Tantalum Resistors: An Audio Advantage?

[Nigel Goodwin]

To actually compare some physical resistors, I'm looking for a circuit that is good at measuring parasitic inductances. I have a function generator and oscilloscope, so I was thinking of using something like a bridge circuit with precision foil resistors as the reference. Any suggestions?

If there was any measurable difference then it would have been used in the advertising for the resistors, rather than the entirely imaginary 'improvements' they claimed.

 

[gophert]

Sound is a funny thing. The people who know this best are the sound effects guys in movies - known as "Foley Artists". They work by the motto, it doesn't have to match the sound of the real event, it only has to Match the viewers expectation of the sound.

This was especially true when I was listening to a 1980s music collection that my sons gave me. Obviously, I had. heard most of these songs on classic rock radio stations on better and better audio systems over the years so most of them sounded "right". A few, however, did not sound "right" because the only time I ever remember hearing them is when I was driving my early 1970s Ford with an aftermarket FM Casette player and crappy 4" speakers.

The guitar riffs just lacked some of the great distortion I remembered. They sounded too clean and weak.

Maybe these songs would have sounded "right" (to my memories) if I played them on a system with Tantalum resistors.

 

[Wirth's Law]

I decided to re-visit this topic. Largely because I still haven't found other sources directly comparing the signal quality or noise of tantalum nitride (TaN) thin film resistors to the more common nickel-chromium (NiCr) variety.

I might not have mentioned it, but there's also the problem of nickel being a "magnetically hard" and strong permanent magnet. A hard magnet will have non-linear behaviors due to the magnetic hysteresis loop, and their current/voltage behavior always "remembers" the previous magnetic field. Hard magnets also experience heating and efficiency losses with AC signals, which is why transformer cores use "soft" iron ferrite rather than pure iron or steel.

Anyway, I decided to put it to the test and try to build a test rig to see if the difference is measurable. I put together this schematic back in August. I've been too occupied with other stuff to put much research into it beyond that, so I may have overlooked some key issues. There are three circuits (see attached)


Going left to right:

The left circuit is a simple voltage divider for an AC signal. I intend to place it between a 50 Ω sine wave source and an oscilloscope. R1A is a precision, non-magnetic, (i.e. expensive) 25 Ω resistor. DUT_1A is the random 25 Ω device under test (DUT). The idea is: if there is no inductance in the test resistor, there should be zero phase difference between the output signals J2 and J3.

The center and right circuits are identical except the center one is for surface-mounted resistors and the right side is for through-hole. Three resistors (R3, R4, R6) are identical, precision, non-magnetic 50 Ω resistors. The final resistor (R_DUT2) is the 50 Ω device under test. The idea is: if the resistor pairs have perfectly balanced impedance, there should be no signal out of J4. If there is a mismatch: capacitors can be added until the impedances are re-balanced. I'm guessing the difference won't be easily measurable except at radio frequencies but that's just my speculation. Any thoughts on this test?

 

[gophert]

Hard magnets also experience heating and efficiency losses with AC signals, which is why transformer cores use "soft" iron ferrite rather than pure iron or steel.

You are confusing soft iron and soft magnets. You are also confused about how heating is avoided in transformers.
Transformers are built with all kinds of materials, hard iron powders, soft iron powders, silicon steel (ferrosilicon steel), laminates, and ferrite. Ferrite is a class of magnetizable ceramics (it is not soft iron). It is mostly iron oxide fired with various other metals.

I think you are also confused about TaN resistors. My understanding is that TaN resistors are self-passivating and impervious to humidity so they retain their stable resistance value for, essentially, as long as the product lasts. There is never a specific mention of other unique aspects such as capacitance, inductance or other that any other thin film resistor can achieve which, is only important for RF and even GHz frequencies. Good luck with your tests.

 

[Wirth's Law]

Incorrect. No one here has said transformers don't use other materials, only that ferrite is one they use. Regarding the heating/efficiency of transformers, I disagree, and I'll refer to an excerpt from my source below:

(Source: R. E. Hummel, Ed., "Electronic Properties of Materials," 4th ed., Springer, 2011, pp. 385-388) "In order to make these devices [transfomers, electromotors, generators, or electromagnets] most energy efficient and economical, one needs to find magnetic materials which have the highest possible permeability (at the lowest possible price). Furthermore, magnetic core materials should be capable of being easily magnetized or demagnetized. In other words, the area within the hysteresis loop (or the coercive force, Hc) should be as small as possible. We remember that materials whose hysteresis loops are narrow are called soft magnetic materials."

