For those (like me) that are interested in doing things the easiest way possible, I should point out that the guys at Twisted Pear offer a Toslink module, per the data sheet, with cap, inductor, and supply regulator, for $15. I've had mine sitting in the basement since, apparently, 2007.
(Off topic, I've instructed my children to offer tours of the basement for $10, with auction to follow, in the event of my demise.
Vendetta is still down there, still waiting to be shipped off to JC for repair.)Yes, there is also a Sipex SP213E inside which does the to RS232 signal level conversion.
For those (like me) that are interested in doing things the easiest way possible, I should point out that the guys at Twisted Pear offer a Toslink module, per the data sheet, with cap, inductor, and supply regulator, for $15. I've had mine sitting in the basement since, apparently, 2007.
(Off topic, I've instructed my children to offer tours of the basement for $10, with auction to follow, in the event of my demise.
The Twisted pear module does only go up to 96kHz, just so everybody know!
TOSLINK tests
The TP module is designed for 5-12V supplies, and has the inductor in place so only suitable "as is" if powering the module supply. You could always leave of the regulator and bridge the inductor if you wanted to direct power from the DAM1021. Otherwise, the layout is a bit ordinary (power traces under the RX module for example, when it should have a dedicated ground plane), but it is a cheap and easy solution.
I've hooked up my Picoscope to the 3.3V supply and GND at the pins of the TORX147 and with the cable connected but not activated in OS X the ripple is about 4-8mV p-p. With music playing the ripple ranges from 4-20mV p-p with just the 0.1uf cap.
The ripple between GND and +3.3V where it enters my input board is around 4-8mV p-p with no music play and the receiver unlocked. With music playing this increases to 14-28mV p-p.
Adding in a 22uH inductor increases ripple in the down stream supply by a factor of 10. I was seeing ripple of 200-250mV with it in place so obviously the onboard regulation doesn't like this. Adding an inductor is a Very Bad Idea™ if you are powering from the DAM1021.
The other things I tried was adding an r/c filter to the 3.3V line. This degraded power line noise by a factor of 2, so not nearly as damaging as the inductor but not useful. removing the resistors and leaving a 56uF Panasonic FC cap in parallel with the 0.1uf cap seems to have made a very small improvement to ripple at the TORX147 - fewer peaks above 20mV was the main benefit.
I thought I should check to se what happens when using SPDIF. Both the +1.2V and +3.3V rails are essentially the same - 4-10mV p-p - with music playing, and slightly quieter - 2-6mV with the occasional spike to 8mV - with no inputs. And to verify my tests I rechecked TOSLINK and got 10-30mV p-p on the 3.3V supply.
The picoscope is in no way a precision instrument, and measures 4-6mV p-p with the clips attached to each other, and this is not influenced by computer activity.
Soooo.... I'd highly recommend if you are powering the inputs from the DAM1021 to stick to the letter of Søren's schematic - it is about as good as it gets for direct hookup.
I'm now wondering what shares the TORX147 supply because the spike in 3.3V noise when using TOSLINK is a bit disconcerting. A dedicated 3.3V supply for the TORX module will be a must do - and here I was thinking it was unnecessary.
cheers
Paul
The TP module is designed for 5-12V supplies, and has the inductor in place so only suitable "as is" if powering the module supply. You could always leave of the regulator and bridge the inductor if you wanted to direct power from the DAM1021. Otherwise, the layout is a bit ordinary (power traces under the RX module for example, when it should have a dedicated ground plane), but it is a cheap and easy solution.
I've hooked up my Picoscope to the 3.3V supply and GND at the pins of the TORX147 and with the cable connected but not activated in OS X the ripple is about 4-8mV p-p. With music playing the ripple ranges from 4-20mV p-p with just the 0.1uf cap.
The ripple between GND and +3.3V where it enters my input board is around 4-8mV p-p with no music play and the receiver unlocked. With music playing this increases to 14-28mV p-p.
Adding in a 22uH inductor increases ripple in the down stream supply by a factor of 10. I was seeing ripple of 200-250mV with it in place so obviously the onboard regulation doesn't like this. Adding an inductor is a Very Bad Idea™ if you are powering from the DAM1021.
The other things I tried was adding an r/c filter to the 3.3V line. This degraded power line noise by a factor of 2, so not nearly as damaging as the inductor but not useful. removing the resistors and leaving a 56uF Panasonic FC cap in parallel with the 0.1uf cap seems to have made a very small improvement to ripple at the TORX147 - fewer peaks above 20mV was the main benefit.
