8412/4 PLL tweaking
The PLL filter for the 8412/4 on the D1 and D1V3 is taken straight from Crystal's datasheet for the 8412. Based on past experience with the 8412 in my old Assemblage DAC 2.0 it's worthwhile applying the "Wildmonkeysects" filter mod.
I was discussing this via email with Skilmek so have cut and pasted the info here:
--
Just checked the schematic for the DIR9001 and see that it has a loop filter on the adapter board. So it seems that applying the wildmonkeysects loop filter will not effect plug and playability.
There were slightly different recommendations for 8412 and 8414:
http://www.audioasylum.com/audio/tweaks/messages/30313.html
http://www.AudioAsylum.com/audio/tweaks/messages/44352.html
So component selection depends on which one you plan to use. The components Spencer has listed on the schematic are the datasheet recommendations for the CS8412, while the values for the CS8414 should be slightly different. If you reclock the dac this mod will probably have no impact whatsoever, but without the reclocker it should be worth doing, as parts cost is minimal.
cs8412
replace C200 with 0.22uF
replace R114 with 470R or 500R resistor
add 3300pf cap between FILT (pin 20) and AGND (pin 21) keeping leads as short as possible
cs8414
replace C200 with 0.20uf or 0.22uF
replace R114 with 270R resistor
add 2200pf or 3300pf cap between FILT (pin 20) and AGND (pin 21) keeping leads as short as possible
WMS adds the note "Presuming it locks, I would try different values of the resistor, 220R, 240R, 270R, 300R and stick with the smallest one that sounds good."
So effectively you add one part, soldered directly to the pins of the socket. If you are using an adapter I'd solder the 2200/3300pf cap between the pins on the adapter to get it closer to the chip or if you are brave use a smd cap and try soldering it as close to the pins on the chip as possible.
--
I haven't done the mod on the my D1V3 as yet but will be making the changes after the easter break..
cheers
Paul
The PLL filter for the 8412/4 on the D1 and D1V3 is taken straight from Crystal's datasheet for the 8412. Based on past experience with the 8412 in my old Assemblage DAC 2.0 it's worthwhile applying the "Wildmonkeysects" filter mod.
I was discussing this via email with Skilmek so have cut and pasted the info here:
--
Just checked the schematic for the DIR9001 and see that it has a loop filter on the adapter board. So it seems that applying the wildmonkeysects loop filter will not effect plug and playability.
There were slightly different recommendations for 8412 and 8414:
http://www.audioasylum.com/audio/tweaks/messages/30313.html
http://www.AudioAsylum.com/audio/tweaks/messages/44352.html
So component selection depends on which one you plan to use. The components Spencer has listed on the schematic are the datasheet recommendations for the CS8412, while the values for the CS8414 should be slightly different. If you reclock the dac this mod will probably have no impact whatsoever, but without the reclocker it should be worth doing, as parts cost is minimal.
cs8412
replace C200 with 0.22uF
replace R114 with 470R or 500R resistor
add 3300pf cap between FILT (pin 20) and AGND (pin 21) keeping leads as short as possible
cs8414
replace C200 with 0.20uf or 0.22uF
replace R114 with 270R resistor
add 2200pf or 3300pf cap between FILT (pin 20) and AGND (pin 21) keeping leads as short as possible
WMS adds the note "Presuming it locks, I would try different values of the resistor, 220R, 240R, 270R, 300R and stick with the smallest one that sounds good."
So effectively you add one part, soldered directly to the pins of the socket. If you are using an adapter I'd solder the 2200/3300pf cap between the pins on the adapter to get it closer to the chip or if you are brave use a smd cap and try soldering it as close to the pins on the chip as possible.
--
I haven't done the mod on the my D1V3 as yet but will be making the changes after the easter break..
cheers
Paul
Thanks, you too! Four days of camping and bird watching in the Victorian Mallee at Hattah-Kulkyne national park. should be nice...
this is how it sounded one windy morning last september, there is a loud call that clips the recording about half way thru so beware!!:
http://www.mactrix.com.au/files/recordings/Hattah_Edit_2.mp3
fwiw the recording was done with a DIY mic preamp - using a walt jung design with THAT1510 chips as a replacement for the original ssm's.
