Carbon film resistors in crossovers

[Kuja]

Hello,

I'm planning to attenuate the tweeters in my speakers.

With my new amp, the upper midgrange and the top end can be a little bit on the cold side on some recordings.


I was planning to use metal oxide resistors, but I got some advices that I could try carbon film instead.

It turned out that some manufacturers are using bunches of 10 or more carbon film resistors connected in parallel:






Upon further reading, I found claims that metal oxide resistors are technically "better" and more precise, but that they can be cold sounding.
Carbon films are technically inferior, but they are warmer, more musical, etc:

Carbon Film vs Metal Film revisited


So guys, do you have some/any experiences with carbon film resistors in speaker crossovers?!

In my country, the choice of high quality components is very limited.
These are the carbon film resistors that I can buy at this moment:
Cinetech Ind. Co., Ltd
http://www.cinetech.com.tw/upload/20...0330164634.pdf

What do you think?

Thanks in advance,

Aleksandar

[Conrad Hoffman]

Metal film, metal oxide and carbon film are three different things. Over simplifying things, metal film is probably the most precise and highly regarded. Metal oxide is often used for power and high temperature resistors. I've found it can have a voltage coefficient at higher voltages. Not much, but they're not as good as metal films. The tempco is probably higher. Carbon films are the cheapest of the cheap. They can perform well, but all properties are inferior. They are sometimes used as fuses to protect circuitry since they burn up easily. I use wire wound resistors for crossovers. There's nothing wrong with paralleling a bunch of metal films either. I only use metal oxide in power supply duty and carbon film for nothing at all. I don't subscribe to the different sounds of resistors unless maybe when used in a feedback network where the characteristics of the resistors are transferred directly to the amplifier, or in certain critical locations in tube circuits.

[DrDyna]

Having a bundle of resistors together might make getting the exact value you need a little easier, but I'd be willing to bet there isn't anyone who could tell the difference between these and a typical good wirewound resistor in an A/B test.

Edit: for the lulz.
http://www.engadget.com/2008/03/03/a...ter-cable-and/

[Helmuth]

I have done both parallel both type's.

I think I like metal film best. Wire wound types sounded bright to a bit harsh in the same application.

A B testing different resistors and capacitors will be hard, but when I do a direct comparison I can hear my preference.

[MartinQ]

Originally Posted by DrDyna

... but I'd be willing to bet there isn't anyone who could tell the difference ...[/url]

I guarantee that different types of resistors, with similar inductance and capacitance, will sound exactly the same. I don't see how it is possible for a difference to be heard.

Post up a theory or some measurements, not placebo anecdotes.

More myths, legends and voodoo ... just like cables. Don't spread the hype.

[davidsrsb]

Paralleled resistors can handle a lot less power than you expect, because their rating assumes the heat is going through the leads into a pcb

[Loren42]

You won't hear any difference.

I can't tell from the picture, but you want to get the crossover inductors as far as you can from the magnet of the driver.

[Helmuth]

Originally Posted by MartinQ

I guarantee that different types of resistors, with similar inductance and capacitance, will sound exactly the same. I don't see how it is possible for a difference to be heard.

Post up a theory or some measurements, not placebo anecdotes.

More myths, legends and voodoo ... just like cables. Don't spread the hype.

Nonsense.

Why bother telling others they can not hear it when they think they can.

And if you think it is nonsense is OK to me. You use what you think is best. Just do not try to say its voodoo when you can not prove it can't be detected by a trained listener.

ABC testing is no use because a dutch mathematicus calculated you have to do about 176 trails to say for sure the results are accurate. When you think you can do so much trails with test persons then you could prove something.

[Loren42]

Originally Posted by Helmuth

0Why bother telling others they can not hear it when they think they can.

It's called the truth and getting an education.

Imagine going to school and the instructors reinforcing every misconception you had with their rational being because you believe it true.

[Loren42]

Originally Posted by Helmuth

ABC testing is no use because a dutch mathematicus calculated you have to do about 176 trails to say for sure the results are accurate. When you think you can do so much trails with test persons then you could prove something.

That claim of 176 trials makes no sense. The point you are trying to make has to do with statistical relevance. However, not every study is the same.

For instance, how many iterations would be necessary to be statistically relevant to determine if the sun was shinning or if it was night?

The statistical relevance of any study depends on a number of factors as well as the ability to resolve or discriminate the changes you are trying to measure.

There is no way that 176 is a magic number in the game of statistical studies.

Lastly, I don't know who initially claimed what here, but whoever makes a claim is the one that must demonstrate proof.

However, my claim is that resistors do not make an audible difference in a passive crossover, so I offer this link as my substantiation.

Passive Crossover Network Design

If you read through the link you will realize there are a number of factors involved, but of all components in a crossover, resistors are the least of ones worries.