Wow, now I'm really excited to see this schematic. What are you going to use as a source for the digital signals? I'm still waiting to hear from my professor, but I should know by the end of the week what I can do.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Multichannel DAC
- Thread starter dunderchief
- Start date
- Status
- Not open for further replies.
Hey, so I just heard from my Uni, and basically they said no to everything. No PCB's, no deal with parts suppliers. So I guess that means we're on our own, which sucks. I'll start ordering samples now I suppose. How's the schematic coming. Any idea on PCB art? Still excited to see what comes out of this.
No worries, I'll be switching jobs soon, so I should have more time to get going on a prototype. I got my samples today. I can't believe TI sends stuff overnight mail. I'm fairly intimidated by the size of most of the components, they're fricking tiny. Any ideas on where to pick up a good surface mount soldering iron. Also, once you get this done, could someone in the US tell me where I can get PCB's printed, I've always had them done through the school, but they won't tell me where they send them.
The Solomon irons that MPJA sells are pretty decent on the cheap end. Someone here was also selling some very nice units a while back. Basically you just need a nice adjustable iron with an isolated tip, and one that has small tips available.
I can't say enough good things about advanced circuits for protos. www.4pcb.com
I can't say enough good things about advanced circuits for protos. www.4pcb.com
the point of this?
Maybe I'm missing something here, but I still can't figure out exactly what is being attempted here.
The only two applications I can see as being even remotely 'mainstream' for a multichannel DAC would be
a) digital crossovers
b) Home theater surround
(or both together, obviously)
Since I actually am 'using'/putting together a PC-based HT system using digital xovers, I think I understand this application.
The problem is that I still don't see where you're going to get either of the above signals in digital form, and until that is established I don't see how you can design a DAC. No affordable HT processor provides digital outputs, and I don't believe any of the typical off0the-shelf digital crossovers provide dig out either.
It seems to me that a computer soundcard is the only viable 'cheap' way to get multiple channels, and aside from the Lynx LS-AES I don't know of any that offer multiple spdif outputs. Hoontech/ST-Audio had one for their DSP-24, but I don't know that it's available anymore. Even then, each SPDIF input will have it's own master clock that you will NOT be able to assume are in sync, so you'll have to route/switch clock as well as signal lines internally.
I guess where I'm confused is why you don't look at something like taking a ST-Audio C-Port system or an M-Audio Delta 1010 (ie an 8 channel soundcard with an external DAC connected by a DB-25 cable) and either simply hack/tweak/upgrade the DAC section or else reverse-engineer the pinouts on the expansion bus if you want to go completely custom. This gives you a well-supported source of digital signals without having to worry about multiple clock source sync.
Maybe I'm missing something here, but I still can't figure out exactly what is being attempted here.
The only two applications I can see as being even remotely 'mainstream' for a multichannel DAC would be
a) digital crossovers
b) Home theater surround
(or both together, obviously)
Since I actually am 'using'/putting together a PC-based HT system using digital xovers, I think I understand this application.
The problem is that I still don't see where you're going to get either of the above signals in digital form, and until that is established I don't see how you can design a DAC. No affordable HT processor provides digital outputs, and I don't believe any of the typical off0the-shelf digital crossovers provide dig out either.
It seems to me that a computer soundcard is the only viable 'cheap' way to get multiple channels, and aside from the Lynx LS-AES I don't know of any that offer multiple spdif outputs. Hoontech/ST-Audio had one for their DSP-24, but I don't know that it's available anymore. Even then, each SPDIF input will have it's own master clock that you will NOT be able to assume are in sync, so you'll have to route/switch clock as well as signal lines internally.
I guess where I'm confused is why you don't look at something like taking a ST-Audio C-Port system or an M-Audio Delta 1010 (ie an 8 channel soundcard with an external DAC connected by a DB-25 cable) and either simply hack/tweak/upgrade the DAC section or else reverse-engineer the pinouts on the expansion bus if you want to go completely custom. This gives you a well-supported source of digital signals without having to worry about multiple clock source sync.