I'll be trying out a new mic rig this trip so hopefully should get some nice recordings.
this is how it sounded one windy morning last september, there is a loud call that clips the recording about half way thru so beware!!:
http://www.mactrix.com.au/files/recordings/Hattah_Edit_2.mp3
fwiw the recording was done with a DIY mic preamp - using a walt jung design with THAT1510 chips as a replacement for the original ssm's.
I'll be trying out a new mic rig this trip so hopefully should get some nice recordings.
Another D1V3 "tweak": Output Relays
Checking Spencer's schematics vs the D1 service manual I've discovered the D1V3 mute relays have contacts in the signal path
, rather than the ground shunt used in the D1 
Should be easy to work around, as my old DAC 2.0 had the same type of bypass surgery done on it. Looks like a job for the post holiday tweak session.
Update: Ironically the assembly instructions direct you to cut the pins required to implement the fix off the relays!! If the pins are left on and soldered to the pads - which are grounded, it's a simple matter of cutting the traces between R19,20,30,40 and the relays, then running short jumper wires from the relay side of R19,20,30,40 to the output pads. In normal operation the relay is not connected, but in mute mode the outputs are shunted to ground.
Checking Spencer's schematics vs the D1 service manual I've discovered the D1V3 mute relays have contacts in the signal path
, rather than the ground shunt used in the D1 Should be easy to work around, as my old DAC 2.0 had the same type of bypass surgery done on it. Looks like a job for the post holiday tweak session.Update: Ironically the assembly instructions direct you to cut the pins required to implement the fix off the relays!! If the pins are left on and soldered to the pads - which are grounded, it's a simple matter of cutting the traces between R19,20,30,40 and the relays, then running short jumper wires from the relay side of R19,20,30,40 to the output pads. In normal operation the relay is not connected, but in mute mode the outputs are shunted to ground.
rewire
Ran out of time to edit the post. I've attached a pic of what needs to be done.
Cut traces on the X. The relay pins that need to be soldered are in the red circles, and the green lines show where the jumpers need to go.
If the pins are already cut on the relay you will need to replace the relays to perform this mod.
While I would not have even contemplated making a DAC like this from scratch without Spencer's boards, I'm finding enough things I would have done differently that I'm now seriously thinking about making my own boards. Gotta love DIY
Ran out of time to edit the post. I've attached a pic of what needs to be done.
Cut traces on the X. The relay pins that need to be soldered are in the red circles, and the green lines show where the jumpers need to go.
If the pins are already cut on the relay you will need to replace the relays to perform this mod.
While I would not have even contemplated making a DAC like this from scratch without Spencer's boards, I'm finding enough things I would have done differently that I'm now seriously thinking about making my own boards. Gotta love DIY
Attachments
Hi,
I have posted the assembly instruction manual and on the page 2 and 3 of first file, there is a more simple way to address this pcb error.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=97504&perpage=20&pagenumber=5
Spencer
I have posted the assembly instruction manual and on the page 2 and 3 of first file, there is a more simple way to address this pcb error.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=97504&perpage=20&pagenumber=5
Spencer
Hi Spencer,
The instructions you refer to in the assembly manual leaves relay contacts in the signal path. Which is precisely what you don't want if you are concerned with good sound, and what the original schematic avoids by using a shunt arrangement.
The description in the assembly instructions show the two pins being taped to AVOID being grounded. I'm saying these pins NEED to be grounded, to allow the output to be shunted to ground. The output from the resistors R19,20,30,40 should be wired directly to the outputs, and the traces connecting R19,20,30,40 them to the relay should be cut. In this arrangement there is always a direct connection to the output jacks with no contacts in the signal path. The output is shunted to ground when the mute engages.
cheers
Paul
The instructions you refer to in the assembly manual leaves relay contacts in the signal path. Which is precisely what you don't want if you are concerned with good sound, and what the original schematic avoids by using a shunt arrangement.
