Why that many DECOU pins?

[roibm]

Chip: TDA1541A
Decoupling pins: 14

The question is this: why that much decoupling? Why is it better(I assume it is better since they use that many)?
I mean, 8 wouldn't be enough? What are the advantages when using 14?

[till]

i suspect they decouple some switched current sources individually

[peranders]

Those are a part of the silicone design.

[roibm]

Originally posted by peranders
Those are a part of the silicone design.

Sure they are, but why would they use 14 and not less?
That's what I try to understand.

Links accepted as well

[till]

Those are a part of the silicone design.

That was a brilliant answer.

[ThorstenL]

Konnichiwa,

Originally posted by roibm
Chip: TDA1541A
Decoupling pins: 14

The question is this: why that much decoupling? Why is it better(I assume it is better since they use that many)?
I mean, 8 wouldn't be enough? What are the advantages when using 14?

Originally posted by roibm
Sure they are, but why would they use 14 and not less?
That's what I try to understand.

The way the TDA1541 works is to use a swiched set of current sources which in effect auto-calibrate the Chip. The thing was called dynamic element matching and is one of the things that make teh TDA1541 so interesting.

The switching invariably generates "glitch" current spikes, which need to be "smoothed out". The number 14 is related to the fact that each channel has 6 active switched current sources for the lowest 6 bits and seventh followed by a passive divider(multiemitter transistor, not R2R - another eason for the unique sound quality) for the remaining 10 upper bits. That requires 7 decoupling points per DAC, as the TDA1541 contains 2 it needs 14.

Hence the answer "Those are a part of the silicone design." is completely correct if terse.

Sayonara

[roibm]

Originally posted by Kuei Yang Wang
Hence the answer "Those are a part of the silicone design." is completely correct if terse.

Peranders answer alone is trivial in the context of my question. I already knew they are used internally in some way, but I didn't know in what way.
Now, thanks to you, I understand it better. Do you happen to have a link to some internal details about this chip?

Arigatou

[till]

Peranders answer alone is trivial in the context of my question.

forgive him for giving a correct but completely useless answer. He needs to make about 100 posts each day on this board to show presence what helps to sell some gainclones or so.

[roibm]

Originally posted by till

He needs to make about 100 posts each day on this board to show presence what helps to sell some gainclones or so.

I wouldn't be that harsh either. The main point is that I'm totally happy with his price for the gc boards. I find it 100% fare.
If it produces the results exposed by the two beta testers I'll be even happier.

[ThorstenL]

Konnichiwa,

Originally posted by roibm
Do you happen to have a link to some internal details about this chip?

No, my info comes from a large number of sheets from Philips from the early 80's showing the various technology behind the DAC and eventually the silicon layout down to individual transistor level. Highly interesting reading, but way too much to scan for me....

Sayonara