Diy-dac 1543

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Yes, it looks quite similar to this which I bought a few months ago: http://cgi.ebay.co.uk/Mini1543-DAC-...hash=item290303676785&_trksid=p3286.m63.l1177

I think it's pretty nice, particularly when you consider the price, and I've been very pleased with the sound I've been getting. The board is good quality, very thick.

I've been using it with an active solid state I/V stage, but this evening I reverted to passive I/V by modifying it along the lines described towards the end of this post by ecdesigns: http://www.diyaudio.com/forums/showthread.php?postid=1813308#post1813308

The result in my system is a very noticeable increase in clarity and realism. A simple mod, well worth doing I think.
 
Hi John

the ss stage is way too complicated for this little DAC and for my skills. The ecdesign seems to be more interesting.

I am not sure how to mod the circuit of the TDA1543 pcb with the ecdesign I/V-stage because i am not an electrician.

Is that the relevant part that has to be done on the pcb? Where have you put the ground of the 3V battery?
 

Attachments

  • onlyiv.gif
    onlyiv.gif
    19.2 KB · Views: 2,292
Tolu said:

Is that the relevant part that has to be done on the pcb? Where have you put the ground of the 3V battery?


Yes that's it. I put the ground of the 3v battery to the -ve pin of the cap that connects to pin 4. Ecdesigns also shows a small 2n2 cap between +3v and earth. I put mine directly across the battery connections.

The ends of the I/V resistors (1k on the stock board, I think) that formerly connected to ground are lifted, then joined and connected to +3V from the battery. If you keep the 1k resistors you will get a slightly higher output than with the 680R shown.

You need to remove the resistor that connects to pin 7.

The output caps can remain as they are.

Hope this helps.
John
 
Hi John

thanks! I think I have understood.

How would you judge the sound quality of the hifidiy DAC in mint condition out-of-the-box, after Leos'/Peter Daniels' mods (8V, 2.7k, 1.2k) and after ecdesign mods (3V battery, 680R) on a absolute scale from 0 to 100?
 
The ends of the I/V resistors (1k on the stock board, I think) that formerly connected to ground are lifted, then joined and connected to +3V from the battery. If you keep the 1k resistors you will get a slightly higher output than with the 680R shown.

1K will result in a distorted output signal, 900 Ohms (2Vpp output) is about maximum with 5V power supply. Vref also has to be increased to approx. 3.6 ... 3.8V when using I/V resistor values higher than approx. 680 Ohms. This could be done using 3 NIMH batteries in series. batteries have to be used because mains-power supplies are to noisy, and they aren't "floating" like with a separate battery.

The advantage of this slightly modified passive I/V conversion is that Vref pin can stay unconnected. The Vref pin will pick-up noise / hum, and route it to highly sensitive DAC analogue circuits. This will cause pollution of the DAC output signals (sound smearing).

Vref was intended for providing a reference voltage for external OP-amps, and for adjusting bias current. In practice this pin is best left un-connected.

The 2n2 is only for decoupling the battery power supply (high frequencies), best caps I tested so far were 2n2 silvered mica caps. ceramic caps may have piezoelectric effects that degrade performance. When using film caps, stacked versions (with lowest inductance) are preferred.

I used small WIMA 1uF / 50V output coupling caps (2.5mm lead pitch), these are small and provide very good performance. They even appeared to produce a slightly more realistic (fuller) sound compared to the expensive Vcap Teflon Tin caps. This test was done using the SD-card player and a 12-crystal lithium-cell referenced precision master clock. I also used a dynamic jitter reduction circuit on the BCK input, and resistive I2S attenuators on both WS and DATA inputs.

DAC sound quality largely depends on I2S signal quality (jitter / jitter spectrum). Attempting to tune a DAC that's fed by poor quality I2S signals is like attempting to tune a phono pre-amp while the turn table speed isn't constant (slipping belt for example). The end result is a flawed output signal, regardless of DAC / phono pre-amp mods.
 
ecdesigns -

That's very helpful, thanks for your input.

Thomas -

No subjective assessment can be 'absolute' I'm afraid. The different approaches will give different results but I can't tell you which is 'better'.
For myself, I shall take note of what ecdesigns has said and not pursue any further tweaking of this dac. I eagerly await the appearance of the SD-card player. :cool:

John
 
I tried this mod and posted some impressions on the AOS forum

I'd like to say a big thank you to ecdesigns for sharing this, IMHO with this mod the TDA1543 is the best I've heard it!!!

Before the mod I was using 2k7 Riken I/V, 1k5 Riken Ref , 8v supplying the TDA1543 using AN8008
I built the TDA1543 + ecdesigns mod on a separate piece of vero and tapped into the I2S lines
 
leo said:
I tried this mod and posted some impressions on the AOS forum

I'd like to say a big thank you to ecdesigns for sharing this, IMHO with this mod the TDA1543 is the best I've heard it!!!

Before the mod I was using 2k7 Riken I/V, 1k5 Riken Ref , 8v supplying the TDA1543 using AN8008
I built the TDA1543 + ecdesigns mod on a separate piece of vero and tapped into the I2S lines


How do you drive the TDA? With 8V or 5V with ecdesign-mod? Did the highs and width of the soundstage improve?
 
Tolu said:
In another thread over at audiocircle I've read that output impedance is 2300 Ohm (with Peter Daniels values). So, this would be a no-go for my system because the DAC is in front of a passive pre (Lightspeed) with just 7 kOhm input impedance.:whazzat:

I think 7K input impedance is a bit low, even with a 680 I/V. I would probably put a buffer at the dac output.
BTW, how does the "Lightspeed" sound.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.