Guitar circuit design: what's your approach on switches?

[stevewf]

To those who already feel confident in designing circuits in electric guitars, here's an open question from someone who'd like to gain some confidence.

Given specific wants/needs for a circuit, what's your mentality on selecting switches when you start designing? And then, on choosing which lugs get which wire?

Some examples, for clarifying the nature of the question:

• How do you decide the number of poles/throws needed for a switching scheme?
• How do you consider a 2P5T "Strat switch"? Do you think of it as one pole for choosing the route of the ground, and the other pole for choosing the route of the hot? I've observed schematics that show a coil's leads attached to the common lug of one pole, and to a throw lug of the other pole; what's the rationale behind choosing that?
• What about a 4P5T "Strat superswitch"? Any additional thoughts on those?
• Similar questions for choosing another common type: the 2P2T switch. Can you summarize why you choose which lug for which wire?

Thanks, from a person who wants to "learn how to fish"!

[sumgai]

steve,


What you're asking for is an exercise in logic. So it turns out to be easy to lay out some "ground rules" (pun intended!), just like any business plan. From those rules, you determine if there's anything common between them, and whatever might be unique that can't be common.

For an example, let's take your Strat switch (the nominal double pole, 5 position variety that shorts between contacts in positions 2 and 4). If we have a business rule that says "select one pickup from among three pickups, and route that to the output", we have an easy time of visualizing that scenario. Oh, wait....did I skip something?

Well, yes, I did. You do need a passing familiarity with a switch's internal logic. By that I mean, you must know which contacts come together in any given position. And that familiarity is what drives the answer to your question about which switch to use for a given job. Of course, I now have to complicate things by saying that you might have a choice in the physical layout of different switches, and still be able to accomplish the logical portion of your design. Other times (of course!), you'll be constrained to use a very specific type of switch to get the job done. It's just part of the territory.

Now that said, if you were to select from only two pickups on board, you can select from a large variety of switches.  I now becomes a matter of physical desires (or even necessities, those often play a part in the selection process). Again by example, about the simplest "logic" design is a Kill switch. This can be accomplished with the most simple switch possible, a single pole, single throw (SPST) unit. But the rub here is that there is a large number of potential switching alternatives:

Should it be a separate item;
Should it be spring loaded (a 'stutter' switch) or should it stay on by itself;
Should it, or could it be incorporated with another switching function (see necessity and 'business rules', already discussed above);
Should it be physically so large that it can be hit with ease (whereby it can be hit accidentally);
Should it be lit with some kind of LED;
Should it be obvious, or should it be 'stealthy', i.e. not obvious to a viewer;
and finally....
Do I already have a switch on hand that will do the job?  The answer to this question dictates the answer to all of the preceding questions about 80 or 90 percent of the time, trust me on that one.

And that's just for starters, asking yourself questions of this nature.  Which brings me to my tl;dr moment:   Even if you have a plan all worked out at the beginning, be highly prepared to modify that plan as you progress towards the finished goal.

One more thing I'll tell you.  As I've told newey and countless others here over the years, Think outside of the box.  If you've been seeing diagram after diagram that does something one way, look to see if you can find an alternative routing that does the same job.  More often than not, this will not be obvious, but the effort just might simplify the overall circuit, and possibly give you a better selection of switches from which to choose.  My point here is two-fold - it'll help you learn what's what in logic reasoning, and you just may become the envy of other members when you show up with a very novel design.

Last thing - break it down into modules.  Start with the most basic, simple thing you can identify.  Build up from there.  And for Gawd's sake, take notes!!  Measure with your meter wherever you might think it'll be a good future reference for troubleshooting purposes, or simply to share with others as they ask questions.  Your brownie point score will increase dramatically whenever that happens. 

I hope that's covered enough of the basics for you.  Good luck!





sumgai

[Deleted]

Tend to go for what you want and then cut down , where switches do the same job or where a switch isn't doing any thing in a state try and marry it up with some thing else .


Go wild go big and then try and cut it down to fit in the hole.(please do not use that as a chat up line, I got slapped last time)

------

Think most switches are easy but the 2p2t on/on/on and the on/off/on ( this one is kinda easy if there is any state that nothing happens the off) can be dam complex

[JohnH]

I spend most of my life designing stuff, or teaching it, or dealing with the consequences of it! It has always puzzled me to figure out how to do it, and I don't really know, despite doing it.

A guitar circuit is a nice compact world in which you can play with design.

A few thoughts:

Almost nothing is truly invented from scratch. Almost all new designs are 95% recycled ideas. So study existing designs to understand them.

Know what the parts are, for the guitars, what switches, pots and other parts are available to use?

Start by imagining how it will be operated, without knowing how it works. What will it do and what knobs will be needed for control?

Then start wrangling parts on paper. To design a circuit, you have to develop a sense of insight into circuit flow and connectivity. This takes time, but start simple.

Then, it's a cycle of considering what you want and then what you can do, and eventually they may converge. Lots of scrap paper needed.

Start an idea in the belief that you will make it work. Then you may be able to find how to do so.

Watch out for designs emerging where all the planets align and the sun shines out! Parts and ideas coming together to resolve each other and solve a great design!

[Deleted]

I wonder if , do a challenge.
Simple two pickup guitar
In a strat body so you have space
Use any switch or pot (in my case no crazy editing) but has to be real.
And show your working,reason and thinking.

