Demystification of ALPS Pots

[Pano]

However, unless there is a centre detent the resetting accuracy makes all this rather academic.

I was thinking the same thing, but apparently it does have a detent.

I've used a number of stepped attenuators bought from eBay. Even had some custom values made in 3 deck models. They worked just fine and were not expensive. I like the little SMD types. But you might not find a good neutral position detent, tho.

 

[JMFahey]

Got this measurement from a tone control unit built with ALPS pots. The pots are in flat position. Never trust anybody! Spec says +-20%. As bad as any other pot. Sad...

You have no clue reading a dB scale, sorry

0.5dB error means 6% voltage ratio tolerance, meaning resistance ratio is 6% off , WAY within 20% rated tolerance.

1dB error means 12% error, still WAY within tolerance.

Sad indeed, but that applies to your Electronics knowledge.

OF COURSE, 1% resistors will be within 1% tolerance , not exactly headline news

 

[JMFahey]

10k pot has 5.017k and 5.309k in center position. Thats the cause.

Your Math skills are even WORSE, my teeth grind at your appalling Math (or plain lack of it) :

5.309 - 5.017 = 0.292 kohm

0.292k/10k (full scale value) = 0.0292 (absolute value) = 2.92% error
WAY WAY within rated 20% tolerance.

 

[diyralf]

Your Math skills are even WORSE, my teeth grind at your appalling Math (or plain lack of it) :

5.309 - 5.017 = 0.292 kohm

0.292k/10k (full scale value) = 0.0292 (absolute value) = 2.92% error
WAY WAY within rated 20% tolerance.

Did I say the opposite?

 

[DPH]

Do you need the full range of the pot? If not, fixing it between some fixed resistors (and scaling appropriately) will pull the difference down dramatically.

And you'd be loathe to know what your inter-speaker tolerance is....

 

[DF96]

+-6dB is not a tone control but a bad cooper wire.

A wire would have to be exceedingly bad (or exceedingly long) in order to achieve 6dB tone control. I assume "cooper wire" is not an oblique reference to superconductors?

Do you need the full range of the pot? If not, fixing it between some fixed resistors (and scaling appropriately) will pull the difference down dramatically.

I thought that. Then I thought some more and realised that the difference would not be dramatic. I said 6dB range limits would only halve the error. Now I've thought some more: adding a 5k resistor at each end of the pot (and adjusting cap values appropriately) would halve the error, but actually give about 10dB range limits. A 10k resistor at each end would give 6dB range limits, and give one-third of the error. Maybe the OP has such strange speakers, or strange room, or strange tastes, that he needs a lot of range on his tone controls?

 

[SamAnytime]

If precise matching is really important to you, go digital. A DSP based attenuator instead of a ganged pot. You might even implement the entire tone control digitally. Texas Instruments' PurePath Studio is a good place to start.

 

[Regenpak]

I happen to have three Alps RK27112-type dual ganged linear pots with center detent. They are marked "510G 10KB x 2". Lacking proper table tools here are the results ("F" unit closest to the shaft, all values in kOhm):

UNIT 1-2 2-3
1F 4.93 4.92
1R 4.92 4.91
2F 4.84 5.07
2R 4.72 4.85
3F 4.70 4.89
3R 4.92 4.75

The third unit is the worst because it is reversed, the first unit is all but exactly identical. Still, the third unit differs only by 7.5% and that's absolute. Pretty good results and resulting in a difference that's absolutely inaudible, even when compared with the TONE DEFEAT position.