Best cheap USB-C headphone dongles?

[markanini]

The 2.5Ohm output impedance of CB1200AU dongles will have substantial FR alteration on some IEMs, to mention another caveat.

 

[imajescape]

Got a nice deal on jcally jm30 on ali for 12 USD (with coins), it's been 6 days they still haven't shipped.

 

[qitamic]

I have some finding on the KT02H20 chip:

Observation

1. With PEQ off, they can output near 1Vrms
2. As long as a PEQ filter is applied, the device will clip at 0.5Vrms at low volume playback
3. By slowly increasing the volume, when the sine wave becoming a square wave due to 0.5Vrms clipping, the device will suddenly output near 1Vrms again
4. After volume has been reduced, the output will be clipped at 0.5Vrms again
5. Firmware from Fiio and TinHiFi explicit the same behavior, and their Noisegate and Limiter of the DRC are no enabled

Test condition

• Connect the device to a 30 Ohm resistive load
• Playback 21Hz sine wave at different volume

Device tested

• JCally JM12
• Fransun KT02H20

Note

• I can't wait to test the Fiio JA11, but it is still on its way
• Output of the Hi-MAX is slightly higher voltage than JA11 and Fransun KT02H20, and do not explicit the 0.5Vrms clipping behavior even with PEQ filter applied

(new member here, don't be harsh on me)

 

[CedarX]

I can't wait to test the Fiio JA11, but it is still on its way

I believe the JA11 hardware is exactly the same as the JM12 or your Fransun (don’t know that dongle). The firmware is the only difference. If you want to re-flash your JM12 or the Fransun with the JA11 firmware, you have in effect a “JA11” that you can test. The only issue is that there is no way back, as the JM12 and, I assume, the Fransun FWs are not available.

 

[jkim]

I have some finding on the KT02H20 chip:

Observation

1. With PEQ off, they can output near 1Vrms
2. As long as a PEQ filter is applied, the device will clip at 0.5Vrms at low volume playback
3. By slowly increasing the volume, when the sine wave becoming a square wave due to 0.5Vrms clipping, the device will suddenly output near 1Vrms again
4. After volume has been reduced, the output will be clipped at 0.5Vrms again
5. Firmware from Fiio and TinHiFi explicit the same behavior, and their Noisegate and Limiter of the DRC are no enabled

Test condition

• Connect the device to a 30 Ohm resistive load
• Playback 21Hz sine wave at different volume

Device tested

• JCally JM12
• Fransun KT02H20

Note

• I can't wait to test the Fiio JA11, but it is still on its way
• Output of the Hi-MAX is slightly higher voltage than JA11 and Fransun KT02H20, and do not explicit the 0.5Vrms clipping behavior even with PEQ filter applied

(new member here, don't be harsh on me)

I believe the best place where you want to continue posting your findings is this thread.

 

[HissingFree]

I've come across with I think a new device, Jcally JM7 Max, with CX31993 + SGM8262:

https://es.aliexpress.com/item/1005009622342774.html

It's great that it also has a lot of measurements from their lab, such as:

And many others. It says it has ~2uV noise, ~105db dynamic range, ~120mW output power in each channel, ~90db SINAD, etc.

The only thing that looks not good to me is the dynamic range compared to the SNR and the noise floor, isn't 105db too low for the supposed low noise and high power?



Interesting comparison with CX31993 + MAX97220:

 

[Tell]

I've come across with I think a new device, Jcally JM7 Max, with CX31993 + SGM8262:

https://es.aliexpress.com/item/1005009622342774.html

It's great that it also has a lot of measurements from their lab, such as:

And many others. It says it has ~2uV noise, ~105db dynamic range, ~120mW output power in each channel, ~90db SINAD, etc.

The only thing that looks not good to me is the dynamic range compared to the SNR and the noise floor, isn't 105db too low for the supposed low noise and high power?



Interesting comparison with CX31993 + MAX97220:

Sure might be objectively a bit better, but the audible difference is just not there. The only way to improve any of these dongles is more power, adding DSP, make them smaller and lower the price.

 

[markanini]

A reasoned analysis on JM7Max would be that it might cut on power consumption over the competition given that it's based on CX31993. Although how many need this much power on the go? The metrics needs to be indenpendatnly validated with realistic loads. Lastly IME there's some slight signal dependent distortion on CX31993, but not exclusive to CX31993: https://audiosciencereview.com/foru...31xx-based-devices-a-comparative-review.63038

 

[HissingFree]

Sure might be objectively a bit better, but the audible difference is just not there. The only way to improve any of these dongles is more power, adding DSP, make them smaller and lower the price.

It literally has more power lol, 120mw at 32 ohm vs 60.

 

[HissingFree]

A reasoned analysis on JM7Max would be that it might cut on power consumption over the competition given that it's based on CX31993. Although how many need this much power on the go? The metrics needs to be indenpendatnly validated with realistic loads. Lastly IME there's some slight signal dependent distortion on CX31993, but not exclusive to CX31993: https://audiosciencereview.com/foru...31xx-based-devices-a-comparative-review.63038

Personally, I'm more interested in the noise reduction, MAX97220 adds a lot of noise and IEMs have notable hissing as a result. If this noise reduction is real, I think it could be an interesting cheaper option without hissing and with good power.
I just don't like the low dynamic range, 105db for a device that has a higher output power, vs close to 120 SNR, what does it indicate that there is such a difference between SNR and dynamic range, anyone can explain to me? does this mean the low noise is only in meaningless conditions?
From what I've read, the difference is that in dynamic range, a really weak signal is played to calculate the noise instead of the device being idle (SNR).
So this would point to the device being garbage in terms of noise in real world conditions.

