[DingoSmuggler]

just putting the last few components in and should have this going.

curious as to what the corner frequency of the analog lowpass filter is, as the 220pf seems like a fairly small value, and my calculations had up rather high (hoping i was wrong)

 

[rededge]

i have soldered up my dac but it is  not getting recognized. all the voltage is fine except the 4.5v is 4.7v.
 
any assistance?
 

 

[jERiCOh]

I had the same problem. I used a better quality usb cable and luckily it started to work... As for the 4V5 reading 4.7V, mine is at 4.75V and works perfectly. I hope changing the cable works for you too.

 

[tomb]

Quote:

i have soldered up my dac but it is  not getting recognized. all the voltage is fine except the 4.5v is 4.7v.
 
any assistance?
 

If it's not the cable and all your voltages are OK, then take a closer look at U1, the PCM2707.  That chip controls all of the USB and connection to the PC.  If you have some bridges or pins that may not be cleanly soldered, that would prevent the PC from recognizing the pupDAC as a legitimate USB device.  I hesitate to suggest re-flowing, but if it all else fails ...
 
You might try posting a pic, too - maybe we'll spot something.

 

[rededge]

i re-soldered re-cleaned and did manage to get it working. so i decide to give it one more clean and when i reconnected it now windows simply does not detect the dac, no sound to indicate a usb device.
 
3v3 next to pcm2707 is less than 0.5v and the +2.5v is less than 0.5v.
 
so what have i blow up.

 

[randytsuch]

Quote:

i re-soldered re-cleaned and did manage to get it working. so i decide to give it one more clean and when i reconnected it now windows simply does not detect the dac, no sound to indicate a usb device.
 
3v3 next to pcm2707 is less than 0.5v and the +2.5v is less than 0.5v.
 
so what have i blow up.

 
On the schematic, U2 is in the upper right of sheet 1, this is the regulator that generates 3.3V for U1.
 
I would check to make sure all pins of U2 are soldered correctly.  You could also check L3 because 3.3V has to pass through L3, but that is a relatively easy part to solder.
 
U10 generates the 2.5v which is used for the opamp for the output, and won't affect the USB device detection.  You still need to fix, and I would look at U10.  I had my share of problems soldering those little regs down too.
 
When you cleaned the board, you probably pushed pins around.  If the pins are not well soldered, then they could lose contact, I would guess that is what happened.  Better to find and fix now than to have it stop working in a month and have you scratching your head.
 
 
 
Another general comment.
The schematic could be more helpful.  For example is says "TP1" on the schematic, but it says "3V3" on the board, so you need to figure out how to relate these.  You also need to cross reference the schematic against the parts list for all of the parts values.
 
I have marked up my schematic to add in parts values, some notes, and part numbers so I don't have to refer to the PL all the time.
 
Randy

 

[christsay]

Just finished up the build and everything is working great.  Thanks for the great design and kit guys!
 
Also, just want to say that customer services at beezar is awesome!  Mouser miscounted and only put five of the ferrites in the bag instead of 6, Tom shipped out two ferrites the day before Christmas and I received them today.
 
-chris

 

[tomb]

Quote:

Just finished up the build and everything is working great.  Thanks for the great design and kit guys!
 
Also, just want to say that customer services at beezar is awesome!  Mouser miscounted and only put five of the ferrites in the bag instead of 6, Tom shipped out two ferrites the day before Christmas and I received them today.
 
-chris

Great!  Glad to hear it and thank you for the kind comments.

 

[tomb]

I've updated the BOM on the website to contain two changes:
 
1. Added the option for using the OPA2836 opamp for U9
2. Included the Mouser version of the 3.5mm jack: STX-3150-3N
 
Both the opamp and the 3.5mm jack have been fully vetted and they work well.
(I built 4 more pupDACs over the holidays using these parts.):
http://www.diyforums.org/PupDAC/PupDAC-bom.php

 

[tomb]

Any thoughts, results, or comments from some of you that have built a production version?

