[kookaudio]

It is correct since the player “sees” impedance of the earphones when it is powered.
But I would not do it.

 

[ThanosD]

It is correct since the player “sees” impedance of the earphones when it is powered.
But I would not do it.

Hmm, why not? I guess because they are for protection right? If I change them for a very low value, let's say 1ohm, I assume it's almost the same as removing them completely?

 

[Ivan TT]

Hmm, why not? I guess because they are for protection right? If I change them for a very low value, let's say 1ohm, I assume it's almost the same as removing them completely?

There’s at least two good reasons for output protection resistors being there:

They protect the output stage from short-circuit, which happens every time one plugs headphones in... or turns Zishan on as output relay shorts outputs to ground during time-delay period (few seconds until it toggles with audible click). Unless the output stage circuit is guaranteed to handle this condition and not to fry (one of the reasons I have OPA1622, indefinite short circuit time)

Headphones and cable appear as a capacitive load to the output stage, not many opamps can handle it well without going into self-oscillation (that you will not hear, but feel as things get hot!), no one really knows if this discrete output stage is prone to this or not.

So removing output protection resistors should be done at ones own risk and ideally taking into account (possible) consequences, again OPA1622 are there in my mod for more than one good reason.

Hope it helps.

 

[ThanosD]

There’s at least two good reasons for output protection resistors being there:

They protect the output stage from short-circuit, which happens every time one plugs headphones in... or turns Zishan on as output relay shorts outputs to ground during time-delay period (few seconds until it toggles with audible click). Unless the output stage circuit is guaranteed to handle this condition and not to fry (one of the reasons I have OPA1622, indefinite short circuit time)

Headphones and cable appear as a capacitive load to the output stage, not many opamps can handle it well without going into self-oscillation (that you will not hear, but feel as things get hot!), no one really knows if this discrete output stage is prone to this or not.

So removing output protection resistors should be done at ones own risk and ideally taking into account (possible) consequences, again OPA1622 are there in my mod for more than one good reason.

Hope it helps.

Damn, thanks a lot. Then I am screwed. I really wanted this dap, but I already have 1 Multi-BA pair, and I am building another single BA earphone. And I surely cannot perform your mods with all this extra jumper wires, and small scale soldering, as I have neither the skills, nor the equipment and time. Also the 4495 is quite pricey, 140€, when I can get the 4497 for 90€. I guess I need to look somewhere else. Thanks again for the help!

 

[Ivan TT]

Damn, thanks a lot. Then I am screwed. I really wanted this dap, but I already have 1 Multi-BA pair, and I am building another single BA earphone. And I surely cannot perform your mods with all this extra jumper wires, and small scale soldering, as I have neither the skills, nor the equipment and time. Also the 4495 is quite pricey, 140€, when I can get the 4497 for 90€. I guess I need to look somewhere else. Thanks again for the help!

Well, you can always give it a go and see what happens, my post above covers the theoretical assumptions, which are not always proved correct in practice.
What kind single BA you are building, if you don’t mind me asking? PM or post in the IEM thread

 

[ThanosD]

Well, you can always give it a go and see what happens, my post above covers the theoretical assumptions, which are not always proved correct in practice.
What kind single BA you are building, if you don’t mind me asking? PM or post in the IEM thread

Nah, unfortunately I cannot afford such an experiment. So I will try to find something more suitable. I'll PM you about the IEM.

 

[Themilkman46290]

Well, I am not sure because I can't check, don't have a oscilloscope, but my ak4490 dsd has no problems with ba drivers, i am sure if to check player.ru someone there has the measurements, but as far as I can tell it has low impedance and the price is lower

 

[ThanosD]

Well, I am not sure because I can't check, don't have a oscilloscope, but my ak4490 dsd has no problems with ba drivers, i am sure if to check player.ru someone there has the measurements, but as far as I can tell it has low impedance and the price is lower

One of the reasons I wanted the Zishan, was the Balanced port. I wanted a source with balanced output, to give it a try. Also the 4490 model is less popular, and with less mods available. And I tried using the Russian forum, but it kinda didn't work for me. It's not very practical, if you want something more technical. It is slightly more expensive too, not by much, but it has an inferior dac chip.

 

[calico88]

what’s best high earphone impedance for 4497 to be exact ?
100 ohms ?

 

[Merlin-PT]

The 2.5mm balanced output amplifier circuit used in zishan ak4497, is a copy from the ak4497 datasheet and I wanted to have at least a superficial understanding of what it does.
This is only a hobby for me and I didn't have any previous understanding of this, so I hope I got it right. It turned out much easier than I thought.

