pupDAC Step-by-Step Build Thread
Jan 29, 2015 at 6:57 PM Post #317 of 359
  Thanks for the advice TomB.. I cleaned the board with flux cleaner and deionised water, let it dry under a bulb for about 12hrs.. the problem is still there.. On the PCM1794, There's a bridge on pins 11 and 12, I've scratched the soldermask on the via next to these pads so it's impossible to remove that bridge now, on the schematic these two pins are grounded (FMT0 and FMT1), So I left the bridge as it is.. I didn't think this could be the problem as those two pins are connected to the same via (one I managed to scrape the soldermask off) that goes to C8 on the other side anyway...
 
I'm very new at this so I don't know how I should go about checking if a cap is shorting, can I just check the voltage across every cap? I'll try to post some photos if I can get something to focus close enough to see the details of the board.. Thanks again!
 
Edit:
 
Pics:
1794
so4tqh.jpg

 
2707
2vw8ep5.jpg

 
Board top
qnpqmt.jpg

 
Board bottom
kq2qs.jpg


To tell the truth, I don't see any issue with the PCM chips.  Another thing that can cause the symptoms you describe is the opamp.  Are you certain everything is clean and dry around it?  The fact that you haven't trimmed any leads is suspect.  Conceivably, those could be touching something and arcing when you're trying to operate the DAC.  Also look for anything that might be shorting right around the opamp.  It looks like you used the OPA2835.  The 2836 is much, much better.
 
Jan 29, 2015 at 8:11 PM Post #318 of 359
 
To tell the truth, I don't see any issue with the PCM chips.  Another thing that can cause the symptoms you describe is the opamp.  Are you certain everything is clean and dry around it?  The fact that you haven't trimmed any leads is suspect.  Conceivably, those could be touching something and arcing when you're trying to operate the DAC.  Also look for anything that might be shorting right around the opamp.  It looks like you used the OPA2835.  The 2836 is much, much better.


I've checked around the opamp and I can't see anything that might be shorting.. I'll trim the leads a bit closer to the joints. Could you please tell me if there should be a voltage across the mica caps? they both have 130mV on both sides of the cap, reference to ground. (and 0V across the caps). I can't think why I used the 2835 :s.. maybe mouser didn't have it instock at the time, but I will replace it with a 2836..
 
Thank you so much for taking the time to assist me :).
 
Jan 29, 2015 at 9:51 PM Post #319 of 359
 
 
To tell the truth, I don't see any issue with the PCM chips.  Another thing that can cause the symptoms you describe is the opamp.  Are you certain everything is clean and dry around it?  The fact that you haven't trimmed any leads is suspect.  Conceivably, those could be touching something and arcing when you're trying to operate the DAC.  Also look for anything that might be shorting right around the opamp.  It looks like you used the OPA2835.  The 2836 is much, much better.


I've checked around the opamp and I can't see anything that might be shorting.. I'll trim the leads a bit closer to the joints. Could you please tell me if there should be a voltage across the mica caps? they both have 130mV on both sides of the cap, reference to ground. (and 0V across the caps). I can't think why I used the 2835 :s.. maybe mouser didn't have it instock at the time, but I will replace it with a 2836..
 
Thank you so much for taking the time to assist me :).


I don't have one apart to measure right now, but any capacitor in a live circuit should read zero DC volts.  That's the purpose in many instances - they block DC, but allow AC current to flow.
 
The 2836 is very much superior, but maybe you should try to get rid of the ticking and noise before changing out the opamp.  The reason I mentioned the opamp is that if it is oscillating, it can also cause the ticking and noise that you describe.  I couldn't see the resistor pads very clearly in your photo and thought you might have some shorting in that area.  You might check to see if the opamp is getting hot?
 
If that doesn't work, I would suggest trying to re-flow the solder on the simpler parts - don't mess with the PCM chips or the TPS regulators.  See if that helps, first.
 
Mar 3, 2015 at 8:16 AM Post #320 of 359
 
I don't have one apart to measure right now, but any capacitor in a live circuit should read zero DC volts.  That's the purpose in many instances - they block DC, but allow AC current to flow.
 
The 2836 is very much superior, but maybe you should try to get rid of the ticking and noise before changing out the opamp.  The reason I mentioned the opamp is that if it is oscillating, it can also cause the ticking and noise that you describe.  I couldn't see the resistor pads very clearly in your photo and thought you might have some shorting in that area.  You might check to see if the opamp is getting hot?
 
