Do resistors age?

[Ivan Vegvary]

Garage sale purchase, box of 100 1M, 1 Watt, 5% resistors. Probably 50 years old.
Thought I'd measure all 100 and single out the one or two that are possibly within 1% of nominal value.
Well, only a single resistor read as low as 1.06M, the other 99 were reading 1.17M, on average.

I know my VOM is fairly accurate, since I can sample new 1/4 watt resistors and get evenly distributed readings on both sides of 1M. Measuring 1% resistors gives me readings of 1.005M or 9.995M, typically.

What gives?

Thank you,
Ivan Vegvary, Oregon

[Phil Allison]

** Resistors do age, some types much more than others.

Carbon composition types are famous for drifting over a period of years or decades to higher values - often much higher.


.... Phil

[Phil Hobbs]

>Garage sale purchase, box of 100 1M, 1 Watt, 5% resistors.  Probably 50 years old.
>Thought I'd measure all 100 and single out the one or two that are possibly
>within 1% of nominal value. Well, only a single resistor read as low
>as 1.06M, the other 99 were reading 1.17M, on average.

Well, the resistance of carbon comp units like those decreases with applied voltage. I bet that if you were dissipating even a sixteenth of a watt in those resistors (250V) they'd be a lot closer to 1M. ;)

Other than that, I have no idea.

Cheers

Phil Hobbs

[ehsjr]

Carbon resistors increase value with age.

Ed

[Spehro Pefhany]

More than we can say for some around here.

--sp

--
Best regards,
Spehro Pefhany
Amazon link for AoE 3rd Edition: http://tinyurl.com/ntrpwu8

[whit3rd]

On Tuesday, November 29, 2016 at 6:14:15 PM UTC-8, Ivan Vegvary wrote:
> Garage sale purchase, box of 100 1M, 1 Watt, 5% resistors. Probably 50 years old.

> Well, only a single resistor read as low as 1.06M, the other 99 were reading 1.17M, on average.
>

> I know my VOM is fairly accurate...
> What gives?

Carbon resistors can be corroded. Oxygen and carbon turns to CO (yeah, carbon monoxide),
and the carbon content will gradually diminish (unless you have ozone around, which
makes the process much quicker). Some very good barrier coatings are available for
carbon film resistors, but there's oxygen everywhere, and it has been decades since
those resistors were labelled.

Carbon-black inks also fade with time, for the same reason; Gutenberg bibles have held
up well, because he printed with metal-oxide inks.

[Robert Baer]

As i remember it (40+ years ago), 5% carbon comps were typically 2% high.

[George Herold]

Don't know, but I had no luck trying to hand select 1% metal film resistors.
I never got a distribution over the whole range, but all clustered near one value,
(sometimes two values.)
I also found that for 10 meg 1% (1/4 W, TH) that when heated (solder them into
place) the resistance would change slightly... on the 0.1%- 0.2% level.

George H.

[Sjouke Burry]

Production runs and mixture batches.

They vary, along with temperature, air pressure and humidity.

So your measurements suggest they all come from 1 production run.

Which side of the distribution curve they hit is a bit of a gamble,
but within the batch they can be quite close together.

[Tim Wescott]

Quite rapidly if they're dissipating several times their rated power.

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com

[tabb...@gmail.com]

On Wednesday, 30 November 2016 18:17:17 UTC, Tim Wescott wrote:
> On Tue, 29 Nov 2016 22:05:30 -0500, ehsjr wrote:
> > On 11/29/2016 9:14 PM, Ivan Vegvary wrote:
> >> Garage sale purchase, box of 100 1M, 1 Watt, 5% resistors. Probably 50
> >> years old.
> >> Thought I'd measure all 100 and single out the one or two that are
> >> possibly within 1% of nominal value.
> >> Well, only a single resistor read as low as 1.06M, the other 99 were
> >> reading 1.17M, on average.
> >>
> >> I know my VOM is fairly accurate, since I can sample new 1/4 watt
> >> resistors and get evenly distributed readings on both sides of 1M.
> >> Measuring 1% resistors gives me readings of 1.005M or 9.995M,
> >> typically.
> >>
> >> What gives?
> >>
> >> Thank you,
> >> Ivan Vegvary, Oregon
> >>
> >>
> >
> > Carbon resistors increase value with age.
> >
> > Ed
>
> Quite rapidly if they're dissipating several times their rated power.

Even at rated for power resistors, some types are only 1000hr rated.


NT

[Bill Martin]

On 11/30/2016 10:17 AM, Tim Wescott wrote:
> On Tue, 29 Nov 2016 22:05:30 -0500, ehsjr wrote:
>
>> On 11/29/2016 9:14 PM, Ivan Vegvary wrote:
>>> Garage sale purchase, box of 100 1M, 1 Watt, 5% resistors. Probably 50
>>> years old.
>>> Thought I'd measure all 100 and single out the one or two that are
>>> possibly within 1% of nominal value.
>>> Well, only a single resistor read as low as 1.06M, the other 99 were
>>> reading 1.17M, on average.
>>>
>>> I know my VOM is fairly accurate, since I can sample new 1/4 watt
>>> resistors and get evenly distributed readings on both sides of 1M.
>>> Measuring 1% resistors gives me readings of 1.005M or 9.995M,
>>> typically.
>>>
>>> What gives?
>>>
>>> Thank you,
>>> Ivan Vegvary, Oregon
>>>
>>>
>>
>> Carbon resistors increase value with age.
>>
>> Ed
>
> Quite rapidly if they're dissipating several times their rated power.
>

Smoke is conductive? Must be so... :-)

[Adrian Jansen]

Depends very much on the materials used to make the resistor, and even
more on the processing used.

Ancient 'carbon composition' resistors could drift ( usually upwards )
as much as several hundred percent per year.

Various brands like Allen Bradley made much better ones, drift maybe
around a few percent or better per year.

Metal-oxide film around less than 1% per year.

Metal film down as low as ppm per year.

I would say that its remarkable that they were all so close together.
--
Regards,

Adrian Jansen

[M Philbrook]

In article <o1neio$eat$1...@dont-email.me>, w...@wwmartin.net says...

depends on color

[Robert Baer]

YES!!!

[Winfield Hill]

Phil Allison wrote...

>
> Ivan Vegvary wrote:
>>
>> Garage sale purchase, box of 100 1M, 1 Watt, 5% resistors.

>> Probably 50 years old. ... a single resistor read as low

>> as 1.06M, the other 99 were reading 1.17M, on average.
>

> ** Resistors do age, some types much more than others.
>
> Carbon composition types are famous for drifting over a period
> of years or decades to higher values - often much higher.

Paul and I have been arguing about carbon-comp,
or CC, resistor drift for years. He claims to
often observe up to 25% change; but measurements
of my extensive 25-year-old collection doesn't
revealed serious drift, with a few exceptions.

It's not surprising Ivan's whole batch has drifted
together, it's be more surprising if they didn't.

Many if not most portions of a design can handle
a 20% variation in value, we're used to that with
caps for example. But if not, one should be using
1% parts. Given the low cost of 1% parts, many
blokes have switched to using 1% as the default.

My CC collection is very well-made Allen-Bradley
types, with crisp molding and bright color bands.

Here's my story: As a pulsed-power-nut, many of my
designs rely on the CC resistor's superior ability
to handle high peak transient power, far more than
the normal 5x spec for most other resistors; and
these parts perform reliably in my designs.

I've previously related here horror stories of
failed commercial products that used ill-advised
CC replacements unable to handle transient power.

Anyway, for the last 10 years I've been moving to
newer type axial-lead bulk-material *pulse-rated*
resistors, and as a result have been *struggling*
with poor distributor inventory of these types,
especially for 1W and higher ratings. Minimum
orders and 16- to 20-week delivery times. :-(

The what-to-use issue is dramatically worse with
surface-mount types. It's especially painful to
try replacing 1W and 2W CC or 2 to 5W WW parts.
Sheesh, forgettabout it. Series/parallel stacks.


--
Thanks,
- Win

[speff]

It would be interesting to do some accelerated aging tests of resistors under bias. I bet I can make them drift a lot.

--sp

[John Fields]

On Fri, 2 Dec 2016 11:45:58 -0800 (PST), speff <spe...@gmail.com>
wrote:

You could, but would the accelerated aging tests correpond to what
happens in real life?

---
John Fields

[Dave Platt]

In article <2f81ce08-e40b-4fbc...@googlegroups.com>,

speff <spe...@gmail.com> wrote:

>It would be interesting to do some accelerated aging tests of resistors under bias. I bet I can make them drift a lot.

There's a common failure syndrome in the Tektronix 2235 and similar
oscilloscopes, in which a set of carbon-composition resistors in the
high-voltage focus chain suffer from value-drift. The scope loses the
ability to focus its spot properly.

I don't know whether the high voltage, or the relatively high power
dissipation in the resistors is tending to cause them to drift more
quickly than would otherwise be the case. Wouldn't surprise me,
though.

My local surplus store did have 510k carbon-comp resistors in stock, I
bought 8, and was able to find five that were close enough to 510k to
work OK. Special high-voltage-rated film resistors would have been
another choice, but would have been much harder to find.

[whit3rd]

On Friday, December 2, 2016 at 1:08:13 PM UTC-8, Dave Platt wrote:
> In article <2f81ce08-e40b-4fbc...@googlegroups.com>,
> speff <spe...@gmail.com> wrote:
>
> >It would be interesting to do some accelerated aging tests of resistors under bias. I bet I can make them drift a lot.
>
> There's a common failure syndrome in the Tektronix 2235 and similar
> oscilloscopes, in which a set of carbon-composition resistors in the
> high-voltage focus chain suffer from value-drift.

