ZolaIII
Major Contributor
- Joined
- Jul 28, 2019
- Messages
- 4,545
- Likes
- 2,696
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. 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!
Best cheap USB-C headphone dongles?
- Thread starter starfly
- Start date
Ok about smartphone's ability to fully power dongles, although in my opinion if we are talking about max power achievable with a dongle, it has to be referred to a fully capable power source, not a phone.
What confused me was this specific sentence:
Since again, K11 is a desktop device, while, if you intended KA11 dongle, as it seems the case, it is only SE and specs are rather different.
abm0
Active Member
- Joined
- May 12, 2019
- Messages
- 163
- Likes
- 73
It's line level for sure, they changed it in the long description text to have a bit of separation by saying "2 Vrms Max" and not follow it directly with @32 but put a different thing in the parenthesis, i.e. "(125 mW @32 ohms)". Here the 125 is probably a copy-paste mistake because they have the same exact SE spec for the TP30 which has additional amplification. So for the TP20 it was probably supposed to read something like "75 mW @32 ohms" (half of their TP20 BAL spec), and in this form it's not quite the same as saying that the 2 Vrms is achieved AT 32 ohms.
They changed this after I got in their DMs to pester them about their "4 Vrms @32 ohms" spec for the TP30 being half a watt and practically impossible LMAO. But no matter how much I tried to clarify they just didn't want put in a clearer separation between the Vrms and the @32 ohms. Or maybe they just didn't want to provide an exact impedance where the 2 Vrms and 4 Vrms were achieved exactly, so it wouldn't affect their marketing.
Then they also had "Output impedance: 32 ohms" and I got them to change that as well, but they didn't agree to simply delete it OR to get the real number from the engineering department, so they just put in another parenthesis to explain how they got that number, i.e. it was only talking about how their max-power tests were done with a 32 ohm load.
ahaha, what a nightmare, kudos to you for actively trying to make them correct but i think it's a Don Quijote fight given the linguistic difference, translation problems and more or less explicit attempt to allure less informed customers.It's line level for sure, they changed it in the long description text to have a bit of separation by saying "2 Vrms Max" and not follow it directly with @32 but put a different thing in the parenthesis, i.e. "(125 mW @32 ohms)". Here the 125 is probably a copy-paste mistake because they have the same exact SE spec for the TP30 which has additional amplification. So for the TP20 it was probably supposed to read something like "75 mW @32 ohms" (half of their TP20 BAL spec), and in this form it's not quite the same as saying that the 2 Vrms is achieved AT 32 ohms.
They changed this after I got in their DMs to pester them about their "4 Vrms @32 ohms" spec for the TP30 being half a watt and practically impossible LMAO. But no matter how much I tried to clarify they just didn't want put in a clearer separation between the Vrms and the @32 ohms. Or maybe they just didn't want to provide an exact impedance where the 2 Vrms and 4 Vrms were achieved exactly, so it wouldn't affect their marketing.
Then they also had "Output impedance: 32 ohms" and I got them to change that as well, but they didn't agree to simply delete it OR to get the real number from the engineering department, so they just put in another parenthesis to explain how they got that number, i.e. it was only talking about how their max-power tests were done with a 32 ohm load.
abm0
Active Member
- Joined
- May 12, 2019
- Messages
- 163
- Likes
- 73
Nah, you know what, I was quite impressed with their auto-translate in the messages section, I got intelligible translations for every seller that was typing in Chinese. I think this was more a question of the person handling the AliExpress side of things being completely a-technical and afraid to change too much in their product listing without really understanding all of that electronics-talk. 
- Joined
- Mar 30, 2019
- Messages
- 1,176
- Likes
- 2,062
UGREEN HiFi Audio measurements
In case anyone is looking for a dongle for car use, here are some measurements of the new UGREEN HiFi Audio double-pronged dongle (say that 3 times fast). What is useful about this dongle is the second, female USB-C connector, which allows you to charge a smartphone while using it. This makes it a good option for long car trips. Overall it is a well-built and nicely designed dongle, especially for the price (~$20).
The regular version of this dongle is the replacement of the highly-regarded and sought-after UGREEN HiFi Audio Pro. Unfortunately it is a downgrade. I am assuming that it performs similarly to my double-pronged version, however I can't state for a fact that it does.
