Post by Yogi B on Mar 13, 2018 1:59:04 GMT -5
THIS IS A DRAFT
The reason I suddenly jumped on this was I was thinking about applying this sort of logic to model the blending of pickups with volume controls.
Introduction & Typical LP Setups
To enable easier comparison between the various blending methods I'm keeping things as simple and generic as possible by using the simple three component model of single coil pickups, and choosing to ignore any differences in output due to mechanical variations (pickup placement, tuning, picking position, etc.). Unless otherwise specified the pickup models are based around single coils, a voltage source with series resistance and inductance of 5kΩ and 2H, respectively and a parallel capacitance of 150pF; the potentiometers used have an idealised 10% log taper and total resistance of 250kΩ, with the tone controls positioned before volumes (as per modern wiring); treble bleed components, when included, will be based around a parallel combination of a 1nF capacitor and a 100kΩ resistor (this value being around 2/5 the of volume pot, so adjust accordingly). In addition external loads of 1MΩ and 500pF are present representing a relatively high impedance input and capacitance of a average 3m (10ft) cable, respectively.
In the following sections there'll usually be three plots -- the first showing the output from the pickup being blended in/out; the second showing the output from the other (fully on) pickup, showing how this changes due to the change in loading; the final plot shows the summed output of both pickups. The plotted steps correspond to reducing the blend of one pickup from 10 down to 0 (as on a Gibson style control knob -- i.e. 30° increments, 1/10 of a pot's full rotation), in spectral order: red → blue corresponding to 10 → 6; purple corresponding to 5; and dark red → dark blue corresponding to 4 → 0.
Now that's out of the way, here's a look at a setup similar to a typical LP (wired with dependent volume controls, with the toggle switch in the middle position) while turning down the volume control for one of the pickups:
There are two things here of particular note, first is a general issue with blending pickups. Notice that when both pickups are at 'full volume' each produces only half (-6dB) the output they would each supply in isolation, this is due to the formation of a voltage divider from the impedances of both pickups: the exact figure of a half is because the pickups are assumed to be identical, thus the load impedance (and therefore the output voltage) is exactly half the sum of the impedance of both pickups. A further consequence of this is that lowering one of the volumes only slightly causes the output of the 'full volume' pickup to jump up to near its full output, this is because we're utilising logarithmic potentiometers -- that slight adjustment from '10' down to '9' has added a large amount of resistance (close to half the value of the pot) in series to the load impedance of the voltage divider we noted above, and therefore the ratio of ZLoad / (ZSource + ZLoad) (and by extension VOUT / VIN) is much closer to 1. If you have used this wiring arrangement in the past, then this is most likely the cause of any difficulty blending pickups you have probably experienced.
The second thing of note, which probably isn't immediately evident when listening to this circuit is that the drop in output from the turned down pickup is far more adverse with the bass frequencies than with the treble frequencies, this is due to the filtering affect the other pickup applies which, being predominately an inductor to ground, forms a high pass filter. Overall this is opposite to what is perceived because it's greatly overshadowed by the output from the other pickup which itself is gradually losing treble content due to the lowering cutoff frequency of the LPF formed by its inductance and the lowering resistance to ground.
Upon reading the above one's first thought maybe along the lines of: "Okay, so we're losing treble when turning down a volume control -- I know how to fix this -- we need a treble bleed!", well let's see. The below images demonstrate the inclusion of treble bleed components (to both volume controls), but are otherwise identical to the wiring present above.
Not much help, and if anything we've made things worse. This is because we're only turning down the one volume control, and it's the wrong one -- as we noted after the previous diagrams: it's not the treble output from the pickup we're turning down which is the issue, it's the overall lower resistance to ground affecting the output of the other pickup.
Unsurprisingly the aforementioned issues are improved by changing the volumes to be wired in an independent fashion, doing so produces the following:
This begs the question question: why aren't more guitars wired this way? Well lets look at how an independently wired volume control works on it's own as, you know, a volume control:
Not so good, the control is quite limited, managing only about 14dB of attenuation with the control at '1', before dropping off steeply toward no output at '0'. That might be enough for you if you only play though a super clean amp, but as you increase the gain (or rather, compression) applied to your guitars signal the volume control will start to act more and more like an on-off switch, making it difficult to clean up a distorted tone by rolling the volume down.
