DAC gallery

Internal view:
 

Attachments

  • 08.JPG
    08.JPG
    106.1 KB · Views: 4,188
  • 09.JPG
    09.JPG
    127.2 KB · Views: 4,068
  • 10.JPG
    10.JPG
    140.4 KB · Views: 4,243
My Scott Nixon USBDKII, based on the TDA1543.
Simple but very good sounding unregulated power supply, integrated volume control (Alps 20k), Sedlbauer toroid transformer 30VA, Wima output caps and the very important Sanyo Oscon for TDA1543 decoupling ;-)
I really like the sound of this dac! It's dynamic and fast, lively, coherent and natural sounding with a remarkably controlled and tight bottom end. Very open and forward, but without being excessive in any area.
But I tested MANY different power supply designs and different output coupling caps before I threw all this into the case...

What about this very nice chassis and the panel knobs ?

matthias
 
Hi,

My own DAC is based on the Satch Kit, which is similar to Thomas's on-line project (where he solicited much advise from me but ignored all of it), but with most of the major design flaws he put into his design (and which all the chinese copycat's promptly copied) already resolved.

In my own unit I re-build pretty much everything.

The digital input section using a circuit that is a mixture of Pat "Jocko Homo" DiGiaccomo's input circuit and an inverted, single supply version of the two transistor buffer Jim Hagerman used in his "Chime".

344cxah.jpg


Receiver is still the CS8414 on an adapter. I have added loads of SMD caps around the CS8414 (all the way to 10uF for supplies), both in PSU decoupling and for the PLL Filter. Around the CS8414 I use TL431 Shunt regulators (also a separate one for the input circuitry) on small PCB's that include the bypass capacitor for the feedback loop, for lowest noise.

fee0ir.jpg


The digital filter and the idiotic asynchronous reclocking have been removed. I used the divider and the crystal position on the PCB to implement DEM reclocking as per the meta-mega-thread on TDA1541 DAC's by ecdesigns. The former digital filter position now holds a small PCB with the I2S attenuators (again, as per ecdesigns).

Overall on the input and digital side I have four individual shunt regulators, one for pretty much every section, each with a CCS. I have two each of Sanyo 2200uF/6.3V Low Impedance Cap's (the green ones) on the input PCB and the Digital PCB respectively plus Os-Con's (including 470uF ones) and SMD Ceramic caps for decoupling.

So for example each supply pin on the CS8414 has one 0.1uF SMD Capacitor in 0603 directly at the IC Body, 1uF & 10uF SMD in 1206 on the Adapter pins, 470uF/6.3V Os-Con plus TL431 (TO-92 case) with ac bypass cap on the voltage set resistors and 2200uF/6.3V Sanyo Low Z.

The digital supply of the 8414 is shared with the "utility" digital stuff, each IC has more of the same decoupling, 0.1uF/0603, 1uF/1206, 10uF/1206 SMD under the PCB and 220uF/10V or 470uF/6.3V Os-Cons (depends what i had around).

The TDA1541A(S2) Uses Wima MKS and MKP capacitors for dem decoupling, with 8pcs 4.7uF MKS on the MSB's (4 per channel) and then 2pcs 2.2uF MKS (1 per channel) and 4pcs 1uF MKP (2 per channel). I would have preferred to use SMD Film Capacitors here, but that was not an option due to the PCB Layout, so I went for the largest value film cap's I could fit. This is an area where further gains can be made (see ecdesigns thread and his most recent TDA1541 DAC Module design.

I used Elna Silmic Capacitors for decoupling plus SMD Film Cap's at the DAC Pins. Again, TL431 Shunt Regulators, actually in this case literally daisy-chained like a zenner string. Fed By CCS. So each TDA1541 PSU Pin has a 0.1uF low inductance SMD Film Capacitor, 470uF/16V Elna Silmic II and the TL431 with ac bypassed setting resistors for the voltage setting.

Analog stage is the diyhifisupply universal tube stage. I added external I/V resistors and the Non-OS high frequency rolloff compensation as per Pedja Rogic. I also replaced the filter resistor in the powersupply that is on the PCB with a choke and added an off-board 33uF/160V film capacitor as the final filter. These two parts take almost the same space as they whole tube stage circuit board though, so I was lucky I had spare space left.