If that doesn't work, I would suggest trying to re-flow the solder on the simpler parts - don't mess with the PCM chips or the TPS regulators.  See if that helps, first.

 
Hi tomb,
 
Apologies for not replying for a long time, I only got around to looking at the pupdac again today. And I figured out what it was. The noise comes from my wifi signal.. As soon as I dropped the output power of the router, the noise from the pupdac was completely gone. If I increase the power, noise comes back, so it definitely is the wifi signal.
 
I don't have the case for the pupdac, but I'm guessing when inside the metal enclosure this won't be an issue. I will try and mount it inside a metal housing to make sure that solves the issue before I order the custom made ones from yourself.  I really appreciate you taking the time to reply to my earlier posts. Hopefully this might help someone with a similar issue down the line.
 
It also picks up noise from a wireless keyboard and a mouse to lesser extent, moving the receivers well away from the dac got rid of a screeching noise it makes when the mouse was moved and a very low ticking noise from the keyboard. It is pretty much dead silent at this point.
 
I've also got the 2836, can't wait to swap it on and see what the difference will be. Even as it is I'm amazed at the sound quality!
 
Best regards,
 
Nivz.
 
Mar 3, 2015 at 3:20 PM Post #321 of 359
 
 
I don't have one apart to measure right now, but any capacitor in a live circuit should read zero DC volts.  That's the purpose in many instances - they block DC, but allow AC current to flow.
 
The 2836 is very much superior, but maybe you should try to get rid of the ticking and noise before changing out the opamp.  The reason I mentioned the opamp is that if it is oscillating, it can also cause the ticking and noise that you describe.  I couldn't see the resistor pads very clearly in your photo and thought you might have some shorting in that area.  You might check to see if the opamp is getting hot?
 
If that doesn't work, I would suggest trying to re-flow the solder on the simpler parts - don't mess with the PCM chips or the TPS regulators.  See if that helps, first.

 
Hi tomb,
 
Apologies for not replying for a long time, I only got around to looking at the pupdac again today. And I figured out what it was. The noise comes from my wifi signal.. As soon as I dropped the output power of the router, the noise from the pupdac was completely gone. If I increase the power, noise comes back, so it definitely is the wifi signal.
 
I don't have the case for the pupdac, but I'm guessing when inside the metal enclosure this won't be an issue. I will try and mount it inside a metal housing to make sure that solves the issue before I order the custom made ones from yourself.  I really appreciate you taking the time to reply to my earlier posts. Hopefully this might help someone with a similar issue down the line.
 
It also picks up noise from a wireless keyboard and a mouse to lesser extent, moving the receivers well away from the dac got rid of a screeching noise it makes when the mouse was moved and a very low ticking noise from the keyboard. It is pretty much dead silent at this point.
 
I've also got the 2836, can't wait to swap it on and see what the difference will be. Even as it is I'm amazed at the sound quality!
 
Best regards,
 
Nivz.

 
Thanks for your kind comments.  I'm glad you found out what's wrong.
 
Just and FYI, but here's what is next to mine (or within 3 feet):
  • PC
  • Two LCD monitors
  • Wireless home phone
  • LAN wireless router
  • Cable modem, also with wireless router
  • Wireless keyboard
  • Wireless mouse
I don't have a problem with the PupDAC in its case.
wink.gif

 
Mar 12, 2015 at 7:09 PM Post #322 of 359
LED on, pupdac not recognized by pc, the following voltage check fails
 
4V5 = 4.7V
-5V = 0V
-2V5 = 0V
 
Any idea?
 
PS: going to check U7/C32-34 Tomorrow, but visual inspection looks fine
 
Mar 12, 2015 at 9:51 PM Post #323 of 359
  LED on, pupdac not recognized by pc, the following voltage check fails
 
4V5 = 4.7V
-5V = 0V
-2V5 = 0V
 
Any idea?
 
PS: going to check U7/C32-34 Tomorrow, but visual inspection looks fine


The 4.7V is fine.  The 5 and 2.5 are not.  Check for shorts around U7 and U8.
 
Mar 18, 2015 at 8:25 PM Post #324 of 359
Hi, I recently built my pupDAC and it's great. I have one issue though, the voltage offset between channels is too much, and one channel is noticeably louder than the other. All of the voltage test points are within .05 volts of stated value, except for the 4.5v one that reads 4.749v. The big issue though is that one channel reads .5mv and the other reads 3.3 mv. I have tested the barrel of the rca connectors, the pads where they mount, and the unpopulated 1/8" output pads. They all read the same values.
 