I thought I should check to se what happens when using SPDIF. Both the +1.2V and +3.3V rails are essentially the same - 4-10mV p-p - with music playing, and slightly quieter - 2-6mV with the occasional spike to 8mV - with no inputs. And to verify my tests I rechecked TOSLINK and got 10-30mV p-p on the 3.3V supply.
The picoscope is in no way a precision instrument, and measures 4-6mV p-p with the clips attached to each other, and this is not influenced by computer activity.
Soooo.... I'd highly recommend if you are powering the inputs from the DAM1021 to stick to the letter of Søren's schematic - it is about as good as it gets for direct hookup.
I'm now wondering what shares the TORX147 supply because the spike in 3.3V noise when using TOSLINK is a bit disconcerting. A dedicated 3.3V supply for the TORX module will be a must do - and here I was thinking it was unnecessary.
cheers
Paul
Last edited:
I think it might be worse than that.
From hifiduino....
Just the supply you want filled with ripple from the TOSLINK RX current draw. And no wonder TOSLINK displays all the signs of higher jitter levels than SPDIF.
Hopefully it's easy enough to fix with a separate regulated supply for the TORX module.
cheers
Paul
Mine is for sale if a french is interrested : 0.02 % tolerance ; reason : not too much time to experiment EQ and digital filters. I will come back later, have 3 others DAC to finish !
230 euros with that (sorry: with VAT !) instead 266.5 euros : I can go to Paris or PP + colissimo recommandé (environ 10 euros) pour le reste de la France.
Let me know in the swap meet section please :http://www.diyaudio.com/forums/swap-meet/269283-sale-soekris-dam1021-0-02-a.html#post4213751
230 euros with that (sorry: with VAT !) instead 266.5 euros : I can go to Paris or PP + colissimo recommandé (environ 10 euros) pour le reste de la France.
Let me know in the swap meet section please :http://www.diyaudio.com/forums/swap-meet/269283-sale-soekris-dam1021-0-02-a.html#post4213751
Søren has mentioned that there is a DC-DC convertor which supplies power to the Spartan-6 chip core, so perhaps the noise generated by the TORX is quite not so bad.
The Si514 has a spec for additive jitter of around 3.5ps for 100mV ripple on the power supply. If the TORX is sharing the same supply as the clock, the increased ripple will influence jitter levels and when we are talking about a clock with sub-picosecond jitter spec even a 1ps rise is a doubling of jitter levels. The FIFO will do naught to solve that.
Last edited:
As a frequency spectral distribution, quaintifyed in level (dB). A phase noise plot.
//
Last edited:
Strange, I find the buffer stage to be very neutral despite its simple powering in relation to the non buffered SE.
have you a hole in the middle (soundstage) between the speakers or not ?
low resolution files stays a little unconfortable to listen to in relation to higher resolution materials ? Does it need NOS filter with FIR E.Q. on te higher fhz (a little à la TOTAL DAC who testimonies about a better result like that because the many 44 K hz material for most of people ?)
Could be interressant to have testimonies about experienced guys about digital filters ? Bunpei members, etc ?
Paul, TNT, do we know the time capacity about the buffer for the jitter ? how much milliseconds ? 5 or 1, less ?
Last edited:
With a 9V R-Core, I had 10 V AC who gave loaded after a diode bridge and filter caps 15.5 V ! bzzzzzzz danger : after 15 V : the dac collapses !
So take care, maybe try it with a load : a shunt resistor ? 5 K ohms ? I don't know, I and electronics make 2
Any one to design a passive symetric DC from a 10 V ac unloaded (9v R-core) ? two diodes bridge, 4 caps and resitors for a Pi ? I didn't success with PSUD2 : predidcted result are not the same as in reality, maybe my bad shematic ? (I know odd demand in relation to the atomic level of DIYA )
So take care, maybe try it with a load : a shunt resistor ? 5 K ohms ? I don't know, I and electronics make 2
Any one to design a passive symetric DC from a 10 V ac unloaded (9v R-core) ? two diodes bridge, 4 caps and resitors for a Pi ? I didn't success with PSUD2 : predidcted result are not the same as in reality, maybe my bad shematic ? (I know odd demand in relation to the atomic level of DIYA
)I have the torroid 7v/15va by Audiophonics linked in first page of this thread : works without problem, if you want i can measure it unloaded if it helps !
The "nominal" voltage of 7.5V is the voltage rating for full rated load current.
The output voltage will drop with higher load current.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.