On the Distortion of Cirrus Logic CS431xx-Based Devices: A Comparative Review

[mc.god]

TRUTHEAR released a dongle with PEQ for US $30.

TRUTHEAR KEYX CS43131 Portable USB DAC/AMP

TRUTHEAR KEYX CS43131 Portable USB DAC/AMP

shenzhenaudio.com: professional headphone, earphone, amplifier, dac, decoder, hifi player, cable supermarket. you will find good price here.

shenzhenaudio.com

Since it uses the CS43131, it is likely to have the "Cirrus Hump," but it seems like a good option because its EQ app (based on CrinGraph) looks easy to use.

I like the form factor, small, angled and no breaking cable, nice product.
If i have to say, I wouldn't be surprised if they managed DRE issue in some way at the source, without leaving the option to the end user, I'd like to see related measurements.

 

[fws]

I like the form factor, small, angled and no breaking cable, nice product.
If i have to say, I wouldn't be surprised if they managed DRE issue in some way at the source, without leaving the option to the end user, I'd like to see related measurements.

Second that. Also specs for 32Ohm are for 1Vrms, but the power is listed as >=50mW, so we could use a clarification on whether CS43131's impedance sensing is working or defeated.

 

[Somafunk]

It seems like a good idea but the extra strain being put through the USB c port would be a worry,

 

[VientoB]

It seems like a good idea but the extra strain being put through the USB c port would be a worry,

I agree. It's going to act like a lever.

 

[mc.god]

It seems like a good idea but the extra strain being put through the USB c port would be a worry,

maybe yes or maybe not, it also depends on the way you normally use the phone. Usually I have it upside down into a purse, with usb-c connector facing upward, so with standard dongles I have a straight connector and a cable hanging from it under the weight of jack plug and headphone plug inserted in it. So I'm using an L-shaped adapter like this so that the dongle cable doesn't bend and can rest on the zipper of the purse, relieving the connector from some strain, or at least i feel like this.
With this Truthear dongle I'll have a similar setup, without the annoyance of an eventually failing cable.

 

[pangcukaipang]

TRN released Wave with dual CS43198, dual RT6863 (sigh), and bluetooth. It's only $50 at local store here.

 

[VientoB]

Bluetooth 6.0 no less

 

[CedarX]

TRN released Wave with dual CS43198, dual RT6863 (sigh)

Dual RT6863… or single opamp for a dual-CS43198 > 3.5 SE?

 

[pangcukaipang]

Dual RT6863… or single opamp for a dual-CS43198 > 3.5 SE?

"TRN Wave is a portable high-definition audio decoder and earphone amplifier featuring the Qualcomm QCC3095 chip paired with dual CS43198 DACs and a BT6863 op-amp architecture."

Hmmm, the description indeed said "a".

 

[RadioRat]

I’m on a hunt for a portable DAC that can actually handle the HD600 and other headphones with a heavy parametric EQ profile. I need more headroom than the Qudelix 5K provides, but my journey through higher powered portables has been a nightmare of software bugs and artifacts.

The Failed List:

* FiiO KA15: Constant software instability, shelving filters didn't function correctly.
* Tanchjim Space Pro: Clean, but the 10 band PEQ lacks shelf filters, making it useless for my sub bass profile.
* iBasso DC04 Ultra: Plenty of power, but major issues. With DRE enabled, I get clicking/clipping on percussive elements. With DRE disabled, the noise floor is audible and distracting.

Needing:

1. Stable 10 band PEQ: Must include functioning high/low shelf filters.
2. Power: Needs to drive 300Ω headphones with enough headroom to handle a significant negative pre amp gain to avoid EQ clipping.
3. Clean Noise Floor: No DRE artifacts or background hiss.

Is there anything that actually works as reliably as the Q5K but with more juice, or is the DSP high power dongle world just not ready yet?

 

[reallystate]

@RadioRat
I recommend reading this article. It explains in detail how to choose an amplifier for high-impedance headphones. Since you plan to use an equalizer, add a few more decibels to the recommended volume levels to account for your preamp.

Keep in mind that while the shelf filters in FiiO devices don't work exactly correctly, it's known how to make them work predictably. To do this, simply multiply the Q-factor of such filters by the square root of two.

