Converting Linear taper pots to audio/log type

[ssstonelover]

This may have been covered (in which case I did a poor search) but as it's thought provoking, I'll risk bringing it up again.

In the article "The Secret Life of Pots" [Copyright 1999 R.G. Keen All rights reserved], the author makes the case that audio/log pots have approximate tapers (due to manufacturing issues) so don't perform nearly as well as they should. 
Secret Live of Pots

(Personally I have certainly noticed/heard pretty radical taper jerkiness in real world application.)

He has a number of charts, which seem compelling, and then suggests a way out of the dilemma by using a linear taper pot but by modifying it, by adding a resistor of approximately 1/4 to 1/5 the measured pot value, between the 1 and 3 legs. According to his math this creates a smooth curve audio/log pot, and would work equally well if there is a need to create a reverse log/audio pot (which would be a godsend as they are almost impossible to find at all, good or bad taper). 

I'm not going to copy the article, charts and math, but it's not a long read. I'm sure some of you gurus have seen the article and can shed some light before I pick up a soldering iron.

Here is my visualization of his ideas, but now extended to include the tone pot and also including a treble bleed. I'd like to get some feedback on the premise, and the extension, before ploughing ahead, thanks!

[Yogi B]

Jul 4, 2021 18:53:54 GMT -5 ssstonelover said:

I'm sure some of you gurus have seen the article and can shed some light before I pick up a soldering iron.

TLDR: you won't be needing that iron.


Also from that article:

Unfortunately there's a gotcha in there ... neither the load seen by whatever drives V_in or the source resistance as seen by the input of whatever is connected to V_out is close to what would exist for a real log pot of value R.

In other words concentrating solely on a single regular volume pot (no treble bleed): with a resistor between the wiper and ground as per Keen's diagram to alter a linear pot to have a pseudo-logarithmic voltage division taper (whereas your tapering resistors seem to be incorrectly placed between both outer lugs) imagine what happens with the volume on full. The wiper shorts lugs 2 & 3 leaving the full value of the pot, 500k in your example, in parallel with the tapering resistor between hot and ground so around 80k to 100k loading the pickup much less than the 500k we'd normally get. That's only if we want a taper of about 25%, to approximate the 10% taper of a standard log pot the taper resistor would need to be in the region of 30k.

As we lower the volume the modified linear probably comes off relatively favourably, due to our old friend/enemy cable capacitance. The combination of resistance in series with the pickup (between lug 3 and the wiper) and cable capacitance forms a low pass filter. The larger the series resistance the more effective the capacitance is at reducing treble. With a standard 10% log pot at 50% rotation we have 450k resistance between the wiper and lug 3, whereas in the modified linear setup this will only be 250k. This is, however, a balancing act between reducing the series resistance and not increasing the load on the pickup (via the requirement that the other branch of the pot, between the wiper and ground, is a small enough resistance to give the required taper). All in all the overall difference isn't huge and probably isn't enough that you wouldn't feel the benefit of a treble bleed, so onto the next section...

Currently our imagined linear pot has a 100k resistor between the wiper & lug 1, if we also add a treble bleed (in the form of a 100k resistor & 1n cap) between the wiper & lug 3 — i.e. we now have two 100k resistors, one each between the wiper and the outer lugs. Therefore at 50% rotation (for most of bass & mid frequency range where the capacitor represents a large enough impedance that it can be ignored), both halves of the voltage divider are now equal: 250k of each half of the linear pot each in parallel with one of the 100k resistors, effectively undoing the action of adding the original tapering resistor. (In truth it's not truly undone, instead of returning to a perfect linear taper you'll get a mirrored "S" taper, where the output stays close to 50% for most of the sweep).

---

This also raises the obvious question: if you're after an approximation of a 20%-25% taper pot that includes a treble bleed, why are you starting with a linear pot? A standard 500k log pot with 220k tapering resistor as part of the treble bleed gets exactly what you're looking for.

i.e. Keen's second example of making a linear pot behave closer to a reverse log (by connecting a tapering resistor between the wiper & lug 3) applies similarly to any taper, not just linear. Therefore adding this type of tapering resistor to a 10% log taper moves the taper closer to a linear taper (and beyond if the resistor is sufficiently small compared to the value of the pot).

