Any good TDA1541A DAC kit?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
P

paul_peraic

Member
Joined 2008
vanofmonks said:
Here is one based on Thorsten's 2005 schematic. The inductors are Panasonic with mu metal shields. To save board space, output capacitors are external of the board.
Click to expand...

Well I will definitely be in on a group buy if necessary (including the tube output stage) very much looking forward to the finished product (hopefully Thorsten will not mind others using his output stage if asked nicely as it would be a great addition :D)

Good idea with the tube output capacitors as they can get very big depending on choice . I for one always have trouble getting those big ones to fit nice .

I have heard putting something like an LT1086 in front of the shunts makes them even quieter can anyone answer if this is true or not ?

Would it maybe be an idea making the on board capacitor diameters just a little larger ?
 
R

regal

Member
Joined 2004
paul_peraic said:
Well I will definitely be in on a group buy if necessary (including the tube output stage) very much looking forward to the finished product (hopefully Thorsten will not mind others using his output stage if asked nicely as it would be a great addition :D)

Good idea with the tube output capacitors as they can get very big depending on choice . I for one always have trouble getting those big ones to fit nice .

I have heard putting something like an LT1086 in front of the shunts makes them even quieter can anyone answer if this is true or not ?

Would it maybe be an idea making the on board capacitor diameters just a little larger ?
Click to expand...

You would want to use a DN2450 casode instead of a LT106 feeding the TL shunt. Much lower noise.

As far as this ube stage I see problems. The inductors are going to pickup lot of hum, the best inductors for this are air coil and they are big.

Thorten's tube I/V stage is much different now than this old design, he now uses a gnome stage:

http://www.diyaudio.com/forums/digital-source/30025-decoupling-tda1541a

http://www.diyaudio.com/forums/tubes-valves/207357-designing-gomes-i-v-stage-tda1541a.html

If one were to make a pcb it would make sense for a gnomes stage but for what is being proposed point to point is better, not many are fond of the 6922/6dj8 anyway with its high H3.

Also Oleg had a webpage that showed one needs to inject 2mA current to offset the TDA1541's dc offset at the top of the i/V resistor, same as what Pedra Rohic does.
 
S

studiostevus

Member
Joined 2009
regal said:
You would want to use a DN2450 casode instead of a LT106 feeding the TL shunt. Much lower noise.

As far as this ube stage I see problems. The inductors are going to pickup lot of hum, the best inductors for this are air coil and they are big.

Thorten's tube I/V stage is much different now than this old design, he now uses a gnome stage:

http://www.diyaudio.com/forums/digital-source/30025-decoupling-tda1541a

http://www.diyaudio.com/forums/tubes-valves/207357-designing-gomes-i-v-stage-tda1541a.html

If one were to make a pcb it would make sense for a gnomes stage but for what is being proposed point to point is better, not many are fond of the 6922/6dj8 anyway with its high H3.

Also Oleg had a webpage that showed one needs to inject 2mA current to offset the TDA1541's dc offset at the top of the i/V resistor, same as what Pedra Rohic does.
Click to expand...

I am using the Gomes stage in my dac, and it sounds excellent. One thing about the 2ma offset: i am using the negative offset as fixed bias to be able to bias the tubes very near to their 0v line (while surely maintaining a negative bias). So not in all cases it's useful to compensate the offset.
 
P

paul_peraic

Member
Joined 2008
regal said:
You would want to use a DN2450 casode instead of a LT106 feeding the TL shunt. Much lower noise.

As far as this ube stage I see problems. The inductors are going to pickup lot of hum, the best inductors for this are air coil and they are big.

Thorten's tube I/V stage is much different now than this old design, he now uses a gnome stage:

http://www.diyaudio.com/forums/digital-source/30025-decoupling-tda1541a

http://www.diyaudio.com/forums/tubes-valves/207357-designing-gomes-i-v-stage-tda1541a.html

If one were to make a pcb it would make sense for a gnomes stage but for what is being proposed point to point is better, not many are fond of the 6922/6dj8 anyway with its high H3.

Also Oleg had a webpage that showed one needs to inject 2mA current to offset the TDA1541's dc offset at the top of the i/V resistor, same as what Pedra Rohic does.
Click to expand...

So then what would be an alternative that is as readily available and as affordable as the 6922/6DJ8/6N23P/6N1P ?
 
Last edited:
P

paul_peraic

Member
Joined 2008
studiostevus said:
I am using the Gomes stage in my dac, and it sounds excellent. One thing about the 2ma offset: i am using the negative offset as fixed bias to be able to bias the tubes very near to their 0v line (while surely maintaining a negative bias). So not in all cases it's useful to compensate the offset.
Click to expand...

Are there any purchasable boards available for this ?
 
A

AudioLapDance

Member
Joined 2003
s3tup said:
Feeding back the bitclock/masterclock/other sync signal to the soundcard is easy, as long as your soundcard supports it.

There are really cheap Envy24-based cards, with Tremor chipset - they go for 5-10$ a card, have tappable i2s lines, and onboard xtals for 44100 and 48000 sampling rates.
You can take the 44100 xtal from the board, throw it on the dac board + reclock the i2s lines with it. Then you feed the Tremor with that clock directly.

Easy, cheap, and you can withdraw all the SPDIF conversions all together, by throwing-in RS485 transceivers on both ends of transmission line.
You get slight ground break, DAC-sourced clock, and no SPDIF.
In the other hand, you get PC-only DAC. High quality soundcard, that's it.


You can use any soundcard, which have SPDIF input, and which is able to lock on external SPDIF, and sync-up with it.
Then you use reclocker/clock distributor on DAC side, and feed the soundcard back with a sync signal in the SPDIF envelope.

SPDIF has simple square wave signal for digital "0", with 64fs rate. You can form it by passig the MCLK thru 1/4 divider + transformer.

Third, but not last solution - is to use Infrasonic Quartet soundcard, which can thake 128fs/256fs sync signal directly, bypassing the spdif thing. It has BNC connectors for masterclock in/out.

All you need to do to eliminate the jitter, is to clock the DAC with local fixed clock (no VCOs), and control the incoming data flow by feeding back the clock to digital source.

You need, as stated several times above, to badwidth-limit the data lines going to the TDA. Attenuation would be nice too.
You'll need to do this on the I2S source end, so the HF signal edges would be passed by BW-limiting capacitors back to the I2S source ground, not the DAC's ground.
Click to expand...

Hey s3tup, can you give some examples of these cards? I did a quick search and didn't see any easy I2S.

Thanks,
Jeff
 
R

regal

Member
Joined 2004
studiostevus said:
I am using the Gomes stage in my dac, and it sounds excellent. One thing about the 2ma offset: i am using the negative offset as fixed bias to be able to bias the tubes very near to their 0v line (while surely maintaining a negative bias). So not in all cases it's useful to compensate the offset.
Click to expand...

The voltage compliance spec for the tda 1541A is 25mv, I understand what you are doing using the offset for negative bias but Oleg had a great tests showing that for passive I/V there was significantly less distortion with a tube stage by injecting the 2mA to the top of the I/V resistor. I wish I could find Oleg's webpage 'cause it was full of some real measurements about passive I/V + tube TDA1541 design. His tube stage was measuring better than most SS I/V stages.
 
A

Aronier

Member
Joined 2008
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom