Out of spec NOS A-B resistors query

[zamproject]

At the end of the day, I think the most simple solution is not to buy and use "NOS" carbon comp res...
Buy new, or use metalfim
Best
Zam

 

[PermO]

I just did a short experiment with a big 3M mechanically trimmed SATOR carbon resistor from the 40's.

Hooked it up to my Fluke wich read 3.328M

Got a heatgun and heated the resistor, I could see the resistance drop while doing it, kept going till it read 3.000M

Left it to cool and now the reading sticks at 3.129M

I'll try again on the same resistor in a couple of hours to see if I can get it down some more...

 

[Disco Volante]

Don't know where you are located, but here in Europe, RS Components has a good selection of good-looking resistors of various chemistries. I prefer new components for my restorations... (Also not affiliated with said company;-)

 

[JohnRoberts]

I just did a short experiment with a big 3M mechanically trimmed SATOR carbon resistor from the 40's.

Hooked it up to my Fluke wich read 3.328M

Got a heatgun and heated the resistor, I could see the resistance drop while doing it, kept going till it read 3.000M

Left it to cool and now the reading sticks at 3.129M

I'll try again on the same resistor in a couple of hours to see if I can get it down some more...

I am not a fan of old stuff just because its old... (I'm old).

A question about your heat experiment is the resistance change permanent or temporary? Resistors can have a temperature coefficient but shouldn't be very large.
======
Today's TMI

Back when I was dealing with large scale manufacturing I ran into a production problem with brittle PC mount potentiometers that were breaking from normal handling on the manual insertion assembly line. After a deep dive researching the problem it turned out that the pot manufacturer was having production problems meeting resistance tolerance. The pot technology was screened resistive ink on a substrate. The green resistance of that conductive ink changes permanently with time in the curing oven at elevated temperature. Apparently the manufacturer was have process management (green ink resistance) issues and spending too much time in the curing oven trying to pull the final resistance into spec. But this extra time in the curing oven dried out the substrate and making it too brittle to survive normal handling.

Ironically or comically the pot manufacturer dropped us because we were "too picky" of a customer. There was a line of other pot makers happy to win our business.

JR

 

[PermO]

It's to soon to tell... it still reads 3.129M now... I just left it sitting there.

I'll heatstress it again in an hour or so.

I don't believe a short burst from a heatgun will drive out accumulated moisture... but who knows ?

 

[Zim]

At the end of the day, I think the most simple solution is not to buy and use "NOS" carbon comp res...
Buy new, or use metalfim
Best
Zam

funny thing is, I actually have all new components as specced in the BOM (I’m doing Rainton’s Analogvibes build) so many of them are still carbon comp rather than metal film, but they are new.

I’d read Rainton’s post about original Allen-Bradley pots having a different, possibly more suitable, log than modern PEC pots so tracked some down and then went down the rabbit hole with about 4 values of resistor. A couple of Ohmite resistors in the BOM are no longer available and the 10% tolerance (rather than specced 5%) alternatives that I ordered yielded only one of appropriate value, even out of the 30 or so I ordered to try and get 4. Hence the original spec Allen-bradley resistor rabbit hole - I had thought the mil-spec should maybe mitigate things like the tolerance problem - clearly not...

I’m still pretty new to this DIY adventure though and, while I’ve now put together a 51x rack, power supply, 4 x 500 series module, a couple of microphones + a bunch of guitar pedals, this is my first ‘proper build’.

As with every project before it, I learn more and more each time and this is my first sortie into sourcing my own parts, hence the initial query and appreciation for everyones input and experience that’s furthering my education.

thanks everyone - this forum’s an invaluable, fascinating and refreshingly inclusive resource

Zim

 

[PermO]

Took another reading and it sits cold at 3.224M now.

I did another round of heatgun to bring it down to 3.000M again

And slowly creeping back up...

 

[JohnRoberts]

I vaguely recall a discussion about "excess" noise in resistors.

[edit- excess noise is the noise correlated to current flow affected by granularity of the resistive media. I saw a campaign arguing that carbon comp was superior in this regard. That was the last I heard of it... not much of an issue since excess noise is down tens of dB below the signal. Perhaps an issue with resistors used with DC bias current. nah... [/edit]

Took another reading and it sits cold at 3.224M now.

I did another round of heatgun to bring it down to 3.000M again

And slowly creeping back up...

is there a published temperature coefficient for that component?

 

[PermO]

Not any that I know of, there was a bit of a war going on at that time in Europe...

It reads 3.127 now at roomtemperature.

I made sure I did not touch th leads or move the clips while doing this, I can just leave it sitting there to tomorrow morning and see what it reads.

Coocking them at a lower temperature for a longer period might be a thing to bring them back.

It's this type;
https://www.ebay.ie/itm/5X-20-kOhm-3W-carbon-film-resistor-Sator-Klangfilm-vintage-/392309828697
I can make money on these !

