ESS Sabre Reference DAC (8-channel)

Re: Buffalo - Level of Complexity

GrahamH said:
I am just getting started into DIY audio and your DAC is just the component I would like to start with.

Hi Graham,

Many people with varying degrees of skill have completed finished DACs based on our modules.

Check out this thread for some more info:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=128137&perpage=25&pagenumber=1

It is probably not the kit for an absolute beginner, but it is not very difficult either.

Cheers!
Russ
 
Re: Re: Buffalo - Level of Complexity

GrahamH said:
Russ,

I am just getting started into DIY audio and your DAC is just the component I would like to start with. I have little skill in electronics assembly and even less understanding, but want to start somewhere fun and interesting. Do you feel your Buffalo kit would be jumping into the deep end or could I reasonably start there?

Cheers,
Graham


[i

Twisted Pear kits are absolutely right for new comers; they are simple to assemble and connect. This assumes that you can solder properly.

Because Russ is prepared to asnwer questions clearly, this makes it easier to commission as well.
 
Re: Re: Re: Re: Es9018

dusfor99 said:
Hi Cauhtemoc,

I can answer "some" questions with no problems, but there are some areas I am not allowed to go into. It never hurts to ask. So what are you curious about? :)

Dustin

I have heard rumours that the new chip is supposedly easier to use in hardware mode (i.e. standalone, or without a microprocessor). Is this true, and if so, exactly how is it easier to use? This question has been itching my mind for a while.
 
Re: Re: Re: Re: Es9018

dusfor99 said:
I can answer "some" questions with no problems
Dear Dustin,

I'd like to understand better the capabilities which are exposed by letting users set the filter characteristics. For example, is there enough flexibility to do a design which has a stereo signal coming in and splitting it for each channel into three (six outputs total) to drive active 3-way speakers? For a given channel how much flexibility is there in the filter. For example, how many taps does it have? What bit length the coefficients and what bit length for the internal arithmetic?

Peter
 
Re: Buffalo - Level of Complexity

fmak and Russ,

Thanks for the advice. I have been practicing my soldering and with the great support in this forum feel comfortable going for it. Since it seems that the revised Buffalo might still be a few days/weeks away, I am going to start with the IVY + LCBPS kits. Practice there and then move on to the Buffalo when it is ready.

Thanks again.

Graham
 
Re: Re: Re: Re: Re: Es9018

Cauhtemoc said:


I have heard rumours that the new chip is supposedly easier to use in hardware mode (i.e. standalone, or without a microprocessor). Is this true, and if so, exactly how is it easier to use? This question has been itching my mind for a while.


Hi Cauhtemoc

THe chip will now autodetect if the input stream on data1 is SPDIF or not. If its not, it will try to see if its I2S or DSD. SO there is not need for a MCU to program it to get it into SPDIF mode.

Dustin
 
Re: Re: Re: Re: Re: Es9018

schro20 said:

Dear Dustin,

I'd like to understand better the capabilities which are exposed by letting users set the filter characteristics. For example, is there enough flexibility to do a design which has a stereo signal coming in and splitting it for each channel into three (six outputs total) to drive active 3-way speakers? For a given channel how much flexibility is there in the filter. For example, how many taps does it have? What bit length the coefficients and what bit length for the internal arithmetic?

Peter



Hi Peter,

The bit lenght for the coefficients is 32, the inter MAC unit is 64 bits. Each of the 8 channels can have its own unique coefficients. ThIs is ideal for making an active Xover that will not have an phase problems. (well other than the speakers phase responce itself). We are almost done with a new GUI that can be used if you have a demo board that actaully has a graphical filter builder built into it. Soon..



Dustin
 
Re: Re: Re: Re: Re: Re: Es9018

dusfor99 said:
Hi Cauhtemoc

THe chip will now autodetect if the input stream on data1 is SPDIF or not. If its not, it will try to see if its I2S or DSD. SO there is not need for a MCU to program it to get it into SPDIF mode.