The text goes on explain that the hysteresis losses of transformers occur when the core is subjected to complete hysteresis cycles, and the magnetic domains of the material dissipate the work done on them as heat. This loss is proportional to the total area enclosed within the hysteresis loop (i.e.: the magnetic softness/hardness). So when I referred to softness and hardness, I was referring to hysteresis loop of the core material. "Steel" was a vague term that could refer to hard or soft steel alloys, so I'll give you that. Heat can also occur due to eddy current losses, but that's a topic for another day. I have no idea if a modern NiCr resistor will easily experience the entire hysteresis loop of Ni or NiCr, so this may be something test if possible.

I think you are also confused about TaN resistors. My understanding is that TaN resistors are self-passivating and impervious to humidity...

Confused where exactly? Today, you've agreed with what I had already posted back on August 22nd about TaN's moisture resistance. My diagram from that same day even shows the passivization layer. So the worst-case scenario is we're both confused.

 

[gophert]

Then please, on with your test.

 

[crutschow]

Remember any A-B test is only really valid if it is a "blind" test with the evaluator not knowing which of the two circuits he is listening to.
For that you need a second person to do the random switching between the two.

 

[MichaelaJoy]

Remember any A-B test is only really valid if it is a "blind" test with the evaluator not knowing which of the two circuits he is listening to.
For that you need a second person to do the random switching between the two.

crutschow: You hit the nail on the head.

You'd have to switch A and B around in a random fashion, so that the test subject doesn't know which side is which.
Also, you need a lot of data points to show any correlation between tantalum resistors and audio quality.

I'm not saying "don't waste your time". I think it's a bigger test than you may realize and monies will have to be spent to pay
participants. You will also have to come up with a "cover story" so as not to give away the purpose of the test.

 

[unclejed613]

the OP is right to be skeptical about this. if an effect can't be measured (distortion, noise, frequency response, phase shift, etc...), then the "results" of testing for equipment reviews that ignore measurements, are completely subjective. there is a body of "audio urban legends" that continues to expand out of control. it began in the 1970s with the "large amounts of negative feedback degrade sound quality" myth. that one actually began because of a limitation in semiconductors as well as circuit design. the myth grew out of the discovery of "TIM" (Transient InterModulation distortion). it was noticed that when a lot of global feedback was used in solid state power amplifiers, certain types of intermodulation (nonlinear mixing in the frequency domain) distortion would result. so, IMD was avoided by improving localized negative feedback, and backing off on global feedback. when it was discovered that IMD was caused by the amplifier not having an adequate slew rate, the limitations on global feedback were no longer necessary because the actual cause of IMD had been found. even though the slew rate problem had been fixed, and global feedback could be used without limitation, the myth still continues to exist that too much global feedback (as well as too much open loop gain) is bad. that particular myth is unique, because it did at one time have some reason to exist, but many of the other ones are just myths based on a poor understanding of physics, or worse, something somebody has thought up to separate people from their hard earned money.

i found this data sheet for TaN resistors **broken link removed**
the primary reasons for using them are high reliability, low thermal drift, environmentally rugged. nowhere does the data sheet mention "happy audio", which seems like it would be a parameter that's difficult to quantify....

there are a lot of mythological audio "problems" out there, and the solution is always expensive, shiny, rare, or obscure. i've seen all kinds of crazy stuff on the internet "audio cables are directional", "wire contains billions of micro-diodes", "power and speaker cables need to be cryogenically de-stressed to work properly" (it may improve the performance of rifle barrels, but i doubt it would have much effect on cables, except that the insulation could crumble apart during the process). again. there's never any hard data to back up these claims. it's always percentages of listening tests (if there's any numeric data at all).

 

[unclejed613]

i'm half tempted to come up with a way to troll "audiophools", but i'm more afraid my bogus claims would become a part of the mythology....
something something noise from cosmic rays something something radiation hardened transistors, blah, blah, blah...

 

[gophert]

i'm half tempted to come up with a way to troll "audiophools", but i'm more afraid my bogus claims would become a part of the mythology....
something something noise from cosmic rays something something radiation hardened transistors, blah, blah, blah...

Thanks @unckejed613 ! I used your post as a reference to my new cosmic Ray Hardened, minimal global feedback, TIM-less® power amplifier. I've already sold three in the first hour.