 

[Lefuneste]

Hi guys.
Just finished mine today!! I have had great time working on it! Spent about 6 hours taking PLENTY of time soldering it all perfectly. It just came to life at once, as usual with all your kits Tom! I may be the first French Puppy breeder as well!! Yeahhh!!
 
I have to say that this thing sounds outrageously good!! I used to have a GrubDAC customized to my needs (caps and clock), but I wasn't suspecting such a big difference in sound with the Puppy.
The sound is definitely more transparent. I started with an opening track of ambient music, where there is a water noise. I know this track very well having heard this record again and again (flac file with Foobar exclusive output mode). The record is called "Voices from the lake" and is from a collective of minimal DJ from Italy. It is really amazing if you're into ambient and minimal stuff. Anyway, this water sounds on the Pup incredibly real compared to the already excellent rendering of the Grub that I knew so well. I also tried it with Anthony Rother's "The Machine Room" album. Synth modulations are freakingly real you can feel the texture of the sound. Overall, the highs are much more detailed and airy. Really amazing how they detach and take room. The low frequencies go really LOW. They are abyssal!!. On my gear they are lacking a very tiny bit of definition, but this is definitely due to my setup or to the tweaks I conducted (see below). I have a tube headphone preamp (Project Sunrise with Amperex Buggle Boy ECC88 from Holland) and I only tried my headphones so far (Beyer DT990Pro 250 Ohms). The sound is incredibly spacious and vast. Almost out of this world!! I have difficulties stopping listening to this DAC. I've been stucked to it for the last 12 hours!! I will try it tomorrow on my Adam monitors.
 
On the list of remarks
 
-I have implemented the improved 2836 Opamp as suggested, not many hesitations about it. Too bad I can't find a simple way to roll the opamps. I haven't found any SMD to DIP8 adapter. If someone knows where to get this please let me know. Maybe a future version will include this. It would be great, but less useful than let's say on a CD player due to the limited number of possible opamp choices for such a low voltage.
 
-Silkscreen may have a little numbering issue for R5 on the back of the PCB, it should read R6 (I thought my eyes were dead but no...). Really not much of an issue if you know what you're talking about when handling the iron, but maybe for the beginners it may lead to confusion. Probably already noticed by someone or it really my eyes that are dying...
 
-For the soldering of the PCM1794 chip, it is quite difficult I have to say as the pins are really close and the chip is high as a cliff. The secret is to have a good flux pen. I accidentally soldered 3 pins well together (Dohh...) with a BIG UGLY drop of solder. I almost thought I had Murphy's law striking... I wicked as much as I could (be careful not to pull the wick on a hard solder or you'll pull the pins out!!), but I couldn't get the bridge to leave because of thick solder texture. Then I grabbed my pen, and Voila!! It just melted the solder perfectly and had it climb up the pins by capillarity just like magic. I touched up all pins afterwards to have a perfect finish. I fluxed every single solder afterwards (Voltage regulator, USB chip even some ceramics). It's just mandatory to have one of these for SMD work. Because after you mess with solder too long it tends to loose it's flux, and behaves like a paste. If you add flux again it goes back to liquid. VERY USEFUL!
 
-I confirm, the reference I found for the mini-jack connector goes perfectly. I had spent a couple of hours reading data sheets to find this one... I am not completely useless Yeah!!
 
-I couldn't help but to change the supplied electrolytics by some Oscons (that's the way I do things sorry, don't blame me!!) I just find the oscons to sound unbelievable and I believe I can hear the difference!! Arf...
 
But Tom, where can I find detailed explanations for the electrolytics choices of values. I would be very interested!!
There are 4 types, One (C22) is the main PSU buffer (upgraded to Rubycon MBZ 1500 uF at 10V), 2 are for output decoupling (C26, C27 upgraded to oscons SEPC 470uF 6.3V), one is not yet understood (C10 upgraded to Nichicon KA 225uf 25V) but seems to have an important role in filtering (I need to read more of the manual and of the blog sorry!), and the last 4 Voltage decoupling for the DAC chip (C2 C9 C16 C19 upgraded to Elna SILMIC II 22uf 25V, but I acknowledge the FM panasonics are really good too).
But Tom, why on Earth did you specify 63V value on these? I mean there must be a reason, but when I look at the diagrams I only see 5V class voltage there. I really scratched my head for good on this... Does 63V value bring something to this decoupling function such as higher ripple tolerance? DAC technical note only states 10uF value there but doesn't recommend higher voltage value so I am puzzled. I am planning to replace these 4 by oscons too by 47uF at 16V or maybe large Tantalums they may be adequate there. Is there a problem going to 16V? I can't see any it works fine at 25V with my ELNAs, but you're the master!
 