Note that these measurements were made using a Google Pixel 6A as the source. Normally I use a Topping HS02 isolator in order to make sure it is the dongle I am measuring but this one is incompatible with it. (It is likely a USB 3.0 dongle.) Perhaps its actual performance is a little better than this, but this is still helpful as a real-life use example. Results are worse on my laptop.
The measurement chain is as follows: Google Pixel 6A - UGREEN HiFi Audio - E1DA Scaler/Cosmos. I measured the voltage at max output and it is 1.004V.
24/48
Pas mal. Seems comparable to the well-respected Apple USB-C dongle. But this one is a much more robust dongle. I'd never use the Apple USB-C in a car!!
Edit: Powering the Pixel at the same time through the second prong does not affect the results much. Honestly, the difference may be my TRSS to dual TS splitter cable, which could be better made. Nicely done.
32/96
With respect to what 96kHz affords, a worse performer. Perhaps my Pixel 6A is down-sampling the file. I'd ignore this one.
The following are mono measurements of the left channel, with stereo measurements of the crosstalk added to them. I do it this way because the E1DA Cosmos does a few dBs better on mono.
This dongle has a mic input. I posted these results before, but as you can see when used for IEM measurements you don't get a flat frequency response. There may be a high pass filter on this, which is not a terrible idea given its intended use for phone calls. Note that the output does not show this so breathe easy. See the results above at max output.
While these results are a downgrade from the Pro version, I think they are without question more than good enough for car use. This gets a recommendation!
In case anyone is looking for a dongle for car use, here are some measurements of the new UGREEN HiFi Audio double-pronged dongle (say that 3 times fast). What is useful about this dongle is the second, female USB-C connector, which allows you to charge a smartphone while using it. This makes it a good option for long car trips. Overall it is a well-built and nicely designed dongle, especially for the price (~$20).
The regular version of this dongle is the replacement of the highly-regarded and sought-after UGREEN HiFi Audio Pro. Unfortunately it is a downgrade. I am assuming that it performs similarly to my double-pronged version, however I can't state for a fact that it does.
Note that these measurements were made using a Google Pixel 6A as the source. Normally I use a Topping HS02 isolator in order to make sure it is the dongle I am measuring but this one is incompatible with it. (It is likely a USB 3.0 dongle.) Perhaps its actual performance is a little better than this, but this is still helpful as a real-life use example. Results are worse on my laptop.
The measurement chain is as follows: Google Pixel 6A - UGREEN HiFi Audio - E1DA Scaler/Cosmos. I measured the voltage at max output and it is 1.004V.
24/48
Pas mal. Seems comparable to the well-respected Apple USB-C dongle. But this one is a much more robust dongle. I'd never use the Apple USB-C in a car!!
Edit: Powering the Pixel at the same time through the second prong does not affect the results much. Honestly, the difference may be my TRSS to dual TS splitter cable, which could be better made. Nicely done.
32/96
With respect to what 96kHz affords, a worse performer. Perhaps my Pixel 6A is down-sampling the file. I'd ignore this one.
The following are mono measurements of the left channel, with stereo measurements of the crosstalk added to them. I do it this way because the E1DA Cosmos does a few dBs better on mono.
This dongle has a mic input. I posted these results before, but as you can see when used for IEM measurements you don't get a flat frequency response. There may be a high pass filter on this, which is not a terrible idea given its intended use for phone calls. Note that the output does not show this so breathe easy. See the results above at max output.
While these results are a downgrade from the Pro version, I think they are without question more than good enough for car use. This gets a recommendation!
Last edited:
I received my Tempotec Sonata BHD (non pro), took early simple measurements with a multimeter to check everything is ok, as it seems to be:
It has 3 manually switchable gain level equally distributed, so we have:
SINGLE ENDED LOW
SINGLE ENDED MID
SINGLE ENDED HIGH
And correctly balanced output gives corresponding doubled voltage
BALANCED LOW
BALANCED MID
BALANCED HIGH
No measurable DC at output.
I'd like to check if the ground pin of 4.4 socket is connected in case one would want to use it as a balanced dac to a balanced amp with a 4.4 to xlr cable but at the moment i I only have a 4.4 iems cable, so i couldn't.
Strangely the gain switch cycle is LOW -> HIGH -> MID -> LOW ecc... so MID is not in between LOW and HIGH, anyway you can confirm it via the led light so no real trouble.
It maintains last gain and volume set at power on.
