KZ ZS6 IEM Modified

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Tipped off by Mr. L at the other place, I got myself a pair of KZ ZS6,
and followed his mod to the last word :

1. Rewaire the two balanced armature in series.
2. Replace the 0402 cap with Panasonic ECPU (I used 470n instead of 680n).
3. Remove the 0402 resistor and rewire with a 7.5R Beysclag MELF.

DIYHiFi.org • View topic - Affordable "High End" IEM with excellent SQ after small mods

Quite fiddly to do, so great soldering skill with SMD required.
You have been warned.
But result is real great sound for very little money.
Highly recommeded.

Cable I used :
https://www.amazon.com/KZ-Dedicated-Upgraded-Replace-Earphone/dp/B07DDHYLP6

Source : Laptop or Pono Player

Amp : XEN ZGF Portable
https://www.diyaudio.com/forums/headphone-systems/286878-xen-zgf-portable-headphone-amplifier.html


Have fun,
Patrick

.
 

Attachments

  • 181103 ZS6 Left.JPG
    181103 ZS6 Left.JPG
    354.5 KB · Views: 633
Very interesting, I have had a couple of pairs of these (my grey pair got nicked at work, so now I have the reds like these :)) and I've been using the Comply wax guards as you mentioned and substituting the mesh grilles for 2.5mm stainless washers. I'll be trying this mod for certain :) can you confirm was it the 0204 0.4w Melfs you went with and which caps as digikey are showing the ECP-u range as obsolete.

Might have to buy another pair of these, someone must be slashing them for black friday :)
 
Just did this mod to one of the 2 pairs my brother brought with him for his visit to the states. Was a little dicey as one of the drivers was rotated to the point where the cap interfered with the dummy screw and boss on the cover. I managed to solder the cap inboard enough where it fit (and put a piece of electrical tape on it just in case.

These sound much better with this mod; before they were "tizzy" sounding in the high end; now they are still a touch bright but quite listenable.

I may just disconnect one of the tweeter BAs on the other pair and see how that sounds. Wiring these in series is the hardest part of the mod, at least to me. I had seen Mr. L's post on this at the other place before and showed it to my brother, who then asked if I would do this. I used the same resistor you listed (also got some 102s), but used the 680nf cap as Mr. L did.

The parts I used (Mouser):
594-MMA02040C7508FB3
667-ECP-U1C684MA5

the 102 resistors: 594-MMU01020C7508GB3
 
Last edited:
> I may just disconnect one of the tweeter BAs on the other pair and see how that sounds.

If you want to do so, IMHO you need to place a resistor to replace the "missing" tweeter.
Else you are still putting full voltage, but now double the current through it.

I find adjusting the damper foam quite effective in taming the brightness.

I did not use 0102, but 0204.


Patrick
 
Good comment, hadn't thought about that. Somewhere I read about someone just disconnecting one of the tweeters, probably on head-fi (which of course is suspect). I'll look into the damping foam... I didn't take these apart any further than I had to.

There is a big thread on HF on KZ devices... I don't hang out there at all anymore but ran across it via google search. Apparently some people are getting drivers from Knowles and doing replacements, but I didn't read far enough to see which KZs they were doing this to. On the one ZS6 that had the driver rotated so that the cap was physically in a bad spot, I did look briefly to see if I could rotate the driver, but these appear to be glued in and I didn't want to break anything, so aborted on that approach.

I was taken aback at how small the existing cap and resistor are... about the size of a grain of sand :) Sure, I could hand solder that...
 
Mine don't have any foam whatsoever in them. Just the screen thing. Eartips are Radioshack Auvio silicone (my brother likes these with his, so that is what is on them).

A question: Mr. L mentioned that 250nf was the ideal capacitor, but that he liked the sound of the 680nf, which is what I used. I noticed you used 470nf on yours. Does the increase in capacitance (680 vs. 470) increase bass (or decrease midrange)? I'm not up on crossover/filter design, hence me asking. The 680 might be a touch bass heavy for my brother, so that is the concern.
 
