Hello Peufeu,
I have been following your thread on DIY Audio, reading up on your posts on your Extremist DACs. I am a total novice to Pro-Audio but am wanting to dabble with DACs. But I could always leverage my analog electronics experience to further my Audio pursuits. Having said that i have a question on your DAC and would appreciate it if you could respond at your own convenience.
1. Is it possible for me to review your DAC schematics at this time?
Just a FYI, I currently have a Spartan3E starter kit board that I could use for the Ethernet interface to the DAC.
Looking forward to hear from you.Thanks in advance,
Ananth
I have been following your thread on DIY Audio, reading up on your posts on your Extremist DACs. I am a total novice to Pro-Audio but am wanting to dabble with DACs. But I could always leverage my analog electronics experience to further my Audio pursuits. Having said that i have a question on your DAC and would appreciate it if you could respond at your own convenience.
1. Is it possible for me to review your DAC schematics at this time?
Just a FYI, I currently have a Spartan3E starter kit board that I could use for the Ethernet interface to the DAC.
Looking forward to hear from you.Thanks in advance,
Ananth
OK, I'm not dead 
I've been to Japan, which was fantastic, really exotic and different, I'll definitely go back there some day ! (check out the photos : http://photo.peufeu.com)
Once back here, I had little time for electronics, many other things to do, but now that the renovations in the flat are well under way it's back to normal.
> what about a TOSLINK input? My old PD-S703 has one of these optical outputs, which I suspect is accurate (I have read your tests about comparing CD output to your original CD-R, by the way, but I don't have any special sound equipment on my PC).
Actually a SPDIF input would be really easy to add (optical or not), however, you get the usual clock synchronization issues, which means the SPDIF source should be slaved to the DAC, or some nifty software controlled VCXO PLL should be used in the DAC, or the other popular option, the ASRC.
The purpose of this project is to use a computer as a source, for several reasons, but mostly for convenience (playlists are cool, it's compatible with any sample bit width and sample rate, etc) and to reap the benefits of easy digital XO, but without the drawbacks of using a computer : cooling fans in the audio room, sub-optimal soundcard quality, etc.
However, I may add a SPDIF input for convenience, to plug in my DVD player, for instance. But then, the computer can play DVDs better than the DVD player (it has built-in progressive scan, deblocking, etc). And the computer can play HD, too. Or I could plug my ADSL TV into the spdif input, but then, the computer can also play the ADSL TV (with VLC on Linux).
I will definitely add a very high quality analog input, probably with the top of the line AKM chip, to rip Vinyl to harddisk, and also to process Vinyl or FM radio with the computer's digital crossover.
So, basically, our requirements are not the same...
> I would prefer a standalone box, that doesn't need an extra computer for working. The Ethernet idea is very nifty, I might add.
> I see that the recent wave of singe-board computers are sporting fast ethernet and rather fast CPUs (e.g. 600+ MHz ARM chips). Wouldn't these systems be able to do all the processing in software, with some assembler libraries thrown in?
Yeah those systems are neat. However, to do active crossover, you need big number crunching power. Some of those new embedded CPUs have good DSP-like capabilities, but only at 16 bit, ie. not for quality audio, but excellent for realtime compression of cellphone voice traffic, for instance. You would need a custom DSP solution, for good sound quality you'll need a 32 bit DSP like the SHARC or Blackfin, and the skill to program it. It should also be doable in a FPGA, but more complex.
I view the PC in this application as a convenient way to get storage, crunching power, and a nice user interface, for a very low price. If the PC is too slow to process 24-96 audio in the digital crossover, upgrading the PC is simple, whereas designing a custom DSP solution with the same price and ease of upgrading would only be possible on a commercial scale. For a small scale project, it just wouldn't provide enough ROI.
> Is it possible for me to review your DAC schematics at this time?