Re: the point of this?
Soundcards with multichannel SPDIF outputs are not much use when you want a multichannel DAC.
There are a number uses for multichannel dacs and there has been at least one previous request in this forum for such a device, the poster wanted to connect an analogue mixer. And for the price of a top end soundcard one can easily put 24 channels in a 1U case by going the DIY route.
ray.
dwk123 said:Maybe I'm missing something here, but I still can't figure out exactly what is being attempted here.
The only two applications I can see as being even remotely 'mainstream' for a multichannel DAC would be
a) digital crossovers
b) Home theater surround
(or both together, obviously)
Since I actually am 'using'/putting together a PC-based HT system using digital xovers, I think I understand this application.
The problem is that I still don't see where you're going to get either of the above signals in digital form, and until that is established I don't see how you can design a DAC. No affordable HT processor provides digital outputs, and I don't believe any of the typical off0the-shelf digital crossovers provide dig out either.
It seems to me that a computer soundcard is the only viable 'cheap' way to get multiple channels, and aside from the Lynx LS-AES I don't know of any that offer multiple spdif outputs. Hoontech/ST-Audio had one for their DSP-24, but I don't know that it's available anymore. Even then, each SPDIF input will have it's own master clock that you will NOT be able to assume are in sync, so you'll have to route/switch clock as well as signal lines internally.
I guess where I'm confused is why you don't look at something like taking a ST-Audio C-Port system or an M-Audio Delta 1010 (ie an 8 channel soundcard with an external DAC connected by a DB-25 cable) and either simply hack/tweak/upgrade the DAC section or else reverse-engineer the pinouts on the expansion bus if you want to go completely custom. This gives you a well-supported source of digital signals without having to worry about multiple clock source sync.
Soundcards with multichannel SPDIF outputs are not much use when you want a multichannel DAC.
There are a number uses for multichannel dacs and there has been at least one previous request in this forum for such a device, the poster wanted to connect an analogue mixer. And for the price of a top end soundcard one can easily put 24 channels in a 1U case by going the DIY route.
ray.
I will also be using this thing in a PC based situation. My thought at the moment was to use multiple cheaper soundcards with SPDIF output, rather than one really expensive one with multiple outputs. I suppose this means even more problems with Clock sync, and to be honest I hadn't considered it. I'm sure DJ would be open to suggestions on a solution to this, or perhaps he already has a solution. I have also thought about going to way of the Delta 410 sound card especially considering it's use of the external breakout box.
dunderchief said:I will also be using this thing in a PC based situation. My thought at the moment was to use multiple cheaper soundcards with SPDIF output, rather than one really expensive one with multiple outputs. I suppose this means even more problems with Clock sync, and to be honest I hadn't considered it. I'm sure DJ would be open to suggestions on a solution to this, or perhaps he already has a solution. I have also thought about going to way of the Delta 410 sound card especially considering it's use of the external breakout box.
The cheaper soundcards, aside from not having external sync options, tend not to have the buffering that lessens the effects of latency, a situation that can only be made worse by multiple soundcards running at once. For multiple I/O, analogue or digital, you are better off with something like a Delta 10/10 or one of the similar level cards.
ray.
Re: Re: the point of this?
Not to say that DIY has to stick to standards, but since I'm assuming that hooking up 4 separate CD transports isn't the goal at some point you have to figure out where you're getting your signals.
To be explicit, this is why I suggested looking at Envy24 based soundcards, since the Envy24 chip runs 4 stereo I2S data lines out of the chip directly along with one I2S clock line shared by all chanels. Since the DACS are external, Hoontech DSP24 expansion connector *probably* has these raw signals directly available which would be ideal for a hooking up a simple DIY dac box.
