[piaseczek]

b0bb,
I have the newest version of the recommended PSU but it's a simplified version. Do you think that it will work as good as yours after the modification?

 

[piaseczek]

More photos.

 

[b0bb]

b0bb,
I have the newest version of the recommended PSU but it's a simplified version. Do you think that it will work as good as yours after the modification?

That is OK.
The first version copied the schematic from the appnotes and it had parts that were optional.

 

[piaseczek]

That is OK.
The first version copied the schematic from the appnotes and it had parts that were optional.

Great news, thanks b0bb.

 

[piaseczek]

[4] Analog post filter.
(Check your actual DAC post filter configuration is similar to my early revision 004)
33nF 63V 5mm lead spacing 1% Kemet PHE426 metallized polypropylene Qty:4
2200pF 100V 5mm lead spacing 1% Vishay KP1830 film/foil Qty:12

b0bb,
Have you seen the 33nF film and foil wimas?
https://pl.mouser.com/ProductDetail/WIMA/FKP2C023301L00HC00/?qs=WI052HLiD5iWRZJ3Tjlxjg==

Isn't it a better option than the kemet phe426?

 

[b0bb]

b0bb,
Have you seen the 33nF film and foil wimas?
https://pl.mouser.com/ProductDetail/WIMA/FKP2C023301L00HC00/?qs=WI052HLiD5iWRZJ3Tjlxjg==

Isn't it a better option than the kemet phe426?

Why do you think it is better?

These have a tolerance of 2.5%, so the frequency response will be out by 5% overall.
Film foil construction will not amount to much if the basic frequency response is not right.

Do you really want to wait 4months (16 weeks) to get them?

 

[piaseczek]

Why do you think it is better?

These have a tolerance of 2.5%, so the frequency response will be out by 5% overall.
Film foil construction will not amount to much if the basic frequency response is not right.

Do you really want to wait 4months (16 weeks) to get them?

Ok, I got it. The film and foil technology isn't as important as the 1% tolerance.

 

[piaseczek]

b0bb,
original, blue MKTs are 5% tolerance? If that is correct I see here an easy upgrade.

 

[b0bb]

b0bb,
original, blue MKTs are 5% tolerance? If that is correct I see here an easy upgrade.

You need to specifically check the ones installed on your DAC, if the are stamped with a "J" then it is 5%, might also be printed on the case.
The ones on mine are 5% Mylar, this is a substantial cost reduction from the 2.5% Polypropylene LKS used on the MH-DA003

 

[b0bb]

Ok, I got it. The film and foil technology isn't as important as the 1% tolerance.

The tight tolerance parts are readily available now.

This will not always be the case, get them while you can.

 

[b0bb]

This is the DS26LV32A I2S receiver on my DAC.
It is not LVDS, instead it runs RS-422, a slightly different protocol.

RS-422 signal levels are higher than LVDS, this might explain the slightly weird termination that LKS used to make it compatible with LVDS transmitters.

The chip tops out at 32MHz, this precludes DSD1024

DSD1024 is just under 50MHz, as such I2S performance is suboptimal on the 004
I replaced this a faster receiver from Sipex/MaxLinear, the SP26LV432CN is designed to operate up to 50MHz

There was an immediate improvement, like a haze was just lifted from the music, everything become clearer.
Soundstage became wider, high notes lost the vestiges of sibilance,especially percussion instruments.
Bass notes went lower without getting muddy.

I can now reliably drop the DPLL setting down to 7 with DSD512, it was not stable with the original receiver chip.
Power consumption dropped between 6-10% with this new receiver chip.

 

[b0bb]

This is a rebuild of the digital power section to increase the current supply to the DAC

All the stock 1.5A regulators were replaced with 3A capable units with transient and noise handling equal to or better than the stock LT1963A

The Nichicon caps have been in continuous use for almost 4 years, these were replaced with 380LX series caps from Cornell Dubilier Mallory (CDM)
Tube amp owners may recognize the Mallory brand.
These caps have 3X lower dissipation loss compared to the Nichicon KG caps, the old caps were running quite hot.
Endurance/lifetime improve by 3X with the 380LX.

Post filter caps are 2700uF 125C automotive grade Nichicon UBY.
The modern regulator replacements allow for the use of relatively large post filter caps to help with the short term current demand

Polypropylene bypass caps replace the original mylar ones. These are the square green boxes from WIMA.

Vishay MBRF10H100 schottky diodes used for the rectifiers.

Caps values left to right 10,000uF 35V x2, 22,000uF 25V x1, 47,000uF 25V x1

 

[b0bb]

The regulator replacements, starting from left to right
Belleson SPX17 5V 3A for the primary 5V regulator

3A LT1764 replaces original 1.5A LT1963
2k and 499R Vishay PTN Tantalum nitiride 0.1% 10ppm resistors used to set the voltage

The 2 Auxillary 5V regulators use LT3083, this has similar performance to the LT1963, outputs 3A current.
Design is similar to the LT3042 requiring just 1 resistor to set voltage
100k Vishay PTN Tantalum nitiride 0.1% 10ppm resistors used to set the voltage.

 

[b0bb]

Initial board prep, remove all the stock parts that are to be replaced leaving just the SMD resistors

 

[b0bb]

LT1763 with solder paste just before hot air soldering

After hot air soldering