Posted this on original thread as well. I had simply replaced the original resistors with the MELFs but it looks like you only want to solder the resistor to a driver on one end, and connect the red wire running to the cap on the other end? I assume the solder connections on the driver are the terminals for that driver?

If so, I'll go back in and redo these.
 
I redid one pair the other day, and listened to it quite a bit. Initially I thought it sounded flatter and more neutral, so I decided to compare the 2 pairs on classical music (Bernstein/CSO/Shostakovich 1 and 7). I had assumed that the correctly modded pair would sound much better; I was wrong. The violins and strings sounded a bit steely. The pair where the 7.5R was basically in parallel with the midrange driver sounded better, and was not too bassy (I'm assuming the driver right over the eartube is the midrange).

The correctly modded pair measures (DMM) around 12.5R on the input whereas the paralleled 7.5R measures ~5R (a bit under) at the input. My iPhone seems to drive these just fine, with plenty of headroom. I'm theorizing that the paralleled resistor would pull the midrange output down some; how much I'm not sure of as I don't have the measurement tools nor the knowledge of how to calculate this. With the resistor in place, the driver measures around 6R.

I may try putting the 7.5R in series as in the properly modded pair, and then put something a bit larger across the midrange (12/16/18/20R).
 
A resistor in parallel with the coil only consumes current.
It does not put more energy into the driver, which is determined by the voltage it sees.
You can leave out the 7.5R in parallel and I bet it sounds the same.

By leaving out the 7.5R in series with the mid-range, you increases the voltage it sees.
So its SPL will increase compare to the other two.
And you are now also driving it with full voltage at low frequencies.
With the 7.5R, it forms a R-L low-pass filter with the driver coil.

But if you like it better, that is the only thing important.


Patrick
 
A resistor in parallel with the coil only consumes current.
It does not put more energy into the driver, which is determined by the voltage it sees.
You can leave out the 7.5R in parallel and I bet it sounds the same.

By leaving out the 7.5R in series with the mid-range, you increases the voltage it sees.
So its SPL will increase compare to the other two.
And you are now also driving it with full voltage at low frequencies.
With the 7.5R, it forms a R-L low-pass filter with the driver coil.

But if you like it better, that is the only thing important.


Patrick
Thanks Patrick. I took a look at the pair I corrected the mod on (thinking I might have put the resistor in series at the point that the green wire connects to the driver), but it was done properly, from the other driver terminal to the red wire.

As for energy into the midrange (I believe), having the paralleled resistor should decrease the energy; same voltage but less current available since some of it will be dissipating in the resistor.

I'll agree that a resistor in series for a filter is desirable... not sure how this is an R-L filter though. Where is the L? Driver coil?
 
After some more work (and burning up one pair), some comments:

The "correctly" modded pair is definitely inferior to the pair with the resistor in parallel to the midrange. I ordered some more MELFs (10, 12, 16, 18R). BTW, you Euros seem to really like your MELFs... why is that? Aren't these basically just metal films without leads? They are certainly magnetic. In retrospect, I would have used a thinfilm 1206 or 0805 0.1% (if available) SMT resistor.

On the pair that I burned up, I had tried to clean up the solder a bit, being the third time I had worked on them, with solder wick, and my iron wasn't heating properly. I wound up taking a pad out on the midrange driver and can't find a connection point. I ordered a new pair to replace this pair. I've rewired the BA tweeters to series and will listen to them like this prior to doing anything else. I'm also trying to draw up a schematic, but lack a datasheet or the like showing the connection points on the drivers (they appear to have solder pads in addition to the actual driver connections). The "cap" in the new pair sure seems to measure like it is a resistor and not a cap. Maybe they changed some things. Once I remove the stock components, I'll try to finalize that.

I also had my son, who is an audio engineering major at Columbia College, Chicago, take a listen to these, and he did not like the correctly modded pair at all. I didn't have him listen to them stock (IMO unlistenable due to the brightness). He thought the pair with just the 7.5R in parallel with the mid driver was dark, but sounded good overall. I added a 7.5R in series as well, and he did not like that as well (I agree). I haven't gotten around to trying other values instead of 7.5R.

I'll report back once I do.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.