I'll agree with this. I mentioned it only because aside from studio links like ADAT, it's the only standard way to easily get multiple digital channels.rfbrw said:
Soundcards with multichannel SPDIF outputs are not much use when you want a multichannel DAC.
Not to say that DIY has to stick to standards, but since I'm assuming that hooking up 4 separate CD transports isn't the goal at some point you have to figure out where you're getting your signals.
To be explicit, this is why I suggested looking at Envy24 based soundcards, since the Envy24 chip runs 4 stereo I2S data lines out of the chip directly along with one I2S clock line shared by all chanels. Since the DACS are external, Hoontech DSP24 expansion connector *probably* has these raw signals directly available which would be ideal for a hooking up a simple DIY dac box.
There are a number uses for multichannel dacs and there has been at least one previous request in this forum for such a device, the poster wanted to connect an analogue mixer.
I wasn't questioning that a multichannel DAC would be useful - I know there are a bunch of ways to use them. My question was along the lines of:
- what would a DIY-er (who can't afford a 1010 or a Lynx) likely be doing with multiple channels such that the output of an M-Audio Revo wasn't good enough?
And for the price of a top end soundcard one can easily put 24 channels in a 1U case by going the DIY route.
ray.
This I'll disagree with somewhat.
First, it depends on what you mean by 'top end' - if you mean a Lynx or a new RME HDSP w/ expansion boards, then maybe. If you mean a Delta 1010 then no way.
Second, I think you're underestimating the cost of DIY a bit, particularly how it will multiply as the channel count goes up. Good results aren't cheap, even if you DIY. Heck - folks are spending $50-$75/channel on MODS to things like a DI/O.
Third, it depends on how you're getting the signals into the DAC - if you need a SPDIF receiver and an ASRC chip for each stereo pair, you're looking at $20-25 per channel before you even get the DAC chips or any other parts.
Once again, my confusion is that anyone 'needing' 24 channels is clearly serious enough that spending $1000 on studio kit shouldn't be a big deal (the software to deal with 24 channels will run $1000's+ anyway, unless you rip it off). OTOH, for the 'starving student' DIY-er the Revo is actually a surprisingly decent card. Since I've yet to see a 'decent' DIY DAC project check in much less than $100 in parts for stereo, I'm certain you won't get 8 channels of anywhere close to the same quality for the under-$100 price that the Revo goes for.
So, given all this, I assume that the target quality is better than the Revo, which probably means at least $50/channel in parts, and could 'easily' be $75-100 with the typical DIY creep. But at this point, you start to hit the price point of the Delta 1010 and other studio solutions, which is why I asked whether the idea of a 'mod' approach had been broached as an alternative to building from scratch.
So, if the intent was a 'cost no object' DIY project that provided 8 channels of SERIOUS d/a, then I'd completely understand since it would be something that isn't really availble off-the-shelf. However, that doesn't fit with the 'low budget' tone of the discussions, and given the relatively remarkable quality available in some $100-$150 multichannel sound cards, it just wasn't clear what the motivation was.
Of course, the answer might just be 'because it would be fun to do", which is really all the motivation you need.....
To be honest I had no idea what I was getting into when I started this. The whole idea behind me starting the thread was that I was building a competition level Car audio system. I also wanted to incorporate a Pc into this system. Car PC systems are notorious for having horrible noise problems relating to the internal sound card. At the beginning I new basically nothing about the SPDIF format, or for that matter DAC design. My basic thought was, wow, what if I could do an external 8 channel DAC and then use digital FIR filters for a 4 way active crossover. So in the end it comes down to "wow, that might be fun". Once again, I had no idea what I was getting into. I'm still gonna build one of these. At this point I'm not sure if it will exactly fit into my plans, but hey, it should be fun.
Re: Re: Re: the point of this?
As soon as multiple channels and external sync enter the scene the price of the soundcard increases rapidly. I would consider the Delta 1010 as a starting point if I was after multichannel I/O and external sync. From there it all depends on the depth of your pockets.
