Rank: Member
Groups: Member
Joined: 5/6/2016(UTC) Posts: 2
|
I'm designing a multi dac build for four way speakers using a minidsp sharc electronic crossover and BIII, IVY, mono sympactico, direct to each driver. Since i'd be running the BIII's in dual mono for each frequency range I'd like to make sure that all of my DACs, the DSP chip, and digital sources maintain word sync. I'm anticipating using an cronos with the RHEA oscillator combo in the main chassis with the DSP Chip and wondering if i can a teleporter for each channel (8 teleporter pairs total) to send the I2S and clock data info to each of the 8 DACs, balanced line, and amp combos which will be broken out independently. The DSP Chip will be outputting 24/96. My understanding is that a Digital input board will be necessary to supply the IS2 to the DSP and that i can use a TP spdif board for this. But the DSP needs 12.288MHz. I'm unclear how to implement the clocking with respect to the DSP chip and digital source. Is it as simple as slaving the spdif board and the dsp board to the cronos? or does the input need to be the master? Is it intended that the cronos be implemented without the Hermes and Amarano as an independent master clock? If i wanted to include asyncronous usb input e.g. the Hermes/Amarano/ Rhea combo, would i need to do this with separate cronos or could one cronos handle all my clocking needs? Can a teleporter transmit the 3.3v clock signals of the cronos alongside a IS2? What outputs would i need to send from the cronos to the teleporter? If so, how can i distribute the output of the chronos to multiple teleporter inputs, so that multiple channels of IS2 going to different DACs will carry the same clock signal? Any guidance would be very helpful. My goal is to keep all of my I2S and interconnects as short as possible, by integrating DSP and amplification into the speaker cabinet. What's the best oscillator combo to use for this type of purpose? Many thanks in advance. Andrew Edited by user Friday, May 6, 2016 10:18:39 PM(UTC)
| Reason: Not specified
|
|
|
|
Rank: Member
Groups: Member
Joined: 8/5/2012(UTC) Posts: 208 Thanks: 4 times Was thanked: 16 time(s) in 14 post(s)
|
Are you planning on putting everything into 1 box or is this a 'distributed' implementation? (as the answer to that question will have a huge impact on your options)
BTW I am doing something similar - quad 3 way mono blocks
|
|
|
|
Rank: Member
Groups: Member
Joined: 5/6/2016(UTC) Posts: 2
|
Thanks Possum. This is a distributed implementation.
I took a look at your posts and I think we're trying to accomplish similar goals. I'm trying to build DSP active loudspeakers and ideally i'd like to have an amp and DSP for each driver designed into the cabinet. At this point, I'm thinking about scraping the minidsp aspect and going with a pc running acourate for x-over and room correction feeding faun i2s bridges to each DAC. In this scenario, i'd just slave each DSP to the output of the faun(s) and explore best how i want to clock those. Does anyone have experience with the Faun distributing I2S over cat6/ RJ45? Are there any concerns or limitations as with the teleporter that Possum noted above?
From a design standpoint, I'd like to have the shortest connection between the Buffalo III, the IV and the sympatico - and i'm wondering if i could mount the buffalo iv combination vertically over the sympatico inputs without encountering any interference from the AMP. From a power stand point, I plan on putting all transformers and power supplies for the sympatico, the IV and the BIII in a separate case and bringing the appropriate current to each component separately for isolation purposes. So really, my concern here would be interference from the amp itself on the BIII and IV.
Then I'd be designing my loudspeaker cabinets, transmission line, to enable the shortest run from each integrated DSP/Balanced Line Stage/ AMP to the driver.
Since i'm using dsp crossovers and room correction, i'm kind of toying with the idea of modular driver and woofer components so that i can mix and match using various FIR filters, and tune each enclosure to it's specific component.
Thoughts?
|
|
|
|
Rank: Member
Groups: Member
Joined: 8/5/2012(UTC) Posts: 208 Thanks: 4 times Was thanked: 16 time(s) in 14 post(s)
|
Yes, our goals are very similar.