Right now I have the schematics for my old DAC (see my website, http://audio.peufeu.com), but this one uses TDA1545, which is really obsolete. I had the chance to compare it with a DIY DAC using AD1955 and the latter one sounded a lot better. It seems the 20 years of R&D between TDA1545 and AD1955 did produce some results !
I will publish the schematics for the FPGA module soon for review, but these don't include the DAC... one step at a time
> Could you please post a block diagram of what you are doing?
Alll the docs for the current FPGA implementation are on my website (dig in the menu, Projects, etc). The FPGA module I'm designing now is quite similar, here are the differences :
- FPGA : Spartan 3-1000 -> now Spartan 3E-1200 or 1600
- LAN : LAN91c111 -> now LAN9117, easier to use chip
- SDRAM : 16 bit -> now 32 bit
Architecture-wise, the Suzaku has a shared bus between the SDRAM and the LAN chip, my module will use two different busses, which will allow a lot more data throughput and power. It should never choke on data.
The original idea was to use a PQ208 FPGA. I have a board routed with this one, but it has several problems :
- The PQ208 package sucks for EMI and parasitics
- There isn't that many IO pins free to use once everything is connected (ie. I can plug some DACs, but the LCD, volume control, etc... no).
- Only the Spartan 3E-500 is available in this package, and I would like to use a fatter FPGA to implement sample rate conversion in the FPGA (since that would use a LOT of CPU in the PC)
So, I wanted to use a bigger FPGA in a BGA package. This is not easy for an amateur, because using a 1mm-pitch BGA means you need pretty tight tolerances on the PCB track/space and vias, which means it's quite expensive. This assembling and PCB problem is almost solved thanks to a benefactor who would rather remain anonymous since he doesn't want to receive tons of requests, lol.
Anyway, I had to remake the design using Altium, which can handle BGA correctly.
I needed a few tries to get into Altium, this software definitely has a
high "bump" for first users. I tried at least 3 times and every time I was
put off by the complexity of the software. But last sunday I tried it
again, and this time I got it. And once past the initial bump, I'm
starting to like this software. It still sucks, but lots less than Eagle.
Anyway, schematic capture is done in Altium, and I have entered all the
models for the parts not in the library. The only thing remaining is
proofing the schematic again and obviously... the PCB layout.
I will post the schematics for review as soon as I have something "showable".
Have a nice day !
I've been to Japan, which was fantastic, really exotic and different, I'll definitely go back there some day ! (check out the photos : http://photo.peufeu.com)
Once back here, I had little time for electronics, many other things to do, but now that the renovations in the flat are well under way it's back to normal.
> what about a TOSLINK input? My old PD-S703 has one of these optical outputs, which I suspect is accurate (I have read your tests about comparing CD output to your original CD-R, by the way, but I don't have any special sound equipment on my PC).
Actually a SPDIF input would be really easy to add (optical or not), however, you get the usual clock synchronization issues, which means the SPDIF source should be slaved to the DAC, or some nifty software controlled VCXO PLL should be used in the DAC, or the other popular option, the ASRC.
The purpose of this project is to use a computer as a source, for several reasons, but mostly for convenience (playlists are cool, it's compatible with any sample bit width and sample rate, etc) and to reap the benefits of easy digital XO, but without the drawbacks of using a computer : cooling fans in the audio room, sub-optimal soundcard quality, etc.
However, I may add a SPDIF input for convenience, to plug in my DVD player, for instance. But then, the computer can play DVDs better than the DVD player (it has built-in progressive scan, deblocking, etc). And the computer can play HD, too. Or I could plug my ADSL TV into the spdif input, but then, the computer can also play the ADSL TV (with VLC on Linux).
I will definitely add a very high quality analog input, probably with the top of the line AKM chip, to rip Vinyl to harddisk, and also to process Vinyl or FM radio with the computer's digital crossover.
So, basically, our requirements are not the same...
> I would prefer a standalone box, that doesn't need an extra computer for working. The Ethernet idea is very nifty, I might add.
