Resistor distortion

[Art M]

How do I go about selecting resistors that exhibit a low Voltage Coefficient of resistance when presented with the large voltage swings present in a power amplifier ? It appears that the distortion problem was discussed about at:

Slones 11.4 'Blameless'

I have not been able to find any further information with references to specific part classes.

This note talks about the problem:

http://www.diyaudio.com/forums/attac...pulse-page.pdf

[jcx]

"Bulk metal resistors have almost unmeasurable voltage
coefficients."

or make all of your feedback dividers out of equal valued, identical type reisistors - they will each have the same power and temp rise - to the degree their individual thermal environments match - then the division ratio is independent of temp, V coefficients

[Damon Hill]

Resistive voltage coefficient, unlike temperature coefficient (and noise), is a little-recognized source of non-linearity; perhaps a very minor contributor overall but worth minimizing since distortion in audio design is otherwise being driven down to very tiny levels. I think it's possible to classify by basic resistor material types, and then by specific manufacturers and products.

Carbon resistors (composition or film) have the worst voltage coefficient, then metal oxide and cermet; metal film, and wirewound and bulk metal foil have the least. Metal film resistors with the lowest Tc often claim a lower voltage coefficient, such as Vishay-Dale's PTF series.

Thus you'd be looking for a Tc of 100 ppm, preferably 25 ppm or better; Vishay's Z-foil resistors are below 1 ppm and probably the most stable resistor made. Unfortunately, they are very expensive and nearly unavailable in resistance above about 250 Kohms. I use the older bulk foil types because they're more commonly available and a bit less expensive. Caddock is a highly regarded manufacturer of high voltage and high resistance parts.

Also, derate resistors by using a higher wattage part where space permits; many film resistors are rated for less than a quarter watt and may be otherwise perfectly suitable for a given application, but a physically larger part with say a half-watt rating will be less stressed by heat and voltage.

Traditionally, many tube afflictionados prefer the "warm" sound of composition resistors and dislike the "sterile" neutral sound of the better quality film and bulk metal types. Vishay's bulk metal foil resistors can be thought of as miniaturized wirewound
with a non-inductive construction. I'm not sure whether inductance really matters in most audio resistor applications; overall quality of construction and perhaps in end cap construction are cited as potential sources of non-linearity, too.

[davidsrsb]

Voltage coefficient gets worse at higher values, so putting a few cheap metal oxide resistors in series wins twice - lower values and less voltage on each part

[bear]

There is/was a discussion of this in the John Curl Blowtorch Pt II thread...

_-_-bear

[AndrewT]

was it 5th element that reported a massive drop in measured (not simulated) distortion when he changed an SMD resistor for another with higher voltage rating?