Fred, those Salas regulators are excellent! You will not be dissapointed with the Twisted Pear equipment. Their DAC is phenominal! You may be dissapointed in not using some tubes in the output though Fred.:razz::wink::cheesygrin:
I'm expecting as much Greg having loosely followed the development of the Buffalo and reading the hype surrounding the ESS chip for some time.Quote:
You will not be dissapointed with the Twisted Pear equipment. Their DAC is phenominal!
In my best Mexican accent. Tubes?We don't need no stinking tubes.:razz:Quote:
You may be dissapointed in not using some tubes in the output though Fred.:razz::wink::cheesygrin:
The PCM1796 DAC chip in the Denon shares some inputs for PCM and DSD so I only have three channels to transmit with no ground needed. Bit clock, Word clock and Data for PCM. DSD is Bit clock, word clock becomes data left and data becomes data right. No ground needed so we have galvanic ground isolation between source and destination components; thus no ground noise issues to worry about.
At first they didn't work for me with DSD or PCM. The DAC wouldn't lock, signal cut in and out, lock LED flickering for both. While poking around, I grabbed the Denon chassis and grounded myself and it worked, solid lock, steady LED in the DAC. I learned the Denon 3910 isn't supplied with a grounding IEC from the factory. After installing one and grounding the chassis to it, I got lock and SACD played wonderfully. PCM played but with a crackling noise. Disconnecting and jumpering the plug on my relay board made the crackling stop. I grounded the modified Sidecar relay board to the back panel where the Teleporter is attached, and now CD's play without crackling through the relay. You'll have to excuse the quick and dirty white ground wire, but we be listening now.:wink::cheesygrin: Enough bench time today.
One could install a Teleporter in a 'putie. With the right sound card, export I2S to a DAC. One would need a Teleporter to transmit and one to receive (transmit/receive done by setting dipswitches).
Rich belts another out of the park! Man you are on the leading edge of DIY quality and ingenuity! Congrats on the teleporter success!!
Thanks Carl! How about for cutting edge a dual mono dac with digital display controlled by a micro controller. Gotta be able to modify/write code to make that all work. Maybe next, after the stealth PTP5 stainless top plate Lenco!:cheesygrin:
Great work Rich,
I've had better luck with the dropout issue. I used the switch setup as per the diyaudio post to select between i2s and spdif adding the auto spdif detect switch. I need to look at grounding more from what you found with the denon. I have a juli@24/192 card in the pc so I need to get i2s off of it next. Looks like it will work. I'll see if it sounds better than the teradac. I2s is the way to go without question! Much cleaner and more open sounding. Not as much glare. Glad you got your teleporters beaming well.
You ever use front panel express software? I just started and don't like it much. I'm software challenged.:eek: I'm going to check with a self employed girl friend (really, just a friend) who has a CNC machine to see if she can cut a front and back panel for me.
He was also concerned about the extra length in the Bit Clock circuit due to the relay to switch between PCM and DSD. As you recall they tell us to keep the I2S wires at equal lengths. Perhaps I should have measured it out and increased the length of the data/word clock wires. I did wrap the small wires with foil that I salvaged from stripping them out of a long video cable.
I?m thinking there are CDP designs with higher voltages than 1.6v to the DAC chip. IIRC the PCM1796 DAC chip has a range from .3v to 6.5v input voltage. Seems like some folks can get away with 1 meter I2S lines from their CDP to the DAC without using a Teleporter, like Mr Majestic with that Denon DVD-1920. As I recall the output from my old Denon 2910 was low, like 1.6v at the RCA out.
It goes on of course how they accomplish that. The link to the white paper.Quote:
III. EXTRACTING THE DATA FROM THE TRANSPORT MEDIUM
A. SPDIF Interface
The SPDIF interface is more complex than the DSD and I2S since it must first derive the embedded clock in the bi-phase encoded data. In fact, experience with many forms of SPDIF decoder suggest that most fail in the presence of high jitter due to the lack of robustness in the clock recovery process. To avoid this potential problem the Sabre SPDIF interface avoids having to extract the clock at all: decoding is done using a method that does not require an explicit measure of the clock frequency.
The point being even using spdif with an ESS Sabre DAC chip, you can get better sound. I think this contributes to folks going to DAC?s using the ESS Sabre DAC chips and liking them. Of course, I'm thinking that it can't fix everything so a lower jitter source will still best the high jitter source.
