Crossover component quality: How do they affect sound?

[scardeal]

So, yesterday I was watching a video that was comparing DIY designs to commercial offerings, and the narrator said several things to the effect that crossover components quality, in addition to the raw design of the crossover, affect sound of the speaker.

For instance, he mentioned 3 things specifically:

1. Inductors - air core vs iron core, gauge of wire used and sheathing used.
2. Resistors - sand cast vs ???
3. Capacitors - I don't remember specifically, but for instance some kits I have seen offer a choice between Audyn or Mundorf Supreme or Dayton Audio or Jantzen, and there's sometimes a significant price difference between the different capacitors.

In particular, I've heard about people replacing capacitors and noting a worthwhile improvement in sound. What I'm curious about is why components of the same specification, or even of the same type and specification, might sound different from one another.

So, why do those same rated components affect sound?
And, bonus question, if I replaced the capacitors in, say, my amplifier with Mundorf Supremes, would that make an improvement/change in the sound quality?
And, bonus bonus question, how much do they affect the sound?

[Chris Roemer]

get ready...

[hitsware2]

http://reliablecapacitors.com/oldRC/www.reliablecapacitors.com/pickcap.html

[craigk]

Recapping an amp is different from a speaker.

[johnnyrichards]

Well... there are differences in components.

1. Inductors: powdered core inductors have a low saturation point, but it is not a cliff like laminate core or even air core. It just means the more power you pour into them, they get progressively farther into saturation. A coils DCR can impact things quite a bit- swap an air core in where a laminate used to be and things will change somewhat dramatically. Some people mention skin effect, but alas that occurs at fairly high frequencies. It is unlikely people can blindly identify core type/gauge in a competently designed crossover.

2. Capacitors: some absolutely sound different, but valid testing is somewhat difficult absent absolute value matching. A drift of 1/2 mF can affect phase at the listening position, for example. It is easy to see the effects of even small value changes in any simulation software. I have as much valid test data as anyone, which has lead me to believe even small differences in capacitor value is a significant contributor to our perception of how different capacitors sound. Some capacitors have lower dissipation factors and ESR, and that is also a contributing factor - ESR perhaps somewhat more than DF but I am not sure if it is as significant as being off by 5% when rolling caps.

3. Resistors: I have no opinions other than the sandcast type are dammed ugly so I stopped using them. I am more concerned with making sure the value is as close as possible to the model, the same with coils and capacitors. I feel there is a more to gain from the more robust leads on the nicer resistors, many of the sandcast have some pretty crappy build quality. The heatsinked type are cool to look at - but require specific mounting to meet their thermal dissipation ratings.

4. Wire: Reactance of a wire is measureable even with hobbyist tools and wires should (imho) be invisible. That being said, some cables are engineered to color the sound and while that may be a pleasing effect, it is nonetheless an example of linear distortion. Changing capacitance, resistance, or inductance will all contribute to a particular sonic signature and it is important to understand there is literally nothing else that a cable contributes. That being said, cheap cables are usually ugly so I use nice looking cables, even when they cannot be seen. Some of them are inexpensive cables that look nice, though.

5. Lastly - crossover design competency. It is my belief that a specific value is more critical than component type, and our best gains as far as crossovers are concerned is to keep refining our design chops. Other than the occasional accidental improvement in linear response or phase due to swapping a different component of a similar value in, it is unlikely the sound we dislike is actually the type of component versus a mistake somewhere in the design process.

Red caps sound better.

[fpitas]

Well... there are differences in components.

1. Inductors: powdered core inductors have a low saturation point, but it is not a cliff like laminate core or even air core. It just means the more power you pour into them, they get progressively farther into saturation. A coils DCR can impact things quite a bit- swap an air core in where a laminate used to be and things will change somewhat dramatically. Some people mention skin effect, but alas that occurs at fairly high frequencies. It is unlikely people can blindly identify core type/gauge in a competently designed crossover.

2. Capacitors: some absolutely sound different, but valid testing is somewhat difficult absent absolute value matching. A drift of 1/2 mF can affect phase at the listening position, for example. It is easy to see the effects of even small value changes in any simulation software. I have as much valid test data as anyone, which has lead me to believe even small differences in capacitor value is a significant contributor to our perception of how different capacitors sound. Some capacitors have lower dissipation factors and ESR, and that is also a contributing factor - ESR perhaps somewhat more than DF but I am not sure if it is as significant as being off by 5% when rolling caps.

3. Resistors: I have no opinions other than the sandcast type are dammed ugly so I stopped using them. I am more concerned with making sure the value is as close as possible to the model, the same with coils and capacitors. I feel there is a more to gain from the more robust leads on the nicer resistors, many of the sandcast have some pretty crappy build quality. The heatsinked type are cool to look at - but require specific mounting to meet their thermal dissipation ratings.