I'm not sure what I should be checking to troubleshoot this, so any help is much appreciated.
 
Mar 18, 2015 at 9:26 PM Post #325 of 359
  Hi, I recently built my pupDAC and it's great. I have one issue though, the voltage offset between channels is too much, and one channel is noticeably louder than the other. All of the voltage test points are within .05 volts of stated value, except for the 4.5v one that reads 4.749v. The big issue though is that one channel reads .5mv and the other reads 3.3 mv. I have tested the barrel of the rca connectors, the pads where they mount, and the unpopulated 1/8" output pads. They all read the same values.
 
I'm not sure what I should be checking to troubleshoot this, so any help is much appreciated.


First, 4.79V is pretty much where it should be for the 4.5V test point.  The 4.5V is a theoretical number, but in practice, it's always worked out to about 4.75V, depending on scatter in the various parts.
 
As for the offset, it's possible that you had some bad luck in the resistor tolerances.  I haven't sold a kit in awhile, so I'm curious - did you purchase and install the 1% V-D RN50 series resistors?  If not, that could be the problem.  The OPA2836 is a very aggressive opamp and if the resistors are not very, very close in value, you could end up with more offset than desired.  It's the reason we chose the 1% resistors, because that lessens the chance for unmatched resistors between the channels.
 
Anyway, look closely at the resistors on the back side of the PCB.  Those are the ones that were so critical they were located on the bottom of the PCB - in order to reduce the lead length.  These are R6, R9, R17, and R21.
 
By the way - just in case, the silkscreen was messed up on the R6 resistor.  It looks more like an "R5," but R5 is clearly labeled on the top.  The resistor on the bottom is "R6."  I would think if you got these two mixed up, it might explain a lot.
 
Anyway, let us know ...
 
Mar 24, 2015 at 3:46 PM Post #326 of 359
 
The 4.7V is fine.  The 5 and 2.5 are not.  Check for shorts around U7 and U8.

C34/35 was dead, replaced the cap, now all voltage check is fine, windows still doesn't recognize the device, but gnd to jack are showing 1.85V. here are some pic of the board itself: http://imgur.com/2vOeYD4,02UIKKL,Vldscce
 
on linux, dmesg gives device descriptor read/64, error -62
 
Mar 25, 2015 at 6:51 AM Post #327 of 359
 


The 4.7V is fine.  The 5 and 2.5 are not.  Check for shorts around U7 and U8.

C34/35 was dead, replaced the cap, now all voltage check is fine, windows still doesn't recognize the device, but gnd to jack are showing 1.85V. here are some pic of the board itself: http://imgur.com/2vOeYD4,02UIKKL,Vldscce

on linux, dmesg gives device descriptor read/64, error -62


Did you solder those top pads for the oscillator (clock chip)? It doesn't show in your pics, but what c an be seen looks awfully bare. Also, one of the two joints showing looks awfully cold. If the clock chip is not working, there's no way to synchronize/hand-shake the USB connection and the PC will not recognize the device.
 
Mar 28, 2015 at 10:38 PM Post #328 of 359
 
First, 4.79V is pretty much where it should be for the 4.5V test point.  The 4.5V is a theoretical number, but in practice, it's always worked out to about 4.75V, depending on scatter in the various parts.
 
As for the offset, it's possible that you had some bad luck in the resistor tolerances.  I haven't sold a kit in awhile, so I'm curious - did you purchase and install the 1% V-D RN50 series resistors?  If not, that could be the problem.  The OPA2836 is a very aggressive opamp and if the resistors are not very, very close in value, you could end up with more offset than desired.  It's the reason we chose the 1% resistors, because that lessens the chance for unmatched resistors between the channels.
 
Anyway, look closely at the resistors on the back side of the PCB.  Those are the ones that were so critical they were located on the bottom of the PCB - in order to reduce the lead length.  These are R6, R9, R17, and R21.
 
By the way - just in case, the silkscreen was messed up on the R6 resistor.  It looks more like an "R5," but R5 is clearly labeled on the top.  The resistor on the bottom is "R6."  I would think if you got these two mixed up, it might explain a lot.
 
Anyway, let us know ...

