ESS ES9039PRO and ES9027PRO announced

[Matias]

New flagship DAC from ESS announced, now with MQA supported variants...

"The ES9039MPRO & ES9039PRO provide 8 channels of Flagship ultra-low noise digital conversion with a dynamic range (DNR) of +132dB, -122dB THD+N per channel and +140B DNR in mono mode.

The ES9027MPRO & ES9027PRO also provides 8 channels of digital conversion with a dynamic range (DNR) of +124dB per channel and -112dB of THD+N."

ESS Technology Introduces New Flagship Performance Audio Digital-to-Analog Converters – ESS Technology, Inc.

www.esstech.com

 

[muslhead]

These continued improvements they are making are impressive.
Its gotta be partially design and partiall fab. Since they are not public and the info is not readily (at least that i could find) does anyone know if they are using TSMC?

 

[boXem]

At a first glance, es9027 THD is meh, ES9038Pro and es9039pro have the same performance. So the main progress would be power consumption.

 

[Music1969]

Infamous ESS IMD hump has been around long enough for ESS to see it discussed.

Hopefully they fixed that crap.

 

[Snoopy]

Nice.. time to wait for a whole bunch of DAC refreshes from topping, smsl, Gustard etc

 

[kchap]

It's probably only a few dollars, but if purchase a DAC containing these chips I will be paying for an MQA license I do not need.

 

[MCH]

Nice.. time to wait for a whole bunch of DAC refreshes from topping, smsl, Gustard etc

Aren't these multichannel? I wish but don't have many hopes topping going there....

 

[Matias]

It's probably only a few dollars, but if purchase a DAC containing these chips I will be paying for an MQA license I do not need.

There are variants with and without MQA support for each chip, so it is up to the manufacturer to choose.

 

[Matias]

Aren't these multichannel? I wish but don't have many hopes topping going there....

Sabre DACs have been multichannel for a long time, at least since ES9018 which has been 10+ years now. The manufacturer can use the parallel decoding to sum and get lower noise and distortion.

 

[maxxevv]

Sabre DACs have been multichannel for a long time, at least since ES9018 which has been 10+ years now. The manufacturer can use the parallel decoding to sum and get lower noise and distortion.

Aren't these multichannel? I wish but don't have many hopes topping going there....

He meant building a DAC with more than 2 channels by Topping since these chips can support up to 8 channels. Something like the Okto 8 DAC with 8 channels would be nice.

But I think a superb performing 4 channel one using 2 channels to sum for each output channel would be highly desirable too. Would make for a nice 2.2 or 3.1 or even a quadrascopic 4 speaker system using active speakers and woofers.

 

[Matias]

Ah, got it. Something like the newly announced RME ADI-2/4 Pro SE, with independent outputs. Topping could have done that with previous generations of ESS DAC chips but they have been hesitant to do so for whatever reason until now. But I agree, hopefully they start getting more into DSP and multichannel.

 

[MCH]

Just a Motu M4 of sorts, but a la topping, + with spdif in and remote control at a similar price point.... (dreaming awake...)

 

[syn08]

Not interested, unless they provide a decent data sheet. I’m sick of the ESS “normal” data sheets, with errors, confusions, incomplete information, all after signing NDAs like for DoD documentation.

I’m fighting the ESS9822PRO ADC chip for over a month now, and I cannot make it work. Hardware is per the data sheet, configuration parameters is per the example in the data sheet (and some data I got directly from ESS, different, of course), hardware wise everything to the last detail seems ok, though the beast spits out only noise floor at some -160dB plus some -100dBFS when input is 1.5Vpk differential on 2.2VDC common mode. I tried a gazillion of parameter combinations, no changes, same behaviour. The data sheet fails to explain the most basic concepts, e.g. what are the chip “common modes” (there are 3 of them), the default values for some registers are different in the register definition and in the register field description, etc... If somebody wanted to create a charade it could not do a better job, and this mess is for the latest data sheet version 0.3, which is called “vastly improved”. BTW, I purchased the chips directly from ESS, I wonder if they didn’t sent me some pre production buggy silicon?

And yes, I am frustrated to the point I’m ready to send a nuke in San Jose. I’m afraid I’ve had enough of this BS company, I hope AKM or anybody else will be back and put these clowns out of their misery.

 

[JSmith]

It's probably only a few dollars, but if purchase a DAC containing these chips I will be paying for an MQA license I do not need.

We all pay regularly (embedded in pricing) and have for years for codec licensing that may or may not be used... think Dolby. I don't see how that is any different. I'm not an regular MQA user, but would rather have more format support than not in my devices.


JSmith

 

[JSmith]

I hope AKM or anybody else will be back

AKM is back.

Audio D/A Converters | Audio Components | Products | Asahi Kasei Microdevices (AKM)

AKM’s audio D/A converters (DACs) provide solutions for high quality audio system with the original digital-analog mixing technology.

www.akm.com

New AKM chips coming.

I guess they've recovered from the factory fire. Time to break new barriers in the SINAD race! https://audioxpress.com/news/akm-develops-advanced-velvet-sound-ak4493s-and-ak4490r-2-channel-premium-audio-d-a-converters

www.audiosciencereview.com

AKM restarting production

https://audioxpress.com/article/the-return-of-velvet-sound-and-news-about-self-powered-speakers

www.audiosciencereview.com

JSmith

 

[Chrispy]

There are variants with and without MQA support for each chip, so it is up to the manufacturer to choose.

So pricing isn't relevant between these?

 

[TheBatsEar]

We all pay regularly (embedded in pricing) and have for years for codec licensing that may or may not be used... think Dolby. I don't see how that is any different.

Dolby adds value.

 

[syn08]

For future reference, the ES9822PRO ADC CHnB n=1,2 channels need to be digitally inverted in the register 0x87 and 0x98 (ADC CHnB DATAPATH CONTROL). Otherwise, the differential digital signals are cancelling each other and the output result is (of course) noise. The default is (what would you think)... not inverted. Why, your guess is as good as mine.

Thanks @phofman for spotting this possible root cause; this information (and God knows what else) is nowhere in the data sheet, and these settings are also not in the sample code in the data sheet, and that I got directly from ESS.

It may sound cryptical for those not involved in building something, but for those that do, I guess it is clear why I'm pissed at these guys and won't further touch any of their products with a 10 foot pole, unless they radically change at least in this (documentation) respect. I checked with AKM, they are not perfect (performance and documentation wise), but they have a very friendly and responsive local representative (helped me a lot before the fire) and the new line of (revived) devices seem competitive.

 

[phofman]

and these settings are also not in the sample code in the data sheet

Well, I did copy them from the example on page 119 of the public version 0.3 of the datasheet

//ADC filter and datapath registers
w 113 0x98; //ADC1A_FILTER_SHAPE = Minimum phase slow roll off
w 118 0x01;
w 130 0x98; //ADC2A_FILTER_SHAPE = Minimum phase slow roll off
w 135 0x01;

Honestly I do not understand their effect since they should apply to outputs of ADC3&4 (wired to CH1 and CH2 pins of the ES9822 package) which should not be used when the 4 ->2 mix is disabled by default, IIUC.

 

[syn08]

In the same ES9822PRO data sheet code example:

w 147 0x98; //ADC1B_FILTER_SHAPE = Minimum phase slow roll off
w 164 0x98; //ADC2B_FILTER_SHAPE = Minimum phase slow roll off

register 147 is Reserved (default value is 0x80, read it myself) and of course undocumented, while register 164 doesn't exist, I2C pukes when addressing it.

I'm about to think this data sheet was put together by some illegal Mexican immigrants, to save a few pennies. Support activities were long time cut. Next step would be ESS to hire some south east asian workforce, to manually carve the chips instead of running through TSMC batches.

-125dB THD @1KHz before H2 and H3 tuning down, BTW.