[Themilkman46290]

On a side note, do you guys know there is a search feature in the top right of the page, if you search through there you can find all this info, i know its hard for noobz, i was a noob too, but many times the answer to my question was allready there, i simply had to look for it. . ...

 

[DBaldock9]

I used 2 in parallel for each side, for a total of 2x220uF = 440uF each side:
https://www.arrow.com/en/products/t55b227m6r3c0025/vishay

I measured the PCB pads distance and PCB space before I ordered them, then I searched for a cap size in arrow that could fit the space.
From the arrow link, this cap measures:
size (mm): 3.5 X 2.8 X 1.9

The larger size is 3.5mm, that fits on top of the PCB pads, with only a tiny bit of the PCB pads still showing.
2.8mm is how tall they will be in the PCB, because they are soldered side way.
1.9+1.9 =3.8mm is the 2 caps in parallel sandwich size, this size fits without any remaining space left in the PCB.

I can't measure them now, because they are soldered, but the datasheet has all the details.
I also found much more expensive options with higher capacity that could fit the space.
The ones I got had a good price/capacity/size relation, they are "ultra low" ESR and I like the brand.

The 100uF stock cap that was there measures 3.45 x 2.88 x 1.92 mm.

The second photo here has a better view of the caps I used:
https://www.head-fi.org/threads/the-zishan-dsds-corner.826185/page-137#post-14816714

Thanks! That seems like a fairly small space, to try to cram 2200uF in to. I found some low ESR 2200uF caps at Mouser, but they're BIG.

 

[Themilkman46290]

So after trying acrylic (plastic doesnt dissipate heat) and wood(also a good insulater) i have come back to metal, it dissipates heat better and takes far less space then wood or plastic and remains strong (dropped my dsd and acrylic simple shattered) so here is my last attempt at a comfortable, durable and most of all cheap!!

 

[Merlin-PT]

I just want to correct some misinformation related to my posts that was left here:

“Before I do the opa1622 buffer mod (that you described in a prev. post.) , I got from arrow real cheap thin film resistors and I'm going to replace all resistors near the opamps. I have no idea if anything will change.”

My text quoted here has nothing to do with opa1622.
I wanted to change the resistors for the same values, but for thin film instead of probably stock carbon. It's just a coincidence I wanted to try this before the opa1622 buffer mod.

I don't want to bring any discussion back, but my text was quoted out of context and misinterpreted.

 

[nesty]

Hard to really say, without mods i would go for ak4495 or even the ak4490 version because they have lower output impedance and go well with everything where the ak4497 really only sounds right with high impedance gear, in stock form i like my ak4490 much better with iems and heaphones like audiotechnica m50x, grado sr80e, or hifiman he4xx. The ak4497 sounds ok with the headphones i mentioned but they get a bit boomy, and distort like crazy if you go above 25-26 out of 31 on the volume, and lose out on sub bass. So if you have high z gear, go for the 4497 or or get it modded to work better with low z gear like some have done, or get the 4495 version that sounds great stock i have heard them all, and i felt 4495 and 4497 are both fantastic. But one thing i can say is if you dont have high z gear and dont want to mod then the 4497 doesnt really make sense, but then again somepeople like b.a. drivers on high z gear, and some love the boomy midbass from low z dynamic drivers on high z sources, and pretty much all heavily marketed gear that you hear about (audiotechnica, sony, fiio, campfire audio, and all iems ) will be low impedance, so a majority of gear will sound different then its supposed to, only high z headphones (beyerdynamic, akg 601,702, sennheiser hd600 and up) and high z earbuds (anything over 100ohms) will work best

As far as what it sounds similar too, well i also couldnt say, but the ak4490 is better sounding (less digital, better soundstage, seperation, and more dynamic) then the hidizs ap60ii, fiio x1mkii,shanling m1, any ipod, and basically any player in its price bracket

Guys, anybody measured the actual stock output impedance/resistance of the zishan dsd? thanks

 

[calico88]

@Merlin-PT
have you tried change the Resistor to Thinl Film ?

