• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

AliExpress Cirmech WM-013 CS4398 DAC - review and measurements

I made some more tests, BTW jitter test result is not bad at all.

Cirmech_jitter.png

So, we have low enough noise, good enough dynamic range, and quite high nonlinear distortion, especially at higher frequencies. It is a kind of fun to make a comparison with $10,995 CD-Media player and D/A processor AVM Ovation MP 8.2. Which has neither lower THD, nor CCIF IMD. The "High-End" world is funny.

AVM Ovation MP 8.2 CD player-D/A processor Measurements | Stereophile.com

Sidebar 3: Measurements I measured the AVM Ovation MP 8.2 with my Audio Precision SYS2722 system (see the January 2008 "As We See It"), using both the Audio Precision's optical and electrical digital outputs and USB data sourced from my MacBook Pro running on battery power with Pure Music 3.0...
www.stereophile.com www.stereophile.com
 
I'm ordering one of these DACs

Is the PSU in the preamp box in post #25 available somewhere? Aliexpress?

Many thanks!
 
  • Like
Reactions: pma
I used to like taking the output of the akm4398 directly in to audio transformers. Gave a nice clean quiet output with auto Se / Bal. Neutrik NTL-1 is very nice. Great simple mod for those who are not so familiar with electronics - just 2 wires in, 2 wires out. Yeah it's expensive, but it has super low distortion, zero self noise and no worries about op-amp stability, capacitor dielectric or ground loops.
 
Last edited:
Last edited:
@pma , what percentage of music you listen, is in 24bits to give so much importance how it measures?
 
IVX said:
oh, there is some old-school muting that also may add tons of distortions. CS43198 has muting too, hence, LPF-opamp is really redundant ;)
Click to expand...
You seem to keep mixing up the CS4398 and the CS43198.;)
Very different parts!

Yes, the CS43198 does not need an op-amp output stage. But the CS4398 needs it, unless you use transformers or something else.

I agree with pma that the distortion is related to the capacitors used. A strange design decision, since most of the capacitors at these values in NP0/C0G should be below 1 cent each in volume.
The components seem to be 0603, so it should be possible to replace them with NP0/C0G in the same size, perhaps with the exception of the 18 nF capacitors. That one is easier to get in 0805.

But of course it requires the right equipment and a steady hand. I find that a microscope is often a useful tool.
 
pma said:
@ssashton : as you know, nothing is perfect.



Like with every link level transformer, distortion gets worse at low frequencies and higher voltage level. And like everything, it needs qualified implementation. And is a little bit out-of-scope the $18 DAC.
Click to expand...
Interesting to see those graphs - they do not match my results at all. I'll see if I can generate a quick measurment later.

I agree the cost doesn't make much sense these days, given for 2 of those transformers you can probably buy a DAC from SMSL or similar. Go back 10yrs and the transformers were £30 each and the budget DAC market far less competitive, while AK / CS 4398 was current..

EDIT: Ahh here you are - 1V rms (same as -0.59dbV in the plots you found).

NTL-1 Transformer Distortion.png



Me: "What is the launch date of the CS4398?"
ChatGPT: "The CS4398 datasheet shows “Copyright 2005–2021 Cirrus Logic” in its header, implying the design dates back to around 2005.

20 Years! I'm older than I realised :-S
 
Last edited:
JensH said:
You seem to keep mixing up the CS4398 and the CS43198.;)
Very different parts!

Yes, the CS43198 does not need an op-amp output stage. But the CS4398 needs it, unless you use transformers or something else.

I agree with pma that the distortion is related to the capacitors used. A strange design decision, since most of the capacitors at these values in NP0/C0G should be below 1 cent each in volume.
The components seem to be 0603, so it should be possible to replace them with NP0/C0G in the same size, perhaps with the exception of the 18 nF capacitors. That one is easier to get in 0805.

But of course it requires the right equipment and a steady hand. I find that a microscope is often a useful tool.
Click to expand...
You are 100.00% right ;) 4398 is a dino from 1990th, it is the same but without HP buffer and no DRE.
 
ssashton said:
Me: "What is the launch date of the CS4398?"
ChatGPT: "The CS4398 datasheet shows “Copyright 2005–2021 Cirrus Logic” in its header, implying the design dates back to around 2005.
Click to expand...
Or, you know, you might just consult the table compiled by someone who's done some poking around with the trusty Internet Archive, and find out that the chip was released in March of 2003. Only a year later the first reviews for the E-MU 1212M using it came out. (If you ever wondered where the legacy of those now-legendary interfaces went post mid-late 2000s, the people working on them seem to have ended up at Universal Audio.)

It wasn't the first 120 dB class DAC for Crystal / Cirrus Logic, being preceded by CS43122 and CS4396/97, but their first mass-market one for sure. That it's still in any way relevant to the mass market is remarkable, and we can assume that there has been a fair bit of work behind the scenes to keep it available over the years. It's still not the longest-serving DAC chip ever (CS4334/5/8/9 made it to 24 years before being retired in 2022, and several Analog Devices multibit R2R DACs are still active including the 1988ish AD1856), but it's getting up there.
 
