A few questions about Carbon Comp (R)

[dumbassbob]

When a CC resistor has a 10% does that mean it can drift from 100K up to 110K and be OK? and is it the same for (R) with 5% does that mean if it is a 100K it can only go up to 105K I bought 20 100K@ 5% and only one was at 100K all the rest went from 98K up to 115...So I found some 10% silver bands....Also what is Drift? how high can you go before it is "bad" I heard of EVH having a 820 that drifted to 1.1K...is it another brown sound myth...

[überfuzz]

Yes to your first query regarding the 10%. I believe the technical term appropriate to more than 10% is, 'quite a drift'. Seriously though, as long as you don't violate the watt-tolerance of resistors it's more a matter of taste than anything else (MHO).

[Tom Phillips]

The tolerance specification only tells you how close the actual resistance was to the stated value when the resistor was new. The manufacturer measured and sorted the resistors after they were made and then marked them as 5%, 10% or 20% parts. In the case of the 100k resistors, the ones that measures between 95k and 105k were marked as 5% tolerance parts. Then the 10% parts were marked etc. This process would leave a hole in the middle of the 10% batch because the 5% parts had already been taken out of the mix.

The tolerance rating does not tell you how the parts will age after they are sold. CC resistors tend to increase in resistance as they age. Depending on the storage conditions a 5% part could end up at a higher resistance after it has aged than a 10% part. They just start the aging process at a different resistance.

[Chuck H]

So, is that to say that 10% tolerance components sold analogous to equivalent 5% components will always be at least 6% off spec? I've never checked for that but it seems likely.

[Tom Phillips]

So, is that to say that 10% tolerance components sold analogous to equivalent 5% components will always be at least 6% off spec? I've never checked for that but it seems likely.

That's the lore and I think it makes sense too. However, since my stock of original Allen Bradly CC resistors is so old and already showing the effects of aging that I don't think I'd be able to make statistical measurements. Might take a look and try though.

[Leo_Gnardo]

So, is that to say that 10% tolerance components sold analogous to equivalent 5% components will always be at least 6% off spec? I've never checked for that but it seems likely.

Maybe "somebody" did it that way but I have batches of 5% and 10% cc (100K for instance) that don't correspond. Gotta batch of 5% from Mouser some years ago, when they claimed to be from one of the legacy brands, Allen Bradley IIRC. Most of them arrived already beyond 105K and the rest have drifted up to around 110K in storage. And I got a batch of 10% AB's in old hundred-paks probably 60's vintage that mostly read dead on the money 100K plus or minus 2%. Neither seems noisy in use as a plate R in Fender & similar amps.

Drift - recently I had an old veteran in, mid 60's BF Fender & it's owner said he loved the tone in the vibrato channel. The first stage plate R supposed to be 100K had drifted up to 180K, wow! Well at least it wasn't making noise so it stayed right where it was.

When you mod or build kit or from scratch you can audition "alternate value" parts & let your ears decide what sounds good for you. A cathode R that drifted up to 1K1 from 820R is no big deal. Deserves a replacement all right to keep things "factory" but like that 180K above, may make for a tiny change in tone you may or may not notice. If "not standard" sounds better, go there, no biggie.

[Tony Bones]

The manufacturer measured and sorted the resistors after they were made and then marked them as 5%, 10% or 20% parts. In the case of the 100k resistors, the ones that measures between 95k and 105k were marked as 5% tolerance parts. Then the 10% parts were marked etc. This process would leave a hole in the middle of the 10% batch because the 5% parts had already been taken out of the mix.

If the manufacturer was makng 5% parts then they were also selling 91k and 110k values. Where did they get those?

Take a look at the standard values for different tolerances. For 10% they are 1, 1.2, 1.5, 2.2, 2.7, 3.3, 3.9, 4.7, 5.6, 6.8, 8.2, and any multiple of 10 (10, 100, 1k, 10k, ...) Notice that the values are spaced about 20% apart; 2.2 + 10% = 2.42 while 2.7 - 10% = 2.43. When making 10% resistors there are no holes. Every part, no matter what the value, falls into some 10% bin.

For 5% parts there are twice as many standard values. So the same pile of potentially random value resistors can be sorted into 10% or 5%, or 20% or 1%. The standard values for each tolerance are chosen that way.