I have been studying the 3910 manual for a while to go I2S, but only with your help I have understood DSD/PCM connections.
I have found points near IC203, where I2S and SPdiF is generated, but then the I2S cables will be longer, and now, what is better, shorter wires and longer board traces or the other way round. With a feedthrough Neutrik connector and a short R45 cable I could mount the teleporter very near IC203.
There are still 2 points of interest:
In the 3910 manual PCM Wordclock and DSD DATA right are the same trace and not DATA left as used on buffalo.
Have you checked left and right with SACD?
Then I have had the idea to tap IC202 pin 75 or CY211 pin 21(connector to display). The blue SACD is switched on this way with a npn transistor. If you use another one to switch your relay, I2S/DSD will be automatic!
I thank you for your work.
Hi rolls! I am glad to have helped you and appreciate your kind words and input. I don't have any schematics for the 3910, but I was going off the PCM1796 DAC chip datasheet.
From page 18 of the PDF datasheet:
"Direct Stream Digital (DSD) Format Interface and Timing
The PCM1796 supports the DSD format interface operation, which includes out-of-band noise filtering using an internal analog FIR filter. For DSD operation, SCK (pin 7) is redefined as BCK, DATA (pin 5) as DATAL (left channel audio data), and LRCK (pin 4) as DATAR (right channel audio data)."
So on the PCM1796 DAC chip, PCM/I2S has pin 4 as WORD Clock, pin 5 is DATA (L and R channel) and pin 6 is BIT Clock. With DSD, pin 4 becomes DATA Right, pin 5 becomes DATA Left and pin 7 is BIT Clock. So all I did was use a relay to switch between pin 6 (PCM BIT Clock) and pin 7 (DSD BIT Clock). On the Buffalo III DAC, input connections remain the same to the Sidecar/BIII. I do have R and L channel SACD working along with PCM. I'll have to pop the cover on the 3910 to see the IC203 point you are referring to and try and understand the clever way of switching you described.
I do use the Neutrik RJ45 connector on the DAC connected with a short Cat5 cable to the Teleporter. That way the Teleporter can be close to the BIII DAC Sidecar input and not exceed the recommended 10cm (4") I2S wires. I see no reason why that couldn't be done in the 3910 as well. IIRC Russ White (or Brian) expressed some concern about using Neutrik RJ45's with the Teleporters due to the extra connections. I figured since the Teleporters allow up to 100 feet of Cat5, a couple of extra connections might not be detrimental to the signal with a short patch cable. I suppose one could cut the plastic away and solder the wires to the Neutrik RJ45 instead of relying on just the push in wire clamp. Of course you would still have the extra RJ45 plug in connections (as in my image in post 43).
One thing I learned about I2S is all the wires should be shielded, including the Cat5 used between the LVDS Teleporters to keep stray RFI from interfering wth the signal. You can get Cat patch cable with shielded connectors as well. If you pick up interference, the I2S could be no better or worse than a SPDIF connection where all the Data is combined.
In regards to my configuration with the BIII used with SPDIF and I2S/DSD inputs; the ESS DAC chip has a SPDIF auto-detect feature. When using high sample rate PCM/DSD input on pin D2 or D4 inputs, auto-detect is prone to false positives and won't remain locked on the signal. Therefore SPDIF auto-detect should be disabled with I2S/DSD. Rather than flip the dipswitch, wire up B5 to a switch and ground. By flipping the switch and grounding B5, you turn SPDIF auto-detect off and I2S/DSD remain unmolested and locked.
Thank you for your kind reply,
If you like I could send you the manual. You can contact me directly, there are 230 pages.
There is still the left/right problem I don't understand.
May I quote you from post 37:
"The PCM1796 DAC chip in the Denon shares some inputs for PCM and DSD so I only have three channels to transmit with no ground needed. Bit clock, Word clock and Data for PCM. DSD is Bit clock, word clock becomes data left and data becomes data right."
Then you said in your latest post:
"So on the PCM1796 DAC chip, PCM/I2S has pin 4 as WORD Clock, pin 5 is DATA (L and R channel) and pin 6 is BIT Clock. With DSD, pin 4 becomes DATA Right, pin 5 becomes DATA Left and pin 7 is BIT Clock."
