diy spdif ad/da circuit???

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mfdu

Well-known member
Joined
May 25, 2005
Messages
153
Location
Melbourne, Australia
hiya chaps.

i'm looking at building a GSSL and would like to build in an AD/DA circuit for use with my digi002.

anything out there? any pcb's? anyone tackled something like this before? i'm sure you have - i'd love to hear about it and see the specs . . .

chris.
 
IMO AD/DA circuits are not really DIY stuff - if you disregard the pre-built evaluation boards from the chip manufacturer.

MCS has done some experiments - but I don't know what he ended up with.

Try a search..

Jakob E.
 
gyraf

point taken. i though you may be interested in the concept - GSSL with integrated ADA.

but i think the cheapest/sanest/best option may be the b%*#$&#er 24/96SRC between GSSL and protools spdif input.

thanks.

chris.
 
I've had the notion that some of the circuitry in the ubiquitous sigma-delta stuff could be profitably done with some discretes. Unfortunately I don't know of any manufacturer who gives access to the decimation hardware, although probably some FPGA hardware could be programmed to do the job.

As far as the SPDIF stuff, that I would not mess with ;-).
 
I don't consider AD stuff to be out of the realm of DIY... sure it's a lot of SMD devices, but I've been doing a ton of SMD DIY!

CPLDs are within my abilities, as are smaller FPGAs. BGA is out of the question at the moment, and I'm not 100% confident in my ability to design a three-voltage supply for fpga, but I just designed my first switcher for another project, and so I may be tackling a PQ208 fpga sometime in the not-too-distant future. I do have experience with VHDL.

AD is something I really want to get to, and soon. I have been mostly thinking I want to stick with ADAT, since I have a hammerfall, but SPDIF might be a better starting place.

I would love to make a digital out option for Green Pres (or any other DIY mic pre), maybe have a little SPDIF module for each one, then a DIY SPDIF->ADAT aggregator. Anyone up for some group design?

I have SMD PCBs under control:
http://gromit.usedforcomparison.com/~tmbg/hybrid_boards.jpg
http://gromit.usedforcomparison.com/~tmbg/arc.jpg
 
I am envious of those that can do the SMD stuff. I am not even sure what 90% of the abreviations stand for and could probably not be able to focus on the pins long enough to solder the damned devices (assuming I could hold the iron steady).

Discrete silicon is as small as I can manage and I'm lucky to do that. I prefer vacuum state not only for the sound, but because it contains parts big enough to see.

Don't get me wrong, I think digital interface and the like certainly have become a "must have", but for the vast majority of DIY projects, this is simply beyond the ability to do well. For those few of you who can do it, please have at it, but every once in awhile explain what the devil you are doing for the fogeys like me, or maybe a link to some site that can explain it.

Now, back to my rocker.
 
as a programmer by trade, digital is a lot of the time more transparent to me. I'm getting the hang of analog in drips and drabs, but I'm definitely envious of the folks like PRR and bcarso who can look at a scheme and say "well duh, even order harmonics are cancelled due to collector current contributing to local feedback..." and then I go crosseyed :/

Digital is more concrete to me, just hook the crap up and deal with the details in software :)
 
Heh. My father predicted that ICs would never be practical because they were too small to work with.

He eventually added a 14 pin DIP to a product, the popular and now discontinued CA3086 transistor array, not because he needed the size reduction or matching properties, but because it made the product look more modern :green:
 
[quote author="thomasholley"]I am envious of those that can do the SMD stuff. I am not even sure what 90% of the abreviations stand for and could probably not be able to focus on the pins long enough to solder the damned devices (assuming I could hold the iron steady).
...snip...
For those few of you who can do it, please have at it, but every once in awhile explain what the devil you are doing for the fogeys like me, or maybe a link to some site that can explain it.[/quote]
I'm new to the SMD game so am probably good from a starter's point of view. I have no idea if the way I'm going about it is the "correct" way or not but can only tell you that it works, seems reliable and looks ok.

The good news is that you probably don't need any specialized equipment to get going. I used a Dick Smith $49 soldering station with a 1.6mm tip to do my first boards. The one thing that I would suggest though is to get one of those magnifying lamps with a circular flourescent tube.

The other thing that is handy but you probably already have around is a pair of tweezers, preferably ones that have a locking action either by latching or by reverse action (release and they close). You want them to hold the component steady while soldering. I've seen suggestions that you can hold them in place with solder paste but I don't use that method. I've also seen it suggested that you need to use eutectic Sn 63% solder (solder that has no 'plastic' state) but I'm using regular 60/40 and it seems to work fine.

Keep the components on the tape and in marked bags. You don't want to have to try reading the markings, especially on caps. Tin one pad lightly, grab your component, line it up and reheat the tinned side to tack it down. Go to the opposite side and do a proper soldering job. Then revisit the first side and give it a nice joint.

I laid out a Wackie "VU" circuit for some projects and used 0805 sized components throughout. I couldn't get certain values so some parts ended up getting doubled up and others put in series but even with a home-etch it worked out pretty well. The only thing I didn't use SMDs for were the LEDs and the electrolytic cap.

Here's a pic of the component side with the tip of the iron for scale:
smd_vu.jpg

*edited for typos
 
just a thought... to keep your SMD usage as low as possible, but at the same time keep your board size down, why not vertically mount normal resistors?

I know they are more fragile, but for most of the applications we have, the device will sit in a rack, and be moved a handful of times in it's lifetime.

Also, just as some food for thought - check out what seattle robotics did with a bog standard electric oven.
http://www.seattlerobotics.org/encoder/200006/oven_art.htm

just get some solder paste and squidge it over the contact pads (in one stroke - no need to be a little blob for each pin) - pop it in the oven - and voila - job done. I will be putting this to the test soon - so I'll have feedback in the next few weeks.

Cheers

R
 
speaking of vertical resistors to save board space, when's the last time you looked at my ARP boards? :D

http://diy.usedforcomparison.com/arp/

9 resistors, three caps, two transistors, and 6 pins worth of header in 0.9x0.9", but it still could have been much smaller using SMD, mostly because you get both sides of the board.
 
Right, and in all those discrete opamps around here (1*7*3*1, 2*5*2*0) I don't believe there are resistors not mounted vertically.

Regards,

Peter
 
http://hepso.dna.fi/misc/pcm2902/PCM2902_usb_dac.html
this proyect is a ad/da usb I don`t know if the conversor are good but you can use external converters.
the pcm2902 is capable of s-pdif in/out too.
I remember an old post about build a sound card this chip could do all more easy . and now usb sound card are popular .
 
[quote author="gyraf"]MCS has done some experiments - but I don't know what he ended up with.[/quote]
I have a working S/PDIF or AES/EBU receiver, 2 ADCs, a DAC and a clock circuit. I also have an AES/EBU transmitter and an ADAT receiver/transmitter, but they are not fully tested yet - updates later this week...

And then I have my partially completed HDRec project... :)

Best regards,

Mikkel C. Simonsen
 
Wow, almost the exact same thing was published in Nuts&Volts.
It is probably as good as it practiacally gets with 16bits at 44k. :thumb:
 
Back
Top