 

[KeepItSimpleStupid]

Yep. A high enough slew rate and having multiple voltage gain stages is the way to go. My pre-amp used 3 attenuators to reduce the gain in each stage, not the overall gain so frequency response is 0 Hz to 100 Khz, or 0.5 Hz to 100 kHz.

My amp is totally nuts. An 800 kHz (0.8 MHz) bandwidth, but intentionally rolled off to 0.5 Hz and 40 kHz. The slew rate is over 100 v/uS.
I did use metal film resistors if I could. Exceptions were the emitter resistor was wire wound. Some were metal oxide (they act as fuses, The output has a carbon resistor with a coil of wire around it.

I hear one amp that choked on a high slew rate signal.

 

[philomel]

Every component in every circuit is a combination of R, C and L. Different components and technologies of them mix different qualities of R, C and L. Hearing is a very elusive sense and can detect many qualities which are pleasing or not. It may be that Tantalum resistors change the shape of the sound signal in a pleasing way like rolling off sharpness of signals by limiting High frequencies in a very subtle way. If it sounds better then it is to someone who likes it. It may be a distributed low pass filter.

 

[unclejed613]

My amp is totally nuts. An 800 kHz (0.8 MHz) bandwidth, but intentionally rolled off to 0.5 Hz and 40 kHz. The slew rate is over 100 v/uS.

that's part of my point. if you have an open loop gain of 80db, and you do your best to tame the distortion, when you close the loop at 20db, the gain margin is 60db. what that gain margin means is your distortion is reduced by that amount (theoretically, but there are a couple distortion mechanisms that result from sources outside the feedback loop), the output impedance of the amplifier is reduced by that amount, as well as other effects on the performance of the amplifier which are all enhanced by that gain margin. the mythology sees this as "bad juju" because you seem to be getting something for nothing. actually you aren't getting something for nothing, because you have 60db of "unused" gain, which is being put to use to linearize the amplifier, reduce noise, control voice coil motion, etc... the mythological position is "that's too easy and inexpensive".... well it's not. just try getting 80db of open loop gain, plus enough bandwidth and high enough slew rate, while maintaining a good phase margin...

Every component in every circuit is a combination of R, C and L. Different components and technologies of them mix different qualities of R, C and L. Hearing is a very elusive sense and can detect many qualities which are pleasing or not. It may be that Tantalum resistors change the shape of the sound signal in a pleasing way like rolling off sharpness of signals by limiting High frequencies in a very subtle way. If it sounds better then it is to someone who likes it. It may be a distributed low pass filter.

the goal in designing an amplifier, as David Hafler said is to make a "piece of wire with gain". amplifiers that add or subtract something from the audio, which have a "sound of their own", are used by musicians to create a unique "sound". that's fine for musicians, because they incorporate the quirks of a particular amplifier into the "total instrument". what this thread is about (at least partially), is whether an amplifier used for listening to music should have a unique "sound" or not. it would be difficult to tell whether you are listening to Frank Marino, Ted Nugent, or Joe Walsh, if you are listening to recordings, and playing them through Ted Nugent's guitar amplifier. the main topic in this thread though is how people get snookered into buying equipment built using super-expensive components, and following bizarre design philosophy (and all the while being convinced that 10% distortion is ok as long as it has a certain mix of even numbered harmonics). the "audiophile" market sells a lot of exotic cables and other paraphernalia, accompanied by technical mumbo-jumbo that has little or no basis in fact. there are myths out there that are just plain ridiculous, like:
a) audio cables are directional, one end of the cable works better as the output than as the input.
b) power cables need to be as heavy gauge wire as is practical so that power flows freely from the wall outlet to the unit's power supply.
c) power supplies in preamplifiers need to be as beefy as power amp power supplies in order to maintain a low impedance ground.
d) speaker cables need to be made of oxygen-free wire, preferably silver or silver plated copper, using many (enough to make the equivalent of 10 or 8AWG wire ) individually insulated small gauge (24, 26, or 28AWG) wires, teflon insulated, and braided. (usually available for up to $30.00 per foot)
e) the best and most natural sounding amplifiers are Single Ended Triode (may be a vacuum tube, or depletion mode MOSFET) running class A
f) audio cables need to be made of exotic metal alloys, because regular wire is made of metal atoms that act like billions of point-contact diode junctions. regular metal cables can have the point contact junctions collapsed by treating the wire cryogenically (the cryo-wire is slightly less expensive than the exotic alloy wire)

there are a lot more myths floating around out there, but those are among the most common ones.