It's my first dongle with volume buttons so don't know if the behavior is normal but in Linux (Pipewire) software volume doesn't function, so I have to use volume buttons that i would have preferred to not touch and leave at max (I don't plan to use it with smartphone).
In alsamixer I have these 2 entry and with these settings it already goes at the max set with hardware buttons:
Luckily volume buttons are not too small and have nice tactile feeling, more importantly they have much more than the 16 levels advertised at Aliexpress (but not on the official site) so, in combination with the 3 gain levels, it results nicely manageable, at least better than what I feared when discovered the software volume issue. Anyway with volume buttons you can't get complete silence, so there is not something like a mute function at last volume step, I'll check again with multimeter the actual number of steps and min levels.
I'll investigate further if something can be set to use software volume leaving hardware volume at max as i do with my other dongles, but in the end it could even turn out for good since now I can disconnect the old SB X-Fi 5.1 pro that I am using just as usb volume encoder and free up a usb port
Ironically i just listened to it few minutes since it's the "boring" part where i don't expect nothing new being it fully transparent
It has 3 manually switchable gain level equally distributed, so we have:
SINGLE ENDED LOW
SINGLE ENDED MID
SINGLE ENDED HIGH
And correctly balanced output gives corresponding doubled voltage
BALANCED LOW
BALANCED MID
BALANCED HIGH
No measurable DC at output.
I'd like to check if the ground pin of 4.4 socket is connected in case one would want to use it as a balanced dac to a balanced amp with a 4.4 to xlr cable but at the moment i I only have a 4.4 iems cable, so i couldn't.
Strangely the gain switch cycle is LOW -> HIGH -> MID -> LOW ecc... so MID is not in between LOW and HIGH, anyway you can confirm it via the led light so no real trouble.
It maintains last gain and volume set at power on.
It's my first dongle with volume buttons so don't know if the behavior is normal but in Linux (Pipewire) software volume doesn't function, so I have to use volume buttons that i would have preferred to not touch and leave at max (I don't plan to use it with smartphone).
In alsamixer I have these 2 entry and with these settings it already goes at the max set with hardware buttons:
Luckily volume buttons are not too small and have nice tactile feeling, more importantly they have much more than the 16 levels advertised at Aliexpress (but not on the official site) so, in combination with the 3 gain levels, it results nicely manageable, at least better than what I feared when discovered the software volume issue. Anyway with volume buttons you can't get complete silence, so there is not something like a mute function at last volume step, I'll check again with multimeter the actual number of steps and min levels.
I'll investigate further if something can be set to use software volume leaving hardware volume at max as i do with my other dongles, but in the end it could even turn out for good since now I can disconnect the old SB X-Fi 5.1 pro that I am using just as usb volume encoder and free up a usb port
Ironically i just listened to it few minutes since it's the "boring" part where i don't expect nothing new being it fully transparent
Last edited:
I made a very quick test on Windows 10 and the Sonata BHD behaves the same way as Linux, no software volume available, setting the windows volume slider at 1 it already goes at max volume you set with hardware buttons. At least with Windows 10 native uac2 drivers, didn't test proprietary drivers since it was my work pc where i can't install software at libidum, but I suppose nothing would change.
I'd like to know from you guys if most of the dongles with volume buttons behave like this or it's more like an exception, thank you in advance.
I'd like to know from you guys if most of the dongles with volume buttons behave like this or it's more like an exception, thank you in advance.
Last addition about Tempotec Sonata BHD (non Pro), i found some measurements taken with RMAA in a youtube video review. I don't know the quality of ADC used since the review is in Russian language so out of my understanding, but graphs are graphs and RMAA summary report has a known series of results, so we have:
translation: 
So it is confirmed to be pretty solid and a very good choice when on sale.
translation: So it is confirmed to be pretty solid and a very good choice when on sale.
HissingFree
Member
- Joined
- Apr 12, 2024
- Messages
- 34
- Likes
- 12
I've noticed that stereo crosstalk is much worse than on other CS43131 reviewed by this user, how much would you say it can be noticeable in real content?
Yes, crosstalk rejection of 3.5 out seems not very good at -60 dB where most other dongles he reviewed got > -75 dB.
4.4 out behaves well at more than -120 dB on par with other balanced dongles so I don't know if there is a design flaw with this 3.5 out of this specific dongle (Sonata BHD Pro measured OK so not something manufacturer related) or maybe something gone wrong with his measurements run.
