• 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!

I think I found the perfect USB C headphone dongle...

Blorg

Senior Member
Joined
Dec 16, 2020
Messages
462
Likes
699
I have an AK4377 EVK which I already lost it.

As long as it can drive 32 ohm to my requirement, then I am very happy. Able to drive 16 ohm properly is a big plus.
Above 2uV noise can be heard with low impedance high sensitivity iem, that's is why I spec thd+n 0.001% at 1mW 32 ohm.
If you want to get down to quoting off the datasheets, the CS43131 in the Moondrop Dawn has a quoted THD+N of -115dB (0.00017%) vs -109dB on the AK4377. It's actually 6dB better, which is not insignificant.

You cannot hear harmonic distortion anywhere near these low levels, it needs to be literally orders of magnitude higher. Try listening here:


Noise floor, which is much more audible, is even better on the Dawn, at full voltage 130dB (unweighted) / 132dB (A weighted), at 50mV 94dB. These are measured numbers from Wolf, not off the datasheet. 94dB is the relevant number for IEMs. To put this in context, this is Amir's chart of 50mV measurements, 94dB puts it very near the top, it would be in the top 10 here. Way over most desktop amps, including very expensive ones that are generally considered excellent. This is a really low noise floor.

index.php


You can't hear any of these differences, even with a low impedance high sensitivity IEM.

I have some devices where I can actually hear the noise floor on sensitive IEMs. So maybe you have this in your head from some other device you've heard that quotes better specs than it actually has, and actually has a high noise floor. But no way are you hearing it on this.

Your original post:
Cannot drive 32 ohm properly. There are others which will perform better.

(1) Have you actually tried this, and if so, what 32Ω IEM did you find it had a problem with, and what was the problem?

(2) What others?
 
Last edited:

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
If you want to get down to quoting off the datasheets, the CS43131 in the Moondrop Dawn has a quoted THD+N of -115dB (0.00017%) vs -109dB on the AK4377. It's actually 6dB better, which is not insignificant.

You cannot hear harmonic distortion anywhere near these low levels, it needs to be literally orders of magnitude higher. Try listening here:


Noise floor, which is much more audible, is even better on the Dawn, at full voltage 130dB (unweighted) / 132dB (A weighted), at 50mV 94dB. These are measured numbers from Wolf, not off the datasheet. 94dB is the relevant number for IEMs. To put this in context, this is Amir's chart of 50mV measurements, 94dB puts it very near the top, it would be in the top 10 here. Way over most desktop amps, including very expensive ones that are generally considered excellent. This is a really low noise floor.



You can't hear any of these differences, even with a low impedance high sensitivity IEM.

I have some devices where I can actually hear the noise floor on sensitive IEMs. So maybe you have this in your head from some other device you've heard that quotes better specs than it actually has, and actually has a high noise floor. But no way are you hearing it on this.

Your original post:


(1) Have you actually tried this, and if so, what 32Ω IEM did you find it had a problem with, and what was the problem?

(2) What others?
Dac amp must be tested with load, not no load conditions. There's no point for you to kept repeating if you don't understand what my requirements for an amplifier. As long as dac amp can match 0.001% at 32 ohm 1mW (which the AK4377 can), then these are good amp by my standard.

What for I go listen other amp that doesn't meet my standard? Spend money just to agree with you?
 

Blorg

Senior Member
Joined
Dec 16, 2020
Messages
462
Likes
699
Dac amp must be tested with load, not no load conditions. There's no point for you to kept repeating if you don't understand what my requirements for an amplifier. As long as dac amp can match 0.001% at 32 ohm 1mW (which the AK4377 can), then these are good amp by my standard.

What for I go listen other amp that doesn't meet my standard? Spend money just to agree with you?

So have you actually tried this with an actual IEM, and do you actually have an issue with audible noise on it?

Which IEM are you talking about?

Which actual device (not chip- device) performs better than this?

You're refusing to answer these questions.

I presume the answer to both is no and you have never actually either heard this and are just hypothesizing off your reading of datasheets- and even there you seem to be mixing up DAC chip datasheets with finished device measurements.

Even looking at the datasheet for the AK4377 chip, it seems to have lower claimed performance than this finished device has actual measured performance. I have a AK4376 device myself that measures much much worse than the chip datasheet.

You specifically said "cannot drive 32 ohm properly" but have provided absolutely no backup for that or even a good definition of what you mean by it.
 

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
So have you actually tried this with an actual IEM, and do you actually have an issue with audible noise on it?

