Pass D1V3 DAC - build thread

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2005
muting relay mod revisited

I had a look at the muting relays yesterday and realised i done something particualrly daft. I'd lifted the pad for the gnd on the right ouput connector so had solder a wire jumper to the nearest gnd - on the muting relay. Looking at this in hindsight it was pretty obvious this meant that the mute on this channel was dumping straight into the outputs. I've rewired the jumper to the star gnd and muting now works as expected. So if the relay mod is done as described it should work fine.

While I had the board out I removed the pulse transformer for the AES/EBU input. I've come across Jocko Homo's SPDIF input circuit that looks worthwhile investigating, so the unused pulse transformer is being repurposed:

http://www.diyhifi.org/forums/viewtopic.php?p=5707#5707

Jocko also has written a short piece on general issues with spdif...

http://home.att.net/~jocko.homo/digi.html
 
Disabled Account
Joined 2005
I don't think there is sufficient interest to warrant making up a VCXO board to be honest, and my personal interest lies in more extensive changes to the dac, so I'm working on a new pcb.

The Jocko Homo spdif input seems to have brought worthwhile improvements despite the fact I'm still using rca digital interconnect so I'm missing most of the benefits. I've bypassed the existing spdif circuits on the D1V3 and connected the new circuit at C203/C204.

The VCXO is being under-utilised at present as it simply supplies a clean clock to the DF chip. To obtain maximum benefit the DAC chips also need to be fed a clean clock. This is not entirely straight forward as the SM5842 (and PMD100?) output a burst clock - the clock pulse stops for part of each cycle - so the VCXO output needs to be modulated to replicate this pattern. According to various sources the data lines also benefit from reclocking.

The extent of the changes required to implement the above means it's not really possible to simply make add on boards.

cheers
Paul
 
Disabled Account
Joined 2005
check with spencer for boards...

http://www.diyaudio.com/forums/showthread.php?postid=1523706#post1523706


for chips... cs8412/4 is available for a number of sources online/fleabay in smd and will work with an adapter board. sm584X is still available but tends to be $$$. For pcm63 it is hard to make any recommendations on where to purchase at present. Yamaha Pro audio apparently have a substitute for the genuine Y grade part available at something like $175AU EACH. most sources are more reasonably priced but are of less certain origin.
 
Disabled Account
Joined 2005
The XO-DAC will work fine for reclocking the DF.

The SM5842 and PMD-100 output a burst clock which means that the bit clock stops for part of each cycle, so my understanding is that you can't simply hook the extra outputs from the XO-DAC to the chip. The reclocking circuit from Marc Heijligers Audio DAC pages includes a circuit (comprised of U4 and U5) which gates the clean VCXO clock to match the output of the SM5842 to work around this requirement.

Guido Tent would be able to give you the final word however.
 
Hello to the D1V3 owners...

I've assembled the unit and tested the voltages connected to +-18V. The regulators U5,6,12,13 for the +-5Vanalog have more than +-0,03V tolerance. The worst value reads -5,1V:xeye: and the others have 5,07V (pos. and neg.).
The specs tell that the range for the analog and digital supply is 5V +- 0,5 V. Therefore I don't expect to damage the PCMs. But what is about performance. Does it make sense to adjust the voltages? I thougt about adding some SMD resistors on the bottom of the board to fine tune the LM3x7s.

Thanks in advance.

Best regards
Josch
 
Hello Joschl,

I would not mind the digital supply voltage. For the analog supply I would adjust it that way, that the positive and negative supply for each channel match (not caring what the absolute value is).
As the PCMs are current output I'm not shure if they are sensitive to the supply voltage but better be safe.

Best wishes
 
Joschl said:
Hello to the D1V3 owners...

I've assembled the unit and tested the voltages connected to +-18V. The regulators U5,6,12,13 for the +-5Vanalog have more than +-0,03V tolerance. The worst value reads -5,1V:xeye: and the others have 5,07V (pos. and neg.).
The specs tell that the range for the analog and digital supply is 5V +- 0,5 V. Therefore I don't expect to damage the PCMs. But what is about performance. Does it make sense to adjust the voltages? I thougt about adding some SMD resistors on the bottom of the board to fine tune the LM3x7s.

Thanks in advance.

Best regards
Josch

You may try fixed regulator for the +5V rails. These are the most important ones for PCM63. Use LT1085-5 if you can afford. It has lower noise and much better rejection rate.

For the +-18V rails, you can try LM4040 instead of LM336.

Also you can use BG N caps for the final output coupling instead of MKPs.
 
Voltage accuracy is not one of the design criteria for the PCM63 supplies.
I think I will leave it as it is. Thanks to all for the information.
I'm just waiting for a 2*9V transformer and then it will play music.:)
For the output caps I have some Wima MKPs. I'm going to try them without shielding as I have no idea where to get some copper foil.

Grüße
Josch
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.