TDA1541A decoupling capacitors value

[Zoran]

I have some questions if You dont mind.
related to decoupling capacitors of the TDA1541A...
*
1. What value and type You can recomend for those caps?
Because different values of caps existing from 0.1uF...
*
2. Are the caps has to be of same values for all decoupling pins?
*
3. In case of paralleling DAC chips, will You preffer, each chip to be decoupled by separate set of caps,
or merging the all decopling pins and aplly decoupling with one set of caps,
where value of single capacitor has to be multiplyed by number of DACs?
*
Thank You very much in advance...
Zoran

 

[zygibajt]

I tried many types and one of the best so far are orange Philips MKT caps.

I observed I prefer higher values of these caps. I think I liked 1uF the most , but can't get my favourite orange Philips of this value any more so I'm using 0,56 now.

 

[Zoran]

Thanks zigibajt
what about
Q2
Q3
?
in case of paralleling the DAC chips
are the pins decoupling pins connected respectably
together and then linked to the decoupling capacitor,
or every dac chip has to be decoupled by own set of caps?

 

[aparatusonitus]

Every dac chip has to be decoupled by own set of caps, all the same values. I use orange 0,1uF Philips MKT caps, but 1uF sound best, as zygibajt noted.

 

[Zoran]

Thanks for reply...

 

[philpoole]

Hi All,

Thought I'd reheat this, as I've been perplexed by this...

Basically, if you follow the TDA1541A datasheet, it suggests 0.1uF for these decoupling caps. Not understanding exactly what these caps are for (I thought they could provide some kind of specific timing for the DEM circuit so the value is perhaps crucial), it is interesting to see other people use, I think I've seen up to 3uF, for some of these caps. AND, only some of them (I think for the most significant bits).
Are these just decoupling capacitors like any other? Is there too big a value (because if they were just decoupling capacitors, why not make them all 100uF and be done with it?)?
Which pins service the most signficant bits?

And also, why do some, sometimes, change the Cosc to 680pF?

Basically, is there any theory behind doing any this? And if so, where can it be found?

Cheers,
Phil

 

[PetrL]

As for single chip: I use 2uF MKT for MSB pins 13 & 18. Higher value is definitely needed, the sound is *much* more lively and authoritative. Don't know why... In my setup going even higher brought no further improvement. 0.1 uF MKP for the rest, no need to increase. I tried 0.47 uF MKT here, but the sound went somehow raw. Not completely bad by itself, but once I heard the smoothness of actual solution, there is no other way.

The only "theory" I can serve is that datasheet solutions tend to impress by economy (small caps). Sound can't be printed.

 

[philpoole]

Hi PetrL,

Thanks for the reply.
Thinking about it, I guess it could be related to the fact that switching the MSB provides the biggest output change, and hence a bigger decoupling cap can help with this?

Good to hear it made a difference for you.

I like the economy theory, I guess some procurement departments would be concerned over the extra 17-odd capacitors needed for the implementation, so making them smaller values may help tip the balance.

Cheers,
Phil

 

[PetrL]

Phil,

In the datasheet of tda1540 there is some functional diagram and explanation of the switching. Several values of caps should be used. Plain nightmare for a production engineer, I guess. The 1541 is probably similar.

The increased MSB cap has certain downside depending on the output stage. My diy DAC is modified one of www.tnt-audio.com The highs became sharp and unbearable, which I solved by LP filter caps parallel to i/v resistors (after some experimets with filtering at the end of output stage).

 

[Kip]

Hi PetrL,
I am using Thomas' DAC ver.3 and purchased a TDA1541A/S2 chip and a CS8412 from him also. I am considering selling the ver.3 dac and building the TNT design.

Can you share your modified TNT schematic with me? I can build off schematics, but can't design. I would rather use a solid state I/V stage and that is one of the reasons why I like the TNT design. I have read some comments that TNT's I/V could be better.

I value TNT's comments. I purchased a Karan integrated amp because of their review of it.

 

[PetrL]

Kip,

although I don't know the Thomas DAC, I would say for the cost vs. sound the (modified) TNT design is hard to beat. Anyway, and most important the sound is very good. My modifications to the TNT schematics are:
-- no decimation
-- DECOU caps: 12 x 100nF MKP, 2 x 2uF (two 1uF) MKT for MSB pins
-- output stage (one channel):
i/v resistor 30 Ohm + 690nF MKT in parallel, LPF cap C407 removed, feedback R421 removed, R422 shortened out.

These modifications resulted after many experiments (I also can't design, only know basic principles), not because I got tired ;-) but I liked the sound at once. So use 2.2uF or so for MSB, if you unlike me have enough space on one side of PCB.

Then there were issues with series resistors of TL431 regulators. For raw 19V I had to decrease R401 to 180 Ohm. On the other hand the TDA is powered from same 19V and the regulators seemed too hot (and you are going to hang S2 on them!), so I increased the series resistors. Don't remember the values, if you are interested, I can check the circuit. Use at least 5W resistors and place them far from TL431s, those get hot!

Of course one can modify the receiver circuit as well. I use basicaly the one of TNT Convertus. I added series resistors (I think 1k) into the 3 i2s lines. And recently used this tweak of filter circuit:
http://db.audioasylum.com/cgi/m.mpl...ighlight=loop-filter+wildmonkeysects&session=
Save the last tweak when you feel to solder something simple yet which again moves the sound to higher level...

 

[Kip]

Thank you,

I think I understand and marked up the TNT schematic. I followed the link for the CS8412 and may have some questions later. If you can give me the values of the dac's series resistors it would help. No big rush, when you get a chance.

There is a large thread on an ultimate TDA1541 dac, but it looks like it is going parallel dacs. TDA1541A S/2 dacs are too costly to give it a shot.

 

[PetrL]

Kip,

The resistors R301, 302, 303 are 120, 220, and 330 Ohm respectively. Maybe now I would choose R303 smaller i.e. TL431 warmer. It's subjective once the regulator works...

 

[Kip]

Got it!
Thanks again

 

[philpoole]

Hi Petrl,

Routing around, I found some standard Elna 2.2uF caps. As a brief experiment, do you think it would matter that they're electrolytic?
I did notice that the MSB pins are actually a negative voltage, so that's not a problem (worth noting if you're in the same situation as me).

Cheers,
Phil

 

[Extreme_Boky]

Hi Zoran,

I obtained the best sound with Black Gates 0.1uF / 50V non polar types.


Regards,
Boky

 

[wrenchone]

I was thinking of using Panasonic V Series stacked film polyesters for decoupling. Has anyone tried these? I have 470n, 1u0 and 2u2 values.

 

[PetrL]

I did no experimets with HQ electrolytics, only foils. I can only say that the type has huge effect on sound, so experiment. For me, the MKT on MSB gave it a push or edge, if you want, whereas the MKP was more (too much for me) laidback. I think I was not at all pleased with MKS.

 

[quantran]

Could anyone advise me if these Siemens caps are good for TDA1541A decoupling?

The part number is B32540 but it is obsolete so I can't find much info about them.

 

[lucpes]

Could anyone advise me if these Siemens caps are good for TDA1541A decoupling?

The part number is B32540 but it is obsolete so I can't find much info about them.

I think they're stacked polyester, medium quality. I would try something better if I were you, but sure, they would work ok. Cannot comment on the sound quality...