C0G/NP0 Parallel-Capacitor-Board

[DPH]

That's a false analogy, though. Upgrading caps willy nilly more like putting in the wrong sized HVAC system in your house and then wondering what happened to all your garage space and bring annoyed by the sound of the circulating fan.

 

[Oneminde]

The usual problem is that equipment is designed to do one thing (which may or may not involve high fidelity sound reproduction) and then claims are made for another thing (which often involves 'better' sound).

Anyway, the point we are making is that you cannot necessarily improve a circuit by simply changing to 'better quality' components. You first have to carefully reverse-engineer the circuit (i.e. do the designer's calculations backwards) to gain understanding, then see what he got wrong, then put it right. This is not possible for someone who appears to be struggling to understand where ripple comes from.

I do get that component type, construction etc have different purposes and swapping one for the other may cause problems. I do get that.

But. if we look at what Bruce is telling us what we should use, in a way, all of you are in a disagreement with him .. BECAUSE no one stepped forward and said; "Yes, he has a point and I agree but have this to add; Be mindful to not introduce other problems"

I asked; Doesn't a circuit benefit from using more accurate values and is more stable - and the dominant answer is NO ... meaning that if a circuit is using a 0201 thick film 200 ppm/C resistor and I change it to 1206 thin film 10 ppm/C .. I am in deep S....

 

[DPH]

But. if we look at what Bruce is telling us what we should use, in a way, all of you are in a disagreement with him .. BECAUSE no one stepped forward and said; "Yes, he has a point and I agree but have this to add; Be mindful to not introduce other problems"

I did just that, and you completely ignored my comment. When you're in Bruce's situation, or Victor's, you measure and find the best part for your application, and on occasion that part surprises you. That is absolutely borne out in that whole presentation you selectively quoted from.

As dmills writes, you pick the right component for the job, and that requires examining in detail.

 

[Oneminde]

Besides the fact that certain components contribute in a good way, such as el-caps, I can only conclude with - so far - that spite the engineers motivation for choosing the component type and secondary values (such as ppm/C), be that cost saving OR actual electrical contribution, in order to use a specific component, lets say thin film vs carbon comp resistor, they dictate they schematic and over all layout of the circuit and there is no other way. Essentially, you have an idea, select component type and then design your circuit.

Got it.

 

[Oneminde]

I did just that, and you completely ignored my comment. When you're in Bruce's situation, or Victor's, you measure and find the best part for your application, and on occasion that part surprises you. That is absolutely borne out in that whole presentation you selectively quoted from.

As dmills writes, you pick the right component for the job, and that requires examining in detail.

To be fair, there is a positive attitude towards considering WIMA FPK vs C0G if there is room to upgrade. But it might not be without some issues due to a potential increase in inductance. Also, I have backed away from PP-film in the PSU and decoupling area and "compromise" by looking at the Hybrid el-caps which is really interesting.

I also do get that one should be careful changing components. In the end, one can always do prototyping and discover things that way.

 

[jan.didden]

But. if we look at what Bruce is telling us what we should use, in a way, all of you are in a disagreement with him .. BECAUSE no one stepped forward and said; "Yes, he has a point and I agree but have this to add; Be mindful to not introduce other problems"

You fall in the same trap again! That cap has a very specific function, I mentioned that, it is part of a deep notch tuning so needs to be a) very precise and stable (you can't just buy a 0.33uF cap) and b) very linear.

There is a very good pure engineering reason why the mini-PCB with the ceramics performs better than the film.

To repeat again: first think about the requirement, then select a device or part that fills it.

I don't know why you say that all are in disagreement with him - I certainly am not! But some things are so obvious that you don't realize it needs mentioning.

Jan

 

[jan.didden]

I have backed away from PP-film in the PSU and decoupling area and "compromise" by looking at the Hybrid el-caps which is really interesting.

Wrong again. You are not compromising, you are optimizing. Selecting a better part for the purpose. Jeez.

Leave that keyboard alone for a few minutes and think about whats being discussed here.