---

As for a tone pot, so far we've been thinking about taper in terms of a voltage divider — however a tone control isn't that, it's a variable resistance using only using two lugs. This is the case that Keen covers in the final graph of the article, in short: a variable resistor can only ever be modified to behave in a more reverse logarithmic manner — pretty much the opposite of what you'd ever want.

My way of rationalising this (without maths) goes something like the following: for an additional parallel resistor to have a significant effect on the total parallel resistance it's value must be small enough to be around or lower than the current resistance, this is a double edged sword. Thus if we want a significant effect on the taper of a variable resistor the tapering resistor much be fairly small, however this necessarily means that for most of the range of rotation, the innate resistance of the variable resistor is significantly higher than that of the tapering resistor and therefore has little effect on the combined parallel resistance. The result is for the majority of the sweep the combined resistance is close to that of the tapering resistor, until sharply decreasing towards zero at the end of the sweep when the variable resistance becomes the lesser of the two.

---

Overall, tapering resistors have their uses in passive guitar wiring mostly as part of a treble bleed, but they're generally much more useful in the context of active electronics where often the differences in loading can be compensated for by adjusting the values of other components of the circuit or 'hidden' behind buffering.

[JohnH]

I've seen these tricks in the circuits of channel fade pots on audio mixers. A linear pot with a resistor from the wiper to ground. They do work well and nice and smoothly, to create a fairly log-like net taper.

But, it has to be in a circuit where we are not worried about the overall impedance, such as within the active circuitry of a mixer. Probably not much good in the delicate passive circuitry of a guitar, unless we can benefit from that effect such as with a treble-bleed circuit.

[ssstonelover]

Thank you Yogi
Obviously I had a critical error in resistor location (duh, should be wiper to ground lug), the then obvious (and negative) interaction with a treble bleed, the difference between a voltage divider (volume pot) and variable resistance (tone pot), how this resistor would be more appropriate for an active circuit where the negative loading effects can be counteracted with additional components, etc...
At the end of the day, the fact it has not popped up on wiring of the sort we do is the tip off and your words and John's too cement this as reasons underpinning this fact.

I will leave my solder iron for useful work LOL.

In passing, and what spurred the initial interest in the subject, is there a pot you know of (I'm thinking especially tone pot here) that has a good useful taper for subtly adding the capacitor into the circuit? I always seem to go from nice and trebly straight over to overly dark without good useful shading in between. Mainly I've been using the Alpha A250K (on SSS setups) which in theory should work OK, but I've seen this on other pots too.

[ssstonelover]

Thank you John.
It's a non-starter as you and Yogi have explained well.
I appreciate your help avoiding this trap.

[Yogi B]

Jul 5, 2021 10:58:35 GMT -5 ssstonelover said:

In passing, and what spurred the initial interest in the subject, is there a pot you know of (I'm thinking especially tone pot here) that has a good useful taper for subtly adding the capacitor into the circuit? I always seem to go from nice and trebly straight over to overly dark without good useful shading in between. Mainly I've been using the Alpha A250K (on SSS setups) which in theory should work OK, but I've seen this on other pots too.

This is why, in general, I now prefer using 250k volume & 250k no-load tone control, where previously I'd've used two 500k pots.

If you're finding that the tone rolled all the way down is always too dark, you could add a smallish (5k-10k) resistor (or trimmer) in series with the tone cap (similar to the Fender Greasebucket), which will limit the range of the pot to what is normally 10 to about 3 with a regular tone pot, thereby giving more precise control over that reduced range. And/or you could try a smaller value tone capacitor.