4. Wire: Reactance of a wire is measureable even with hobbyist tools and wires should (imho) be invisible. That being said, some cables are engineered to color the sound and while that may be a pleasing effect, it is nonetheless an example of linear distortion. Changing capacitance, resistance, or inductance will all contribute to a particular sonic signature and it is important to understand there is literally nothing else that a cable contributes. That being said, cheap cables are usually ugly so I use nice looking cables, even when they cannot be seen. Some of them are inexpensive cables that look nice, though.

5. Lastly - crossover design competency. It is my belief that a specific value is more critical than component type, and our best gains as far as crossovers are concerned is to keep refining our design chops. Other than the occasional accidental improvement in linear response or phase due to swapping a different component of a similar value in, it is unlikely the sound we dislike is actually the type of component versus a mistake somewhere in the design process.

Red caps sound better.

That's a very sensible summary. Especially about red caps. I keep some by my speakers, even though it's a fully active setup.

[scardeal]

Well... there are differences in components.

1. Inductors: powdered core inductors have a low saturation point, but it is not a cliff like laminate core or even air core. It just means the more power you pour into them, they get progressively farther into saturation. A coils DCR can impact things quite a bit- swap an air core in where a laminate used to be and things will change somewhat dramatically. Some people mention skin effect, but alas that occurs at fairly high frequencies. It is unlikely people can blindly identify core type/gauge in a competently designed crossover.

2. Capacitors: some absolutely sound different, but valid testing is somewhat difficult absent absolute value matching. A drift of 1/2 mF can affect phase at the listening position, for example. It is easy to see the effects of even small value changes in any simulation software. I have as much valid test data as anyone, which has lead me to believe even small differences in capacitor value is a significant contributor to our perception of how different capacitors sound. Some capacitors have lower dissipation factors and ESR, and that is also a contributing factor - ESR perhaps somewhat more than DF but I am not sure if it is as significant as being off by 5% when rolling caps.

3. Resistors: I have no opinions other than the sandcast type are dammed ugly so I stopped using them. I am more concerned with making sure the value is as close as possible to the model, the same with coils and capacitors. I feel there is a more to gain from the more robust leads on the nicer resistors, many of the sandcast have some pretty crappy build quality. The heatsinked type are cool to look at - but require specific mounting to meet their thermal dissipation ratings.

4. Wire: Reactance of a wire is measureable even with hobbyist tools and wires should (imho) be invisible. That being said, some cables are engineered to color the sound and while that may be a pleasing effect, it is nonetheless an example of linear distortion. Changing capacitance, resistance, or inductance will all contribute to a particular sonic signature and it is important to understand there is literally nothing else that a cable contributes. That being said, cheap cables are usually ugly so I use nice looking cables, even when they cannot be seen. Some of them are inexpensive cables that look nice, though.

5. Lastly - crossover design competency. It is my belief that a specific value is more critical than component type, and our best gains as far as crossovers are concerned is to keep refining our design chops. Other than the occasional accidental improvement in linear response or phase due to swapping a different component of a similar value in, it is unlikely the sound we dislike is actually the type of component versus a mistake somewhere in the design process.

Red caps sound better.

So, for #2, would better capacitors have tighter tolerances and lower ESR (resistance apart from capacitance)?

For #3, is there a danger of cheap resistors having inductance? And better ones again having tighter tolerances?

For #4, is thicker gauge wire always better, given the same type/material of wire?

I'd take #5 as a given...

[johnnyrichards]

2. I would argue better caps have better specs. It is surprisingly independent of price.

3. Most resistors will have inductance. Usually it is pretty minor.

4. Yes.

5. You'd be surprised.

[Millstonemike]

2. I would argue better caps have better specs. It is surprisingly independent of price.

I'm confused by that statement. I would think better caps would cost more.

[dynamo]

I'm confused by that statement. I would think better caps would cost more.

I think the point is that a cheap cap like the Q4 or Dayton that is tightly wound on a big machine that does only this may be more consistent and measure better than a pricey boutique cap made by hand or with lesser equipment. This may pin the cheaper cap as the better (more consistent and tighter tolerances) cap vs the boutique. I have no data to back this. Sorry if this isn’t what you meant JR but that’s how I read it and if so I agree.

I love experimenting with different caps, but just cuz it’s more expensive doesn’t mean it sounds better, just different.

[dynamo]

I think non-inductive resistors are pointless. The little bit of inductance is irrelevant in a crossover and even if it were enough to take 20k down a touch, that sounds like a benefit to me..

[Kornbread]

2. I would argue better caps have better specs. It is surprisingly independent of price.

To and extent, I'd wager this a true statement in many things of audio. Just because it costs more does not inherently designate it any better than a 'lesser' part it replaces. This could be applied to a lot of things in life.

... and thanks johnnyrichards that was very informative and well written.

[TN Allen]

"Some people mention skin effect, but alas that occurs at fairly high frequencies."

I've wondered if skin effect is a significant concern in component choice. At what audio frequency is it likely to become relevant, and how might it effect or alter the current and sound?

[djg]

Wonder no more.

Understanding Skin Effect and Frequency

[philthien]

Where people cannot perceive a difference in sound quality, they tend to substitute price as the determining factor.