Which IEM are you talking about?

Which actual device (not chip- device) performs better than this?

You're refusing to answer these questions.

I presume the answer to both is no and you have never actually either heard this and are just hypothesizing off your reading of datasheets- and even there you seem to be mixing up DAC chip datasheets with finished device measurements.

Even looking at the datasheet for the AK4377 chip, it seems to have lower claimed performance than this finished device has actual measured performance. I have a AK4376 device myself that measures much much worse than the chip datasheet.

You specifically said "cannot drive 32 ohm properly" but have provided absolutely no backup for that or even a good definition of what you mean by it.
The EP-630. Low impedance (12 ohm I think) and very sensitive. I already lost it so If you want to listen to noise, you may get something similar in the market.

Performs better that what?
I already have an AK4377 EVK, for 16 ohm load:
THD+N at 1mW 0.00095%
THD+N at 4mW 0.00066%
THD+N at 16mW 0.00051%
Best THD+N 0.00047%
Crosstalk -78dB at 4mW

For 32 ohm load:
THD+N at 1mW 0.00077%
THD+N at 4mW 0.00062%
THD+N at 16mW 0.00050%
Best THD+N 0.00048%
Crosstalk -84.6dB at 4mW

The datasheet claim -107dB 32 ohm, equivalent to 0.00045%.
I measured 0.00048%, so their EVK implementation is good.
AK4377 IC have a HPGND pin which improves crosstalk. Normal non-inverting amp I can get -70dB for 32 ohm.
This DAC amp can push to -84dB, makes my music more enjoyable.

Also, this DAC amp can drive 1.09V but distorts with 32 ohm load.
So in theory increasing the supply from 1.8V to 1.9V might help, another possibility is increasing the C1P and C1N capacitor to boost more current driving capability.

So you still believe I only read datasheet? :p

So what AK4376 device you have?
Why not read the reviews here and get one that can drive good quality down to 16 ohm?
 

Blorg

Senior Member
Joined
Dec 16, 2020
Messages
462
Likes
699
The EP-630. Low impedance (12 ohm I think) and very sensitive. I already lost it so If you want to listen to noise, you may get something similar in the market.
The Creative EP-630? $20 earbuds from 2005? Seriously? Which you have conveniently "lost".
1.jpg


You still haven't answered if you actually tried these with the Dawn, the device this thread is on, that you made the allegation "Cannot drive 32 ohm properly."

Now it's 16Ω, not 32Ω. Generic 16Ω earbuds from 20 years ago. This is supposedly the fatal flaw in this device, that it can't drive $20 earbuds from 2005 for some reason.

So what's the issue with 32Ω exactly? What is the 32Ω device that it has problems with?

As these are attached wire earbuds with a 3.5mm termination you wouldn't be able to use them on the Dawn, which is 4.4mm only. I guess that would mean it "can't drive them properly", with the stock termination it can't drive them at all. You wouldn't hear hiss. You wouldn't hear anything. But I don't think that's what you meant.

As the Dawn is 4.4 balanced, with two entirely separated channels and no common ground, it will also have substantially better crosstalk performance, essentially none. I don't think this matters, but it's going to be better.

Spec on these earbuds on 16Ω and 106 dB/mW. RAA measured them at 112dB/mW. Whichever of these it is, this is moderately sensitive but not extraordinary or radically out of line for earbuds/IEMs. I have IEMs that are substantially more sensitive than these, and there is zero hiss on the Dawn.

You would not hear any noise whatsoever with the Dawn, if you did actually cut the 3.5mm off and rewire them balanced with a 4.4. You didn't do any of this, but if you did, you wouldn't hear noise on the Dawn.

I remember 2005, I remember the devices we had then, and the sort of very slight hiss you could get. This thing, the noise is so far down, it's just utterly impossible that you would hear any hiss.

And Hiss is not THD, it's noise floor.

THD is substantially LESS audible than hiss. I linked you some listening tests on that, to see what level of THD you can actually hear. You could have THD probably several orders of magnitude worse than the Dawn and you still wouldn't hear it. You certainly won't hear it on the Dawn.

You would not be putting 1V into these either. RAA measured 130.1 dB/V SPL. That's louder than you want to listen.

The AK4376 device I have is the Fiio BTR3. I have several other AKM devices as well, Fiio K5 Pro, UTWS5, LC-BT2, these all use AKM chips and I have nothing against AKM. I don't have strong feelings about DAC chips either way, I also have several ESS and several Cirrus Logic devices. They are all fine.