Dustin

Thanks. This is exactly what I had hoped for, and it's a feature that should clearly have been in the old DAC.

What can you tell us about the new ADC? Could you tell us the DNR for example, as this is something that appears to be missing from the product brief on your website. It also appears that the ADC does not require a microcontroller, and all configuration is done via pin-strapping, is this correct?
 
Re: Re: Re: Re: Re: Re: Es9018

dusfor99 said:




Hi Peter,

The bit lenght for the coefficients is 32, the inter MAC unit is 64 bits. Each of the 8 channels can have its own unique coefficients. ThIs is ideal for making an active Xover that will not have an phase problems. (well other than the speakers phase responce itself). We are almost done with a new GUI that can be used if you have a demo board that actaully has a graphical filter builder built into it. Soon..



Dustin

Hi Dustin,

How long is the FIR filter (or is it IIR filter?)

Where can I download a product datasheet of this new chip with all the details for design. Or how can I get it or sign to get it.
Is also data of the new ad converter available?

Some time ago I did sent an email to the dutch/belgium representative but did not get an answer yet. Someone told me they are not very active. I do not know if this is true, but anyhow I did not get an answer.

Ronald
 
Just completed most of the support PCB for my Buffalo. One LCDPS for the digital supplies, two LCBPS (my own PCB and parts choice) for the upcoming CounterPoint ver2, the digital input PCB with AES/EBU, True 75 ohms BNC and optical SPDIF using the TORX142L (192Khz speed support). The RCA is not mounted because I'll use the USB input card for the forth input into the CS8416 MUX.

There is also the small Simple Reg PCB for the Buffalo analog reference modification. The MUX is all assembled as well as the Buffalo. I just need to complete the USB card and wait for the Counterpoints.
 

Attachments

  • all support pcb_s.jpg
    all support pcb_s.jpg
    81.9 KB · Views: 1,323
How long is the FIR filter (or is it IIR filter?)
In fast roll-off mode: The filter is a two stage FIR. The first stage is 512 taps (cosine symmetry). The second stage can be up to 28 taps (either sine or cosine symmetry).

In slow roll-off mode: The filter is now a single stage FIR with 64 taps (no symmetry required).

Where can I download a product datasheet of this new chip with all the details for design. Or how can I get it or sign to get it.
You'll have to get in touch with marketing at ESS for a full data sheet. Alternatively, the new Sabre32 GUI help files contain all of the register definitions (if that's what you're looking for). An alpha version is available on the website. http://esstech.com/index.php?p=support_downloads
 
I downloaded the new GUI, but it looks like I can't play with the filter designer unless I have a new demo board attached.
Hi rossl, there's a new version posted that has the filter designer as a stand-alone application. You should now be able to use it without a demo board attached.

http://esstech.com/index.php?p=support_downloads

Does that mean that we now have another source of insider information?
I don't know if 'insider information' is the right phrase ;)
 
Hi Russ. I assembled the Simple Reg modification PCB for the Buffalo and I'm testing it right now. In your latest schematic with notes, you mentioned that the Vref can be changed by changing the R3 value.

Reference to the enclosed schematic with my measurements results.

I can only get 2.8V at Vout, changing R3 cause small variation from 2.75 to 2.85, but that about it. All the other currents you recommended (R4 > 4ma, R6 > 1ma, R1 >4ma) are meet. The two leds reference source D1-D2 (green leds) is 3.9V. The way I see it, R2,R3 form a simple voltage divider, not loaded by the high input impedance of IC1, so the ratio of Vled X (R3/R2+R3) should fix the Vref value. I should be able to adjust R3 to get Vref 3.3V but I cannot. By the way I'm using the circuit with jumper (short) used for Re and Rf.

Any suggestions?

Is the value of 2.8V enough for the Vref value on the Buffalo PCB, or is it too low?

Thanks for any help you can provide.
 

Attachments

  • simple reg_with voltages.jpg
    simple reg_with voltages.jpg
    35 KB · Views: 2,010