Potentiometer matching gig

[Silvas]

Hello !

Has someone attempted making a test jig for matching tapers on potentiometers ? Some pots can be very off from each other but i can imagine myself matching pots for a dual preamp based on its travel vs amplitude graph....for example.

What do you think ? Good idea ? Bad idea ?

 

[warpie]

Is it not just a matter of measuring the total resistance of each pot?

 

[Rochey]

No - they may not have a matching curves.

For instance, on a stereo pot, there may be up to 6dB difference at 11 o'clock.

 

[warpie]

ah OK, thanks.

EDIT: But the usual 10 or 20% tolerance refers to the total resistance, right?

 

[Rochey]

Far as I know, there's the tolerance, and then there's the channel to channel matching. two different metrics.

 

[gyraf]

..stepper motor, reference pot and test pot readout.

Not easy, but absolutely doable in e.g. a simple Arduino setup..

Jakob E.

 

[Rochey]

Gyraf, that was my first thought too...

But your not interested in the resistance at 11 oclock, then at 12 o'clock.
Your interested in the matching between left and right gangs on the pot.

Therefor, isn't it simply a case of comparing channel A to channel B constantly, and making sure your not out by a certain %/dB?

 

[JohnRoberts]

Hello !

Has someone attempted making a test jig for matching tapers on potentiometers ? Some pots can be very off from each other but i can imagine myself matching pots for a dual preamp based on its travel vs amplitude graph....for example.

What do you think ? Good idea ? Bad idea ?

Many years ago, when I was selling an analog 24 dB/oct loudspeaker crossover (LOFT), the frequency pots were the best 4 gang Alps pots I could source, but I was still seeing more deviation between the gangs, than I considered acceptable.  I used a circuit topology where the pot was out of the circuit when full up and full down, then added a resistor shunt trim from wiper to either endpoint to make them match at 50% rotation.

This is yet another reason why modern DSP is more accurate...  If you really need precision use a multi-position switch.

JR

 

[Rob Flinn]

Even some of the more expensive pots have a 20% tolerance spec between gangs.............

 

[JohnRoberts]

Even some of the more expensive pots have a 20% tolerance spec between gangs.............

Bulk resistance spec is typically 20%, tapers can be more accurate, so when possible use them potentiometrically (variable ratio) not as rheostats (variable resistance).

Variation for screened ink manufacturing process is related to ink formulations and how the resistance changes while final cured in an oven. This process is better or worse managed by different vendors. I had one pot maker quit us (Peavey) declining our business saying that we were too picky. It turns out their factory was over cooking the pots to make them meet resistance specs but the overcooking weakened the substrate so we were getting too much breakage just from normal handling in production.

JR

 

[Rochey]

[I had one pot maker quit us (Peavey) declining our business saying that we were too picky.

Wow, I'll bet that REALLY pissed off the purchasing team!

 

[JohnRoberts]

[I had one pot maker quit us (Peavey) declining our business saying that we were too picky.

Wow, I'll bet that REALLY pissed off the purchasing team!

As I recall, it took us all by surprise. I was working with the vendor, trying to help them clean up their process, which in my judgement was too sloppy regarding green resistance of the screened resistance elements, so they were damaging the parts trying to fix it after that step by over cooking them... Their initial response was that we were mishandling the parts, but I could pretty easily find a range of parts from them discolored by the over heating, with all the breakage in the obviously "cooked" parts.

Their decision to walk away from a pretty large account (there were multiple part numbers affected by this decision) was a small hassle, but there were other vendors happy to step up and make those parts. Purchasing was probably happy to have the other vendors kissing their a** for the new business...  I have to admit it was odd, to have a vendor quit us, and for that reason... usually it was becuse we were hammering them over price. ;D  Price may have factored into the decision to walk away from that part, but it wasn't because it couldn't be made profitably, just that they couldn't make it profitably (As I recall it was a US company trying to get a Taiwan factory up to speed and running smoothly. Never an easy task, while Taiwan was easier than mainland China. ) 

Working for a large scale manufacturer caused me to learn way too much about several vendor's businesses.

JR

 

[RuudNL]

For stereo pots I could think of two solutions:

- Measure the difference between the two sliders (kind of 'Wheatstone bridge' principle).
- Connect both resistive tracks to the same input signal (tone generator) and observe the output signals from the wipers on a (dual needle) millivoltmeter. In this way you can get a quick impression of the tracking.

For single (mono) pots the prodedure would be much more difficult.

 

[PRR]

To my small mind: the quick/best way would be to run mono music through both pot gangs to a stereo headphone amp and headphones.

First trim for "center of head" effect when full-up (or pot bypassed).

Then as you turn-down, the mono music should stay center-of-head down to quite low levels.

This won't give exact-number error results. However some pots will be rock-solid center-of-head, and others will wander left or right. Three piles: solid, less-solid, and jumpy.

The mechanics of potentiometer attenuation suggests the worst-case no-taper is around -6dB, so a fade to -20dB is plenty. If you start loud, it's still clearly audible at -20dB.

"Taper" pots are generally two tracks end to end. Spliced at -15dB or -20dB. You could have a worst-case at -5dB and others at -15dB and -26dB. You might want a 20dB boost so you can skim the 0-5 half of the rotation.

 

[keefaz]

There is the baxandall volume control by Douglas Self, maybe better tracking with linear potentiometers

http://www.groupdiy.com/index.php?topic=46637.0

 

[abbey road d enfer]

Hello !

Has someone attempted making a test jig for matching tapers on potentiometers ? Some pots can be very off from each other but i can imagine myself matching pots for a dual preamp based on its travel vs amplitude graph....for example.

What do you think ? Good idea ? Bad idea ?

As JR suggests, there are some tricks that can be used to improve a situation.
I have been faced with this problem with parametric equalizers, where reverse-log pots are traditionally used. I think I use the same technique as JR, using lin pots, which track much better to start with, and define the top and bottom range with fixed resistors and adjust the mid-rotation with a trim pot. This is not a new technique; that's derived from the technique used to calibrate superhet radios.
This technique can be used for volume control, for frequency control of state-variable filters, but is not really applicable to Baxendall-type tone controls, or rheostat-type applications.

 

[Silvas]

Thanks for all the info !

I think the stepped motor and readout is indeed a great idea....now, time to study picbasic again....

 

[Rob Flinn]

One option is you want really good tracking might be to get a single pot to control a pair of vca's, since the vca's can be made to track together very well .................

 

[sodderboy]

I use 1k into both stacks, out to A-B inputs on a scope. Use the flattest line pots. Use X-Y if no A-B on the scope and go for the least movement across the rotation.

Mike

 

[Nele]

easy and cheap:

- hookup each pot to an analog input of an arduino
- read and compare the numbers the analog inputs read in your serial monitor as you turn or slide the pots. even better: let the arduino program compare the numbers and print the difference. you can even cook up a simple processing sketch which will plot the taper..