So I finally had some time to mess about with this. I tested these 4 resistors and they're all spot on. Tested with a Fluke 87V 
 
R6 - 1.499k
R9 - 10.2
R17 - 1.499k
R21 - 10.2
 
I reflowed the OPA 2836 for sanity's sake even though the joints looked nice. I looked over the rest of the joints using a 100mm macro lense and zooming in, which really helps show those chip leg joints, and everything looks nice. After putting the resistors back in, I'm still getting the a high offset.
 
Any other ideas?
 
Edit: To add a bit more, I tested the other resistors, but in circuit cause I didn't feel like pulling them all. Obviously some are going to read different than their values in circuit, but I figured it would let me compare both channels to each other. Everything matched up with it's respective pair. The only deviations from the listed values were R7 and R18 read 94 instead of 100 ohms, and R8 and R19 read 113 instead of 1.5k.
 
Mar 28, 2015 at 10:51 PM Post #329 of 359
 
 
First, 4.79V is pretty much where it should be for the 4.5V test point.  The 4.5V is a theoretical number, but in practice, it's always worked out to about 4.75V, depending on scatter in the various parts.
 
As for the offset, it's possible that you had some bad luck in the resistor tolerances.  I haven't sold a kit in awhile, so I'm curious - did you purchase and install the 1% V-D RN50 series resistors?  If not, that could be the problem.  The OPA2836 is a very aggressive opamp and if the resistors are not very, very close in value, you could end up with more offset than desired.  It's the reason we chose the 1% resistors, because that lessens the chance for unmatched resistors between the channels.
 
Anyway, look closely at the resistors on the back side of the PCB.  Those are the ones that were so critical they were located on the bottom of the PCB - in order to reduce the lead length.  These are R6, R9, R17, and R21.
 
By the way - just in case, the silkscreen was messed up on the R6 resistor.  It looks more like an "R5," but R5 is clearly labeled on the top.  The resistor on the bottom is "R6."  I would think if you got these two mixed up, it might explain a lot.
 
Anyway, let us know ...

So I finally had some time to mess about with this. I tested these 4 resistors and they're all spot on. Tested with a Fluke 87V 
 
R6 - 1.499k
R9 - 10.2
R17 - 1.499k
R21 - 10.2
 
I reflowed the OPA 2836 for sanity's sake even though the joints looked nice. I looked over the rest of the joints using a 100mm macro lense and zooming in, which really helps show those chip leg joints, and everything looks nice. After putting the resistors back in, I'm still getting the a high offset.
 
Any other ideas?

 
Unnecessary lead length?  Are the resistors flush to the PCB.  For the four you measured above, are they on the bottom of the PCB, next to the opamp?  What about the mica capacitors?  Those have to really be pressed into the PCB to keep from sticking way up over any other part.  The added lead length could be offsetting things.
 
Have you tested the voltage into the opamp?  These would be pins 8 and 4.  See if those are the same voltage.  Pin 8 should be the same as the 2.5V test point coming from U10.  Pin 4 should be the -2.5V test point coming from U8.  See what those voltages are at the pins on the opamp, though - relative to Ground.
 
I can't think of anything else.  I would suggest that 3.3mV is not that much offset in the scheme of things, but you are claiming to hear a sound level difference between channels.  How about the rest of the through-hole resistors?  Can you compare their resistances, left channel to right channel and see if there's something not matched?
 
Mar 28, 2015 at 11:06 PM Post #330 of 359
   
Unnecessary lead length?  Are the resistors flush to the PCB.  For the four you measured above, are they on the bottom of the PCB, next to the opamp?  What about the mica capacitors?  Those have to really be pressed into the PCB to keep from sticking way up over any other part.  The added lead length could be offsetting things.
 
Have you tested the voltage into the opamp?  These would be pins 8 and 4.  See if those are the same voltage.  Pin 8 should be the same as the 2.5V test point coming from U10.  Pin 4 should be the -2.5V test point coming from U8.  See what those voltages are at the pins on the opamp, though - relative to Ground.
 
I can't think of anything else.  I would suggest that 3.3mV is not that much offset in the scheme of things, but you are claiming to hear a sound level difference between channels.  How about the rest of the through-hole resistors?  Can you compare their resistances, left channel to right channel and see if there's something not matched?

Everything is flush to the board, the mica caps do stick up maybe 1-2mm since that was as far as I could get them. Voltage into pins 4 and 8 match up.
 
As for the other resistors, I actually edited my last post with those readings.
 

Users who are viewing this thread

Back
Top