> I see that the recent wave of singe-board computers are sporting fast ethernet and rather fast CPUs (e.g. 600+ MHz ARM chips). Wouldn't these systems be able to do all the processing in software, with some assembler libraries thrown in?
Yeah those systems are neat. However, to do active crossover, you need big number crunching power. Some of those new embedded CPUs have good DSP-like capabilities, but only at 16 bit, ie. not for quality audio, but excellent for realtime compression of cellphone voice traffic, for instance. You would need a custom DSP solution, for good sound quality you'll need a 32 bit DSP like the SHARC or Blackfin, and the skill to program it. It should also be doable in a FPGA, but more complex.
I view the PC in this application as a convenient way to get storage, crunching power, and a nice user interface, for a very low price. If the PC is too slow to process 24-96 audio in the digital crossover, upgrading the PC is simple, whereas designing a custom DSP solution with the same price and ease of upgrading would only be possible on a commercial scale. For a small scale project, it just wouldn't provide enough ROI.
> Is it possible for me to review your DAC schematics at this time?
Right now I have the schematics for my old DAC (see my website, http://audio.peufeu.com), but this one uses TDA1545, which is really obsolete. I had the chance to compare it with a DIY DAC using AD1955 and the latter one sounded a lot better. It seems the 20 years of R&D between TDA1545 and AD1955 did produce some results !
I will publish the schematics for the FPGA module soon for review, but these don't include the DAC... one step at a time
> Could you please post a block diagram of what you are doing?
Alll the docs for the current FPGA implementation are on my website (dig in the menu, Projects, etc). The FPGA module I'm designing now is quite similar, here are the differences :
- FPGA : Spartan 3-1000 -> now Spartan 3E-1200 or 1600
- LAN : LAN91c111 -> now LAN9117, easier to use chip
- SDRAM : 16 bit -> now 32 bit
Architecture-wise, the Suzaku has a shared bus between the SDRAM and the LAN chip, my module will use two different busses, which will allow a lot more data throughput and power. It should never choke on data.
The original idea was to use a PQ208 FPGA. I have a board routed with this one, but it has several problems :
- The PQ208 package sucks for EMI and parasitics
- There isn't that many IO pins free to use once everything is connected (ie. I can plug some DACs, but the LCD, volume control, etc... no).
- Only the Spartan 3E-500 is available in this package, and I would like to use a fatter FPGA to implement sample rate conversion in the FPGA (since that would use a LOT of CPU in the PC)
So, I wanted to use a bigger FPGA in a BGA package. This is not easy for an amateur, because using a 1mm-pitch BGA means you need pretty tight tolerances on the PCB track/space and vias, which means it's quite expensive. This assembling and PCB problem is almost solved thanks to a benefactor who would rather remain anonymous since he doesn't want to receive tons of requests, lol.
Anyway, I had to remake the design using Altium, which can handle BGA correctly.
I needed a few tries to get into Altium, this software definitely has a
high "bump" for first users. I tried at least 3 times and every time I was
put off by the complexity of the software. But last sunday I tried it
again, and this time I got it. And once past the initial bump, I'm
starting to like this software. It still sucks, but lots less than Eagle.
Anyway, schematic capture is done in Altium, and I have entered all the
models for the parts not in the library. The only thing remaining is
proofing the schematic again and obviously... the PCB layout.
I will post the schematics for review as soon as I have something "showable".
Have a nice day !
Phew Peufeu, (see what I did there?)
I thought we had lost you, maybe not from the physical realm but from the diy audio realm! I'm sure I wasn't the only one. Let me say how great it is that neither of these scenarios came to pass.
I believe there is huge interest in this concept - just look at the number of views of this thread!
I sent you an email at the end of Sept wondering if I (or some of my students) could help in any way - you seem to have a lot of the areas already covered but email me if there is some area you feel might be useful!
Japan, looks fantastic - good photography BTW.
John
I thought we had lost you, maybe not from the physical realm but from the diy audio realm! I'm sure I wasn't the only one. Let me say how great it is that neither of these scenarios came to pass.
