Resistor Distortion

BinaryMike

Pelagic EE
I've been playing with my new distortion analyzer and decided to test a few resistors. Raw data like this could be interesting to AKers, so here's a dump.

The attached ZIP file has 23 PNG images taken with a very simple test fixture that puts a 10K test object (DUT) in series with a high-quality 1K wirewound resistor and applies 12.86Vrms at 1KHz across the DUT (14.14Vrms total). The analyzer looks at voltage across the 1K resistor, which is an accurate function of current through the DUT. The file names include a rated wattage field (.25W, .5W, etc), a type field where:

cc = carbon composition
cf = carbon film
mf = metal film
mof = metal oxide film
ww = wire wound


... and a sample number. The file designated 'bypass' was obtained with generator and analyzer directly linked and with the generator output dialed down to 1.286V. All of the carbon comp resistors are old stock.
 

Attachments

  • Rtest 3-2-17.zip
    878.4 KB · Views: 202
Looks like everything but the carbon comp is very comparable, and pretty close to invisible from a measurement POV. The carbon comps all look to suck to some degree, with the larger ones being worse.

Thanks for the data.
 
Mike, great job on this! Real science! Now I am convinced that carbon film resistors are just fine for almost everything, except high power. One comment, you should have put the reference "bypass' reading up first in line, to show the noise floor. But all (except C. Comp.) are at or slightly above the noise floor.
Perhaps the reason Carbon Comps. generate harmonics is because they are packed carbon, with little gaps internally that the electrons have to flow around.
Just a theory.
 
The carbon comps all look to suck to some degree, with the larger ones being worse.
The 1/4W carbon comps are definitely the worst in this batch, but I was surprised at how poorly the 2W carbon comps performed -- worse than the 1/2W samples. If this isn't an anomaly (all of my 2W CC samples came out of the same lot), then it calls the old strategy of using oversized resistors into question.
 
What I read in that is that film resistors work better than CC, which is fine by me. More reason to not use them :)
 
I would like to send my batch of Mouser carbon film resistors vs my cheap radio shack resistors. The mousers sounded awful for some reason????
 
I've been playing with my new distortion analyzer and decided to test a few resistors. Raw data like this could be interesting to AKers, so here's a dump.

The attached ZIP file has 23 PNG images taken with a very simple test fixture that puts a 10K test object (DUT) in series with a high-quality 1K wirewound resistor and applies 12.86Vrms at 1KHz across the DUT (14.14Vrms total). The analyzer looks at voltage across the 1K resistor, which is an accurate function of current through the DUT. The file names include a rated wattage field (.25W, .5W, etc), a type field where:

cc = carbon composition
cf = carbon film
mf = metal film
mof = metal oxide film
ww = wire wound


... and a sample number. The file designated 'bypass' was obtained with generator and analyzer directly linked and with the generator output dialed down to 1.286V. All of the carbon comp resistors are old stock.

Well, this proves what was well known - there is no place for carbon composite resistors in any audio gear. Other types are practically indistinguishable - all are offering very low distortion. Now someone needs to measure noise level and behavior at very high voltage (hundred volts) or high current (several amperes).
 
Well, this proves what was well known - there is no place for carbon composite resistors in any audio gear. Other types are practically indistinguishable - all are offering very low distortion. Now someone needs to measure noise level and behavior at very high voltage (hundred volts) or high current (several amperes).
I don't think carbon film resistors can be construed as entirely blameless, because they're used by the dozens in typical circuits. The simple fact that CF resistor distortion is visible at all when testing a single example is cause for alarm.

The generator section of the SR1 analyzer is maxed out at 40Vpp or 14.14Vrms. I would like to conduct resistor testing at 5~10 times greater voltage because it's relevant to tube amplifier design, but I haven't thought of a simple way to get that much signal voltage with sufficiently low distortion. One scheme that has some potential is to use a good audio op-amp driving a high-voltage SEPP output stage with tons of feedback, followed by a very sharp low-pass filter made with excellent passive components. Does anyone here have experience with something like this?
 