If you're fine with the taper when at full volume, but are finding that the tone pot taper becomes less usable at lower volumes when the volume in question is paired with a parallel treble bleed — that's an unfortunate side effect of the treble bleed, but one that can be fairly effectively countered. The fix is to route the treble bleed capacitor through usually unused half of the tone pot, similar to the wiring of the bright cap on tweed amps. Due to the four alternative ways to wire a standard tone control, that description may not be obvious, so here's a schematic:

Note: This will not work with a (single-gang) no-load pot, and I'd recommend using 250k pot. Or if you're willing to go dual-gang and do some pot surgery, I'd use: a log 250k pot (possibly no-load) in series with the tone cap; and a linear 500k pot to blend out the treble bleed cap.

(There's also a different dual-gang concept I thought about a while back, A Dual-Gang Tone Control: Putting Cable Capacitance to Work, which offers an improved taper as well as slightly different sweep of tones.)

If all you really want is a steeper taper, then the first thing to note is that sometimes you'll get people selling pots originating from Asia as "A" taper meaning a 15% taper rather than the usual 10%, as that's the naming scheme used by the Japanese Industrial Standards (this is the reason you'll come across "D" taper pots in, for example, Ibanez guitars — "D" taper is the JIS name for a 10% taper). This being the case I'd check the actual taper A250k pots you have with a meter. If they are in fact 15%, then switching to 10% might be enough of an improvement. If not, (Alan) Entwistle has "developed" Smoothtrak pots which, as hinted at within elaborate marketing gubbins, are just 5% taper (JIS's G taper) pots.

(Alpha's website has a taper curves page handily gives the alternative JIS taper names.)

[ssstonelover]

Hi Yogi,

Just saw your reply, sorry about that. I've been busy (a good thing mostly but ....), I'm sorry as I unintentionally missed seeing what was beginning to be some really good education (which I sorely needed regarding "taming the tone control"). I want to keep going on that please! My silence was purely accidental.

You mention a number of approaches based on the symptoms. Actually I try to use tone while the volume is at "10" but certainly it would be best if it works smoothly an any volume setting (so closer to what I would call your second scheme/version). Many of my circuits don't have treble bleed.

Accordingly I have translated the electrical schematic to a diagram for your checking (as that is more my speed) though I certainly need to be more conversant on EE language, so please see how I did on my translation...



Also I checked your other helpful links, suggestions, the Alpha site, and contacted Entwistle. I think his pots may work... it would be nice if there were some reviews out there, but his marketing has a siren's call to me so I might take the plunge.
The other idea that occurs to me (assuming most of the tone shading occurs between, say 5k-50k, is to make a no-load 50k linear pot, so as to jump into a long stretch of low resistance where tone shading resides. Does that make sense or do I have it wrong?

Again sorry to have given the impression I was not responsive, as on the contrary I most certainly am.

[Yogi B]

Jul 9, 2021 19:34:55 GMT -5 ssstonelover said:

Accordingly I have translated the electrical schematic to a diagram for your checking (as that is more my speed) though I certainly need to be more conversant on EE language, so please see how I did on my translation...

The first version is fine. On the second the treble bleed cap is correct, but you want the resistor between lugs 2 & 3 of the volume pot (the same as the first version) & the blue linking wire from volume lug 3 to the centre lug of the tone pot (again the same as the first version).

---

Another way to visualise this is, by referring to your drawing of the first version (ignoring the 5k to 10k GreaseBucket resitor) as the typical wiring, we want everything the same except to add the resistance of the unused half of the tone pot in series with the treble bleed capacitor.

Per the typical wiring the treble been capacitor is connected between lugs 2 & 3 of the volume pot, so either we need to introduce this extra resistance between lug 2 & the cap, or between lug 3 & the cap. This means that, of the two lugs connected to the unused tone pot resistance (the tone pot's lug 2 & 3), one must be connected to either of the volume pot's lug 2 or 3.

With the typical wiring it so happens that the tone pot's lug 2 is already connected the the volume pot's lug 3 via the blue wire. Therefore all we need to do is disconnect the lead of the treble bleed capacitor currently connected to the volume's lug 3 (and indirectly the tone's lug 2), and reconnect it to the tone pot's lug 3.