As to 8 or more channels in 1U for the price of a Delta 1010, I think, as a diy project, it is well within reach. Having done something not dissimilar in the past , I am certain I could hit a sub Delta 1010 budget target without too much penny pinching.
ray.
dwk123 said:
This I'll disagree with somewhat.
First, it depends on what you mean by 'top end' - if you mean a Lynx or a new RME HDSP w/ expansion boards, then maybe. If you mean a Delta 1010 then no way.
Second, I think you're underestimating the cost of DIY a bit, particularly how it will multiply as the channel count goes up. Good results aren't cheap, even if you DIY. Heck - folks are spending $50-$75/channel on MODS to things like a DI/O.
Third, it depends on how you're getting the signals into the DAC - if you need a SPDIF receiver and an ASRC chip for each stereo pair, you're looking at $20-25 per channel before you even get the DAC chips or any other parts.
Once again, my confusion is that anyone 'needing' 24 channels is clearly serious enough that spending $1000 on studio kit shouldn't be a big deal (the software to deal with 24 channels will run $1000's+ anyway, unless you rip it off). OTOH, for the 'starving student' DIY-er the Revo is actually a surprisingly decent card. Since I've yet to see a 'decent' DIY DAC project check in much less than $100 in parts for stereo, I'm certain you won't get 8 channels of anywhere close to the same quality for the under-$100 price that the Revo goes for.
So, given all this, I assume that the target quality is better than the Revo, which probably means at least $50/channel in parts, and could 'easily' be $75-100 with the typical DIY creep. But at this point, you start to hit the price point of the Delta 1010 and other studio solutions, which is why I asked whether the idea of a 'mod' approach had been broached as an alternative to building from scratch.
So, if the intent was a 'cost no object' DIY project that provided 8 channels of SERIOUS d/a, then I'd completely understand since it would be something that isn't really availble off-the-shelf. However, that doesn't fit with the 'low budget' tone of the discussions, and given the relatively remarkable quality available in some $100-$150 multichannel sound cards, it just wasn't clear what the motivation was.
Of course, the answer might just be 'because it would be fun to do", which is really all the motivation you need.....
As soon as multiple channels and external sync enter the scene the price of the soundcard increases rapidly. I would consider the Delta 1010 as a starting point if I was after multichannel I/O and external sync. From there it all depends on the depth of your pockets.
As to 8 or more channels in 1U for the price of a Delta 1010, I think, as a diy project, it is well within reach. Having done something not dissimilar in the past , I am certain I could hit a sub Delta 1010 budget target without too much penny pinching.
ray.
Re: Re: Re: Re: the point of this?
I'm in complete agreement, which is how I ended up with the 1010. Still, it's 'good' rather than 'state of the art',
8 channels I'd agree with - I was reacting to the 24 number - you'll be pinching pennies to get 24 channels under $1k or so, I would think. I agree completely that targeting 8 channels with say a $600 budget should result in something that is a solid step up from the Delta, which is my entire reason for posting to this thread. I'd love to see a 'Delta 1010 upgrade' movement afoot, but I doubt that will happen.
Since the DSP24 is basically the same architecture as the 1010 and is cheaper AND available without the external DAC unit for ~$200 or so, I think is is a pretty reasonable starting point for this type of project. If it was a direction that a few people wanted to try together, I might be interested enough to pick one up to pitch in.
rfbrw said:
As soon as multiple channels and external sync enter the scene the price of the soundcard increases rapidly. I would consider the Delta 1010 as a starting point if I was after multichannel I/O and external sync. From there it all depends on the depth of your pockets.
I'm in complete agreement, which is how I ended up with the 1010. Still, it's 'good' rather than 'state of the art',
As to 8 or more channels in 1U for the price of a Delta 1010, I think, as a diy project, it is well within reach. Having done something not dissimilar in the past , I am certain I could hit a sub Delta 1010 budget target without too much penny pinching.
ray.