Once more we also have heavy treble loss, even when turning down the volume only slightly. We could attempt to combat this with the addition of treble bleed components, here I've added the same values as before (1nF capacitor in parallel with 100kΩ resistor) -- as we can see this does give some minor improvement:
It does, however, also come with a fairly major caveat, let's look at what happens when we go back to blending two pickups:
Shucks! Our independent controls are no longer as independent -- as we reduce the volume control it begins to act more like a '50s wired tone control, albeit with a cut-off frequency that's several hundred Hertz higher than normal.
At this point you should have come the the realisation that neither setup is really ideal for an LP setup where the volume controls are used to control individual volume and parallel blending. What's best for you is clearly a matter of personal opinion but for me the winner is dependently wired volumes.
Jazz Bass Setup
Another option to consider is the wring used in '60-'61 Fender Jazz basses -- with the stacked volume and tone controls. (Note: the more modern volume-volume-tone arrangement is essentially the same as the independent volumes section above.) The original JB wiring uses dependently wired volumes, however it also includes separating resistors between the output terminal of each volume control and the output jack. Here's a plot of the original values specced by Fender, i.e. 250kΩ pots with 220kΩ separating resistors:
As you can see this gives good separation but at the cost of loss of treble, again due to the added resistance lowering the cut-off frequency of the low pass filter formed with the cable capacitance. This can be improved by lowering the value of the separating resistors, but that obviously decreases the separation between the volume controls. Therefore if opting for the inclusion of separating resistors, the resistor value chosen has to balance these two factors in accordance with personal taste. For comparison here's the same thing but with 22kΩ separating resistors:
Blending In Parallel With An Always-On Pickup
A common arrangement where this situation can be found is on Strats and similar guitars whereby one of the tone controls is replaced by a blend control which blends in the neck pickup at positions 1 & 2 and the bridge pickup at positions 4 & 5.
Doing this usually employs wiring the neck tone control as a master tone then wiring the lower half (terminals 1 & 2) of what was the middle tone pot in between the neck and bridge output wires. Thinking about this you' should realise that this means that the blend control works backwards compared to the LP middle situation, i.e. turning down this control blends in the connected pickup. I suspect this is done for to main reasons: firstly it means that with all the controls on 10 the guitar will function like a normally wired Strat; secondly it enables reuse of the existing tone pot, which ordinarily should be a log taper pot. Thus in order to reverse the action of the blend control, you'd need to use a reverse log taper -- note that this is the opposite of what is used in the LP middle position, a regular log taper, that accounts for the poor blending ability we noted earlier.
Anyway here's a plot of this wiring method with the relevant components. To be consistent with the other plots I've flipped the direction of the control as previously discussed.
You'll note the much more consistent volume reduction of the blended pickup than when we were blending with a log pot, however there is still the issue that the affect on the bass frequencies is far more than that on the treble frequencies. To help remedy that will require upgrading our blend control to a dual gang pot, the first gang being the reverse log taper, the second being a regular log taper: the first gang wired as above; the second is to be wired as a variable resistance to ground from the blended pickup, to achieve this we can utilise the now unused half of the 5-way switch (which was previously used to switch the separate tone controls).
With the introduction of a dual gang pot, we have a new kink: both gangs loading the signal, instead of just one. There's two obvious ways to combat this, there's not a great deal of variation between the two but I figure I may as well include plots of both. Firstly maintaining 250kΩ pots we can simply leave the reverse log portion ungrounded, here's the result of that:
Alternatively we could leave both gangs grounded, but double up the potentiometers' resistances up to 500kΩ, this gives the following:
Blending In Parallel Using A Blend Pot
Since we know blending two pickups with individual volume controls is tricky another option is to use a dual-gang blend pot and a master volume. This way we have both pickups at full (or nearly) full volume at the centre position (often marked with a detent in pots designed for this purpose), turning one way to mute one pickup, the other to mute the other. This makes the control symmetrical about the midway point, and will make the results equal -- if, as I am, you're considering not considering the affect on tone due to pickup position and the electrical properties of pickups to be equal. As such the results I'm showing in this section begin at the centre point and step in twelfths.
Such blend pots typically come in two flavours M & N taper and A & C taper, though retailers are often poor at describing what they are offering.
A & C tapers are the standard logarithmic and reverse logarithmic tapers, though in truth it's really two A tapers one of which (typically the bottom gang) is mounted upside down, reversing the terminal ordering and relative wiper position -- thus making it equivalent to a reverse log taper. Retailers describe this type of blend pot as log pot, or note that at the centre position it doesn't produce the full output but rather 70-80%.A more in depth look by ChrisK at the taper of both types of blend pot can be found here. The terminology used therein equates as follows: a (true) blend pot is an M & N taper pot; a pan pot is an A & C taper pot.