---

So the only component which needed rewiring, you did so correctly, but unfortunately you didn't stop there.

Entwistle. I think his pots may work... it would be nice if there were some reviews out there, but his marketing has a siren's call to me so I might take the plunge.

It's annoying that they're seemingly the only 5% taper pots that are even somewhat available in non-industrial quantities — Alpha, CTS & Bourns all theoretically offer 5% tapers, but I don't want a thousand of the things thanks.

The only vague experience I have with Entwistle is with a set of XS65 (single coil) pickups (more or less identical to Artec's SBBN pickups), which I got after someone found out the hard way that although the pole pieces are slotted they are NOT adjustable.

I haven't gotten around to trying them in anything yet, and my current plan to 'fix' them involves using regular humbucker poles instead, that or just giving up and using AlNiCo rods.

The other idea that occurs to me (assuming most of the tone shading occurs between, say 5k-50k, is to make a no-load 50k linear pot, so as to jump into a long stretch of low resistance where tone shading resides. Does that make sense or do I have it wrong?

That should work, but I'd worry that there'd be too much of a jump — it'd roughly be equivalent to jumping from 10 to 6/6.5 on a 250k pot (or from 10 to 5 on 500k).

Again sorry to have given the impression I was not responsive, as on the contrary I most certainly am.

No need to apologise. You, or anyone else, will reply as an when they get around to it — even if that's never, that's fine. And to be fair I, myself, am not always the best at responsiveness.

[ssstonelover]

Thanks Yogi for all the help and advice and here is the revised drawing:


Entwistle: I ordered some of his "Smoothtrak" 5% taper pots (mostly AE250K but also 1 x AE500K). They must have a very sharp initial ramp down from 250K (or 500K) to achieve 5% over the rest of the taper but the way I am thinking about this is that this may be the best compromise compared to all the alternatives...I'll try to give feedback down the road....

Interestingly in his pdf single coil wiring he echoes the 5K resistor grease circuit (at least for neck/mid tone pot) and also mentions 100K pots (as an taper pot alternative). He obviously has not updated his published material on this link to include his Smoothtrak....LOL

As far as these being the only 5% taper around.... yeah.... or you try to fake it with 100K linear pots and a 5K resistor possibly and live with some reduction is highs at the "10" setting or do some kind of scraping or tone disconnect switch for more of a no-load feel.

The other thing is the pot has a 18 tooth split shaft 6mm type which does not bother me, but may be an issue if you try to combine it with a 24 tooth 6.35mm spline split shaft "vintage taper" volume pot which seems to only come that way from Mojo Tone, etc. (made by CTS?) and want to precisely match the knobs (a small issue really).

This leads to the next obvious question, the ideal volume taper.... I'm curious to know if you have any favorites or if it's not an issue as far as you're concerned, and most any 'log/audio taper' will do for the volume.

[Yogi B]

Jul 14, 2021 14:00:46 GMT -5 ssstonelover said:

This leads to the next obvious question, the ideal volume taper.... I'm curious to know if you have any favorites or if it's not an issue as far as you're concerned, and most any 'log/audio taper' will do for the volume.

I prefer around a 20%/25% taper, whether that's actually the taper of the pot itself or a more commonly available taper paired with the appropriate treble bleed. (That is: a 10% 500k pot with ~390k resistor in parallel with 330p cap; or a 15% 250k pot with ~680k resistor in parallel with 330p cap.)

Edit: corrected the above values, that serves me right for writing them from memory.

Though really I don't mind all that much, I'm fine with anything from linear to 10% log. I find the question of "what is the best taper?" to be more situational for volume than with tone.

As an actual volume control when playing through something pretty clean I favour linear. However when plugging into something with a lot more gain/compression, where the volume is more of a saturation control, I favour logarithmic taper as it cleans up much quicker and thereby gives a larger rotational range to dial in a clean-ish setting.