5ectop.jpg


The actual powersupply circuit board of the Satch was also completely re-structured. The Power transformer in the Satch kit has two windings, one center-tapped to feed a lot of stuff, including the op-amp output stage and the DAC and another single winding.

I used this single winding and schottky diodes to make the basic supply for the digital side. This is followed by an RC/RC/RC/RC Filter, the capacitors are some Nichicon 4,700uF/25V, I think I used 3.3 Ohm resistors, in both "ground" and "+V" line.

As the Satch PSU has ton's of regulators I re-purposed them to give a chain of regulating the negative line with an LM337 (already in place) and using LM317 as current sources in the positive lines, all links to the main PCB carrying the voltages and grounds were replaced with Chokes, 2.2mH, lowish DCR.

I build another such supply for the DAC, this time using 3,300uF/50V Elna Silmic capacitors (so again schottky rectifiers -> RC -> RC -> RC -> RC) followed by an LM317 as positive voltage regulator and a second LM317 set up as current sink. This feeds the Zenner string again via chokes, 10mH this time I think.

2pyxcnc.jpg


Both basic supplies are completely separate and have a fairly high amount of filtering for both common mode and differential noise already. The passive RC filtering already attenuates the 100Hz by around 50dB and together with the LM317 regulators removes any noise from that source.

So the TL431 Shunts only need to deal with the noise from the LM317's (not a lot) and the AC signals dumped into the powersupplies by the actual circuitry.

Most will likely find my approach overbuild and unneccessary complex, but it works well.

Before anyone asks for schematics, sorry, there are non and unless I take my own working DAC apart and reverse engineer it I could not even draw a schematic. Most of the work was done without anything written down or drawn out.

So this is only meant to give some ideas on how extreme one may mod a simple kit.

Ciao T
 
Last edited:
strange there isnt more in here actually isnt it? given how busy this section of the forum is these days.

heres 2 of the 4 channels of my dac in testing phase. 4 channels used for digital XO

in this pic its directly connected to the BAL-BAL version of the wire headphone amp with the input resistors removed so acts as a buffered IV stage that can drive headphones directly; and boy does it!!
 

Attachments

  • fifo-akd12p testing lock but no audio_09May2012_0936.jpg
    fifo-akd12p testing lock but no audio_09May2012_0936.jpg
    418 KB · Views: 1,858
  • fifo-akd12p testing lock but no audio_09May2012_0937.jpg
    fifo-akd12p testing lock but no audio_09May2012_0937.jpg
    358.8 KB · Views: 1,916
yeah probably, so why dont you report the post using the exclamation mark button on the left and ask if we can have it as a sticky? i'll post some more pics in the next week that are in focus; with the various other IV stages and i2s setups i'm testing out
 
strange there isnt more in here actually isnt it? given how busy this section of the forum is these days.

heres 2 of the 4 channels of my dac in testing phase. 4 channels used for digital XO

in this pic its directly connected to the BAL-BAL version of the wire headphone amp with the input resistors removed so acts as a buffered IV stage that can drive headphones directly; and boy does it!!

Hi Qusp

Some great work posted on this thread, well done everyone.

How much did Ians FIFO boards make a difference?

Those boards with the LT1764 regs on them - Where did you get them from if you dont mind me asking?

What are you using for a transport? (bottom right corner?)

What power supply setup do you have for your DAC?

Great work by the way.

Ryan
 
Hey Ryan, soz mate, just saw your post, i'll post detail tomorrow but figured I should post here so I remember and leave the window open. loving the fifo, but i'm going to have to give it up for a while as i'm about ready to put the 2 dacs together for 4 channel 2 way digital active XO and the fifo doesnt fit in the scenario. Ian will be doing a multichannel soon enough, but till then it'll be limited to my headphone rig. anyway more detail tomorrow I gotta crash.
 
Hey Ryan, soz mate, just saw your post, i'll post detail tomorrow but figured I should post here so I remember and leave the window open. loving the fifo, but i'm going to have to give it up for a while as i'm about ready to put the 2 dacs together for 4 channel 2 way digital active XO and the fifo doesnt fit in the scenario. Ian will be doing a multichannel soon enough, but till then it'll be limited to my headphone rig. anyway more detail tomorrow I gotta crash.

Hi Qusp,

Still very interested in seeing some more photos/details of all your work...:eek:

Ryan