Modding the DirectStream DAC MKI

Haha thanks :slight_smile: And apologies for my forum ignorance here. I’ve just been super busy and haven’t been able to read up on everything.

These things are not obvious, you are far from ignorant. :slight_smile:

1 Like

Modification of the top plate to get better cooling.

  1. Opening milled.
  2. Output plate.
  3. Cork top plate

4 Likes

Thanks Elk :smiley: My dual overhead gerbil wheels plead the fifth on that :wink:

2 Likes

Next round of wine is on T . . er, wait, what forum is this? :thinking: :sweat_smile:

1 Like

Hahaha… Reaching for my wallet already!? I see how you guys are :wink:

2 Likes

I think I will also try the EMI/RFI shielding inside at Tedā€˜s previous recommended places as I have some Sillpoint ERS and mumetal on hand.

In contrary to the 3M paper, the Stillpoint ERS is effective down to the 10MHz range, which according to Ted’s info is partly relevant (see chart below).

@tedsmith: in case I could place it with some isolation (e.g. behind the display board), will the ERS or the mumetal be more effective in the 10-1000 MHz range?

SPECIFICATIONS
ERS shielding effectiveness (in db attenuation):
10 Mhz - 49
100 Mhz - 48
200 Mhz - 62
400 Mhz - 53
800 Mhz - 54
900 Mhz - 50
1 Ghz - 49
2.4 Ghz - 53 (802.11B)
3.5 Ghz - 54
5.8 Ghz - 57 (902.11A)
7.5 Ghz - 61
10 Ghz - 66
17 Ghz - 55

Mu metal and ERS do different things, they could easily be independently useful. On the other hand most of the DS (except for the output transformers) is pretty insensitive to magnetic fields (I explicitly try to keep loop area small, etc.) Mu metal needs to be a closed cage to be effective so it’s hard to shield things that weren’t designed to expect to use it. Similarly you’d need a Faraday cage to nuke conductive noise. But using things like ERS to attenuate RFI is effective without being a closed surface. The display board is indeed a good place to put some ERS to separate it from the rest of the interior. If you don’t use the Bridge take it out. If you do, perhaps some ERS to separate it from the rest of the box might be helpful. But note, ERS is much more effective near either the source or the receiver of the RFI, not just as a wall between them.

5 Likes

Thanks, that’s what I also wanted to ask…a sheet of ERS directly above the bridge card? I still use it currently.

And does a stripe of ERS behind the output transformers, on the bottom side of the digital board make sense? That’s the only place with parts on the analog board below.

Didn’t realize that mumetal doesn’t work as sheet as many years ago I got rid of some hum just placing a sheet at a certain place, but also understood now that it’s just against magnetic fields.

The bottom of the digital board is good, just don’t short any vias there. If you have the XS4400 transformers, they stick up thru the digital board and IIFC that gap probably isn’t wide enough to shield with ERS (I could be wrong.)

4 Likes

Correct it isn’t there at all the gap. It’s just possible to shield the space behind, not the transformers themselves. Except if it would make sense to shield the iron around the coils.

In case the ERS could theoretically short anything I could use a one sided adhesive tape on the one side of an ERS stripe and a double sided on the other and stick them together all around and stick the double sided to the upper board.

Hi all. Is anyone going direct to amp with this mod? Seems like some hint of negativity possibly if you do not have a pre in the mix?

I can fully confirm your impressions. I also have done the Farrad + XS4400 mod. Together with Windon its a ā€œdream teamā€. In my experience, every of the components needs about 300 hrs burn in. Specifically the Farrad even more. Not sure how much you have on the clock, but there may be more to come …

1 Like

I can just tell in my case with preamp, I usually had lifelike level (means the loudest level I’d like to listen to for some time, which is probably louder than what most others would prefer) between 10 and 11 o’clock (high gain recordings) and between 1 and 4 o’clock (low gain recordings).

After the transformer mod it seemed like a wider macro dynamic range. Low volume parts of the music are lower in level, loud parts louder (but only the maximum peaks). This led to the impression that for most of the dynamic spectrum I now have to turn up to 11-2 o’clock (high gain recordings) and don’t have enough range anymore for the really low gain recordings for maximum volume (but I still reach fairly high levels).

An example for a very low gain recording is this 24/176,4:

Without a preamp in my case the DS wouldn’t be able anymore to drive to loud levels, but this wasn’t possible before, too, with low gain recordings.

I use balanced connection, so I don’t have the -6dB problem.

5 Likes

Interesting! Thanks for sharing.

Btw, the heat sinks now are cold enough to be touched for 30sec. Before 1sec.

2 Likes

update in update.
Belleson has been added to lps.
https://www.belleson.com/
@wctaudio

3 Likes

At about 100 hours burn in.
Not going to really wax poetic in audiophile terms. :grinning:

This transformer mod is very good. Clearly an improvement across the spectrum and great relative value.
Thanks to all who explored this and provided info and service.

3 Likes

Thank you everyone for this great forum!
I would also like to do the FARAD (and transformer) modification (with the TE Connectivity connector No. 350428-1). Mattijs from FARAD asks in this context: Do you also know the pinout, where should plus go and where GND? (For assembly with a level 2 cable connection GX16-4)?
@ Ted Smith: Can you answer that, please? Thank you in advance!

2 Likes

Always follow the red cable for plus and use a multimeter for checking cabling. Red is pin as usual. But Ted will tell better I’m sure.