Trick to mass market adoption: Be cheap (relatively), be easy to implement (relatively), be good enough. It's relatively common to get 0.001% - 0.002% level of performance out of this IC, and there isn't really any real practical reason to need higher than that. Inside integrated amps as an afterthought, inside Xonar DX... where you need something that's good enough to justify its existence while not willing to needlessly spend a fortune on those AD and Burr-Brown.

1760234734924.jpeg


You know the DAC's old when people still thought big XO = better sound.

The time for CS4398 will come (as in die off) when as-cheap-as-possible solutions start giving better performance, and we have reached that point with CS43131 and Sabre DACs in the dongle space right now. If the board costs like $15 the enclosure usually costs just as much so let's say $30 (and I'm generously ignoring the price of the power transformer)...

This thing is going for $40 on Amazon US right now powered by ES9028
www.audiosciencereview.com

NEOHIPO DC10 Stereo DAC Review

This is a review and detailed measurements of the NEOHIPO DC10 budget stereo DAC. It was sent to me by the company and is on sale for US $48. Suggested retail price is $80. The two switches feel solid and frankly all that is needed in a simple DAC. Back panel follows the same rule: Let's...
www.audiosciencereview.com www.audiosciencereview.com

It's still more expensive than dirt cheap but having both SPDIF and USB inputs costs money.

Measures better? Yea. Necessary upgrade? Nah. I'll be running my CS4398 until the integrated amp it lies inside of eventually dies. Which may be never. Or until I migrate and sell that chonker off.
 
  • Like
Reactions: pma
ssashton said:
Interesting to see those graphs - they do not match my results at all. I'll see if I can generate a quick measurment later.
Click to expand...

They had old equipment. Please do not show 1kHz plot, but 50Hz plot.

Below measurement of Neutrik NA2M-D2B-TX adapter that uses NTE1 transformer.

NA2M_THDfreq_-3dBu.png



However, link transformer topic is OFF-Topic in this thread.
 
AnalogSteph said:
Or, you know, you might just consult the table compiled by someone who's done some poking around with the trusty Internet Archive, and find out that the chip was released in March of 2003.
Click to expand...

Am I the only one who doesn't know of this table? Why didn't anybody tell me?!
 
ssashton said:
FWIW NTE-1 is a totally different animal than NTL-1. I have both, NTE-1 has much higher distortion. I agree though, it's a bit off-topic
Click to expand...
Sure. However, low frequency distortion is an issue of every link level transformer. The only question is the voltage level when it starts to be a real problem. This is directly related to core size. Small core ==> lower voltage level where the low frequency distortion starts. BTW, I measured a custom made link level transformer with low low frequency distortion. Weight? Not much below 1kg. LF distortion - still nothing to call home about.
 

Attachments

  • trafo_mirek_thd_freq.png
    trafo_mirek_thd_freq.png
    86.5 KB · Views: 23
Last edited:
pma said:
Sure. However, low frequency distortion is an issue of every link level transformer. The only question is the voltage level when it starts to be a real problem. This is directly related to core size. Small core ==> lower voltage level where the low frequency distortion starts. BTW, I measured a custom made link level transformer with low low frequency distortion. Weight? Not much below 1kg. LF distortion - still nothing to call home about.
Click to expand...
Very nice result!

Since it is still on my desk, here is the frequency vs. disto for NTL-1.

Warm lows, clear highs? :)

1760270229302.png


I'm not saying transformers are the ultimate solution - only that it's a very easy way for beginners to upgrade gear that has issues such as the wrong capacitors or badly designed output stage without needing to solder lots of tiny SMD parts or try to hack in a new circuit without causing new ground loops etc.
 
Would not you like to use even bigger font size???

Back to the topic, below is a THD+N and THD measurements at 1kHz of a preamp which has this Cirmech DAC used as an S/PDIF input. It is quite acceptable and it benefits from higher impedance at the preamp input, compared to input impedance of the Cosmos ADC based measuring system. Please note the H3 dominance - a logic result of high K ceramic capacitors nonlinearity (capacitors used in the DAC output LPF).

Cirmech_preamp_THDNlevel_1k.png


Comparison of analog and digital inputs:

Cirmech_preamp_THDNlevel_1k_analog-digital.png


This is the preamp wiring diagram:

BUFPRE_wiring.png
 
Last edited:
You must log in or register to reply here.

Similar threads

pma
DIY preamp with multiple inputs - project
Replies
4
Views
845
pma
pma
pma
AliExpress cheap low distortion analog sine generator - review and measurements
2 3 4
Replies
76
Views
14K
pma
pma
R
Any Balanced S/PDIF DACs that are similar to TOPPING D10B, but S/PDIF instead of USB input?
Replies
8
Views
2K
rcstevensonaz
R
audio_tony
Measurements of LJM CS4398 DAC kit
Replies
4
Views
2K
audio_tony
audio_tony
NTTY
  • Poll Poll
SMSL D200 DAC Review
2
Replies
36
Views
14K
jmc06
J
Back
Top Bottom