Jan

 

[DF96]

I asked; Doesn't a circuit benefit from using more accurate values and is more stable - and the dominant answer is NO ... meaning that if a circuit is using a 0201 thick film 200 ppm/C resistor and I change it to 1206 thin film 10 ppm/C .. I am in deep S....

You appear not to understand what we are saying. If you could change a component for one which is identical in all respects except temperature behaviour and the new one is better then it is true that in almost all conceivable circuits this would do no harm; however, in most circuits it would do no good either. That is not what you were proposing, so not really relevant to this discussion - except possibly because it gives us an opportunity to clear up some confusion you may have.

I also do get that one should be careful changing components. In the end, one can always do prototyping and discover things that way.

In order to correctly discover things from prototyping you need a good understanding of how the circuit works and how components work. If not, it is easy to be led astray.

 

[Oneminde]

You appear not to understand what we are saying. If you could change a component for one which is identical in all respects except temperature behaviour and the new one is better then it is true that in almost all conceivable circuits this would do no harm; however, in most circuits it would do no good either. That is not what you were proposing, so not really relevant to this discussion - except possibly because it gives us an opportunity to clear up some confusion you may have.

Confusion... maybe, one cannot exclude such things. The initial post, the very start, showed potential impact on distortion figures using different capacitor construction / material. When it comes to me mentioning replicating an existing product but "improve" (as I saw it) the component accuracy, I stated from the very beginning that the values and schematic would remain intact, perhaps this was not clear enough and if so, then I do apologies. I wrote this while you made a reply:

To be fair Jan, optimizing is many times more challenging than originally conceptualized. To clarify: A more accurate component, such as going from ex: 1% 100 ppm/C to 0.001% 10 ppm/C (same resistance value) is not as straight forward as initially assumed after listening to Bruce Hofer and debating this topic in this thread. While this thread started with C0G ceramic vs PP-Film/foil capacitor, the initial idea was that better accuracy could only be seen as beneficial to the over all performance (such as obtaining lower distortion figures) and not something that could potentially be bad or outright create instability issues or any form of problem.
Changing values is one thing, that isn't something you do on a whim...But, changing to more accurate components and that this was included as part of potential issues was completely new to me.

To improve or optimize the schematic or function of a circuit, yes, then one would REALLY need to know and understand on a engineering level as proposed in this thread. For me, selecting more accurate components was a different beast, which is another form of optimizing as Jan points out ...

What can I say....My intention was not circuit/schematic modification or optimization, it was as at the start, focused on component optimization. And I don't get why more accuracy is - generally - seen as non beneficial. After all, the initial post forwarded evidence of a significant lower THD levels going from one capacitor type to another.

But I guess we'll have to leave it there, no point in going around in circles and I hope I corrected or clarified some issues that I added to the debate.

 

[jan.didden]

For me, selecting more accurate components was a different beast, which is another form of optimizing as Jan points out ...

No that's not what I said. Optimizing is selecting a part that makes the circuit do better what it is supposed to do.

Blindly selecting a part with less tolerance thinking it makes the circuit do better what it is supposed to do is not optimizing, it is trial and error and hoping for the best.

 

[DF96]

To improve or optimize the schematic or function of a circuit, yes, then one would REALLY need to know and understand on a engineering level as proposed in this thread. For me, selecting more accurate components was a different beast, which is another form of optimizing as Jan points out ...

Yes, you still have not understood what we are saying. To select 'more accurate' components (even when this might be helpful) requires engineering knowledge. In many cases the 'more accurate' component will be physically larger, so have greater stray capacitance which can lead to more RF pickup or even instability. We have already mentioned that the 'more accurate' component may have too small ESR to damp resonances. If the designer has chosen 10% tolerance as good enough, then in most cases you do not improve the circuit by putting in 1% components; you merely make it more expensive, although this alone could boost sales in some markets.

 

[Oneminde]

Yes, you still have not understood what we are saying. To select 'more accurate' components (even when this might be helpful) requires engineering knowledge. In many cases the 'more accurate' component will be physically larger, so have greater stray capacitance which can lead to more RF pickup or even instability. We have already mentioned that the 'more accurate' component may have too small ESR to damp resonances. If the designer has chosen 10% tolerance as good enough, then in most cases you do not improve the circuit by putting in 1% components; you merely make it more expensive, although this alone could boost sales in some markets.