 

[Alou]

I have both the KA15 and the DC04,they don't have any problems at all or any noise artifacts.Everything works as it should.I only find that the KA15 amplifies the sound a bit more like a loudness plugin,turning off gain and going to H mode sounded more normal but this also personal taste.

 

[Klingsor]

Is it possible to change registry settings with WISCE software? So that I can change ADPT_PWR = 111 into ADPT_PWR = 001?

 

[Klingsor]

I don't understand why enabling the DRE mode should be responsible for the distortions. The root cause seems to be though the switching power supply rails of the charge pump.

 

[tccalvin]

I have both the KA15 and the DC04,they don't have any problems at all or any noise artifacts.Everything works as it should.I only find that the KA15 amplifies the sound a bit more like a loudness plugin,turning off gain and going to H mode sounded more normal but this also personal taste.

I think @RadioRat hears noise and artifacts because they lower the gain to make sure their filters don't clip. Still, they shouldn't be that audible.

I don't understand why enabling the DRE mode should be responsible for the distortions.

We don't understand why either. But it does. Turning off DRE fixes the problem, even though the PSU stays the same.

Is there anything that actually works as reliably as the Q5K but with more juice, or is the DSP high power dongle world just not ready yet?

Take a look at the Fosi DS3 and the NiceHCK Octave. There are no independent measurements AFAIK, but the ES9039Q2M is a very, very solid chip.

 

[Klingsor]

We don't understand why either. But it does. Turning off DRE fixes the problem, even though the PSU stays the same.

Maybe turning DRE off (and also enabling NOS) consumes less power.

Compare the graph and table 4-2 of the datasheet:

 

[gtg465x]

I’m on a hunt for a portable DAC that can actually handle the HD600 and other headphones with a heavy parametric EQ profile. I need more headroom than the Qudelix 5K provides, but my journey through higher powered portables has been a nightmare of software bugs and artifacts.

The Failed List:

* FiiO KA15: Constant software instability, shelving filters didn't function correctly.
* Tanchjim Space Pro: Clean, but the 10 band PEQ lacks shelf filters, making it useless for my sub bass profile.
* iBasso DC04 Ultra: Plenty of power, but major issues. With DRE enabled, I get clicking/clipping on percussive elements. With DRE disabled, the noise floor is audible and distracting.

Needing:

1. Stable 10 band PEQ: Must include functioning high/low shelf filters.
2. Power: Needs to drive 300Ω headphones with enough headroom to handle a significant negative pre amp gain to avoid EQ clipping.
3. Clean Noise Floor: No DRE artifacts or background hiss.

Is there anything that actually works as reliably as the Q5K but with more juice, or is the DSP high power dongle world just not ready yet?

For the KA15, can't you just multiply Q by √2 on the shelving filters to get them to function as expected?

And for the DC04 Ultra, isn't there a third DRE mode called "Optimized"? Presumably that one should fix both the clicking and the noise floor.

 

[Gracchiolo]

I don't understand why enabling the DRE mode should be responsible for the distortions. The root cause seems to be though the switching power supply rails of the charge pump.

I doesn't. You get a distortion measurement because you measure distortion, but it is misleading. It is a time domain defect. See post #37. It produces a "click" when the class G amplifier switches power supply voltage. The software sums up the harmonics different than 1 KHz and tells you the distortion is xx% which is correct, but it can't tell you that that number is actually a form of noise.

 

[tccalvin]

I doesn't. You get a distortion measurement because you measure distortion, but it is misleading. It is a time domain defect. See post #37. It produces a "click" when the class G amplifier switches power supply voltage. The software sums up the harmonics different than 1 KHz and tells you the distortion is xx% which is correct, but it can't tell you that that number is actually a form of noise.

Yeah. It's not harmonic distortion. It would be more accurate to say that the noise level gets extremely high for a split second, AKA impulse or click. To be fair, the measurement does say THD+N. In a broader sense, you could call it non-harmonic distortion.

 

[Klingsor]

The Class H power supply mode (register setting ADPT_PWR = 111) is a power saving mode. It is made for mobile devices to consume less power. The noise will disappear if you change Class H mode into Class AB mode (register setting ADPT_PWR = 001).