New ESS Sabre32 (ES9018) based DAC from Audio-gd....the NFB-1 - Page 37

Quote:

Originally Posted by glt 

 

Quote:

Originally Posted by audio-gd 

Quote:

Originally Posted by borrego 

Quote:

Originally Posted by tim3320070 

Why not get the DI for the NFB-1, it's only $60 more over the built-in version?

For one, the additional I2S input capability on the digital board. One might be able to tap wires to it and connect the I2S output from a DAP. This is of course provided Audio-gd has fixed the I2S / Sabre32 clock issue.

The problems is on the ES9018 , while it accpet the I2S, it had very rigorous request the I2S signal because its I2S input pots are without the master clock input.

I asked an engineer work for  another manufacturer, he said in their gears , the I2S input will change to SPDIF then feed to the ES9018.

So I don't think the ES9018 can support I2S input as well.

But we have find the way to fix this problems. 

 The ES9018 must need  the input master clock and built in a master clock selector can automatically choice between the I2S master clock and the 80MHz main clock.
 

What is the problem with the I2S exactly?

 

While interfacing a Musiland device to the Buffalo DAC, I experienced dropouts with I2S. The drop outs went away after increasing the DPLL bandwidth. One would think it is because of jitter. The funny thing is that with SPDIF input, you can use the lowest DPLL bandwidth setting with no drop outs whatsoever.

 

The same problem (drop outs in I2S and no drop outs in SPDIF) has been reported for the TI PCM2707-based TPA USB interface and also the Teralink USB interface. The only reported exception is the HiFace.

 

I have programmed the chip where I can select the bandwidth of the DPLL, so for me it is easy as I just turn a knob, but it can also be done automatically by detecting lock errors and backing out on the dpll bandwidth setting until no lock errors are experienced.

We also have the software for the ES9018 once we start test the ES9018. The software can doing anythings with the ES9018.

We have try to adjust the DPLL to max but only slight improve.

And we also resetting the Registers 11 address, from 0x85 change to 0x9D, but also slight improve and can't fully fix.

But if change the Registers 11 address, the ES9018 working a bit steady but sound quality slight degrade, maybe just my mentation.

Hmmm interesting.  (Also I set reg 25 to "allow all settings")

In my experiment, I didn't have to use "highest". "Medium" bandwith eliminated the dropouts. Another thing I did was to switch to manual spdif (reg 17) and 24 bit I2S (reg 10). I don't know if reg 17 and reg 10 would make any differences.

In theory, higher bandwidth allows more jitter to go through, so ideally "lowest" bandwidth is best, and in my case this means spdif input.

Edited: had an error in setting register 11 that I removed.

Accept I2S with the master clock, there is without any problems. But too bad the ES9018 only accept higher than 192X master clock while I2S input mode.

In this mode, the ES9018 can setting at the highest performance mode.

Some time I think the ES9018 just a beta version, ESS maybe consumate it become perfect at future.

Quote:

Originally Posted by Kingwa 

Accept I2S with the master clock, there is without any problems. But too bad the ES9018 only accept higher than 192X master clock while I2S input mode.

In this mode, the ES9018 can setting at the highest performance mode.

Some time I think the ES9018 just a beta version, ESS maybe consumate it become perfect at future.

Yeah, too bad. Seem to be some bug in the I2S interface. I appreciate you sharing the information. Merry Christmas!

There has also been reports of users of the Hiface EVO having problems outputting its I2s signal into Sabre DACs...they were blaming the HiFACE EVO as an unfinished product.  So it turns out the Sabre was at fault.

Interesting turn of events...., the Sabre chip plot thickens.

ESS seems to have a pretty fast product cycle, I wouldn´t be suprised if there is a 9018 turbo edition out next year or so :) Seems like a fine DAC chip though, too bad the I2S side isn´t ready for business.

While true, a user on TPA from Korea states that it was solved. It was indeed user error and perhaps some blame on the Evo as well.

Linky:

http://www.twistedpearaudio.com/forum/default.aspx?g=posts&t=1260

WOW, it is gone.!!!!! Discontinued!! 

go check audio-gd

that is fast!!!

Where did you see that?

Working perfectly or not, who ever got these Sabre Dacs from AGD now have some rare items.

The big question:

What is next?

Rare as in obsolete and not serviceable.

Audio GD is hurting their cause by creating so many rare products in such a short time. Perhaps they should learn the current cutting edge design and strike their own path.

Put out some industrial statements we can live with for a while.

You have a valid point....

Short product life cycles is not a sign of stability in a company.  However a one-off great product is still nice to have.

So are they going Wolfson in the end or back to PCM1704?

I wouldn't mind if they go back to Reference series.

Or maybe they have new chip to build DACs with? Or they may didn't get enough customers passion for the NFB series, and want to go back to R2R.