It all started back in January sometime where I’d had ideas of designing my own DAC and been doing a bit of reading on the subject, and it was around that time that I had an idea of what parts to use. For one, there was the tda1541 and co. DAC's that were all the rage, yet the parts were all discontinued and I didn’t like my chances of getting all the parts for it, especially in Australia. and then I saw the pcm1704, what looked like nearly the best audio dac ever, on paper anyway and it is actually in production and easily available to the diy’er, and if just had to use them.
Garbs had been asking a few questions about a dac he was doing on here a while ago (and he still is working on it
) and one part that came up was the cs8416, a feature upgraded version of the cs8412 that was real popular by those building nos DACs, and with this one it had mux input for 4 separate inputs, and plenty of pins to setup in hardware mode that later on put a lot of doubt when the dac didn’t work, but it was a nice looking receiver that I added that in too.
The rest of the parts were pretty simple to add in, the reg02's got borrowed from the guzzler USB dac, isolators were mentioned in the datasheets and I liked the idea of a completely separate analog and digital sections so they got in too, and in designing the analog psu section, I was trying to come up with +/-15v and +/-5v and the complete lack of options for regulators that came negative voltages I just settles with the good old lm78** and lm79** that was a tack onto the end.
One goal of mine, after my headphone amp that I had done a few months earlier back then that I was going to design this board to fit in a case instead of just making a board and then worrying about mounting it in a case later. The Hammond cases had been a bit of a theme through all my builds so I was going to make it fit in one of them, it was a bit of a tight fit but planning for it this time it worked.
A guy on another local forums here was getting a big batch of pcb's done through sparkfun.com and offering to anyone who wanted to add something to his order, so I took up the offer and after quite a few delays I eventually got a board back that was very well done, the sort of quality you see on any professionally built board.
I had ordered all the parts in and they arrived a few weeks before the board, so I decided I would spend that time on the casework first and just slot the dac in when I’m done and it would be all good, works in theory but something always goes wrong
Everything coming off the board was on pin headers, so everything could just plug into it (and that would have been one of the smartest things I had ever done in this project) so making the casework with leads coming off I even printed out a board on paper to size everything up and after squishing everything on the back panel it came out very nice (pics of casework and layout
here)
The board arrived a few weeks later, at the same time as the sample order of cs8416 that I had to get and I was over the moon with joy, I was hanging out to sit down and solder That baby up and jumped right into it and after only a few hours I had stuffed everything into it and plugged it all in and hoped it all worked first go, but of course it didn't
So in comes design fault 1, the digital power supply didn’t work at all. Tracing the voltages on the board, I ended up finding out that the 5V line was shorting to ground somewhere and on a board of
this complexity, it doesn’t just show itself too easily. so that lead me to start removing components off the board one at a time until I could find it and after about 3/4 of the digital side components were removed I came across this -
in designing of the board I got heaps of errors (~100) come up in my placement of vias, pads and stuff because some overlapped and in doing so I didn’t check the errors that seriously and overlooked that simple mistake. So out came a screwdriver and I gouged into the board cutting the pad off that trace and there went my short.
Yet powering it up again, I was getting very weird voltages out of the regulators now, as the psu goes, the AC comes in, through a bridge rectifier out comes ~10-12v and as the reg02s had a max limit of 10, I pre regulated it down with an lm317 to 5.5v before them to take all the heat load off them (being only smd) which is well within their dropout voltage, only needing 5.1v in yet I was getting out 4.7 on the 5v and 4.1 on the 3.3v regs which was way out of spec for these regulators, which comes in as design fault 2.
not wanting to deal with them now, I just used some lm78**'s to get a temporary psu going for them as I want to at least get some sound out of them and deal with it later. when I did come back to them, after staring at the board for ages and comparing to the guzzler dac I had successfully working with these chips I noticed I have the Vin and Vout the wrong way *slaps head* so the only thing to do was to swap them back, with some very classy soldering I bent up the pins off the board, attached leads from them to the rightful locations, powered it up and I got the right voltages at last.