This is the first time anyone mentions: "have too small ESR to damp resonances.". My initial response is WTF .

Schematics deal with ideal components, a capacitor has capacitance only. In the real world, there will be ESR and ESL in a capacitor and I have learned that it is undesirable and one want these values to be generally low.

Have you all assumed that using ESR to dampen resonances is in the public domain and every knows ?

 

[jan.didden]

You haven't been paying attention.

Lossy elcaps help with absorption of the sharpest charge peaks, but film caps are a wrong choice for a PSU, except maybe in high frequency switched-mode power supplies. But we're not talking about that here.
Better spend that money on a good coupling cap.

Anecdote: many people buy a superreg in the diyaudio store, and have the great idea to replace the output elcap with a film cap. Only to have the supply oscillate. Here, the small ESR of an elcap is critical for the circuit stability. It's mentioned in the write -up, but hey, who reads that??

In fact, if you look into the data sheet of say an LM7815, there is a minimum loss resistance (ESR) that is required at the output for stability.

Sometimes loss is great, especially with stuff you don't want!

Jan

 

[jan.didden]

Have you all assumed that using ESR to dampen resonances is in the public domain ...

Yes it is.

... and every knows ?

No, that's why we repeatedly explain it. It's part of an advanced process called 'learning'. You should give it a try someday.

Jan

 

[Oneminde]

Thanks Jan, you are so kind .... "slowly inhale and slowly exhale ... breath".

What is the point in debating if the n00b constantly is being picked at. "Oh look (you idiot) you missed the point or you are not paying attention. How the FFFFFF am I suppose to learn and be able to advance and productively contribute as a human being if I get one chance only, miss that chance and I might as well blow my brain out.... OMG... This is one of the most anal threads I've been a part of... I'm out.

 

[jan.didden]

Nobody is picking on a noob. Its just very frustrating for people spending considerable effort to explain stuff only to see that it isn't even read. Or the noob is extremely dense, but that's normally not the case.

Jan

 

[DF96]

This is the first time anyone mentions: "have too small ESR to damp resonances."

No. Have you actually been reading this thread?

My initial response is WTF

Ignorance is bliss - for a while.

Schematics deal with ideal components, a capacitor has capacitance only.

No. The designer knows about parasitics too, although these will not usually be explicit on the schematic.

In the real world, there will be ESR and ESL in a capacitor and I have learned that it is undesirable and one want these values to be generally low.

No. Sometimes they should be low, sometimes middling, sometimes highish.

Have you all assumed that using ESR to dampen resonances is in the public domain and every knows ?

Good engineering is not a secret, unlike alchemy. Some audio designers prefer alchemy, as it sells better.

 

[irribeo]

Did you read what the designer of your ampboards said in general about switching amps?
Switching Amplifier (Class D) Basics | Audioholics

Then you could try to read some of the Ncore white papers and or AES papers study few years till you can follow the calculations and you might produce something that isn't much worse than the original

 

[DF96]

This is the first time anyone mentions: "have too small ESR to damp resonances.".

Let us do some counting:

Those electrolytics need some loss to work their job well.

(post 12)

Actually, some lossy elcaps are the best, because that loss absorbs some of the ripple, turning it into a bit of heat, instead of turning all of it into ground currents.

(post 15)

parasitic resonances benefit from parasitic losses to damp them.

(post 16)

Film caps in power supplies can lead to worse results just because they are lossless.

(post 19)

Here, the small ESR of an elcap is critical for the circuit stability. It's mentioned in the write -up, but hey, who reads that??

(post 22)

For example, sometimes overly low ESR is NOT desirable

(post 24)

Like others mentioned, high Q is not your friend here necessarily.

(post 30)

We have already mentioned that the 'more accurate' component may have too small ESR to damp resonances.

(post 71)
I make that eight mentions of this point. Did you miss all these posts?

 

[Oneminde]

Last comment. I am at fault here and I have a pile of electronics engineering books piled up which I will consumer. Until that is done, I will take a break from public appearance - Much better that way.