Anyway, the PupDAC is a really wonderful piece of equipment I have to say probably sounding like multi KUSD DAC!! I will keep you updated for more tweaks in the future.
 
All the Best!!
 

Close match on all caps
 

That's a close one, perfect fit with Rubycon MBZ 1500uF 10V
 

It's ALIVE...

 

[tomb]

Quote:

Hi guys.
Just finished mine today!! I have had great time working on it! Spent about 6 hours taking PLENTY of time soldering it all perfectly. It just came to life at once, as usual with all your kits Tom! I may be the first French Puppy breeder as well!! Yeahhh!!
 
I have to say that this thing sounds outrageously good!! I used to have a GrubDAC customized to my needs (caps and clock), but I wasn't suspecting such a big difference in sound with the Puppy.
The sound is definitely more transparent. I started with an opening track of ambient music, where there is a water noise. I know this track very well having heard this record again and again (flac file with Foobar exclusive output mode). The record is called "Voices from the lake" and is from a collective of minimal DJ from Italy. It is really amazing if you're into ambient and minimal stuff. Anyway, this water sounds on the Pup incredibly real compared to the already excellent rendering of the Grub that I knew so well. I also tried it with Anthony Rother's "The Machine Room" album. Synth modulations are freakingly real you can feel the texture of the sound. Overall, the highs are much more detailed and airy. Really amazing how they detach and take room. The low frequencies go really LOW. They are abyssal!!. On my gear they are lacking a very tiny bit of definition, but this is definitely due to my setup or to the tweaks I conducted (see below). I have a tube headphone preamp (Project Sunrise with Amperex Buggle Boy ECC88 from Holland) and I only tried my headphones so far (Beyer DT990Pro 250 Ohms). The sound is incredibly spacious and vast. Almost out of this world!! I have difficulties stopping listening to this DAC. I've been stucked to it for the last 12 hours!! I will try it tomorrow on my Adam monitors.
 

 
Many, many thanks for relaying your opinions.  Cobaltmute and I have known this for some time, but it's nice to hear it all confirmed by someone not connected.  Things have been kind've slow for the pup, actually.  It doesn't have the razz-ma-tazz 24-bit USB to attract all the unwashed masses, but I don't think it needs it.  Some people say the 24-bit thing is over-rated, too, compared to a good-quality DAC that's been implemented well, and this is definitely one! 

 

On the list of remarks
 
-I have implemented the improved 2836 Opamp as suggested, not many hesitations about it. Too bad I can't find a simple way to roll the opamps. I haven't found any SMD to DIP8 adapter. If someone knows where to get this please let me know. Maybe a future version will include this. It would be great, but less useful than let's say on a CD player due to the limited number of possible opamp choices for such a low voltage.
 

 
I hear you - but it's going to be a bit difficult with the faster opamps that we've listed and vetted.  As you noticed below, cobaltmute even put some of the resistors on the back side to cut down on the trace and lead length.  The potential for instability is definitely there, as some of us experienced during the prototyping stage.  There are such adapters, I think, but I doubt that they would be a good idea.
 

-Silkscreen may have a little numbering issue for R5 on the back of the PCB, it should read R6 (I thought my eyes were dead but no...). Really not much of an issue if you know what you're talking about when handling the iron, but maybe for the beginners it may lead to confusion. Probably already noticed by someone or it really my eyes that are dying...
 