I googled for opamps and found the 3 building blocks that make this circuit:
- High Pass Filter (left of the red line)
- Low Pass Filter (right of the red line)
- non-inverting operational amplifier (the configuration of the opamp used in the LPF)

The original circuit from the AK4497 datasheet is in BLACK color and is the balanced Left channel. It represents the components around only 1 of the 2 op275 opamps. The right channel is a duplication of this circuit built with the 2nd op275 opamp.

Spoiler: Photo of both channels

The differences from the datasheet circuit to the zishan are in RED color.
Note: zishan uses 390ohm resistors instead of 400ohm, it's almost the same so I did't mark it different.
My comments are in GREEN and RED.

To the left of the vertical RED LINE there is a High Pass Filter, as the name says it only pass frequencies above a cut value.
Freq_HPF = 1/(2*pi*R*C)
=1/(2*PI()*10*10^3*100*10^-6) = 0,15915 Hz

To the right there is a Low Pass Filter using the opamp for gain and as the name says it only passes frequencies bellow a cut value.
Freq_LPF = 1/(2*pi*sqrt(R1*R2*C1*C2)))
=1/(2*PI()*SQRT((22*22*27*10^-9*56*10^-9))) = 186046,5 Hz = 186 KHz

The non-inverting operational amplifier gain amplifies the sound.
Gain = 1 + R2G / R1G
=1+400/400 = 2
Note: the datasheet circuit has a gain of =1+400/200 = 3, zishan uses a gain of 2.

So to sum it up, the zishan balanced output amplifier circuit, is a band pass filter, passing all freq. from 0,15915 Hz to 186Khz and it amplifies the sound (increases voltage) 2 times.

As I said I didn't knew this before I google it, so please feel free to correct me or add anything, I hope I understood it right.

Random googled sources used:
https://www.electronics-tutorials.ws/filter/filter_3.html
https://www.electronics-tutorials.ws/filter/second-order-filters.html
https://www.electronics-notes.com/a...-amplifier-op-amp/non-inverting-amplifier.php
https://www.circuitlab.com/editor/#?id=me84mf

 

[Themilkman46290]

what’s best high earphone impedance for 4497 to be exact ?
100 ohms ?

Well i can say that hd600,hd650,dt770 pro 250 ohm, and akg k701 are the best i heard as far as fullsize and seahf f400s, k's samsara, ty hi-z hp150ts, toneking to200,to400,and if you want budget buds nicehck eb200, monk light spc 120ohm

Those are a few that i have either sampled or i own and they all sounded very good, basically anmything over 80ohms will sound good with the ak4497

 

[Ivan TT]

The non-inverting operational amplifier gain amplifies the sound.

I believe the function of the socketed opamps is a substractor, it substrats two signals (positive and negative) and this produces gain increase of x2, but the difference gets attenuated, thus improving signal to noise ratio. It also acts as an extra pole in terms of LPF filtering due to 1nf caps, to reduce delta-sigma modulator noise even further (as the cutoff frequency is way outside of the audio passband even at higher sampling frequencies.
Since tgecresistors are of the same value (1k), it has gain of 1, so any opamps used in dip8 socket should be unity gain stable (eg 627, but not 637 which may still work due to caps in the fb loop).
Just my two cents worth.

 

[Themilkman46290]

Ak4490 with a diy dip8 lead extension on a diy acrylic endplate, after many hours of obscene and explicit language and a bit of elbow grease, got a way to test opamps before I decide what to try permanent wile I wait for a pair of ad8620arz for my ak4497 , so far my two lme49710ha are really blowing me away, the v5i close second

 

[Themilkman46290]

Thanks to Dbaldock9 for suggesting the 2 lme49710ha, they are my favorites so far

 

[Merlin-PT]

I believe the function of the socketed opamps is a substractor, it substrats two signals (positive and negative) and this produces gain increase of x2, but the difference gets attenuated, thus improving signal to noise ratio. It also acts as an extra pole in terms of LPF filtering due to 1nf caps, to reduce delta-sigma modulator noise even further (as the cutoff frequency is way outside of the audio passband even at higher sampling frequencies.
Since tgecresistors are of the same value (1k), it has gain of 1, so any opamps used in dip8 socket should be unity gain stable (eg 627, but not 637 which may still work due to caps in the fb loop).
Just my two cents worth.

Thanks, that explains the 1 socket opamp circuit that is after the 2 soldered opamps.
For anyone that wants to see the full circuit with all the opamps, it's in page 38 of the ak4497 evaluation board:
https://www.akm.com/akm/en/file/ev-board-manual/AK4497EQ.pdf
EDIT: https://velvetsound.akm.com/content...dac/ak4497eq/ak4497eq-en-evaluationmanual.pdf

The resistors used around the socket opamp in zishan are 1.5k (620ohm in the eval board) and the caps are 0,1nF (1nF in the eval board).