ESS Sabre Reference DAC (8-channel)

khundude said:
Russ, that is a great idea with the LED to switch the OTTO! Do you have any suggestions on other pure DSD players to mod?

One or two more things about transformers for the power supplies. Are the ones you sell able to be wired for 230v operation as I will be back in Asia in four months? Same with the Denon player, will it be easy to convert to 230v or do I buy one with 230v?

Thanks. :)

As for the SACD players to mod, so far the DVD-1920 is the easiest by far. I don't have a Pioneer to check. Basically any SACD player that has easy access to the DSD lines in its DAC should work fine.

The DVD-1920 is pretty cheap these days. I have seen them in good working order for < $100 on ebay.

The trafos we stock are 120/240 and will work fine with 230V.

Cheers!
Russ
 
eclectic2k said:
Thanks Russ for all the great info and the good work.

I've been following this thread since early on, and am very excited to try out the buffalo. I'm interested to see how well it handles jitter (no I2S here - just spdif ;-) and how well it compares overall to my Lavry DA10.

In the process I've discovered a few other items of use on the twisted pear site (darwin, OTTO, & the JT attenuator) I'm saving my pennies and making a shopping list. :D

Any solid plans yet on when the JT & the buffalo will be back in stock?

Thanks!
-CK

Building these DACs is a painstaking process. Brian just finished shipping a round, and probably needs a bit of a breather to get the new Buffalo modules going.

JT has been redesigned (just slightly) and will be ordered with the next batch of boards. It will then just need a quick test and it is good to go.

I would never hesitate to use SPDIF input on the Buffalo. It is quite good. The only slight difference in PCM and SPDIF input that I have noticed is that it is helpful to run the DPLL at slightly wider bandwith then you would with PCM. The reason is at very low DPLL bandwith the SPDIF can take a couple seconds to fully lock. I adjust for this in the basic firmware provided on the buffalo.

Still, a design note, I am informed that the higher the DPLL bandwidth the lower the jitter attenuation of the DAC. So, if you want he absolute lowest jitter you will probably want to use I2S input with the DPLL set to "LOWEST" which is the way the buffalo firmware is setup. This is the way I am using the DAC now. SPDIF into one of my Wm8804 modules, then I2S into the DAC. This is my preferred setup for now. Plus it makes it very easy to switch in DSD using the OTTO. Since the DAC is already in PCM/DSD mode, so no switches or jumpers to change. And since its SPDIF I can run a 15ft cable with no issues. ;)

Cheers!
Russ
 
khundude said:
Russ, how difficult would it be to change the voltage on the Denon to 230V?

Have you compared SACD pure DSD and SACD PCM converted to see if you can hear the difference?


Well it seems to me you have at least two simple choices:

1) Get the 200-240V version of the DVD-1920
An externally hosted image should be here but it was not working when we last tested it.

2) Get a 220/110 travel converter.


My only fear with players that convert DSD to PCM is that it is reported that they sometimes , at least apparently in the case of a few Oppo players, tend to get the level wrong so you either lose dynamic range, or you get clipping. I would rather take straight unmolested DSD if I can. Especially since its easy to make go 6-10ft via SDIF2. The Sabre chip sounds incredibly nice with DSD input, so I am a bit hooked. :)

Now keep in mind some of the issues with the Oppo can probably be fixed with firmware, but I don't know if they have been. I have not done any serious comparison of the two type of players.

Cheers!
Russ
 
Russ White said:

The only slight difference in PCM and SPDIF input that I have noticed is that it is helpful to run the DPLL at slightly wider bandwith then you would with PCM. The reason is at very low DPLL bandwith the SPDIF can take a couple seconds to fully lock. I adjust for this in the basic firmware provided on the buffalo.

Still, a design note, I am informed that the higher the DPLL bandwidth the lower the jitter attenuation of the DAC. So, if you want he absolute lowest jitter you will probably want to use I2S input with the DPLL set to "LOWEST" which is the way the buffalo firmware is setup. This is the way I am using the DAC now. SPDIF into one of my Wm8804 modules, then I2S into the DAC. This is my preferred setup for now.


I have found that SPDIF into a WM8804, then I2S into the Sabre DAC works great for the CD/DVD players and the two networked media sources ( SMCWAA-G and Popcorn-Hour ) that I have tried. I use lowest DPLL setting for those. It sounds great.

However, at lowest DPLL, I get intermittent loss of lock when I use the Plasma TV as a source. The Sony DVR is even worse. It needs the HIGHEST DPLL setting to remain in lock. Otherwise I get drop outs and it never does fully lock with a low DPLL window. The jitter coming from my digital cable must be horrid for the audio packets. (The HDTV video looks OK)

I would be interested to hear if anyone else is having the same symptoms with cable TV sources. It's not just a long delay in locking, it keeps going in and out of lock.

Another interesting observation is that if I input the TOSLINK from the Sony DVR to the WM8804 -> I2S -> Sabre, it won't remain in lock unless I use the highest DPLL window. But if I send video/audio from the Sony DVR to the Plasma via HDMI first, then listen to the TOSLINK from the Plasma, I can use Medium-High DPLL. Apparently the Plasma re-clocks the audio and makes it slightly better.