 
Dang!  I've built 5 of the new PCBs and never noticed this.  It's not exactly a "5", but it's for sure a piss-poor "6."  I will highlight that this weekend on the website to make certain that no one makes this mistake.  Thanks for pointing it out!!
 

-For the soldering of the PCM1794 chip, it is quite difficult I have to say as the pins are really close and the chip is high as a cliff. The secret is to have a good flux pen. I accidentally soldered 3 pins well together (Dohh...) with a BIG UGLY drop of solder. I almost thought I had Murphy's law striking... I wicked as much as I could (be careful not to pull the wick on a hard solder or you'll pull the pins out!!), but I couldn't get the bridge to leave because of thick solder texture. Then I grabbed my pen, and Voila!! It just melted the solder perfectly and had it climb up the pins by capillarity just like magic. I touched up all pins afterwards to have a perfect finish. I fluxed every single solder afterwards (Voltage regulator, USB chip even some ceramics). It's just mandatory to have one of these for SMD work. Because after you mess with solder too long it tends to loose it's flux, and behaves like a paste. If you add flux again it goes back to liquid. VERY USEFUL!
 
-I confirm, the reference I found for the mini-jack connector goes perfectly. I had spent a couple of hours reading data sheets to find this one... I am not completely useless Yeah!!
 

 
Yeah, it's really no different than the old PCM2702 on the AlienDAC and the BantamDAC - or the PCM2704/5 used on the SkeletonDAC, but yes - it's probably an order of magnitude more difficult than the square PCM2706/7.  A flux pen was talked about and recommended all the way back during the AlienDAC.  I bought one and have used it ever since - it's still half-full.  So, the investment is really cheap considering how long it lasts.  I've never done any SMD without it.
 
EDIT: Oops - forgot that it was scootsit that found the Mouser mini-jack!  It worked out well on the last 4 I built and Mouser will probably be the primary source after this first run of kits.  The ones supplied in the kits came from DigiKey, though.  I had purchased 100 of them well before I ordered the rest of the kit parts from Mouser.

 

-I couldn't help but to change the supplied electrolytics by some Oscons (that's the way I do things sorry, don't blame me!!) I just find the oscons to sound unbelievable and I believe I can hear the difference!! Arf...
 
But Tom, where can I find detailed explanations for the electrolytics choices of values. I would be very interested!!
There are 4 types, One (C22) is the main PSU buffer (upgraded to Rubycon MBZ 1500 uF at 10V), 2 are for output decoupling (C26, C27 upgraded to oscons SEPC 470uF 6.3V), one is not yet understood (C10 upgraded to Nichicon KA 225uf 25V) but seems to have an important role in filtering (I need to read more of the manual and of the blog sorry!), and the last 4 Voltage decoupling for the DAC chip (C2 C9 C16 C19 upgraded to Elna SILMIC II 22uf 25V, but I acknowledge the FM panasonics are really good too).
But Tom, why on Earth did you specify 63V value on these? I mean there must be a reason, but when I look at the diagrams I only see 5V class voltage there. I really scratched my head for good on this... Does 63V value bring something to this decoupling function such as higher ripple tolerance? DAC technical note only states 10uF value there but doesn't recommend higher voltage value so I am puzzled. I am planning to replace these 4 by oscons too by 47uF at 16V or maybe large Tantalums they may be adequate there. Is there a problem going to 16V? I can't see any it works fine at 25V with my ELNAs, but you're the master!
 

 
OK - You scared the you-know-what out of me with that pic of the giant power cap.  I wasn't reading too closely at first and thought maybe Mouser had messed up and supplied a 63V cap instead of a 6.3V cap, there.  C22 is definitely a 6.3V cap and that's what's packaged in the kits. (whew!)  I assume what you photographed is your Rubycon upgrade and that's fine if it fit for you.  As for the others, it's simply impossible to find smaller uf-rated electrolytics in very low voltages.  The caps are not very big as they are and they are what cobaltmute wanted.  That said, I can't imagine that your Oscons would ever be a downgrade.

 
 

Anyway, the PupDAC is a really wonderful piece of equipment I have to say probably sounding like multi KUSD DAC!! I will keep you updated for more tweaks in the future.
 