I don't think carbon film resistors can be construed as entirely blameless, because they're used by the dozens in typical circuits. The simple fact that CF resistor distortion is visible at all when testing a single example is cause for alarm.

The generator section of the SR1 analyzer is maxed out at 40Vpp or 14.14Vrms. I would like to conduct resistor testing at 5~10 times greater voltage because it's relevant to tube amplifier design, but I haven't thought of a simple way to get that much signal voltage with sufficiently low distortion. One scheme that has some potential is to use a good audio op-amp driving a high-voltage SEPP output stage with tons of feedback, followed by a very sharp low-pass filter made with excellent passive components. Does anyone here have experience with something like this?

There are OpAmps available for high voltage. LTC6090 can do 140 volts PP. There are hybrid devices available for up to 400 volts rails like MSK131 (though you will not like the price).
 
A DC offset, which would add self-heating, may also be an influence - but not sure that just raising the temp of the part would have the same effect.
 
There are OpAmps available for high voltage. LTC6090 can do 140 volts PP. There are hybrid devices available for up to 400 volts rails like MSK131 (though you will not like the price).
The LTC6090 has 0.01% THD at 100Vpp. Perhaps my best bet is to find a beat-up old SS power amp with an extremely low THD spec. I don't care about its listening qualities, but the THD spec needs to be well below 0.001% in the midrange. Output voltage should reach +/- 80V or more, which suggests a power rating of 300~400W at 8R. Got any ideas?
 
A lot of Crown gear has pretty low distortion specs and can be had for not stupid money. Not sure if they are quite that low though.
 
Doesn't this thread call into question what are you actually trying to do? If you are trying to restore Fisher, Scott, Pilot, etc. gear, then what is wrong with using new carbon comps to replace old carbon comps? It won't sound any worse than it was originally intended. If you are looking for resistors for a new build, or are trying to improve/mod. the original design of a vintage amp then by all means go with the least noisy.
 
Doesn't this thread call into question what are you actually trying to do? If you are trying to restore Fisher, Scott, Pilot, etc. gear, then what is wrong with using new carbon comps to replace old carbon comps? It won't sound any worse than it was originally intended. If you are looking for resistors for a new build, or are trying to improve/mod. the original design of a vintage amp then by all means go with the least noisy.

He tries experimentally find what would the best resistor type for specific application: low current, high current, low voltage, high voltage etc. This is a good experiment. There were similar done and results published for capacitor types.
 
He tries experimentally find what would the best resistor type for specific application: low current, high current, low voltage, high voltage etc. This is a good experiment. There were similar done and results published for capacitor types.
Absolutely, I downloaded BinaryMike's file. I'll have to look for the capacitor results.
Not criticizing, just saying that changing the type of resistor from the original may change the sound of a vintage piece of gear.
 
Would be interesting to see how vintage CC's compare to new ones. I wonder if they have gotten worse with time. That might actually be telling, if 50 year old resistors are worse than they were 50 years ago, it would be a good argument for replacing them.

Personally I'm not really into the "sound" of components. If its changing the sound, its distorting the signal in some fashion. I realize not everyone shares that opinion, and I don't wish to get into a long devolved conversation about it but thats just my opinion on the matter and you know what they say about opinions.
 
I don't mess around with vintage tube gear any more, but if somebody wants to send me an assortment of new 10K carbon comp resistors, I'll measure them in the same way and dump the data here for comparison.
 
Well, this proves what was well known - there is no place for carbon composite resistors in any audio gear. Other types are practically indistinguishable - all are offering very low distortion. Now someone needs to measure noise level and behavior at very high voltage (hundred volts) or high current (several amperes).
Actually, there are (only) two reasons I can think of to use Carbon Comps: 1. Radio circuits-Carb. comps. have very low inductance which won't detune circuits.
2. Grid stoppers - installed as close to the socket pins as possible, again , for low inductance into a high impedance tube grid.
 
Back
Top Bottom