For the PCM1796 Dac, word clock becomes data right is correct, but I don't know what happens in the buffalo.
In the manual and teleporter, word clock becomes data 1, which normally means left. Do I miss something?
rolls, duh, my apologies. It just hit me what you said, the BIII DSD R & L inputs are listed as different in the integration guide. I obviously missed that. That means if the guide is correct, my SACD R & L channels are reversed at the BIII inputs. I would need to add two more relays to flip the DSD channels in the 3910. Good catch, thanks!
How are you with FrontPanelExpress software?:smile:
With the help of the manual you could also look for the DSD right and left traces in the neighborhood of IC203,
before they meet with I2S, then you could unsolder the resistors there and solder new ones, not SMD, left to right, and right to left.
This way you don't need 2 more relay, but the original Stereo output would be reversed when you play SACD.
But as buffalo owner you won't use the Denon DAC any more.
Currently I am pulling the I2S/DSD signal off the board before the SMD resistors that connect to the DAC chip inputs. I don't want to reverse the stereo signal, so relays appear the best way to accomplish the channel swap. I need to get an enclosure designed/made and then I will revisit the DSD issue.
I don't have that many SACD's, but obviously need the option and want it right.
All problems solved at the moment,
I think I can start I2S as soon as possible, with confidence.
The Denon sounds fantastic right now, because it is Lampizator modyfied (SPdiF with pencil tube).
Sadly my Teac P2-s with new clock has no chance any more.
Thanks for your help!
The use of multiple shunt regulators for the digital and analog section certainly has audible benifits but results in a lot of excess heat and my single rack height chassis can get very warm after long listening sessions.
In fact it gets warmer than a couple of my high power Hypex (pure:smile:)Class D amplifiers.I may try to source a bigger or better ventilated chassis to give it more room to breath.
I have been keeping an eye on(more like drooling over) TP's development process of this DAC over the last few years and appreciated how they were striving to optimize each performance area for maximum sound quality.(For instance the use of such high quality voltage regulation for the DAC chip itself is a rarity). I owe a big thanks to my bud Rich for a very generous gesture that steered me towards building one of these wonderful sounding devices.
That was one of the first things I looked for when choosing a chassis for my DAC (second to size) was breathablility. There are vent holes all around the edges and top. It keeps it pretty cool.
Post some pice when you have a chance Fred. Interested in seeing what your DAC is like.
I have been searching for the perfect output tube como for my DAC recently. I know you shun the valves, but the tube output sounds great with the Buffalo DAC. I listened to my rig with having installed one RCA long black plate 12AX7 and one Valvo Hamburg 45 degree halo getter long plate tube per channel. The combination has a lot of detail and resolution, but still retains some of the warmth and midrange liquidity of the RCA black plate tube. Good combo!! I know this is getting crazy (somewhat) doing tube combinations, but you should hear the output!:wink::biggrin:
Hey Fred! Glad you are liking it. The Salas are great performers and those boards look awesome. Shunted power supplies are the way to go to get all the bang out of the BIII. The IVY III output board sounds really good. The Legato is a bit different flavor, some like it better.
I need to get cracking on an enclosure myself. I started plans for one using FrontPanelExpress. I'd like to have the front and rear panel holes CNC'd with some nice silkscreen lettering. The DAC is worth it.
I was wondering if you got my PM?
Hi Jeff. I've been busy with work and don't check in much, but I don't see one. IIRC new members can't send PM's here until they have 10? posts? I sent you one with some alternate contact info.
rolls, I haven't been using the Denon 3910 except to test stuff in the shop. While I wasn't working on a multiple relay board to correct the reversed SACD channel problem, Twisted Pear was and came up with an I2S/DSD switching solution. Meet OTTO2.
PCM is connected to B1; PCM Bit clock, Word clock, Data and GND. DSD is connected to B2; DSD Bit clock, Data left and Data right (which share the PCM Word and Data circuits) are reconfigured on the B2 side so the SACD channels are no longer reversed.:cheesygrin: OTTO has the grounds tied so no need to jump GND from B1 to B2. OTTO is controlled by the switch I had previously installed. It passes the selected signal to the Teleporter which sends it to the DAC.
Rich, your a sick puppy....and I mean that in a good way.
Nice frickin' job bud, that baby looks bad ass. Wish I had the patience for this stuff. Good work !!!