 

[jkim]

I just don't like the low dynamic range, 105db for a device that has a higher output power, vs close to 120 SNR, what does it indicate that there is such a difference between SNR and dynamic range, anyone can explain to me? does this mean the low noise is only in meaningless conditions?

The only explanation is that THD is substantial even at -60 dBFS since DR is based on THD+N. If the published specs (2.5 microV noise & 115 to 118 dB SNR at full scale) are correct, SNR at 50 mV should be 83 to 86 dB, which would be quite decent.

 

[markanini]

Personally, I'm more interested in the noise reduction, MAX97220 adds a lot of noise and IEMs have notable hissing as a result. If this noise reduction is real, I think it could be an interesting cheaper option without hissing and with good power.
I just don't like the low dynamic range, 105db for a device that has a higher output power, vs close to 120 SNR, what does it indicate that there is such a difference between SNR and dynamic range, anyone can explain to me? does this mean the low noise is only in meaningless conditions?

Both 105 and 120dB fall outside audibility thresholds. I wonder if you could realiably ABX test signals with 120 vs 105dB noise floor, let alone musical content. Most likely its easier to hear the signal-dependent distortion that DACs produce when employing tricks to increase the measured metrics.

 

[HissingFree]

The only explanation is that THD is substantial even at -60 dBFS since DR is based on THD+N. If the published specs (2.5 microV noise & 115 to 118 dB SNR at full scale) are correct, SNR at 50 mV should be 83 to 86 dB, which would be quite decent.

Ok so it is the distortion part that causes the huge discrepancy and not the noise probably. Is hissing in headphones usually caused by high noise floor and not by distortion? meaning the 2.5 uV vs 5-6uV should be relevant in reducing the hissing on sensitive IEMs, even if the distortion is still relatively high?

 

[jkim]

Ok so it is the distortion part that causes the drop and not the noise probably, hissing in headphones is usually caused by high noise floor and not by distortion?

Correct. But I missed the fact that the noise figures are all A-weighted. My SNR estimates are then 80 to 83.5 db at 50 mv unweighted.

 

[HissingFree]

Both 105 and 120dB fall outside audibility thresholds. I wonder if you could realiably ABX test signals with 120 vs 105dB noise floor, let alone musical content. Most likely its easier to hear the signal-dependent distortion that DACs produce when employing tricks to increase the measured metrics.

It's the hissing in IEMs that is audible to me, it always coincides with a device with higher noise and lower snr, such as JM6 Pro or Tinhified JM12 (iirc measured 102db vs 110db for ja11 firmware).
So the noise floor in such devices is very much audible to me when using sensitive IEMs (but not extremely sensitive ones either).

 

[markanini]

It's the hissing in IEMs that is audible to me, it always coincides with a device with higher noise and lower snr, such as JM6 Pro or Tinhified JM12 (iirc measured 102db vs 110db for ja11 firmware).
So the noise floor in such devices is very much audible to me.

Most recorded music is max 90-96dB(16 bits). Maybe you should create som test files for you own pupose to determine where the threshold is for you, on your best DAC.

 

[markanini]

I received a Faaeal ALC5686 based dongle and played back the "cirrus hump" test clips: dune soundtrack, rumble, Cmaj. I can share measurements if anyone is interested, but it audibly crackles. So far KTO2H20 with TinHifi FW is the only 1V chip that's free from audible crackling in my tests, which include CX31993 and ALC5686. Samsung dongle doesnt crackle but has a unique distortion discovered by @danadam https://audiosciencereview.com/forum/index.php?threads/53071

 

[gaz]

JCALLY JM6 3.5 mm - I just received this and it works perfectly in my Oneplus Nord CE 2 phone and Lenovo WIndows 10 laptop - just plug in.

A nice looking dongle with gold-plated contacts, warranty, and well-packaged. I was looking for a non-amplified, non-Apple dongle and this is it.

 

[HissingFree]

I received a Faaeal ALC5686 based dongle and played back the "cirrus hump" test clips: dune soundtrack, rumble, Cmaj. I can share measurements if anyone is interested, but it audibly crackles. So far KTO2H20 with TinHifi FW is the only 1V chip that's free from audible crackling in my tests, which include CX31993 and ALC5686. Samsung dongle doesnt crackle but has a unique distortion discovered by @danadam https://audiosciencereview.com/forum/index.php?threads/53071


View attachment 469147

What measurements can you share? for example, I would like to know power/output voltage at 32 ohms, noise floor or at least SNR at 50mV.

 

[jkim]

I received a Faaeal ALC5686 based dongle and played back the "cirrus hump" test clips: dune soundtrack, rumble, Cmaj. I can share measurements if anyone is interested, but it audibly crackles. So far KTO2H20 with TinHifi FW is the only 1V chip that's free from audible crackling in my tests, which include CX31993 and ALC5686. Samsung dongle doesnt crackle but has a unique distortion discovered by @danadam https://audiosciencereview.com/forum/index.php?threads/53071


View attachment 469147

Thanks for sharing this info. So, the only 1V dongles that have been tested free of any peculiar distortion artifacts are the JA11/JM12 (KT02H20) w/ TinHifi FW and Hi-Max CB1200AU (see tests here and here)---have not posted all the tests of the Hi-Max, though.

Of these two, the Hi-Max is on top since it has slightly better noise performance and supports onboard 8-band PEQ. Oh, by the way, the JA11/JM12 produces random distortion in response to test signals, regardless of firmware (as shown in the tests linked above).