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

[deton24]

JCally JM20/JM20 Max - “still produce distortion when multitone is increasing in a certain condition.”, but -
at least some versions starting from December 2025 are [allegedly] free of the clicking - Hitsujiomeguruboken
Can't remember when I saw that. Probably below one of sharur1's videos.

 

[KDRN]

is it just me or is the Jcally Jm20 had gotten harder to find. most of the stores in aliexpress are no longer offering the jm20's. in amazon, theres none. only the Jm20 Max are available in amazon. i wanted a cheap DAC to pair it with my L30ii and the jm20 is a good (enough) dac to pair it with.

 

[ATTI]

Now that I know how the Dune soundtrack would normally sound, I checked this issue on my TempoTec V1 which also use dual CS43131. When I use it as a DAP, playing the Dune soundtrack, this crackling noise issue happened.

Hi! No updates about firmware fixes in V1? If you are still using the device, is it very noticeable in regular use?

 

[tccalvin]

Hi! No updates about firmware fixes in V1? If you are still using the device, is it very noticeable in regular use?

I haven't listened to the TempoTec but I have tried multiple devices with a "bad" implementation of the CS43131. if you're just listening to audio (no crazy eq/dsp) you will not notice the clicks ~99% of the time.

 

[pangcukaipang]

Hi! No updates about firmware fixes in V1? If you are still using the device, is it very noticeable in regular use?

Gotta be honest, it's only noticeable on that Dune soundtrack. I 100% couldn't notice on any other tracks I listened to.

 

[Mixzvx]

About the PEQ. I saw on Reddit that the web app can unlock the hardware PEQ.You can try it.

https://www.reddit.com/r/iems/s/HdicmSN30p

 

[CedarX]

About the PEQ. I saw on Reddit that the web app can unlock the hardware PEQ.You can try it.

https://www.reddit.com/r/iems/s/HdicmSN30p

Unrelated with the WebApp (I’m sure it’s great ), the reference to Savitech is incorrect for these DSP-enabled chips.
But it doesn’t really matter…

 

[Mixzvx]

Unrelated with the WebApp (I’m sure it’s great ), the reference to Savitech is incorrect for these DSP-enabled chips.
But it doesn’t really matter…

I asked TTGK's customer service, and they told me that both Moondropdspmini and JM20Pro use the CB5100 chipset.I don't think he's making it up.

 

[CedarX]

I asked TTGK's customer service, and they told me that both Moondropdspmini and JM20Pro use the CB5100 chipset.I don't think he's making it up.View attachment 550897

I believe this is correct. TTGK / CBHT Audio chips are custom versions of the RISC-V based SPVxx chips from SpaceTouch (https://www.spacetouch.co).
Unless they recently purchased these companies, Savitech has nothing to do with this.
But again, not really important.

 

[Jalan]

is it just me or is the Jcally Jm20 had gotten harder to find. most of the stores in aliexpress are no longer offering the jm20's. in amazon, theres none. only the Jm20 Max are available in amazon. i wanted a cheap DAC to pair it with my L30ii and the jm20 is a good (enough) dac to pair it with.

JM20 seems still possible to find to me on Aliexpress but less cheap. That's the same for JM30 that is the same product with another case for what I know.

 

[RedCometZero]

Just wanted to weigh in on the Crinear Protocol Max, a high powered 3.5mm/4.4mm USB DAC with 10-band PEQ from Crinacle. It uses dual CS43198, and dual SGM op-amps.

It's a good looking device with a great feature set, but unfortunately it has the Cirrus hump and audible distortion at low volumes. It has a number of customizable settings (Class AB/Class H, Low/High gain, LP/MP fast/slow and NOS filters, and an Eco/Boost switch on the side), but the only setting that gets rid of the hump and the low-volume distortion is the NOS filter.

@crinacle, can you take a look and consider the firmware fix that others DAC companies have implemented to fix this?

 

[crinacle]

Just wanted to weigh in on the Crinear Protocol Max, a high powered 3.5mm/4.4mm USB DAC with 10-band PEQ from Crinacle. It uses dual CS43198, and dual SGM op-amps.

It's a good looking device with a great feature set, but unfortunately it has the Cirrus hump and audible distortion at low volumes. It has a number of customizable settings (Class AB/Class H, Low/High gain, LP/MP fast/slow and NOS filters, and an Eco/Boost switch on the side), but the only setting that gets rid of the hump and the low-volume distortion is the NOS filter.

@crinacle, can you take a look and consider the firmware fix that others DAC companies have implemented to fix this?

Likely placebo. The Protocol Max does not have the IMD hump and was addressed during R&D.

 

[carlcamper]

Likely placebo. The Protocol Max does not have the IMD hump and was addressed during R&D.

Hi Crin, he is actually referring to the Cirrus hump which is audible on all Cirrus CS43198 and CS43131 based DACs unless a DRE disable feature is turned on or NOS filter is selected. You can play back the Cmaj flac file in the first post of this thread and confirm it yourself. I’ve actually tested your Dusk DSP cable and confirmed this issue. Hope you can contact your firmware teams for a fix!

 

[crinacle]

Hi Crin, he is actually referring to the Cirrus hump which is audible on all Cirrus CS43198 and CS43131 based DACs unless a DRE disable feature is turned on or NOS filter is selected. You can play back the Cmaj flac file in the first post of this thread and confirm it yourself. I’ve actually tested your Dusk DSP cable and confirmed this issue. Hope you can contact your firmware teams for a fix!

Yes, I am aware of the Cirrus hump and it has been addressed. Please refer to the measurements linked in my previous post.

 

[CedarX]

I’ve actually tested your Dusk DSP cable and confirmed this issue.

Where did you get the information that Dusk DSP was based on CS431xx chip?

I was under the impression it used a CX31988 (which may have DRE-like issues as well…).

 

[RedCometZero]

Likely placebo. The Protocol Max does not have the IMD hump and was addressed during R&D.

Hi Crin. Whether it's the IMD hump or another issue, it's definitely not placebo, and it should be easy for you to replicate by listening to this track at low volumes with the Crinear Reference and Protocol Max, which is what I used.

I guess it's possible that my specific unit is defective, but I think it's more likely that your graph does not capture every manifestation of the Cirrus DRE issue. The reason I think it's the Cirrus DRE issue is because this track is known to surface the issue, and, like other examples of this issue, it goes away when set to NOS filter, which disables the DRE feature.

 

[tccalvin]

I think it's important to distinguish between the Cirrus hump and DRE artifacts (clicks). A device can exhibit no Cirrus hump and still have the artifacts. The only (known) way to get rid of those is disabling DRE completely, although we've seen that it is possible to reduce the artifacts even when DRE is on by adjusting some parameters (I'm too ignorant to know which).

So yeah, no hump doesn't necessarily mean no distortion.

 

[MaxwellsEq]

I think it's important to distinguish between the Cirrus hump and DRE artifacts (clicks). A device can exhibit no Cirrus hump and still have the artifacts. The only way to get rid of those is disabling DRE completely, although we've seen that it is possible to reduce the artifacts even with DRE is on by adjusting some parameters (I'm too ignorant to know which).

So yeah, no hump doesn't necessarily mean no distortion.

It's become unhelpful to group the CS (accidental?) design flaw under the term: "CS IMD Hump". The audible design flaw is the switching ticks. The IMD may be inaudible and might be fixable through parallel paths, whereas the ticks requires DRE to be disabled (which results in adequate, but less impressive SINAD), or set to a lower value.

 

[VientoB]

Yeah nobody can hear the IMD hump, surely?

 

[jf127ime]

By playing DSD (DoP), there are no problems at all.