A story of mysterious interaction between a diy tda1543 dac setup and the human body.

[Bartb]

@Abraxalito: (woke up,) rewired and works correctly now! Thanks! (will listen to it when day dawns though ) Will also try this circuit with 24x TDA1543 and see how it sounds, asap.

 

[Bartb]

@mmerrill99 Hello! this is the ebay auction number for the SD player: 122618274108
There's no pictures of the remote control so I guess the sellers are just too lazy
My guess is that all of these players come with a remote. If you don't get a remote, will see what I can do to get codes.
This is the picture of the SD card. I think it's fine and within those specs?

 

[mmerrill99]

Yea that as card should be fine but have you tried another one, just in case? As I said I know of two with the exact same SD player that have no problems - one other possibility, there is mention of firmware update on some sites that sell this but no idea how you get the firmware or instructions on how to update? They seem to make a big deal about the card having the latest firmware so maybe your problem is the sump prom of the old firmware?

 

[mmerrill99]

Let me correct my last post - it seems that there can be white noise but it seems to be track dependent - seems to be related to how the track was ripped? I don't know what the conditions are that give rise to this but it seems DbPoweramp is one of the offending rippers that causes this - maybe it's track offsets or some such condition?

 

[Bartb]

Hello! Yes! That's exactly the ripper I'm using.
I'll look into this and also check dbpoweramp's forum.
Kind of annoying after I've ripped my entire collection
Thanks for rhe tip.

 

[mmerrill99]

Yep, it's a known problem - see here dating from 2009 & 2014 - haven't read all the posts
Hissing noise at the end of WAV tracks
clicks and burst of white noise at end of rips

 

[Bartb]

@mmerrill99 Hello. did some digging+experiments
This particular player doesn't recognize the (music) file's header, it thinks it is sound hence the burst of static.
It is possible (with dbpoweramp) to rip tracks and disable "WAVE ID Tagging".
I have verified that static disappears when this is done.
( For the record: > Audio Codecs>Wave ID Tagging>Tag Creation>None )
So if you want to get this particular player I see two alternatives.
You could rip CDs without a header (but then the files will be incomplete so not recommended).
Or else: rip CDs with header to your HD. The moment you need some album on SD card you perform a "Batch Conversion" (from WAVE) to WAVE without tags and destination SD card. Simply disable tags the moment you do this conversion.

@Abraxalito Meanwhile you circuit performs very well. Really really like it !

 

[abraxalito]

I used to use dBpoweramp for ripping. I got a loud click at the start of .wav files (played back on a QA550 SD player), but no static. But then for some reason they started charging money for the app. Now I use freac (free audio converter) and rip to .flac. If I need .wav the same app can transcode.

Glad you're enjoying the sound! Good that we've cleared this up about the player, now I can order one. I am assuming that it does indeed play .flac?

 

[Bartb]

Hello!
just checked, yes it plays FLAC files as well (note photo). A few more notes..
The software is a bit buggy.
Please note (photo) that it does not order the songs correctly (just in FLAC).
The folder name appeared in chinese in previous screenshot.
Sometimes a song title is not displayed correctly (in WAV, but song plays okay).
No static burst in FLAC even though I used default settings (in dbpoweramp).
Player has two clocks for 44.1Khz and 48Khz (photo) .
( I wonder if you'll want to upgrade those. What I also have in my mind is placing a very good 44.1khz clock right next to the DAC and feeding the player with it - losing 48k probably)
PS yes, 80s and 90s oriented

 

[mmerrill99]

@mmerrill99 Hello. did some digging+experiments
This particular player doesn't recognize the (music) file's header, it thinks it is sound hence the burst of static.
It is possible (with dbpoweramp) to rip tracks and disable "WAVE ID Tagging".
I have verified that static disappears when this is done.
( For the record: > Audio Codecs>Wave ID Tagging>Tag Creation>None )
So if you want to get this particular player I see two alternatives.
You could rip CDs without a header (but then the files will be incomplete so not recommended).
Or else: rip CDs with header to your HD. The moment you need some album on SD card you perform a "Batch Conversion" (from WAVE) to WAVE without tags and destination SD card. Simply disable tags the moment you do this conversion.

@Abraxalito Meanwhile you circuit performs very well. Really really like it !