Also televisions, and computer video monitors; anything with a CRT and focus circuit.

> I don't know whether the high voltage, or the relatively high power
> dissipation in the resistors is tending to cause them to drift more

It's partly the high ohms value that does it. For a carbon film resistor, the high ohms
means a thin film (so a few microns of surface oxidation does major damage). The
situation for composition resistors is similar, you cannot get high value in composition
resistors without small particle sizes.

[Phil Allison]

Winfield Hill wrote:

> Phil Allison wrote...

> >
>
> > ** Resistors do age, some types much more than others.
> >
> > Carbon composition types are famous for drifting over a period
> > of years or decades to higher values - often much higher.
>
> Paul and I have been arguing about carbon-comp,
> or CC, resistor drift for years. He claims to
> often observe up to 25% change; but measurements
> of my extensive 25-year-old collection doesn't
> revealed serious drift, with a few exceptions.
>

** Only last week I had a 50 year old valve amplifier on the bench ( Rogers Cadet II) that used all CC resistors. Most had drifted high by 45% to 300%. A couple were open cct. All eight UK made Mullard valves were original and OK, so the amp had not seen excessive use.



.... Phil

>
> My CC collection is very well-made Allen-Bradley
> types, with crisp molding and bright color bands.
>

** I have a quantity of one like that - banded 100ohms, 5%.

All of them show 109 ohms +/- 1 ohm on my Fluke DMM.



.... Phil

[Winfield Hill]

Phil Allison wrote...

>
>Winfield Hill wrote:
>
>> Phil Allison wrote...
>>
>>> ** Resistors do age, some types much more than others.
>>>
>>> Carbon composition types are famous for drifting over a period
>>> of years or decades to higher values - often much higher.
>>
>> Paul and I have been arguing about carbon-comp,
>> or CC, resistor drift for years. He claims to
>> often observe up to 25% change; but measurements
>> of my extensive 25-year-old collection doesn't
>> revealed serious drift, with a few exceptions.
>
> ** Only last week I had a 50 year old valve amplifier on
> the bench ( Rogers Cadet II) that used all CC resistors.
> Most had drifted high by 45% to 300%. A couple were open
> cct. All eight UK made Mullard valves were original and
> OK, so the amp had not seen excessive use.

Our observed drifting CCs were "new" in drawers.
It appears that higher values suffer higher drift.


--
Thanks,
- Win

[circuitmaker]

Winfield Hill wrote:

> But if not, one should be using
> 1% parts. Given the low cost of 1% parts, many
> blokes have switched to using 1% as the default.

But the tolerance itself doesn't change anything in
the context of aging, so it must be another factor.
The 1% parts are mostly metal (oxide) layer, so is
it that? Does metal (oxide) layer not age noticeably?
What else does not age? Mica caps?

Another thing to mention: I have a 2k/0,02% resistor
older than 30 years and it still gives 2,000 readings on
all my meters. It is big and sealed - what can be inside?

c.

[Sjouke Burry]

Either conductive bulk, or conductive layer, or fleas
transporting charge packages.

[whit3rd]

On Wednesday, December 7, 2016 at 3:04:12 AM UTC-8, circuitmaker wrote:

> Another thing to mention: I have a 2k/0,02% resistor
> older than 30 years and it still gives 2,000 readings on
> all my meters. It is big and sealed - what can be inside?

Probably a card sloppily (low tension) wound with manganin or
constantan wire, possibly with a series or shunt trim select-according-to-test
resistor. I'd wager (if manganin) it's good for +/- 10 C temperature
range for that accuracy. To exclude oxygen (so the wire doesn't corrode)
there may also be an oil bath or potting material in addition to the sealant.
Lacquer on any uninsulated surfaces is likely.

It's possible to make metal films quite well nowadays, but annealed
bulk wire has superior aging (a metal film on ceramic substrate can
undergo mechanical stress as the dissimilar materials temperature-cycle).

[Phil Allison]

circuitmaker wrote:
>
> Winfield Hill wrote:
>
> > But if not, one should be using
> > 1% parts. Given the low cost of 1% parts, many
> > blokes have switched to using 1% as the default.
>
> But the tolerance itself doesn't change anything in
> the context of aging,
>

** The tolerance spec of a resistor includes effects from soldering and aging for a reasonable lifespan of the part. This is why the value when new is well within the marked percentage.

1% resistors normally have much less aging than 5% or 10% types - they would be useless as precision parts if they did not.

An exception is WW types, where aging is very low but tolerances are rarely better than 5% cos of the very different manufacturing process involved.


... Phil