I believe there is huge interest in this concept - just look at the number of views of this thread!
I sent you an email at the end of Sept wondering if I (or some of my students) could help in any way - you seem to have a lot of the areas already covered but email me if there is some area you feel might be useful!
Japan, looks fantastic - good photography BTW.
John
I'm interested but I'm also a bit worried about the hardware. Have you considered using a board like the Nexys from Digilentinc.com ? I have one right here that has an S3-1000 and a high-speed hirose connector. It would seem prudent to design a secondary board that plugs into this one including the ethernet and audio i/o, due to the sheer cost of prototyping a 6 or 8 layer BGA board. What do you think? I'm not trying to change your design, I'm just saying it would probably be a lot quicker and easier for the first round.
OK, I'm learning Altium (still no luck with the autorouter...), I did some tries at BGA escaping/routing and it should be OK. I'm checking and re-reading datasheets (especially the FPGA) to make sure the schematics are correct... it isn't that complicated but I always feel I'm forgetting something somewhere...
jkeny: I got your email, sorry for not replying (yet...) basically when I get a module working, we'll talk about a common effort on software.
> Have you considered using a board like the Nexys from Digilentinc.com ?
Yeah but I'm not that comfortable with the MAC core from Xilinx (esp. the price, is it 15K or sumthing) and the opencores MAC core has been described by some as a nightmare to work with... so...
jkeny: I got your email, sorry for not replying (yet...) basically when I get a module working, we'll talk about a common effort on software.
> Have you considered using a board like the Nexys from Digilentinc.com ?
Yeah but I'm not that comfortable with the MAC core from Xilinx (esp. the price, is it 15K or sumthing) and the opencores MAC core has been described by some as a nightmare to work with... so...
Hi peufeu,
I cannot PM you as I am new to this forum and have not posted enough. But I would be more than glad to review your schematics and provide any constructive feedback if needed. If you want to do this offline, you can email me the schematics as a zip file to vsap78@yahoo.com.
Ananth
I cannot PM you as I am new to this forum and have not posted enough. But I would be more than glad to review your schematics and provide any constructive feedback if needed. If you want to do this offline, you can email me the schematics as a zip file to vsap78@yahoo.com.
Ananth
You got mail
Meanwhile, please check this out for the next step in the project :
http://www.diyaudio.com/forums/showthread.php?s=&threadid=114227
Meanwhile, please check this out for the next step in the project :
http://www.diyaudio.com/forums/showthread.php?s=&threadid=114227
mako1138 : I'll send you schematics (but not now, I'm making some modifications, see long post above).
TNT : For the shield I put a ferrite that can be soldered or not. However, just use UTP cable (the most common type of Cat-5 cable) which is Unshielded Twisted Pair, hence no ground connection via the shield. STP is Shielded Twisted Pair. The connectors are different, UTP is plastic, STP has a metal shield on the connector.
TNT : For the shield I put a ferrite that can be soldered or not. However, just use UTP cable (the most common type of Cat-5 cable) which is Unshielded Twisted Pair, hence no ground connection via the shield. STP is Shielded Twisted Pair. The connectors are different, UTP is plastic, STP has a metal shield on the connector.
mako1138 : I'll send you schematics (but not now, I'm making some modifications, see long post above).
TNT : For the shield I put a ferrite that can be soldered or not. However, just use UTP cable (the most common type of Cat-5 cable) which is Unshielded Twisted Pair, hence no ground connection via the shield. STP is Shielded Twisted Pair. The connectors are different, UTP is plastic, STP has a metal shield on the connector.
TNT : For the shield I put a ferrite that can be soldered or not. However, just use UTP cable (the most common type of Cat-5 cable) which is Unshielded Twisted Pair, hence no ground connection via the shield. STP is Shielded Twisted Pair. The connectors are different, UTP is plastic, STP has a metal shield on the connector.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.