A AK4377 evaluation kit is not a finished product. But I'm sure it is possible to implement an AK4377 into a perfectly good product. By the numbers though, there are chips from ESS and Cirrus Logic that have lower noise and distortion. There are chips from AKM that have lower noise and distortion as well, it's one of their lower end chips. It's fine though.

It's just that you seem to have this weird idea that this specific one entry-level AKM AK4377 DAC is somehow uniquely remarkable in a way that others are not, and feel compelled to dump on other stuff you have never actually heard, and stuff that objectively measures excellently as well.
 

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
The Creative EP-630? $20 earbuds from 2005? Seriously? Which you have conveniently "lost".

You still haven't answered if you actually tried these with the Dawn, the device this thread is on, that you made the allegation "Cannot drive 32 ohm properly."
Once lost then lost, I already got other new IEM which sounds better. Sadly the EP-630 is a good iem to test noise.

Why must I try the Dawn?
As long as the graph shows it does not meet my requirement then I'll give a pass.

So what's the issue with 32Ω exactly? What is the 32Ω device that it has problems with?
It does not meet my requirement of 0.001% HD+N at 1mW or below. Your disagreement is your issue, not mine.

And Hiss is not THD, it's noise floor.

THD is substantially LESS audible than hiss. I linked you some listening tests on that, to see what level of THD you can actually hear. You could have THD probably several orders of magnitude worse than the Dawn and you still wouldn't hear it. You certainly won't hear it on the Dawn.
Go read my previous post, I state THD+N. The N stands for noise.
It doesn't matter if I don't hear hiss, objectively the graph did not meet my requirements for no hiss.

The AK4376 device I have is the Fiio BTR3.
THD+N -82dB at 32 ohm 1mW

It's just that you seem to have this weird idea that this specific one entry-level AKM AK4377 DAC is somehow uniquely remarkable in a way that others are not, and feel compelled to dump on other stuff you have never actually heard, and stuff that objectively measures excellently as well.
I already mentioned again and again, you must use load on a DAC amp. Dawn already shown some weakness for 32 ohm, 16 ohm already losses to AK4377 with increase in distortion. Base on measured data, I don't need to spend money on devices that do not meet my requirements. Gotta save money for my 9th property purchase.

As for AK4377 device, I'm still waiting for measurements on the 7Hz 71.
If data is ok, then it will replace my favourite EVK.
 

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
1679646176330.png

It's above 0.001%, reject.

If it's line touches 0.001%, then I will look into it's 16 ohm capability.
It's not capable to drive 16 ohm below 0.001%, reject.

This is the dac amp that I would prefer:
 

Blorg

Senior Member
Joined
Dec 16, 2020
Messages
462
Likes
699
@jasonhanjk the Topping G5 measures excellently and I am sure sounds great but the specific measurement you are flagging, from the same person measuring, Wolf at L7AudioLab, is significantly worse than the Moondrop Dawn.

1mW THD+N
Topping G5:
~0.0035% @32Ω, ~0.005% @16Ω
Moondrop Dawn: ~0.001% @32Ω, ~0.0017% @16Ω

This is meaningless because both are well below the threshold of audibility and you can't hear the difference, but the Dawn is 3.5x better on the exact metric you are concerned about.

The G5 is a larger device with much more power. But this specific metric, it's 3.5x worse and "not capable to drive 16 ohm below 0.001%" at 1mW. So presumably, you reject. You'll be rejecting a lot, as this is a top measuring device.

1660640135-THDN-Ratio-vs-Measured-Level.jpg
 

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
@Blorg
You can't even read the graph properly.
The ASR measurement is base on low gain and you go compare graph base on a different gain.
ASR low gain max power is 5mW. Wolf's result is above 1W.
The only comparable result of 1W in ASR is high gain.

You also have a reading deficiency, 0.001% THD+N only applies to 32ohm 1mW; not 16 ohm.
16ohm I uses 5mW and it must be below 0.001%.
 

Blorg

Senior Member
Joined
Dec 16, 2020
Messages
462
Likes
699
@jasonhanjk I'm comparing both based on Wolf's measurements. Moondrop Dawn has a low gain mode as well, you can switch it in the app between 2V/4V.

But you can't hear any of this. This is all so low as to be totally inaudible. And you're saying you can hear these minute differences between SOTA stuff on a crappy $20 IEM from 20 years ago... that you lost. If you care about sound quality, focus on the transducer, it makes far more of a difference.