The Teleporters may work perfectly for you (they just didn't for me and I could not solve the issue. It may well be a wiring issue in my apartment - who knows?).
My first suggestion is to try a pair - they're only $90.
My second suggestion is to seriously look at using AVB - you only need a single ethernet cable to exchange a large number of sync'd bit-perfect high res audio channels and I find the intelligence of AVB very useful as you can diagnose connection issues with software (I am using Motu's software and their AVB UltraLite device). If you're a muso then AVB wipes the floor with the other solutions as you can easily patch in MIDI, mics, guitars etc to your sound system and remix all the signals - and record of course.
The Motu AVB software also easily lets you 'mix n match' any channel to any other. You can also code this into the AVB endpoint. The Motu windows driver provides you with up to 48 channels in and out. If you're doing all your dsp in your pc then all you need is an AVB UltraLite and 2 AVB endpoints. If you use a Mac then you dont even need the Ultralite as AVB is directly supported via the Thunderbolt connection (not sure about the routing software support required tho). That's straight from your Mac to your speakers - no soundcard in between.
The faun i2s bridges may also be a good option. I don't know, but I doubt it will be as flexible as AVB.
PS: I would strongly recommend using the Legato over the IVY
|
|
|
|
Rank: Member
Groups: Member
Joined: 8/5/2012(UTC) Posts: 208 Thanks: 4 times Was thanked: 16 time(s) in 14 post(s)
|
Also, you need to be sure you can actually slave the DSP output. The MiniSharc DSP, for example, is master I2S out (like most things). You will need something like the Metronome to async reclock it (but the Metronome is only stereo).
You are embarking on a very complex project. Using software to do the xover processing and AVB probably offers you the simplest starting point that is likely to work.
|
|
|
|
Rank: Member
Groups: Member
Joined: 8/5/2012(UTC) Posts: 208 Thanks: 4 times Was thanked: 16 time(s) in 14 post(s)
|
Correction - If you are including the DSP in the speaker then definitely try the Teleporter solution first. It would offer by far and away the most simplest solution (mine is more convoluted :) )
|
|
|
|
Rank: Administration
Groups: Administration, Customer Joined: 10/24/2006(UTC) Posts: 3,979 Location: Nashville, TN
Thanks: 25 times Was thanked: 89 time(s) in 83 post(s)
|
The current teleporters are designed to support shielded Cat5/6 Make sure you connect the shield to GND at one or both ends depending on your setup.
|
|
|
|
Rank: Member
Groups: Member
Joined: 8/5/2012(UTC) Posts: 208 Thanks: 4 times Was thanked: 16 time(s) in 14 post(s)
|
Yes, I noticed this improvement to the Teleporters. However, I was using the Neutrik RJ45 chassis sockets (which are grounded/shielded) with shielded cat 6 and this made no difference for me - and I was only using 1.5 m runs!
I have been using an XMOS AVB solution as an alternative for a month now without any interference probs at all. The one downside of this solution however is if you need to tap into an I2S master - the Metronome works in slave slave mode (I am MCLKing it with a Cronus) but it is not multichannel.
A multichannel asynch reclocking capability would be very cool for this kind of scenario (hint hint :-) )
|
|
|
|
Rank: Administration
Groups: Administration, Customer Joined: 10/24/2006(UTC) Posts: 3,979 Location: Nashville, TN
Thanks: 25 times Was thanked: 89 time(s) in 83 post(s)
|
Likely its the way that it was grounded that caused the problem - it can't just be to chassis gnd - it has to be to the digital pwr gnd. Edited by user Sunday, May 29, 2016 6:26:27 PM(UTC)
| Reason: Not specified
|
|
|
|
Rank: Administration
Groups: Administration, Customer Joined: 10/24/2006(UTC) Posts: 3,979 Location: Nashville, TN
Thanks: 25 times Was thanked: 89 time(s) in 83 post(s)
|
The reason I say this is that I have only ever run into problems when the GND points have different DC bias.
|
|
|
|
Forum Jump
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.