The description in the assembly instructions show the two pins being taped to AVOID being grounded. I'm saying these pins NEED to be grounded, to allow the output to be shunted to ground. The output from the resistors R19,20,30,40 should be wired directly to the outputs, and the traces connecting R19,20,30,40 them to the relay should be cut. In this arrangement there is always a direct connection to the output jacks with no contacts in the signal path. The output is shunted to ground when the mute engages.
cheers
Paul
Attachments
Re: rewire
Hi Paul, I decided to take your advice and get the relay out of the signal path. The supplied relays from spencer (Takamisawa relays R5YW-K) are hard to locate. I believe I found a substitute for anyone else interested in this mod and need to secure two more relays if the two pins were cut for your kit; you can get Omrons from Mouser for 2.95 ea.
Here is an attachment if you are interested at looking over the data of both relays and how they match up w/ a Mouser link for the part.
Stan
spzzzzkt said:
Hi Paul, I decided to take your advice and get the relay out of the signal path. The supplied relays from spencer (Takamisawa relays R5YW-K) are hard to locate. I believe I found a substitute for anyone else interested in this mod and need to secure two more relays if the two pins were cut for your kit; you can get Omrons from Mouser for 2.95 ea.
Here is an attachment if you are interested at looking over the data of both relays and how they match up w/ a Mouser link for the part.
Stan
Corrected link for the Takamisawa RY5W-K data sheet:
http://www.fujitsu.com/us/services/edevices/components/relays/signal/tmpl_parts_ry-5w-k.html
http://www.fujitsu.com/us/services/edevices/components/relays/signal/tmpl_parts_ry-5w-k.html
There's the Nais/panasonic too:
http://www.relays-r-us.co.uk/webpages/products/nais_cat_index.htm#
look at the 'DS' (2e)series.
http://www.relays-r-us.co.uk/webpages/products/nais_cat_index.htm#
look at the 'DS' (2e)series.
Hi Paul,
When you mentioned:
“For instance I discovered that the transistors for the mute relays aren't shown in his pics, nor mentioned in the notes but the outputs are muted unless they are installed”.
Are your referring to the 2SJ148 and the 2SK982? And if so are the IRF mosfets (9610 and 610) that are parallel just another option if you don’t go w/ the 2SJ148 and 2SK982???
Stan
When you mentioned:
“For instance I discovered that the transistors for the mute relays aren't shown in his pics, nor mentioned in the notes but the outputs are muted unless they are installed”.
Are your referring to the 2SJ148 and the 2SK982? And if so are the IRF mosfets (9610 and 610) that are parallel just another option if you don’t go w/ the 2SJ148 and 2SK982???
Stan
Hi Stan,
Yes Q49,50,51,52. The parallel transistors do seem to represent an either/or option - I used the IRF's because I had a small pile of them and they work fine.
Thanks for the heads up on the DS2E relay option. Was able to find the panasonics at a local surplus parts supplier. Also ordered a handful of 8TQ100 schottky rectifiers to upgrade the PSU's.
cheers
Paul
Yes Q49,50,51,52. The parallel transistors do seem to represent an either/or option - I used the IRF's because I had a small pile of them and they work fine.
Thanks for the heads up on the DS2E relay option. Was able to find the panasonics at a local surplus parts supplier. Also ordered a handful of 8TQ100 schottky rectifiers to upgrade the PSU's.
cheers
Paul
Relay Mod
Output relay mod is done. Apart from a minor "foxes paw"* of connecting + of right channel to gnd and wondering why I had a what sounded like a major ground loop the procedure was straight forward and as described.
It is easiest to jumper from relay side of the output resistors to the output connector pads. I also changed the digital rectifiers to 8TQ100 schottky so not sure what mod contributed what. There seems to be a minor improvement to the precision of imaging, which is similar to the results I got performing the same mod on the DAC2.0. The improvement isn't earth shattering, but still worthwhile I think.
Reclocker
It seems the bias voltage applied to the PPL loop in the original D1 is partly required bias the Fujitsu Lithium Tantalate VCXO used to closer to the middle of it's pull range. I pretty sure this will not be required when using the Tent Labs VCXO.
*faux pas
Output relay mod is done. Apart from a minor "foxes paw"* of connecting + of right channel to gnd and wondering why I had a what sounded like a major ground loop the procedure was straight forward and as described.
It is easiest to jumper from relay side of the output resistors to the output connector pads. I also changed the digital rectifiers to 8TQ100 schottky so not sure what mod contributed what. There seems to be a minor improvement to the precision of imaging, which is similar to the results I got performing the same mod on the DAC2.0. The improvement isn't earth shattering, but still worthwhile I think.