You are in all these threads on excellently measuring stuff quibbling over supposed differences in noise and harmonic distortion that are -100dB down even at low IEM power.

You don't have any APx555 measurements of your supposed AK4377 evaluation kit either so you don't even know how it actually performs. No actual device made with that chip measures as well as your supposed evaluation kit does, but we are meant to believe that for some reason. The spec sheet on the chip is significantly worse than a lot of this other stuff you are criticising.

None of this matters or makes any actual audible difference. You cannot hear these differences in THD+N. The Topping G5 looks like a great product though so if you are happy with that, go for it.
 

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
By the way, here is a cheap 7Hz 71 dongle (bought for $19 including shipping) measurements that have one of the lowest thd+n with load:

300 ohm
Gain -1dBr
THD+N 1mW 0.0004%
THD+N 2.5mW 0.0005%
Best THD+N 0.0004%
Crosstalk -106dB
Noise 0.9uV (0.7uV Aw)
Max Power 2.5mW X2

32 ohm
Gain -1dBr
THD+N 1mW 0.0009%
THD+N 5mW 0.0006%
Best THD+N 0.0005%
Crosstalk -81dB
Noise 0.9uV (0.7uV Aw)
Max Power 20.5mW X2

16 ohm
Gain -1.2dBr
THD+N 1mW 0.0012%
THD+N 5mW 0.0008%
Best THD+N 0.0007%
Crosstalk -75dB
Noise 0.9uV (0.7uV Aw)
Max Power 24mW X2

Issue with this dac amp:
1. Higher thd at 20Hz and 6kHz, indicating a possible dc blocking cap. Unsure why a rise of 3rd harmonics 6kHz though.
2. Lower output power than datasheet, it does not provide 2.2uF capacitor / capacitance as required in datasheet.
3. Did not program it's firmware to use +2dB hence not able to reach 1Vrms.

1680245852687.png
 

jasonhanjk

Active Member
Joined
Mar 24, 2019
Messages
252
Likes
180
@jasonhanjk I'm comparing both based on Wolf's measurements. Moondrop Dawn has a low gain mode as well, you can switch it in the app between 2V/4V.

But you can't hear any of this. This is all so low as to be totally inaudible. And you're saying you can hear these minute differences between SOTA stuff on a crappy $20 IEM from 20 years ago... that you lost. If you care about sound quality, focus on the transducer, it makes far more of a difference.

You are in all these threads on excellently measuring stuff quibbling over supposed differences in noise and harmonic distortion that are -100dB down even at low IEM power.

You don't have any APx555 measurements of your supposed AK4377 evaluation kit either so you don't even know how it actually performs. No actual device made with that chip measures as well as your supposed evaluation kit does, but we are meant to believe that for some reason. The spec sheet on the chip is significantly worse than a lot of this other stuff you are criticising.

None of this matters or makes any actual audible difference. You cannot hear these differences in THD+N. The Topping G5 looks like a great product though so if you are happy with that, go for it.
I did not have AP555. Haha
Objectively AK4377 wins in the THD+N department.
 

Blorg

Senior Member
Joined
Dec 16, 2020
Messages
462
Likes
699
If you can get your required performance at $19 with the 7Hz 71 and are only looking to use it with sensitive IEMs, why would you want to spend $299 on a much bulkier Topping G5? The G5 does have way more power, if you had a harder to drive over-ear. But you will never need that power with an IEM, and particularly not the sensitive IEMs that you are talking about 1mW of power and complaining about tiny differences in THD+N that are way down. I don't doubt the 7Hz 71 is a very decent device too, I haven't heard it or seen third party measurements but it wouldn't surprise me if it is totally audibly transparent as well, a lot of dongles are these days.
 

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
2,969
Likes
2,606
Location
Nashville
I agree with the concern for a fixed cable. My L&P W2’s cable is not fixed, and I’ve gone thru 2 Lightning OTG cables already, but I didn’t have to replace the dongle. When used for portability the cable an see a lot of stress. I probably should’ve ordered an extra.

By the way, if I were in the market for a dongle today, the HiBy FC4 would get my most serious look, and if I wanted Bluetooth it would be the Fiio BTR7.
I have too many dongles at this point but went ahead and ordered the Hiby FC4. Let's just say I never use it-still default to the 9038D. Underwhelmed.
 

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
2,969
Likes
2,606
Location
Nashville
Top Bottom