Problem using Sound Card on Roboard...
Sound Card
by bluecat
Thu Aug 26, 2010 2:20 pm
I will be interested in one of your sound card boards, if it will materialize 
by PaulL
Sat Sep 04, 2010 6:59 pm
Bluecat- You'll be first on the list IF it does materialize and if you're up for the cost, but I don't want to start "pre-orders" for what is definitely new territory for me.
Simply put, it's not exactly easy (or cheap) to build your own PCI sound card. That said, it's the only conclusion I can come to that will solve this problem, and I've proven that it will, using the IM380A and a "standard PCI" sound card.
If someone else produced one, I'd buy it in a heartbeat.
On to progress:
I have received a small handful of chips, enough CS4281 chips for 7 boards. The chips SEEM new, but I won't know if they work until I either build the board or swap the chips into a sound card that uses these. They could be rejects or frauds, but I'm hoping they're just leftovers. If the controllers work, I should be OK. If the CS4297 codecs are bad, there are others out there that should work. These chips are old enough that XP has built-in drivers for them, so no trouble there if I can get the board right.
I have most of the schematic design work done using the CS4281 audio controller and CS4297 audio codec. I even have a sound card off Ebay that uses these exact chips on a 2 layer board, so I have a working example that I simply have to “shrink” into a mini-PCI form factor.
Cost for the board is going to be a major issue. Such low volume quantities for the capabilities needed will be expensive:
* 1mm board thickness (this seems to be the one that causes the price to go way up)
* 2 sides
* Solder Mask / silk screening
* Gold Plating of card edge
* Decent trace / space minimums
* Microvias would be nice, if not required.
In short, with what I've found so far, the PCB's themselves are NOT going to be cheap, around $60 to $80 each in small quantities. Price drops for more boards, but finding more chips is going to be difficult if not impossible. Prototypes will be MUCH higher ($250 to $500), I’m trying to skip even doing prototypes (not the best way to go, but having a proven design in hand, I have more confidence in success out the gate).
I even checked into building my own (I've done this in the past, was looking for better methods), but the quantity of vias required, gold plating, etc are too much to reasonably do myself.
I need to keep looking, but I doubt it’s going to get much cheaper. Right now, I’m leaning towards using Sunstone, they seem to have the best prices / capabilities I can find.
Things to do from here:
* Find the best capability / price board shop for low volume quantities and prototypes.
* Finish up the schematics for the board. This includes adding an audio amp and whatever else makes the most sense to add for the PCB space.
* Generate a Bill of Materials, determine device packages for board design.
* Buy a hot-air reflow station (like the one on Sparkfun's site).
* Lay out the board, get boards made, populate, and test.
Another option would be to build a MiniPCI prototyping board that is less custom and more likely to sell in volume, and build the sound cards as .062” thick add-ons for those boards. My problem here is choosing a connector that would be appropriate.
In Summary:
I have some chips, I have a working example of exactly what I’m trying to do in a full-sized PCI card using those exact same chips, and I have all the datasheets / reference design / documentation needed to answer any and all questions for this project.
All I need now is money and time… Time first, then money.
Take Care,
Paul
Simply put, it's not exactly easy (or cheap) to build your own PCI sound card.
That said, it's the only conclusion I can come to that will solve this problem, and I've proven that it will, using the IM380A and a "standard PCI" sound card.
If someone else produced one, I'd buy it in a heartbeat.
On to progress:
I have received a small handful of chips, enough CS4281 chips for 7 boards. The chips SEEM new, but I won't know if they work until I either build the board or swap the chips into a sound card that uses these. They could be rejects or frauds, but I'm hoping they're just leftovers. If the controllers work, I should be OK. If the CS4297 codecs are bad, there are others out there that should work. These chips are old enough that XP has built-in drivers for them, so no trouble there if I can get the board right.
I have most of the schematic design work done using the CS4281 audio controller and CS4297 audio codec. I even have a sound card off Ebay that uses these exact chips on a 2 layer board, so I have a working example that I simply have to “shrink” into a mini-PCI form factor.
Cost for the board is going to be a major issue. Such low volume quantities for the capabilities needed will be expensive:
* 1mm board thickness (this seems to be the one that causes the price to go way up)
* 2 sides
* Solder Mask / silk screening
* Gold Plating of card edge
* Decent trace / space minimums
* Microvias would be nice, if not required.
In short, with what I've found so far, the PCB's themselves are NOT going to be cheap, around $60 to $80 each in small quantities. Price drops for more boards, but finding more chips is going to be difficult if not impossible. Prototypes will be MUCH higher ($250 to $500), I’m trying to skip even doing prototypes (not the best way to go, but having a proven design in hand, I have more confidence in success out the gate).
I even checked into building my own (I've done this in the past, was looking for better methods), but the quantity of vias required, gold plating, etc are too much to reasonably do myself.
I need to keep looking, but I doubt it’s going to get much cheaper. Right now, I’m leaning towards using Sunstone, they seem to have the best prices / capabilities I can find.
Things to do from here:
* Find the best capability / price board shop for low volume quantities and prototypes.
* Finish up the schematics for the board. This includes adding an audio amp and whatever else makes the most sense to add for the PCB space.
* Generate a Bill of Materials, determine device packages for board design.
* Buy a hot-air reflow station (like the one on Sparkfun's site).
* Lay out the board, get boards made, populate, and test.
Another option would be to build a MiniPCI prototyping board that is less custom and more likely to sell in volume, and build the sound cards as .062” thick add-ons for those boards. My problem here is choosing a connector that would be appropriate.
In Summary:
I have some chips, I have a working example of exactly what I’m trying to do in a full-sized PCI card using those exact same chips, and I have all the datasheets / reference design / documentation needed to answer any and all questions for this project.
All I need now is money and time… Time first, then money.
Take Care,
Paul
by PaulL
Fri Jan 14, 2011 12:43 pm
I'm thinking of taking a "shortcut" in building the sound card.
I need to add circuitry for microphone preamp, output amp, etc, but these things may need some tweaking, and a one-shot board run may be problematic.
SO, I was thinking of building the sound card w/ audio controller and codec chip, bringing signals out to some pads, and add a "breadboard"-style solder pad arrangement on it. This would allow prototyping WITH the sound card. I believe this to be the most risk-free (and easiest) method of implementation at this time.
That means that the preamp / amp selection is up to whomever, possibly with board space left for other small circuits.
I need to take another look at the size the card could be on the back of Roboard, and play around w/ laying out a proto area.
This would be a fast way to get the board made, as there isn't much work in laying out prototyping pads...
And it adds flexibility for preamps and amps, or whatever else.
Take Care,
Paul
I need to add circuitry for microphone preamp, output amp, etc, but these things may need some tweaking, and a one-shot board run may be problematic.
SO, I was thinking of building the sound card w/ audio controller and codec chip, bringing signals out to some pads, and add a "breadboard"-style solder pad arrangement on it. This would allow prototyping WITH the sound card. I believe this to be the most risk-free (and easiest) method of implementation at this time.
That means that the preamp / amp selection is up to whomever, possibly with board space left for other small circuits.
I need to take another look at the size the card could be on the back of Roboard, and play around w/ laying out a proto area.
This would be a fast way to get the board made, as there isn't much work in laying out prototyping pads...
And it adds flexibility for preamps and amps, or whatever else.
Take Care,
Paul