TDA1541 whups 1543's butt

[nl]

Reading the various comments by people who have done "zero-oversampling" DIY DACs using both TDA1543 and the TDA1541, I got the impression that the difference was one of those things that only audiophools could get excited about.

Wrong!

After three years of using the TDA1543, and tweaking it with various different battery powersupplies, etc., I finally decided that it was unacceptable due to biting sibilants and a hard, glare-y upper midrange especially apparent on vocals. This wasn't too bad when using speakers, but on headphones it was like chewing glass. This was the famous "mellow" sound of the zero-oversampling TDA1543, presumably the basis of the glowingly-reviewed and big $$$ 47 Labs DAC?

So, I put together a somewhat common TDA1541 setup, using TL431 shunt regs and a CCS-loaded 5842 (417A) as an output stage, direct coupled. I ended up spending almost two months debugging it, fixing the problems only by rebuilding it completely from scratch. This seems like a touchy chip.

Well, now we're getting somewhere. The difference is astounding. Finally, the sound I've been looking for lo these many years. As Herb Reichert might say, it expresses the emotional content of the music. Rather than a marginal improvement, I feel like I'm now getting 70% of the music, compared to 25% before. My headphone (regular listening) setup is VERY revealing, a natural feature of headphones exaggerated by the direct-coupled, battery-powered, battery-biased topology using DHTs. I've long thought about "smoothing it out" using coupling caps and other tricks, since it's so hard to find or build an adequate front end, not to mention the shortcomings of the recording process. Now I'm willing to go with it a while longer.

The conclusion is: don't waste your time on the 1543.

[Terry Demol]

Originally posted by nl
Reading the various comments by people who have done "zero-oversampling" DIY DACs using both TDA1543 and the TDA1541, I got the impression that the difference was one of those things that only audiophools could get excited about.

Wrong!

After three years of using the TDA1543, and tweaking it with various different battery powersupplies, etc., I finally decided that it was unacceptable due to biting sibilants and a hard, glare-y upper midrange especially apparent on vocals. This wasn't too bad when using speakers, but on headphones it was like chewing glass. This was the famous "mellow" sound of the zero-oversampling TDA1543, presumably the basis of the glowingly-reviewed and big $$$ 47 Labs DAC?

So, I put together a somewhat common TDA1541 setup, using TL431 shunt regs and a CCS-loaded 5842 (417A) as an output stage, direct coupled. I ended up spending almost two months debugging it, fixing the problems only by rebuilding it completely from scratch. This seems like a touchy chip.

Well, now we're getting somewhere. The difference is astounding. Finally, the sound I've been looking for lo these many years. As Herb Reichert might say, it expresses the emotional content of the music. Rather than a marginal improvement, I feel like I'm now getting 70% of the music, compared to 25% before. My headphone (regular listening) setup is VERY revealing, a natural feature of headphones exaggerated by the direct-coupled, battery-powered, battery-biased topology using DHTs. I've long thought about "smoothing it out" using coupling caps and other tricks, since it's so hard to find or build an adequate front end, not to mention the shortcomings of the recording process. Now I'm willing to go with it a while longer.

The conclusion is: don't waste your time on the 1543.

Absolutely agree, great sounding DACs

Looks like you'll have to search for some single or dual crowns.
WRT 417A, good but I think the 6C45 is better. You can run it at
30mA or more. What else.... use as small an I-V R as you can get away with, most people I know are using 25 or 33R with great results.
I still plan on doing some low noise discrete shunts to replace the
TL431's as they can be improved a fair bit.
What else.... get on the I2S bus or implement one of Mr Tents
XO-DAC to take care of jitter.

Cheers,

Terry

[nl]

Speaking of problems, I'm still dealing with what appears to be digital noise. On the first version, there was a static-like noise, but which was not random but reflected the music. The static-like noise was not level dependent, always the same volume in loud or soft passages, and went away between tracks. After many attempted fixes, I eventually rebuilt the whole thing (point to point), with special attention to a compact layout and short current loops.

The new version has a different noise, a hum around 600hz or so between tracks, and an intermittent noise during music. Once again, not level-dependent. Fortunately, the level is about 10db down from the earlier version, so it is tolerable, but far from ideal.

The 5842 is a touchy tube, maybe more so as it is loaded with a solid-state current source using Gary Pimm's BMCCCS design.

Anyone else with these kinds of problems?

[Guido Tent]

Originally posted by nl

The conclusion is: don't waste your time on the 1543.

spot on !

Yes, the 1541 is more difficult to apply, but worth the trouble !!!

regards

[Andypairo]

Originally posted by Guido Tent


spot on !

Yes, the 1541 is more difficult to apply, but worth the trouble !!!

regards

IMHO the 1543 is a fairly good performer thanks to its ease of implementation, but a properly built 1541 DAC (which needs adequate-quality implementation and components) is at a different level.
It will cost 5 to 10 times (or more..) the other tough.

After all the former is a "economy" chip....

Cheers

Andrea

[Terry Demol]

Originally posted by nl
Speaking of problems, I'm still dealing with what appears to be digital noise. On the first version, there was a static-like noise, but which was not random but reflected the music. The static-like noise was not level dependent, always the same volume in loud or soft passages, and went away between tracks. After many attempted fixes, I eventually rebuilt the whole thing (point to point), with special attention to a compact layout and short current loops.

The new version has a different noise, a hum around 600hz or so between tracks, and an intermittent noise during music. Once again, not level-dependent. Fortunately, the level is about 10db down from the earlier version, so it is tolerable, but far from ideal.

The 5842 is a touchy tube, maybe more so as it is loaded with a solid-state current source using Gary Pimm's BMCCCS design.

Anyone else with these kinds of problems?

As with all high gm tubes, grid stoppers are a must.
Also 417A is notorious for being microphonic.

Cheers,

Terry

[t.]

Hi guys,

Has anybody built this one
http://hjem.get2net.dk/torbenk/dualdactube.htm

I have a couple of crown dacs and 6c45pi tubes and would be interested in building it

[lucpes]

Regarding the TL431 shunt regs, an extra cap (C3 in the link) really helps:

http://www.tnt-audio.com/clinica/reg..._noise3_e.html

Just stick a 47uf/16V or bigger quality cap onto the reg, between ref and cathode.

[cmt42001]

Hi,

I have built one using two parallel TDA1541a and experiment with various tube output options using 6DJ8 SRPP, 6C45, 2C51 and WE404a connected as triode. Now I settled on the WE404a coupled with parafeed transformer out. BTW, all the tube options sound very good.

One thing I noticed is that putting an asychronous reclocker on the I2S signal between the CS8412 and TDA1541a really help out the soundstage and focus.

[Guido Tent]

Originally posted by cmt42001

One thing I noticed is that putting an asychronous reclocker on the I2S signal between the CS8412 and TDA1541a really help out the soundstage and focus.

sure, the 8412 output jitter is way too high for seious conversion at low distortion levels.............

enjoy