ESS THD ‘Hump’ Investigation

[maxxevv]

I have a reflow oven and heat gun; the problem is to get hold of the DAC. It's available at Alibaba, but only in quantities of ten.

This seller seems to do single piece orders though.

https://www.aliexpress.com/item/new...lgo_pvid=e404ebfe-192d-4be3-9fbe-9dd11f0e2d0c

 

[VintageFlanker]

So the hump is only for 9038 generation chips?

9038 Q2M I think. The Oppo UD-205 with 9038 PRO does not have the hump.

 

[mahirzukic2]

9038 Q2M I think. The Oppo UD-205 with 9038 PRO does not have the hump.

The guy in the previous page just explained it that PRO version is not affected by this hump, meaning this only relates to the non-PRO versions of the ESS DAC chip.

 

[Veri]

The guy in the previous page just explained it that PRO version is not affected by this hump, meaning this only relates to the non-PRO versions of the ESS DAC chip.

Yeah well the SMSL D1 has it and it uses two 9038 Pro chips. Also the Oppo also shows the hump, it's just the least "humpy" variant measured so far. You can still see the upwards slide, albeit slight.

 

[jackenhack]

This seller seems to do single piece orders though.

https://www.aliexpress.com/item/new...lgo_pvid=e404ebfe-192d-4be3-9fbe-9dd11f0e2d0c

Thanks for the heads up. Will order a couple when my cash flow is in better shape.

 

[la2ygoo]

TOPPING NX4 DSD & DX7s both use es9038q2m，nx4 do not have THD ‘Hump’,but dx7s have.（the offical pic）
it is hard to understand.

Is that mean NX4 do not have imd ‘Hump’ too？

 

[caguilar91]

TOPPING NX4 DSD & DX7s both use es9038q2m，nx4 do not have THD ‘Hump’,but dx7s have.（the offical pic）
it is hard to understand.

Is that mean NX4 do not have imd ‘Hump’ too？

On the NX4 DSD x-axis shows Vrms and on the DX7s x-axis shows dBFS, different measuring criteria in the end.

 

[amirm]

The Topping DX7 (note: no "s") did not have this either: https://www.audiosciencereview.com/...s-of-topping-dx7s-dac-and-headphone-amp.2378/

Indeed that is the first time I discovered this issue. I think the DX7 used the ES9018K2M?

Hard to imagine ESS solving this problem then losing it.

 

[la2ygoo]

[QUOTE =“caguilar91，post：134749，member：4189”]在NX4 DSD x轴上显示Vrms，在DX7s x轴上显示dBFS，最后有不同的测量标准。[/ QUOTE]

the dx7s pic，blue line mean 0dbfs = 13Vrms, -10dbfs = 4Vrms, -20dbfs = 1.3Vrms,-30dbfs = 0.4Vrms , and so on

 

[caguilar91]

[QUOTE =“caguilar91，post：134749，member：4189”]在NX4 DSD x轴上显示Vrms，在DX7s x轴上显示dBFS，最后有不同的测量标准。[/ QUOTE]

the dx7s pic，blue line mean 0dbfs = 13Vrms, -10dbfs = 4Vrms, -20dbfs = 1.3Vrms,-30dbfs = 0.4Vrms , and so on

I simply wonder if it might "soften" using the Vrms vs dBFS, with the NX4 DSD reaching some 4.5 Vrms prior to clipping. Seems to me it shouldn't be that much different but well, sometimes a slight difference with measuring parameters might show some nicer graphs.

Also, even though the hump it seems preferable the one with the DX7s, might have to do when reaching the lowest (best the DAC can give) that THD+N, IMD and stuff get humpy. And yet that ES9018 from the DX7 seems better than the ES9038 from the DX7s at least in that graph above.

 

[amirm]

As an added data point, here is the IMD of the Melokin DA9.1 ES9038Pro:

Doesn't seem to have the ESS Hump.

 

[amirm]

Spoke too soon. Looked up the Melokin review and realized that one channel has the hump. Here is the update measurement:

So we have the identical DAC but one channel has the hump and the other does not. Seems to be a circuit issue, not DAC.

 

[gvl]

So we have the identical DAC but one channel has the hump and the other does not. Seems to be a circuit issue, not DAC.

A counterfeit 9038Pro would explain the missing hump and the so-so measurements

 

[Veri]

A counterfeit 9038Pro would explain the missing hump and the so-so measurements

Would seem crazy to have these things counterfeit on the market honestly

 

[Krunok]

Indeed that is the first time I discovered this issue. I think the DX7 used the ES9018K2M?

It did, one per each channel.

 

[Pluto]

Would seem crazy to have these things counterfeit on the market

While my knowledge of the state of the market in counterfeit semiconductors is hardly ‘in depth’, there might be scope for crooks re-marking cheap devices as expensive ones. Obviously, this can only work when there is total compatibility between the two, and the difference in performance is subtle and difficult to measure (i.e. hump or no hump).

Key question – what is the cost differential between the standard and ‘pro’ versions of the device in question? Sufficient to make it worthwhile for crooks to fake the latter from the former?

 

[rajapruk]

Spoke too soon. Looked up the Melokin review and realized that one channel has the hump. Here is the update measurement:

View attachment 19807

So we have the identical DAC but one channel has the hump and the other does not. Seems to be a circuit issue, not DAC.

Or could it be unit-to-unit-variations from the production? (incl channel-to-channel-variations)

 

[miero]

Will they fix the hump?

The ES9068Q will offer Best-in-Class performance based on the SABRE DAC architecture with HyperStream® II modulation and can be used as a functional upgrade to the successful ES9018K2M, ES9028Q2M, and ES9038Q2M products. The ES9068Q provides a Dynamic Range of 128dBA and total harmonic distortion of -120dB THD+N. The DAC can support a wide range of input formats including: PCM up to 32bit 768kHz, DSD1024, DoP, SPDIF, and MQA.

More: http://www.globenewswire.com/news-r...CODEC-with-Hardware-MQA-Hi-Res-Rendering.html

 

[maxxevv]

That means an ES9068PRO is somewhere in the pipeline as well.

 

[Veri]

Will they fix the hump?

More: http://www.globenewswire.com/news-r...CODEC-with-Hardware-MQA-Hi-Res-Rendering.html

38- to 68-? thanks for sharing!