Carbon comp resistors - shifting value

[Mick Bailey]

I've never noticed this before, but just got in some sets of new carbon comp resistors and some give an instant stable measurement on my Fluke DMM, but others creep up slightly, as in a 220k starting 217.3 then going up 217.4, 217.5, 217.6.... until a stable value is reached - like measuring a resistor across a cap. Mainly affects values of 100k or more, but not all of them do this.

I checked a much older batch and they also do the same thing. Tested some old ex-equipment resistors and they don't do this. What's the deal?

[loudthud]

When I worked at Tektronix they told us that a 5% carbon comp resistor has a tolerance of about +/- 15% unless you dried them in an oven for something like 30 minutes. They do absorb moisture. Make sure you are not changing the part's temperature by touching it. Use clip leads so you get a firm connection without touching the part.

[soundguruman]

The do absorb moisture, which may partially account for why they are so noisy.
Also CC is made with glued in leads.

There are huge amounts of CC resistors costing 1/2 cent each or less.
And to sell these, people want you to think that CC has a special "tone"
or a vintage "tone." Which of course they do not.

We should be using metal film or bulk metal foil, which is far superior.
But it's going to take the amp makers a few years to catch on to what's better.

[Mick Bailey]

I agree with metal film being technically superior, and very much less sensitive to temperature and climate. Carbon comps had almost disappeared in the UK, but have in recent years made a comeback for pulse applications, so the major electronics industrial suppliers are now stocking Arcol carbon comps in 1/4W and 1/2W. It comes down to where they're being used; I think that in repairing vintage gear, or re-creating vintage circuits, it's nice to use components that look like the originals.

If moisture is the cause of the effect, then I'm not sure why the resistance would change in one resistor and not another from the same sealed pack. I get the temperature aspect and have tried measuring again, but without handling the component beforehand or during test (like matching transistor pairs for ladder filters), but get the same results. The interesting thing is that if the probes are removed and quickly reapplied, the resistance 'resets' to the original value and climbs again.

[Austin]

I feel like carbon comp can withstand overload without burning out better than other kinds, but maybe just because they look beefier. Do you guys think this is true?

[Tom Phillips]

I feel like carbon comp can withstand overload without burning out better than other kinds...

To my knowledge that has always been a positive characteristic of CC resistors. They do withstand heavy momentary power dissipation events better than many other types of resistor construction but I don't believe they are not necessary better than all other kinds such as wire wound types. A lot depends on the quality of the part too. The original hot molded Allen Bradly CC resistors for example, were much better CC types than less expensive competing CC types. This was true in the old days and now. I believe the ABs were superior in many respects from the finish to the composition of the resistive core to how the leads were attached.

[Gingertube]

Carbon Comp does have superior voltage withstand, particularly high voltage pulse withstand but they are noisy and drift badly with age. The other thing some folk like about them is that their resistance has a fixed component and a much smaller voltage dependent component. The result of this (when used for anode loads) is some additional 2nd harmonic distortion introduced by that voltage dependence.
Their noise is current dependent so you don't want them in the high current paths which generally means you should not use them as cathode bias or anode loads.

All film resistors, carbon or metal film are manufactured by depositing the film on a "former" and then the resistance value is trimmed by cutting a spiral through the film. Arcing across this cut is what limits their voltage withstand.

I prefer to series connect 2 metal films of half the value to get the voltage withstand I need. Some folk (HiFI nutters in particular) do like to use carbon comps for grid stops, I never understood why???

I have used carbon comp recently - in the day job for a laser Q-Switch driver where I was generating -3600 volt pulses. I used 6 off 2 watters in series across the output so each was seeing 600 volt pulses. I would have needed 12 metal films to do that job.

Cheers,
Ian

[trobbins]

Ian - check out PRO2 range - they have excellent surge power and voltage specs - and don't look too out of place in vintage gear :-)

[loudthud]

Some folk (HiFI nutters in particular) do like to use carbon comps for grid stops, I never understood why???

Carbon comps have lower inductance and end to end capacitance.

[soundguruman]

Carbon comps have lower inductance and end to end capacitance.

Since we now have (metal film) resistors with:
A. much lower thermal noise
B. Non inductive construction
C. Welded on leads (not glued)
D. Very low cost (even ones from China work better than CC)
E. Superior thermal durability
F. Wide availability

I am at a loss to explain why builders still use carbon resistors "at all."
And now that I can buy bulk metal foil (aerospace grade) for $10 each, I am thinking that the advantages in noise reduction (for very high gain circuits) makes it just about worth it.

[Mick Bailey]

All of the carbon comps I have use a moulded-in lead that is very firmly attached.

If the arguement for low noise/efficiency was so compelling, a 5E3 built on a credit-card sized, through-plated PCB with metal film resistors and Mosfets instead of valves would outsell any eyelet board amp - it will be cheaper, lower noise, more reliable and more efficient, with pots and sockets mounted on the board. Kind of like Behringer building a Tweed Deluxe versus Victoria or Clark.......

I notice that the voltage rating for metal film types has reduced over the years due to the components getting smaller, which means that higher wattage (hence larger) resistors now have to used in certain areas of amp construction.

[Steve Conner]

Well, I bet solid-state amps built on PCBs do outsell tube amps many times over.

I have blown the 1/4 watt size metal film resistors with overvoltage before. I like to use several in series, because the voltage rating doesn't go up with size as much as I'd like.

[Enzo]

If the arguement for low noise/efficiency was so compelling, a 5E3 built on a credit-card sized, through-plated PCB with metal film resistors and Mosfets instead of valves would outsell any eyelet board amp - it will be cheaper, lower noise, more reliable and more efficient, with pots and sockets mounted on the board. Kind of like Behringer building a Tweed Deluxe versus Victoria or Clark.......

Nah. People don't buy a 5E3 because it is efficient ( or inefficient for that matter), nor do they buy them because they are particularly low in noise. They buy them because they want a 5E3. You can build that "better" amp, and it may sound just fine, but it won't kill the 5E3 market. If Outback Steak House opens a new restaurant next door to McDonalds, do you think McDonalds gets worried?

Those surface mount pc board amps sell rather well too. They sport names like Line 6, or as you said, Behringer. There is a market for that stuff TOO, Fender even makes them. But Fender continues to make the tube stuff for people who want tube stuff. And a new Deluxe costs a lot more than ther equivalent solid state thingie.

[Mick Bailey]

This is the point I was making - that low noise/efficiency aren't the primary considerations when it comes to carbon comps.

I still didn't figure out why the value on some creeps up a little when measured. Tried keeping them warm/dry for several days and still the same characteristics, so I'm thinking it's not moisture related. Looks like some experimentation is needed to establish the cause.

[R.G.]

... or carefully examine the datasheets. Carbon comp resistors have a temperature coefficient of resistance, a voltage coefficient of resistance, and moisture sensitivity. They are also history-sensitive, in that soldering, pulse currents, and simply aging change their values. The old IBM component guide had a whole chapter on what to watch for with carbon comp as regards drift and reliability. After being forced to read it several times, the message got through - carbon comp drifts with everything.