 
Many, many thanks!!  Cobaltmute and I have toiled over this DAC for a few years, now.  You have no idea how gratifying it is to hear this. 

 

All the Best!!
 

Close match on all caps
 

That's a close one, perfect fit with Rubycon MBZ 1500uF 10V
 

It's ALIVE...

 
YAY!!!

 

[Lefuneste]

Hi Tom and thanks for your long reply.
 
-On the PCB print, I agree with you, it is not a real 5 nor a real 6, something in between. It's not really that important, but I believe you should correct it for begginners out there as R5 and R6 have different values.
 
-For the 63V value, I expected that this was the reason. It doesn't hurt actually to have thicker foils in life!! As decoupling of other DAC chips often relies on 63V values for MKS caps, I had to ask.
But you didn't answer my question about the 120 uF cap, is its value of some importance in filtering something or can I upscale it without problem? I believe it's not really a problem but I prefer to make sure.
 
-For the 24 bit stuff Ahaha don't make me laugh, you should visit France someday, I'll make you listen to my TDA1541 pushed to the limits on a 25 year old Sony CD player... CD specification only talks about 16 bit, the rest is BS as it relies on data manipulation except if you produce music in 24 bits which some people do, but usually they think their gear is the best out there and only realy on latest solid state technology, not taking seriously DIYers and tube afficionandos.I thought the same before starting to build a Project Sunrise preamp... Leave the extra bits to others (for the moment)... There are countless reports of gizmo-bits not being even statiscaly noticeable (check SACD listening tests on Wikipedia). Furthermore one extremely important aspect of your DACs (Grub & Pup) is the fact that they don't need a driver AT ALL. I've used them on Linux, Windows, Mac without ever wondering what ASIO horrible driver I had to install. I LOVE it this way. I have no words to describe this simple pleasure of plugging this killer of a DAC straight away from my pocket and killing everybody out there, wether at a party or in anyones place. I've got a huge collection of useless soundcards that I've used thoughout my long DJ life all of them eating 24 bits. I was possibly the first in France to mix MP3 in 1998 and at that time, you had to rely on terrible equipement, that got even worse in my opinion. You still mix 16bits 44.1kHz tracks you know!! Pfff... You DAC is the best price/performance/convenience out there, no doubt about it! And you have ferrites all over, not the case of many DIY DACS out there. This kit looks and sounds like a real achievement for the community. Not even talking about the uber-coolesque case design!! I actually start to thing to implement it in my car system...
 
-For the opamp, I'll leave it like this as I actually really love the sound of this 2836, not too harsh, quite analog. Perfect match, I haven't tried BOM caps, but with caps upgrades I did, the DAC sound plenty analog and detailed. Not harsh AT ALL (oscon effect possibly). Again trebles are airy, bass is really deep to the cave. Response seems linear, but not boring at all. You really rediscover your tracks and can tell difference between bad and good MP3 encoding at once. A good sign certainly. I can't stop listening this Pup!!
 
Thanks Tom!!
 
BTW did you notice that Right and Left are reversed when looking at the DAC to plug a cable? Lol... I just realised that building a passive input switch todayusing same Hammod case and a DPDT Russian military switch!! SWEET!!
 

 

[cobaltmute]

BTW did you notice that Right and Left are reversed when looking at the DAC to plug a cable? Lol... I just realised that building a passive input switch todayusing same Hammod case and a DPDT Russian military switch!! SWEET!!

This was by design to keep the straightest path from the PCM1974 to the outputs. They follow on the same side as they are on the chip. Otherwise I would have had to cross traces on the board, and I didn't want to do that for various reasons.

As for the values of the caps, so long as they are above what will be seen by the cap, you can choose any value you want.

 

[scootsit]

I used a toaster oven and did the top side reflow-style, it came out AWESOME!
I have a bridge or two, but everything is centered and straight (except U10, but it's not too bad).
 
I'm going to touch it up tomorrow and then get started on the bottom.
 
Using solder paste and the oven, the top took under 1hr to fully populate.