M & N tapers consist of a conductive strip for half the the pots rotation and a linear resistive element for the remaining half, and again the bottom gang is made a mirror of the top by orientating it upside down. Retailers often describe this type of blend pot as a linear pot, or note that it produces 100% of the output at the centre position. However beware that dual-gang pots with the regular linear taper -- that spans the whole element, not half -- also exist and may also be called a linear blend pot.
First we'll look at the M/N case, here both pickups are at full volume at the centre position and turning either way decreases the volume linearly. Thus this works in the same way as the independent controls we examined before, but being a linear taper the volume rolls off at a rate somewhere between the log and reverse log versions we've seen before.
Next I'll cover the other common case, the A/C blend pot. As the pot uses the standard log tapers, at the centre position the resistance between the input and output terminals of both gangs will be at around 10-15% of their maximum value, depending on the exact ratio of the taper. This means you'll lose a bit of treble and a couple of decibels of output at the centre position -- unlike the MN blend pot.
A very modest improvement to the losses in the centre position -- which is caused by the additional resistance in series with each pickup forming the upper part of a potential divider -- can be had
by increasing the impedance of the lower part of the potential divider. This could, for example, be done by upping the value of the volume control to 500kΩ and using a no-load tone control.
Another potential improvement applicable to both M/N and A/C designs is to leave the normally grounded terminals of the control ungrounded, reducing the total load on the pickups. This would mean a pickup could never be fully off, but that can be remedied by making the pot go open at the relevant end of each conductive element. (Coincidentally that's the same end of the taper as a regular no-load tone pot.) Thus below are plots of an improved A/C case with the modded blend pot, 500kΩ volume and no-load tone. Finally to compensate for the minimum loading this entails, which may make the result sound overly shrill, a resistor can be placed directly in parallel with each pickup (i.e. before the blend control) to tame resonant peaks the pickups, here I've used 470kΩ.
Again neither of these options are perfect: the A/C variety gives better control, but has the treble/volume loss issue; the M/N variety avoids that but has most of it's action occur at the centre, which -- since that's also the point where the centre detent (if present) limits the ability to make minute adjustments -- is especially unfortunate. To combat these issues woodstockwizzard came up with the following solution using a tapped blend pot:
And plotting that out gives this:
Although never specified I assume WoodstockWizzard intended the use of tapped A & C taper elements for this design, as it is what gives the best results -- in fact potentially the best results of this section. There is, however, the question of actually sourcing such a pot, as tapped pots aren't the most common thing in the world, especially in guitar compatible formats & values. But if failing to actual get your hands on one, you could theoretically hack together such a thing with some advanced pot surgery.
Blending In Series
The following roughly correlate to the options presented in JohnH 's Blending coils in series.
The first option uses a potentiometer simply as a variable resistance in parallel with the blended coil, this is commonly used for spin-a-split (A.K.A. variable coil split) wring with humbuckers. As with the Strat blending before there's two symmetrical ways to wire such a control: 'forwards' with a log pot, coil fully bypassed with the knob turned to 0; 'backwards' with a reverse log pot, coil fully bypassed with the knob turned to 10. Here we'll be looking at the former:
The second variation uses the potentiometer wired as a potential divider, like a regular (dependent) volume control, let's look at that, again with a log taper pot:
We can see here that the output of the blended pickup reduces at a fairly consistent 4dB per 10th of rotation, similar to how the same pot would operate as a volume control -- this consistency in operation between series blending and volume control makes series blends more intuitive to dial in, than the traditional parallel counterpart.
This behaviour continues with third variation and the introduction of treble bleeds. In particular, the introduction of a parallel type treble bleed having the additional benefit of making the apparent taper of the pot more linear, which helps the combined signal blend more smoothly between its extremes. Furthermore, if using this wiring for a spin-a-split control, I'd recommend using a smaller valued tapering resistor than usual -- around 1/5 of the pot value, rather than the regular 2/5 -- in order to make the taper even closer to linear. Thus, here's the same as above with the addition of a treble bleed consisting of a 1nF cap and 50kΩ resistor in parallel.
Alternatively one might prefer to start with an actual linear taper potentiometer and use a Kinman-esque treble bleed. Here are plots for that, the treble bleed values as before (1nF cap, 50kΩ resistor) though arranged in series this time:
I welcome other's input for presentation, corrections and such. As well as any blending schemes I'm missing, so far I plan to add: parallel with independent volumes, JH's series blending options (spin-a-splits), linear & log dual-gang blend pots (done for parallel), and woodstockwizzard's tapped blend pot (hopefully better than last time I looked at it).