I can't tell how much this can be audible, I took a quick test:
- set Sonata BHD to high gain and maxed out volume with button
- disconnected my Truthear Gate left earpiece from cable to avoid damage
- set Audacity like this:
I couldn't hear any sign of the left test tone in the right earpiece but honestly I don't know if this is a valid test or maybe it just set right channel into muting, plus I'm into a slightly noisy environment at this moment.
- Joined
- Mar 30, 2019
- Messages
- 1,176
- Likes
- 2,062
JCALLY JM6 PRO measurements
Here's another one from a brand with a good reputation here. From my understanding the "pro" in the name comes from the use of the MAX97220 chip that affords it more power than the standalone CX31993 dongles have. I measured the voltage at max output and it is 1.455V.
The measurement chain is as follows: Windows PC - Topping HS02 isolator - JCALLY JM6 PRO - E1DA Scaler/Cosmos.
Please note that the output was set to 24/48. The RMAA measurements are of the left channel taken with the Cosmos in mono mode as it performs better that way. They are supplemented by the stereo crosstalk results.
Do note I am now going to use the AES17-2015 standard notch setting for distortion results on REW. The first measurement was posted so that it can be compared apples to apples with my previous ones. The difference is only a dB here but I prefer to be upfront about these things.
SINAD of about 96.8 dB, which meets spec. In all this gets the job done. The 1kHz sine shows a wide skirt with some hash, but it should be inaudible. These are workman-like results that are good for the price and product position, but less so each passing day. To be sure, the JM-20 is the one to get, but anyone that has this and is happy should remain so.
I am hoping to build my own custom alligator clip cable so that I can use my E1DA loading board in order to take measurements at specific loads. But I am guessing this performs well enough for IEMs and more efficient headphones.
Addendum: results at 50mV.
Here's another one from a brand with a good reputation here. From my understanding the "pro" in the name comes from the use of the MAX97220 chip that affords it more power than the standalone CX31993 dongles have. I measured the voltage at max output and it is 1.455V.
The measurement chain is as follows: Windows PC - Topping HS02 isolator - JCALLY JM6 PRO - E1DA Scaler/Cosmos.
Please note that the output was set to 24/48. The RMAA measurements are of the left channel taken with the Cosmos in mono mode as it performs better that way. They are supplemented by the stereo crosstalk results.
Do note I am now going to use the AES17-2015 standard notch setting for distortion results on REW. The first measurement was posted so that it can be compared apples to apples with my previous ones. The difference is only a dB here but I prefer to be upfront about these things.
SINAD of about 96.8 dB, which meets spec. In all this gets the job done. The 1kHz sine shows a wide skirt with some hash, but it should be inaudible. These are workman-like results that are good for the price and product position, but less so each passing day. To be sure, the JM-20 is the one to get, but anyone that has this and is happy should remain so.
I am hoping to build my own custom alligator clip cable so that I can use my E1DA loading board in order to take measurements at specific loads. But I am guessing this performs well enough for IEMs and more efficient headphones.
Addendum: results at 50mV.
Last edited:
Only issue with the MAX97220 chip is that it raises the output impedance to 2Ohms over the standalone CX31993 0.5 Ohm. Other than that it's a great cost effective option with more power than Apple dongle.
daxliniere
New Member
Hey everyone, first post here.
CS43131 seems to be favourable - does anyone know if it works properly with Google Pixel 8 phones or where to find compatibility information, please?
All the best,
Dax.
CS43131 seems to be favourable - does anyone know if it works properly with Google Pixel 8 phones or where to find compatibility information, please?
All the best,
Dax.
Should be fine.
daxliniere
New Member
Thanks @markanini. I've seen a bunch of (older) threads saying that P8 and P8P don't play well with UAP2.0 USB dongles so I was hoping to get a definitive yes/no.
This lists JCally AP10 and iBasso DC04PRO among tested DACs that work, which uses CS43131 Maybe order with a return policy if it's a different CS43131 DAC.
Just got the "generic" GraveAudio DA06 with CX31993 for €5.50.
Name is not very reassuring, but whatever.
This has no additional max97220 or other opamps since line level output voltage sits at exact 1 V.
Build is mixed up, plug and case appears solid metal but cable is super filmsy and the 3.5 socket doesn't give a very solid feel when plugging\unplugging a jack.
Testing with Truthear Gate plenty loud at 35% OS volume, just the same i set with the Sonata BHD in low gain mode (same 1 V max out), no hiss, no strange emi noise.