8 channels I'd agree with - I was reacting to the 24 number - you'll be pinching pennies to get 24 channels under $1k or so, I would think. I agree completely that targeting 8 channels with say a $600 budget should result in something that is a solid step up from the Delta, which is my entire reason for posting to this thread. I'd love to see a 'Delta 1010 upgrade' movement afoot, but I doubt that will happen.
Since the DSP24 is basically the same architecture as the 1010 and is cheaper AND available without the external DAC unit for ~$200 or so, I think is is a pretty reasonable starting point for this type of project. If it was a direction that a few people wanted to try together, I might be interested enough to pick one up to pitch in.
Re: Re: Re: Re: Re: the point of this?
24 channels for $600. Easy. TI have some nice 8ch dac chips.
ray.
dwk123 said:
8 channels I'd agree with - I was reacting to the 24 number - you'll be pinching pennies to get 24 channels under $1k or so, I would think.
24 channels for $600. Easy. TI have some nice 8ch dac chips.
ray.
I am kind of thinking about a similar project. The DAC part should be OK; there are plenty of very good integrated DAC around. But the PCI computer connection part is my concern, assuming the idea here is to use the computer to control, equalize, and modify the sound. How easy it is to design a PCI sound card that has 16 channels output? Do one uses one PCI controller and four codecs? And do you need to write custom drivers for such a card?
Good news is my final test is in a few hours and after that I'll have more time for this project. The basic schematic for each channel is done, so I just need to add a selector/multiplexer circuit. That is if I still understand correctly what you want to do with it Steve. Perhaps you should put up a block diagram of the system this project is destined for so I can get a better idea of what else I need to include.
A PC in the car will add a lot of noise to an already noisy setup, I have a better idea if you're interested. Still basically a PC, but much less noisy.
I'll be back.
DJ
A PC in the car will add a lot of noise to an already noisy setup, I have a better idea if you're interested. Still basically a PC, but much less noisy.
I'll be back.
DJ
Yeah, the whole PC thing is basically a side project. The overall project I was looking at was basically software crossovers using FIR filters. That was the idea, so this seemed like what I was looking for. I'm definately interested in what you have to say about a PC in a car environment for audio. Perhaps a new thread. Lemme know, I'm still very interested in this project although, to be honest, I'm not exactly the most qualified person to be building your prototype.
Interest might have waned on this topic a bit, but I recently came across something that might be of interest.
I just got a Panasonic XR25 receiver, which uses digital amp sections based on the TI Equibit chipset. You can think of these as high-power DAC chips that drive the speakers directly - an integrated DAC and power amp in one stage.
Based on a quick listen, these are VERY VERY good when driven by an unprocessed spdif signal. Since the Equibit modules use I2S internally, if you can come up with a set of multichannel I2S signals, it should be possible to feed the amp sections more or less directly.
Since the XR24 only costs $300 and features 6x100w output, this could be an extremely low-cost way to get a VERY high quality multichannel digital system.
I just got a Panasonic XR25 receiver, which uses digital amp sections based on the TI Equibit chipset. You can think of these as high-power DAC chips that drive the speakers directly - an integrated DAC and power amp in one stage.
Based on a quick listen, these are VERY VERY good when driven by an unprocessed spdif signal. Since the Equibit modules use I2S internally, if you can come up with a set of multichannel I2S signals, it should be possible to feed the amp sections more or less directly.
Since the XR24 only costs $300 and features 6x100w output, this could be an extremely low-cost way to get a VERY high quality multichannel digital system.
Wow, not exactly what I was going for, but definately sounds like an interesting product. Could you point me to a product reference. I'll do some poking around myself. This sounds a lot like DJ's digital speaker project. I'm not sure what DJ is up to these days, but I'm still very interested to see what he comes up with.
- Status
- Not open for further replies.