It is interesting to experiment with the DPLL setting while listening to a high quality DVD-Audio 24/96 track. As I step through from low DPLL window to high DPLL window, I can hear the soundstage close down.

Note that I don't hear any noticable increase in distortion when I do this. It is only a loss of the pristine stereo image. The precise location of instruments fade away.

It really doesn't bother me that a high DPLL window setting for the cable TV sources lose the spectacular soundstage that the lowest DPLL setting gives. After all, it's only television. <grin> Most of the TV audio lacks dynamic range anyway.

But since I do want to change the DPLL setting depending on the input, I have decided to design in a front panel control for DPLL on the DAC that I am designing. The front panel momentary switch will go into a PIC micro input that wakes it from sleep mode when pressed. (IOC input). Then the PIC will step through the settings with each keypress. I will also have some LEDs to show the status. Then the PIC will go back to low power sleep mode after setting the Sabre's register.

My DAC design keeps on getting more complex. Sometime soon I will have to stop designing and build it. :D
 
Hi Ross,

Thats an interesting experience you are having. I can't say I have seen the same. But we have two different circuits and different sources etc. It almost sounds like your SPDIF receiver may be having issues?

I have had no issues locking onto any SPDIF source at all. Cable box, TV, Receiver, Squeezbox, PC, DVD, CDP. This holds true both directly into the Buffalo, or VIA my own Wm8804 board. I wonder if you have some other issue? If the WM8804 is showing locked, the PCM stream should be good.

I certainly have never seen any of my buffalo lose SPDIF lock. The only thing I have ever noticed is it taking a few seconds to settle down(not so much acquire the lock as get the DPLL settled), but ounce its settled, its golden. I have yet to have perceived any need for DPLL bandwidth > "Medium", and in fact have seemed to notice a distinct loss of quality above that. This could just be in my head.

BTW, all of the DAC features will be made available in the "uber" controller firmware I am testing (RE and IR input with VFD display) which is targeted for the Buffalo. So if people feel the need to experiment they can.

Cheers!
Russ
 
Hi Russ,

Maybe I should explain a little more about the Sony DVR.

If you go over to avsforum.com and look at the thread

http://www.avsforum.com/avs-vb/showthread.php?t=537711

There are over 11,700 posts in the thread about the Sony DHG-HDD250. I'm not suggesting that you read them all, :D but if you were to do a search you would find dozens of complaints that somebody's AV receiver will not lock on the TOSLINK.

Neither of the other two DACs that I have in the house will lock on the Sony's TOSLINK output. The only way that I could do it before I got the Sabre demo board was to use HDMI to the plasma, or listen to the analog outputs of the Sony.

The Sabre demo board is the first DAC that I have seen that will actually lock on the Sony. That, in itself, is an accomplishment, because this Sony is a horribly designed piece of equipment. Read through that thread on AVS forum and you will soon discover that.

It's a worst case scenario.
 
rossl said:
Hi Russ,

Maybe I should explain a little more about the Sony DVR.

The Sabre demo board is the first DAC that I have seen that will actually lock on the Sony. That, in itself, is an accomplishment, because this Sony is a horribly designed piece of equipment. Read through that thread on AVS forum and you will soon discover that.

It's a worst case scenario.


Ah, OK. Perhaps I just don't own sufficiently sucky equipment to test the full extent of the DAC's ability to lock onto a horrid source. :D ;)

Cheers!
Russ
 
This may seem like a noob question but is the IVY Active I/V Stage necessary for the buffalo DAC if I am using a tube buffer after the DAC chip? The tube buffer is the promitheus audio brand with output transformers and a tube power supply. It is very picky about any source noise. My squeezebox 3 even with the DAC going directly to the RCA jacks is way too noisy for this buffer, but my ipod is dead silent. Would this DAC work well with the buffer or would I be better off to purchase the OPUS dac? Thanks
 
matejS said:
Hi!

A short questions...
... what is output current swing of a single DAC (or x4 for single channel)?

Thanks,
Matej


The differential output is 0.924*1.65V (3.1746 Peak) + 1.65 DC bias and the series output resistance is nominally 195 Ohms with 4 channels in parallel. This works out as 18.14ma p-p centered at 9.43ma unless I messed up with my calculator.

Cheers!
Russ
 
tripathdude said:
This may seem like a noob question but is the IVY Active I/V Stage necessary for the buffalo DAC if I am using a tube buffer after the DAC chip? The tube buffer is the promitheus audio brand with output transformers and a tube power supply. It is very picky about any source noise. My squeezebox 3 even with the DAC going directly to the RCA jacks is way too noisy for this buffer, but my ipod is dead silent. Would this DAC work well with the buffer or would I be better off to purchase the OPUS dac? Thanks

I don't really have enough information to answer your question. Strictly speaking you should not need the IVY. It just depends on what kind of performance you are after. If you don't use the IVY you will likely need AC coupling caps to the input your buffer.

Cheers!
Russ
 
Is the buffalo dac a current or voltage output dac? Is there anyway to find out in the design if my tube buffer can accept a current output dac directly. I know output transformers will eliminate any dc offset so I am guessing output capacitors are not required. Thanks for your input on this.