Testing with Thieaudio Ghost (60 ohm, 91 db), they go pretty loud at 55% OS volume, at least for my preferences, so there's even good margin for eq with negative preamp.
Plugging trrs mic cable activetes the ADC and OS recognize the input device, flawless.
I bought it just for a nice to have gadget to leave at my office pc but honestly with my iems and headphones i could perfectly live with this alone, €5.50 (thin metal case and usb-c\usb-a adapter included) and dac\amp resolved, jaw dropping
Only strange thing is that Linux recognizes it supporting "only" max 96 Khz, nothing more:
Also Windows 10, the first time i plugged it showed max 32/96 in audio control panel, but unplugging\replugging reinitialized it and then showed 32/384.
Honestly it doesn't make a difference to me, just sharing for completeness.
P.S.
As you see, usb communication is in SYNC mode, that matches with the large jitter skirt I've seen in other CX dongles review.
Name is not very reassuring, but whatever.
This has no additional max97220 or other opamps since line level output voltage sits at exact 1 V.
Build is mixed up, plug and case appears solid metal but cable is super filmsy and the 3.5 socket doesn't give a very solid feel when plugging\unplugging a jack.
Testing with Truthear Gate plenty loud at 35% OS volume, just the same i set with the Sonata BHD in low gain mode (same 1 V max out), no hiss, no strange emi noise.
Testing with Thieaudio Ghost (60 ohm, 91 db), they go pretty loud at 55% OS volume, at least for my preferences, so there's even good margin for eq with negative preamp.
Plugging trrs mic cable activetes the ADC and OS recognize the input device, flawless.
I bought it just for a nice to have gadget to leave at my office pc but honestly with my iems and headphones i could perfectly live with this alone, €5.50 (thin metal case and usb-c\usb-a adapter included) and dac\amp resolved, jaw dropping
Only strange thing is that Linux recognizes it supporting "only" max 96 Khz, nothing more:
Bash:
TTGK Technology Co.,Ltd CX31993 384Khz HIFI AUDIO at usb-0000:00:1d.0-1.1.1.1, : USB Audio
Playback:
Status: Stop
Interface 1
Altset 1
Format: S16_LE
Channels: 2
Endpoint: 0x01 (1 OUT) (SYNC)
Rates: 8000, 16000, 32000, 44100, 48000, 96000
Bits: 16
Channel map: FL FR
Interface 1
Altset 2
Format: S24_3LE
Channels: 2
Endpoint: 0x01 (1 OUT) (SYNC)
Rates: 8000, 16000, 32000, 44100, 48000, 96000
Bits: 24
Channel map: FL FR
Interface 1
Altset 3
Format: S32_LE
Channels: 2
Endpoint: 0x01 (1 OUT) (SYNC)
Rates: 8000, 16000, 32000, 44100, 48000, 96000
Bits: 32
Channel map: FL FR
Capture:
Status: Stop
Interface 2
Altset 1
Format: S16_LE
Channels: 1
Endpoint: 0x81 (1 IN) (SYNC)
Rates: 8000, 16000, 32000, 44100, 48000, 96000
Bits: 16
Channel map: FL
Interface 2
Altset 2
Format: S24_3LE
Channels: 1
Endpoint: 0x81 (1 IN) (SYNC)
Rates: 8000, 16000, 32000, 44100, 48000, 96000
Bits: 24
Channel map: FLAlso Windows 10, the first time i plugged it showed max 32/96 in audio control panel, but unplugging\replugging reinitialized it and then showed 32/384.
Honestly it doesn't make a difference to me, just sharing for completeness.
P.S.
As you see, usb communication is in SYNC mode, that matches with the large jitter skirt I've seen in other CX dongles review.
Basic Channel
Active Member
- Joined
- Mar 22, 2024
- Messages
- 296
- Likes
- 274
Interested to see if you have the issue I do moving around. If you have the chance to try of course. Good to know it works on linux as I have not plugged it into the PC yet.
I just plugged it to my phone and playing via UAPP, made some steps around my house, moved and rotated the phone in my hand in all ways and directions, nothing happened.
Only issue playing around in UAPP with hardware volume slider it stopped playing a couple of time and I had to unplug\replug to make it come back working, but I'm not much accustomed to UAPP usb settings since i usually use my phone 3.5 audio out.
Last edited:
Similar threads
- Poll
