Resistor Sound Testing

[DragonMaster]

Each persons having different opinions of different types of resistors, I decided to try them myself. I'd like to AB Metal Film, Metal Oxide, Carbon Composition and Carbon Film types.

For that, I'll use a pair of AKG headphones, a CD player and a *simple* headphones amp - where the different resistors will be tested.

To hear the sound of the resistors, I suppose I need to put an high value resistor in the signal path. But, what for the rest of the circuit?

BTW, simple means something similar to this:
http://www.geocities.com/tomzi.geo/2tr_amp/2tr_amp.htm

[Curmudgeon]

That circuit won't take you where you want to go. To list just two problems, the capacitors will influence the sound more than the resistors under test. And the biasing arrangements are flakey to the max. That is the sort of circuit you saw in the earliest transistor days, when circuit design was still influenced by tube design.

I'm not an expert on headphone amp design, but over at Head-Fi Forums there are many.

Being mildly compulsive, I would change all the resistors in the circuit; but changing the feedback resistor pair will have the greatest effect. However I'd actually use a different approach. I'd make a 20 dB amplifier, probably using one of the Burr Brown/Texas Instruments 2604 family (look for the opamp-rolling articles) with a 20 dB multistage attenuator -|-|-| and not change your headphone amp circuit at all. I'd use Vishay bulk film resistors, from Michael Percy, for the amplifier circuit, and the attenuator would be made up of the resistors under test.

For an inverting circuit, and purchasing convenience, I'd use series 10K, Shunt 10K, series 10K, Shunt 10K, series 10K (into the amp inverting input) and feedback of about 80 K around the amp. This will increase the effect of the resistor type. I'd insert this in series with your headphone amp input, and your headphone amp of course needs to be as good as you can manage.

[I_Forgot]

Will you be setting up a mechanism to switch the resistors without you seeing which is in the circuit? Will you be listening to more than a single sample of each brand/value? And will this test occur with statistically significant sampling? If not, save yourself the time and effort, and us the undoubtedly long rambling account of the test and just tell us which resistor you think looks nicest...

I_F

[SY]

Mark, you're very cynical.

I like that.

[chipco3434]

n = 2
p=0.5

[I_Forgot]

At the very most, even if you detect some difference in the "sound" of some resistors, it is going to be at most very subtle. Will it allow you to predict anything? Will your results be applicable to every batch of resistors made by all the manufacturers? Will it apply to every resistor in every circuit, or just to certain types of positions (opamp feedback, for example)? Will the result be valid for all types of circuits or even all circuits of the same type?

I'm not meaning to be cynical. It's just that doing a test like this properly (whatever THAT means) requires so much time and energy, and the results are of such limited applicability, that it seems to me that it would be far better to spend the time studying electronics and learning something of value instead. Get a copy of "The Art of Electronics" and study it. You'll learn many far more valuable things from it in an hour than you will by listening to every resistor in the world.

In spite of what the marketing guys tell you, audio is not magic. It is engineering.

I_F

[Curmudgeon]

It's your time, money, interest and above all your hobby. I'd suggest enjoying it in any way you wish.

And this does make a nice little project to learn a bit about electronics too.

"The Art of Electronics" will take you as far, and as fast, as you wish to go.

[serengetiplains]

Originally posted by I_Forgot
At the very most, even if you detect some difference in the "sound" of some resistors, it is going to be at most very subtle. Will it allow you to predict anything? Will your results be applicable to every batch of resistors made by all the manufacturers? Will it apply to every resistor in every circuit, or just to certain types of positions (opamp feedback, for example)? Will the result be valid for all types of circuits or even all circuits of the same type?

I'm not meaning to be cynical. It's just that doing a test like this properly (whatever THAT means) requires so much time and energy, and the results are of such limited applicability, that it seems to me that it would be far better to spend the time studying electronics and learning something of value instead. Get a copy of "The Art of Electronics" and study it. You'll learn many far more valuable things from it in an hour than you will by listening to every resistor in the world.

In spite of what the marketing guys tell you, audio is not magic. It is engineering.

I_F

Oh my, what a soap box!

I once changed a single resistor in a conrad johnson tube preamp and heard a difference for the better. Resistors add measurable noise and so far as I've lately understood, noise ain't music. Ask Mr. Vishay, he could probably tell you.

DragonMaster, do do your test. Ain't no measuring equipment anywhere can tell you, in the final instance, and for subtleties like resistor sound, whether an audible change may have resulted, "audible" being the key word. And, hey, if you also read The Art of Electronics, you'll have an edge for knowing to correlate ears to EE sense.

[DragonMaster]

No, I'm not going to use ABX. I've heard bad things about all of them and think that they all sound the same so I don't think it's necessary.

[I_Forgot]

Originally posted by chipco3434
n = 2
p=0.5

I just realized what this means! Very good! Hah!

Originally posted by DragonMaster
No, I'm not going to use ABX. I've heard bad things about all of them and think that they all sound the same so I don't think it's necessary.

In this case I don't know whether you are trying to be funny or not... if you are, VERY GOOD! If you aren't, well...



TD