Hello!
just checked, yes it plays FLAC files as well (note photo). A few more notes..
The software is a bit buggy.
Please note (photo) that it does not order the songs correctly (just in FLAC).
The folder name appeared in chinese in previous screenshot.
Sometimes a song title is not displayed correctly (in WAV, but song plays okay).
No static burst in FLAC even though I used default settings (in dbpoweramp).
Player has two clocks for 44.1Khz and 48Khz (photo) .
( I wonder if you'll want to upgrade those. What I also have in my mind is placing a very good 44.1khz clock right next to the DAC and feeding the player with it - losing 48k probably)
PS yes, 80s and 90s oriented

Thanks Bart - good info & glad it all worked out

As far as the clocks are concerned - you need both clocks - the ARM processor uses one of these audio clocks as its system clock & switches between them depending on the samplerate of the audio file being played.

Why not use two 'good' audio clocks & Flip Flop to reclock the I2S stream just before DAC input. Feed the clock signals back to the correct on-board solder pads & also connect each clock enable signal from the correct solder pad to the correct clock.

 

[mmerrill99]

In that picture - Bottom right corner pad of each osc is clock signal output
Top left corner pad is enable signal pad for each clock

 

[kp93300]

Can i replace the BC 807 in the single pnp output stage by a BC547?

 

[abraxalito]

Not and have it work, no. The BC807 is PNP, the BC547 is NPN. If you wish to substitute transistors you need to ensure they're the same type (NPN for NPN and PNP for PNP) as an absolute minimum.

BC557 would probably be OK.

 

[Bartb]

In that picture - Bottom right corner pad of each osc is clock signal output
Top left corner pad is enable signal pad for each clock

Thanks for the tips ! I was wondering at some point how the clocks are activated/alternated, this clears it up.
At some point I'd also like to place a minidsp-minisharc between the dac and the player. But that will be much later when everything else is understood and working

 

[Alexandre]

Will also try this circuit with 24x TDA1543 and see how it sounds, asap.

The power supply becomes a bit of a problem. Especially if you leave it on 24/7. Remember TDA1543 pulls 50mA each.

In my opinion the paralleling of many dac chips is overrated. I actually get the best results with one selected TDA1543. They all sound a bit different from one another, especially different batches.

-Alex

 

[abraxalito]

Yeah I agree Alex, simply paralleling DAC chips is over-rated, if the aim is to get better SQ.

What I plan to do is to try out paralleling complete DAC modules (based on TDA1387s) - that is DAC chips with their own dedicated power supplies (shunt regs giving nice low output impedance) and see how that sounds. Simply multiplying DAC chips puts more onus on the power supply.

 

[Bartb]

The power supply becomes a bit of a problem. Especially if you leave it on 24/7. Remember TDA1543 pulls 50mA each.

In my opinion the paralleling of many dac chips is overrated. I actually get the best results with one selected TDA1543. They all sound a bit different from one another, especially different batches.

-Alex

Yes power supply becomes a problem.
Anybody paralleling 24x @9V: this is the *required* heatsink.
Heatsink measured at +9 degrees C over room temperature

 

[Bartb]

Yeah I agree Alex, simply paralleling DAC chips is over-rated, if the aim is to get better SQ.

What I plan to do is to try out paralleling complete DAC modules (based on TDA1387s) - that is DAC chips with their own dedicated power supplies (shunt regs giving nice low output impedance) and see how that sounds. Simply multiplying DAC chips puts more onus on the power supply.

Interesting... how many modules would you parallel ?

 

[abraxalito]

I was thinking 4 initially. The aim of the experiment would be to see if the bass improves still further. If the bass is limited by low frequency noise then paralleling should give a reduction in noise. The result of listening to 4 paralleled modules then would give an indication whether its worth trying more...

Already with my lingDAC the bass to my ears sounds better than any other DAC I've heard, but only with a specially low impedance, low noise supply (well under 1ohm Zout).

 

[Bartb]

I was thinking 4 initially. The aim of the experiment would be to see if the bass improves still further. If the bass is limited by low frequency noise then paralleling should give a reduction in noise. The result of listening to 4 paralleled modules then would give an indication whether its worth trying more...

Already with my lingDAC the bass to my ears sounds better than any other DAC I've heard, but only with a specially low impedance, low noise supply (well under 1ohm Zout).

Let us know how that works out..
If you are focusing on bass, maybe you will also like this "capacitor map" that I found on lampizator.eu 's site.

At this point I'm curious to hear the TDA1387 as well
With reference to your circuit (Single PNP transistor in common-base mode.) the TDA1543 is replacable with the TDA1387 without modifications?