 

[Merlin-PT]

Yes, I changed _all_ the resistors around the opamps to thin film, I didn't knew what to expect and it gave me a cleaner brighter sound.

 

[Ivan TT]

Yes, I changed _all_ the resistors around the opamps to thin film, I didn't knew what to expect and it gave me a cleaner brighter sound.

My 5 cents worth if you don't mind?
AK4497 datsheet specifies 0.1% precision resistors, and in my experience they insure ultralow DC offset, which may not be much of an issue (well, DC offset negatively impacts dynamic range regardless) if output decoupling caps are used, but if you plan to remove them it becomes a necessity. Using AD797 I have under 1mV DC offset, which is not bad at all
Also, I used 0805 resistors instead of 0603 as apparently Johnson noise is lower in 0805 sized ones.
And finally sourcing 0.1% resistors may be tricky, especially if one wants something like Susumu, but Panasonics are easy to obtain and that's what I and I believe @Merlin-PT are using.

 

[Merlin-PT]

I got SEI 1% only because they were the cheapest, at 0.01USD each, I only wanted to see the difference between carbon and thin film.
I've also read it's important and recommend to use 0.1%, maybe I'll change them to 0.1%, depends on the price and availability.

 

[Ivan TT]

I've also read it's important and recommend to use 0.1%, maybe I'll change them to 0.1%, depends on the price and availability.

"A resistor that has 0.1% or less absolute error must be used for external LPFs" - that's from datasheet.
I used ERA-6ARB series but they are not cheap at around 38 cents each. Yet another good reason to get rid of LPF altogether, then you need only 8 1kOhm resistors
But obviously then Panasonic PPS caps ECH-U1H series around DIP socket become a must too, just over 40 cents a pop, but they are much smoother sounding...

 

[Merlin-PT]

I wanted to experiment different values and I needed more quantity.
Then I planned to use 0.1% in the subtractor after I finished the tests.
I'm going to keep the LPF for the 2.5" output, I like the 2.5" sound output.

I already removed the output transistors, but I don't use low impedance headpones, my motive was only the tonality and maybe higher current and I wanted to try other buffers. If I ever use low impedance, I'll try the additional steps you recommended.

One thing I had already tried, bypassing the LPF, but now it's more evident, I get much better sound when connecting the subtractor to the DAC. I want to change the subtractor resistor/caps values to the same as the eval board, to see if it makes a difference connected to the LPF.

I didn't tried the buffers yet and can't say how this will end for my player.
I want to experiment with my player and try different configurations.

 

[Themilkman46290]

Do you guys think the to600 or the k's 600 will sound good with the ak4497, or should i stay in the 300-400ohm range?
My wife is really liking my samsara so i thought to get something on that level

 

[buzzbuzz]

Zishan Z3 with the new AK4493 is "on sale" at AliExpress for $44.69 usd. It is about $4 more than the older AK4490 version.

 

[Ivan TT]

Zishan Z3 with the new AK4493 is "on sale" at AliExpress for $44.69 usd. It is about $4 more than the older AK4490 version.

... and its not too bad, slightly different sound signature compared to AK4497, but responds to "standard" mods well

Spoiler

(OPA1622, no output resistors/caps, replaced 3.3V LDO and clock, extra capacitance on VREF, AD8066 after reducing rails to +/-10V, I don't want to post a photo as it's DSD thread after all, sorry).

Subjectively AK4497: natural, organic laid-back tone, wide soundstage extending far, exquisite details noticeable on reverb tails and cymbals decays.
AK4493: forward, "in-your-face" energetic sound, tightly wrapped 3D soundstage (not worse than 4497, just different), great but not excellent level of detail.

 

[Merlin-PT]

I have one in my cart waiting for Ali anniversary sale in 6 days.