like a voltage reference for the current source. these normally get decoupled with a capacitor for the full audio range plus a bit extra....
Beautiful work Stijn!
Some inspiration to tidy up my own nest of wires for sure....
Can I ask how your output is set up? I don't see any caps or transformers there as I know them, but the red box with the rotary control, is that a TVC?
Hi James, Thanks.
My output is NEG/POS connected at the dac & directly connected to a shunt stepped attenuators by Khozmo (Z-foil shunt, takmans REX stepped) without a output cap or transformer as I have no real DC to worry about.
Are you not concerned about the mixing of audio channels running this way or do you have an amp setup with separate channel grounds?
Without separation, each channel will be made up of 50% POS (the true channel signal) and 50% NEG which will be an average of the 2 L&R channels. So you can say that with this method, you're getting 25% bleeding of the channels.
I ran like this for a while, but hadn't appreciated how it effected the separation in the stereo image...
Try the test from my post here http://www.diyaudio.com/forums/digi...pcm1794-waveio-usb-input-228.html#post3992086
James, I’ve just tested what you described and as you explained there is some crosstalk across the ground plain of a single ended amp when connecting POS/NEG of the DAC. Which is pretty logical, but it never crossed my mind to try and test it.
I have a pair of headphones connected directly to the DDDAC as well, and it drives these surprisingly well. When I disconnect my interlinks to my amp I can make the crosstalk go away but the difference in real music listening is very small. This was easily tested listing with headphones plugged in and plugging/unplugging the interlinks to hear the difference. I’ve tried a few pieces of music and can only barely hear the crosstalk with some recording (studio mixes mainly).
There are people (which I hold in very high regards) that feel mono is still superior to stereo. J
I’ll leave it as is for the moment. But this might well be the excuse I needed to build a new balanced amplifier.
I have a pair of headphones connected directly to the DDDAC as well, and it drives these surprisingly well. When I disconnect my interlinks to my amp I can make the crosstalk go away but the difference in real music listening is very small. This was easily tested listing with headphones plugged in and plugging/unplugging the interlinks to hear the difference. I’ve tried a few pieces of music and can only barely hear the crosstalk with some recording (studio mixes mainly).
There are people (which I hold in very high regards) that feel mono is still superior to stereo. J
I’ll leave it as is for the moment. But this might well be the excuse I needed to build a new balanced amplifier.
Last edited:
Sounds like a perfectly sensible suggestion to meor just two isolated mono amplifiers to have true Stereo again
I've got back to work today and this is sat waiting on my desk
An externally hosted image should be here but it was not working when we last tested it.
80va R-core, 2 x 12v centre tapped with static shield and grounding wire.
I'll get that swapped into my unregulated power supply asap
Hello James,
You are the first one that is going to use an R core for the Dac. ( oh no i did buy one but dont have a dac and the famous Mr Hiraga is using too)
It is nice that your other transformers are not that big too so you can actually compare two different technologies and see the benefit of a screen as well.
If it is a big leap forward you can go for a bigger one?!?
Today i did ask some questions regarding the chokes at Lundahl by email. I will share the answers here.
I hope soon i will be able to make a test with my future supply to find out wether the tension will be high enough or if i have to skip some parts on the circuit. If i am right Stefan did end up with 10,5 volts?? But he is still having holidays so we will have to wait .
Sincere greetings, edward
You are the first one that is going to use an R core for the Dac. ( oh no i did buy one but dont have a dac and the famous Mr Hiraga is using too)
It is nice that your other transformers are not that big too so you can actually compare two different technologies and see the benefit of a screen as well.
If it is a big leap forward you can go for a bigger one?!?
Today i did ask some questions regarding the chokes at Lundahl by email. I will share the answers here.
I hope soon i will be able to make a test with my future supply to find out wether the tension will be high enough or if i have to skip some parts on the circuit. If i am right Stefan did end up with 10,5 volts?? But he is still having holidays so we will have to wait .
Sincere greetings, edward
Question.
As described before by Doede and commented on in the latest DDAC documentation (ver4.1), additional High frequency filtering can be done by adding two (or four depending on where you put them & which main board you have), caps. Doede write:” value not critical. Approx 4,7nF to 10nF per Deck”.
I had not added these caps before and did add Wima’s 0,01uf’s ( a factor 4 lower then suggested for a 4 board stack) last weekend and what an improvement they made! The high frequencies became much detailed and open, what I’d been looking for for a while. I’m now thinking about getting some nice Mica’s, but which value’s to get?
Does “per deck” imply to use 40nf for a 4 board stack? Does increasing the value not lower the roll-off frequency. Does this mean that by adding more boards the quantitisation/high frequency noise comes down in frequency, or am I missing something?
What value have you guys used, what are your experiences with adding these caps?
As described before by Doede and commented on in the latest DDAC documentation (ver4.1), additional High frequency filtering can be done by adding two (or four depending on where you put them & which main board you have), caps. Doede write:” value not critical. Approx 4,7nF to 10nF per Deck”.
I had not added these caps before and did add Wima’s 0,01uf’s ( a factor 4 lower then suggested for a 4 board stack) last weekend and what an improvement they made! The high frequencies became much detailed and open, what I’d been looking for for a while. I’m now thinking about getting some nice Mica’s, but which value’s to get?
Does “per deck” imply to use 40nf for a 4 board stack? Does increasing the value not lower the roll-off frequency. Does this mean that by adding more boards the quantitisation/high frequency noise comes down in frequency, or am I missing something?
What value have you guys used, what are your experiences with adding these caps?
Last edited:
thanks for feedback Stijn,
more decks means lower output impedance. to keep the HF roll off constant you than need to increase the value of the capacitor.
the range I suggest is more like, do you like 50kHz roll of, or rather 100kHz ?
effect depends heavily on the amplifier loudspeaker chain after the dac, hence it is an option to tweak the dac further to your situation or taste....
Thanks Doede. I knew there had to be something I missed, that makes sense.
I'm still trying to get a handle on how to determine the optimum value.
How would you envisage the downstream chain or amp/speakers would dictate the choice of an optimum HF filter cap value?