 

[zamproject]

Carbon comp and carbon film are two different stories !

Best
Zam

 

[PermO]

I'm not shure it's film, the picture is correct for the resistors I have, Its a ceramic tube, with a layer of carbon, wich has machined grooves to tune it to it's value and it's then dipped in paint.

I avoid the Allen Bradley resistors, I don't like them.

 

[zamproject]

For what I know it's what we call "film", a ceramic core, with a carbon film around, trimed with a "groove" as you say.

To my observations, they don't drift due to substrate like carbon comp, BUT It happen the mechanical junction to the legs (crimped copper ring) crack or break, giving wrong or inconsistent value (the contact surface being random)

Best
Zam

 

[mjrippe]

I have slowly been getting rid of all my carbon comp resistors, replacing with metal film that should last much longer and hold their spec. However, I have heard that you can bake them at low temp for several hours and then dip them in shellac or varnish to seal them against moisture.

 

[scott2000]

I know some old military stuff came packaged with those humidity indicators. Just actually got a couple indicators shipped from Mouser with some op amps in anti static bags iirc..Never seen that before. Maybe I'm mistaken but I'm pretty sure...

These

https://staticcontrol.descoindustri...ssories/Humidity-Indicator-Cards/51060HIC125/

 

[bOTL]

Purchase some new Carbon Film or Metal Film do not waste your time with ancient Carbon Comp.

 

[Whoops]

just got 20 x 2.7k carbon comp resistors and all are measuring somewhere around 3.3k. the colour codes are correct for 2.7k so the values have obviously drifted.

Tolerances of +- 20% for Carbon Composition resistors was normal and standard back in the day.
That's basically what you have, +20% tolerance.

"In view of the manufacturing tolerances, it was very difficult to manufacture them to a given value. Instead they were sorted after manufacture. Typical tolerances were only ±20%, although ±10% and occasionally ±5% versions were available."

 

[Tubetec]

Hi Permo ,
I have a small handfull of those Sator Austria carbons(used ,recovered from old gear ) in the 2kohm values , just out of curiosity I put the meter across one that came to hand to check how it spec'd out , 2025ohms so within 1.5% ,even though no tollerance is noted on the item . Considdering these things are 50 or more years old and had a lifetimes use already , thats good going .

The Welwyn panclimatic are a favourite of mine , the end caps can become an issue with careless handling but apart from that I dont think I ever found one out of spec . Ive had excellent results with 100k panclimatic as anode load in U47 type tube mic circuits.

 

[MisterCMRR]

This is a memorable subject for me! One of my first jobs in electronics was as a test tech in a General Electric plant called the Neutron Devices Dept (I couldn't enter the factory, but they paid me for 3 months while the FBI investigated me and I got a "Q" or top-secret security clearance). After almost another year there, an engineer confided in me that they were making the neutron trigger generator that set off an H-bomb. Understandably, reliability of every little component had to be incredibly high. I would get tasks like temperature cycling and salt-spray before/after tests on resistors, capacitors, etc. I once rejected a batch of carbon composition resistors (the Allen-Bradley types that were industry-standard in the day ... this was 1966). Anyway, I was told to re-test them ... PROPERLY! READ THE SPEC SHEET. Hot-molded "carbon" resistors have "interesting" characteristics, as you'll find on page 9 (page 6 of the file) of the attached data sheet. Measuring with an ordinary (low voltage) ohmmeter will always give quite high readings. For example, 100 kΩ resistors must be measured at 100 VDC to give their correct value. Maybe this strong voltage-coefficient of resistance is what gives them unique distortion characteristics in audio circuits. In a class-A amplifier, it would produce musically-pleasing even order distortion - likely 2nd would be dominant because of the smoothness of the E/I plot's curvature. Maybe those ancients were smarter than we think!

 

[Tubetec]

Cheers Bill ,

 

[MisterCMRR]

I vaguely recall a discussion about "excess" noise in resistors.

[edit- excess noise is the noise correlated to current flow affected by granularity of the resistive media. I saw a campaign arguing that carbon comp was superior in this regard. That was the last I heard of it... not much of an issue since excess noise is down tens of dB below the signal. Perhaps an issue with resistors used with DC bias current. nah... [/edit]

You're right John that "excess noise" applies only when there's DC current flow in a resistor. My recall of several discussions about it (in books by Motchenbacher and others) is that carbon composition resistors are the worst in this regard, metal and metal oxide films better, but the best, having essentially no such noise, are solid metal wire-wound or foil resistors. In terms of thermal agitation noise (or "Johnson" noise, the kind most associated with resistors in audio circuits) all resistors are exactly the same, regardless of construction. "Audiophile" designers waste money when they put exotic resistors in signal paths where no DC voltages exist. But things like the 6k81 phantom-power resistors will contribute excess noise into the mic signal, so metal-film (at least) resistors are appropriate.