Reclocker
It seems the bias voltage applied to the PPL loop in the original D1 is partly required bias the Fujitsu Lithium Tantalate VCXO used to closer to the middle of it's pull range. I pretty sure this will not be required when using the Tent Labs VCXO.
*faux pas
DAC analog regulator mods
Just done a nice mod on the DAC analog regulators based on those described on Martin Clark's pages:
http://www.acoustica.org.uk/t/3pin_reg_notes1.html
I've replaced the each of the 300R resistors R105,109,126,127 with a unit made up of 2 parallel 3.3V zeners and a 1N4148 diode. With the 100R resistor Spencer uses this is pushing the limits giving 5.3-5.4V, tho' this is still within the supply specs from the PCM63. Probably better to change the 100R to 120R or use 3V zeners to give an output closer to 5V.
If you try this make sure you pull the DAC chips and verify the voltages are within spec. I hooked up the negative unit incorrectly and had -17.5V on pin 28, which would have definitely killed the DAC if it had been left in place!!!
I've nabbed the correct diode orientation from this pdf:
http://www.diyaudio.com/forums/attachment.php?s=&postid=1099238&stamp=1168083193
I haven't implemented the 10uF (or bigger) cap from adj to gnd yet. This is on the to do list.
The result of the mod should be a quieter analog supply to the DAC, which seem to have clean things up slightly. It's hard to be certain as I can't A/B the changes but I seem to be getting slightly better detail and sharper imaging. Instruments stand out from the background a little better than previously using my reference recording - Miles Davis' "Bitches Brew". "Spanish Key" is definitely sounding better than I've heard it before.
I'm also thinking about doing the same mod on the digital supplies, and replacing the LT1117 regs with LM340-T5 smd regs. The LT's have far worse ripple rejection and output noise specs than the lowly 340's so this seems like a no brainer but a little fiddly as a replacement job!!
cheers
Paul
Just done a nice mod on the DAC analog regulators based on those described on Martin Clark's pages:
http://www.acoustica.org.uk/t/3pin_reg_notes1.html
I've replaced the each of the 300R resistors R105,109,126,127 with a unit made up of 2 parallel 3.3V zeners and a 1N4148 diode. With the 100R resistor Spencer uses this is pushing the limits giving 5.3-5.4V, tho' this is still within the supply specs from the PCM63. Probably better to change the 100R to 120R or use 3V zeners to give an output closer to 5V.
If you try this make sure you pull the DAC chips and verify the voltages are within spec. I hooked up the negative unit incorrectly and had -17.5V on pin 28, which would have definitely killed the DAC if it had been left in place!!!
I've nabbed the correct diode orientation from this pdf:
http://www.diyaudio.com/forums/attachment.php?s=&postid=1099238&stamp=1168083193
I haven't implemented the 10uF (or bigger) cap from adj to gnd yet. This is on the to do list.
The result of the mod should be a quieter analog supply to the DAC, which seem to have clean things up slightly. It's hard to be certain as I can't A/B the changes but I seem to be getting slightly better detail and sharper imaging. Instruments stand out from the background a little better than previously using my reference recording - Miles Davis' "Bitches Brew". "Spanish Key" is definitely sounding better than I've heard it before.
I'm also thinking about doing the same mod on the digital supplies, and replacing the LT1117 regs with LM340-T5 smd regs. The LT's have far worse ripple rejection and output noise specs than the lowly 340's so this seems like a no brainer but a little fiddly as a replacement job!!
cheers
Paul
Spencer,
I think the 47uH inductors you have inline with the PCM63 supply rails might be a concern. These will form a resonant circuit with the 47uF capacitors, resonant frequency will be 3.386kHz right in the audio band.
The characteristic impedance is 1 ohm so the DC resistance of the inductor and/or ESR of the cap might be enough to damp it properly but could be worth double checking.
Dave
I think the 47uH inductors you have inline with the PCM63 supply rails might be a concern. These will form a resonant circuit with the 47uF capacitors, resonant frequency will be 3.386kHz right in the audio band.
The characteristic impedance is 1 ohm so the DC resistance of the inductor and/or ESR of the cap might be enough to damp it properly but could be worth double checking.
Dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.