Back to getting the dac to play, I set it up and plugged in an optical source off a DVD player, and had a listen, and got an earful of white noise and static
but one small light of hope was that it was from the digital input as pausing it on the player would stop the noise so I concluded that it was making it all the way through just that it was getting muddled up along the way.
Putting it down to the df1704 digital filter and processor chip to at fault here, as there was nothing on the pcm1704's you could do to upset the input signals and having triple checked the cs8416 over and over for the right settings on its pins I gently lifted some pins on it (which being a ssop smd part is not something I would ever recommend to anyone) and still didn’t get anything out of it, and this time instead of static, came nothing at all
I replaced the df1704 thinking that I had killed it, I put a new one in and pre bent my pins I didn’t want connected this time and soldered it on (and after lifting some tracks had to do some repair routing to some of them as well) I had another go, but was still left with nothing.
By this point I was completely out of ideas and the dac sat there for a few months, when out doing stuff with my mother, we stopped by at one of her friends places as she needed to do some stuff there, and the husband of my mothers friend was a retired telecom engineer who had some very very cool gear (he was into digital TV, had 3 satellite dishes in the yard ranging from one around 1 meter in diameter up to this massive 3.5meter one where he could pick up all the free to air stations in Australia, with a separate broadcast for each state [having 7 here in aus] so missing a show in Melbourne, could pick it up 2 hours later when it came to air in Perth, as well as getting channels from Europe where they were really into watching the cycling from there and also NASA's channel) but of all the things he had I had my eye on the oscilloscope, after talking about my somewhat failed dac, he invited me to bring it back one day to have a look.
Few weeks later I take it round and we spend the day going over it, and seeing what came up on the scope, and followed the digital signal from input right up to the pins of the pcm1704, and everything managed to clock right through up until there which made me feel quite a bit happier that it got that far, just left stumped that nothing came out of the dac chips.
I went over and over it again of the schematics at home and everything came to the conclusion that these dac chips were busted, but at nearly $100aud just to replace them I wasn’t in much of a hurry to jump on that.
After needing to order some parts from digikey for another project, I bit the bullet and ordered 2 more pcm1704's as well. so I replaced one of them, turned it on and with very little confidence of it working was absolutely blown away when I was getting sound out, that I was left sitting at my bench in shock, that’s where I got up to post this thread and tell the few guys on msn who had been following my progress of this. After calming down, I went and replaced the other channel, and that worked fine as well.
Now being on holidays from uni, back at my parents place, I didn’t bring home with all my headphone equipment that is still sitting back in my other place in the city, so the only source I got is the vcr/dvd combo using the coax out of that with my parents not liking the idea of me just taking that so I’ve only had a short listening session of it with a min amp and my etymotic er4's but from that I can defiantly say that it sounds dam sweet, full of detail and has a nice bass kick that is very tight and accurate, though from only 10 min of listening it cant say much as I feel a little biased that of course I can only say it sounds good after working on this project for 9 months now
Some better analysis will come later when I go back to the rest of my gear, so that will have to just wait for now.
Next is to fix the USB receiver on it, where I managed to mix some pins up on that as well and the hack job in getting that to work was unsuccessful in having the computer recognise that but with still a few chips I can play with that later. Also the input selection isn’t working with the flip flops not behaving nicely so using a real binary counter IC on Some proto board to tap into the input selection is probably in order which I should have used a counter IC instead of flip flops on there in the first place.
A few have been asking about wanting to build a dac project on here, so maybe this will be a start for that with a group buy maybe in the future for the pcb's but there are many changes needing to be made to this before I would build another one, but that will be for another day.
Hope you all enjoyed reading all this, it has been quite an experience for me.
![DaKi][er](https://cdn.head-fi.org/avatars/m/6/6700.jpg?1273083357){kind=avatar}