----

1st thing I'd do is give each end of the pickup a common on a switch and then do every combo (N, B, N+B, N+b, NxB, Nxb, kill maybe)
I normally like to tie a end as means I need less poles. Neck + to output that way I can phase the Bridge.
-n 1)GND 2)OUT 3)GND 4)GND 5)+B 6)-B 7)N/C
As I would like kill IE ground and output link as a stage I'll put Output and Ground as a common
OUT 1)N/C 2)+B 3)+B 4)-b 5)-B 6)+B 7)GND
GND 1)N/C 2)-b 3)-b 4)+B 5)N/C 6)N/C 7)OUT
So I need 3P7T ,what isn't easy
www.arrow.com/en/products/m378636-0198/grayhill
3P8T means I gotta add one more state if I can think of it or just add another kill say at the start

-------
And if force to use a lever switch.. go for the max and work down due to cost.
4p5T
Here I can go wild a bit
And have each bit of the pickup on a common ( I would like to half of as makes it cheaper )
N+ 1)Out 2)Out 3)Out 4)N/C
-n 1)GND 2)GND 3)+BO 4)N/C
+BO 1)N/C 2)Out 3)-N 4)Out
-bO 1)N/C 2)GND 3)GND 4)GND
With +B and -b being on a 2P2T phase switch giving your -bO and +BO. I had forgotten about the kill switch state so will have to change things around, as you see N+ as a lot of Output so tie that to output and freeing up a pole or I could just change the N+ and Out around on the switch and make 5)GND

----

I think Yogi B and MattB thinking would be interesting too as they approach things differently to me I think

[asmith]

Aug 11, 2022 13:02:38 GMT -5 JohnH said:

Start by imagining how it will be operated, without knowing how it works. What will it do and what knobs will be needed for control?

+1. For me, everything grows from this root.

[MattB]

I agree with everything posted so far, in particular the advice to study existing designs and understand how they work.

For complex designs I usually start by hand drawing a basic block diagram.  I'll draw the parts I know how to do, and then look for ways to add the rest. This is how I try to figure out how many poles I neeed for each switch, and where they need to go.

Once I have a basic layout I think might work I'll draw in individual connections, and try to add any bits I left out. If I get stuck and can't see a way round the problem I'll usually just add more poles wherever I need to. If I get everything working I can look for ways to simplify things and hopefully get rid of the extra poles.

This is the initial diagram I drew for the designs I posted in this thread: 

guitarnuts2.proboards.com/thread/9879/challenge-coil-select-phase-series?page=2

Each of the triangles represents one pole of the 5-way superswitch, with the common connection at the point. At this point I'm not thinking about indiviual switch positions or specific pickup selections. I just assume anything connected to a 5-way pole can be switched on or off as required.

There is a 5-way pole on either side of the parallel/series switch, to control what pickup selections go through it. There is a pole connected to the output, to avoid pickups hanging from hot. And there is a pole to split the middle pickup, and connect it to either hot or ground.

Before I drew this, I didn't know how to connect the middle pickup. I drew the top part, with the neck and bridge pickups and parallel/series switch, then looked for how I could connect the middle.

There might be other layouts that could work. At this point I'm not sure exactly what each switch will be doing. But laying it out like this it seems clear to me that I need something on either side of the parallel/series switch, and I need something to select the middle coil, and that means I need at least three poles on the pickup selector.

I didn't add the phase switch, because this diagram was enough for me to see that it would work. This is where having a good understanding of switch logic and basic modules comes in handy. I know how a DPDT centre-on works. I know how a phase switch works, and how to combine it with a coil split. So I just need to see the basic layout, and I know that it will work without having to draw out every connection.

If you compare this diagram to the first schematic I posted, you'll see that I had to add some extra connections to make the parallel/series switch work. If you compare it to the second diagram, you'll see I had to move some parts around to cut the parallel/series switch down to two poles. So my initial diagram wasn't perfect, but for me it's a lot easier to start with something that sort of works and try to improve it.

[stevewf]

Very useful, folks!

Some things that said here that are sinking in, not in any intended order:

• Get familiar with guitar wiring by looking at existing solutions. Through this, we can build a repertoire that includes things like like pickup selection, serial/parallel circuit routing and phase swapping.

• Get familiar with the switch components themselves (and with potentiometers, etc, not mentioned in the thread title, but seems to apply).

• Think about what features you want first, without think about how to get it. In early stages, think big.

• Create a high-level overview of the circuit at first. Since you've built a repertoire, you can summarize or even omit those in the early drafts.

• Try to reduce the number of poles as you go. Consolidating switches and/or tying a coil's lead directly to ground/output can help.

• Think outside the box. View from a different angle.

And here's some smoke. Smoke is usually what results when the gears in my brain begin turning and y'all have done gone started that, so here:
Viewpoint

• Player's view. Imagine the way it'll feel to use the controls that you're proposing. Is it intuitive? Too fiddly? Novel/fun?
• Mechanical view. Can you cram all of that proposed circuitry into that guitar? Is there room to spare?
• Electrical view. Does it offer enough desired features? Is it too hard to solder that many wires onto that lug?

Thank to answers here, I've thought more about my own mental workings in this scope... (ironic? Thinking about mentality?).
Poles and throws. A pole's relationship to a throw can be like the pole is deciding "which signal source shall I convey onward?" or "to where shall I convey the signal?".

[newey]

Aug 13, 2022 22:59:18 GMT -5 stevewf said:

Poles and throws. A pole's relationship to a throw can be like the pole is deciding "which signal source shall I convey onward?" or "to where shall I convey the signal?".

Yes. And each (unique) throw is one possible road in or out of the pole.

Consolidating switches and/or tying a coil's lead directly to ground/output can help.

Ideally, to ground; to output if it can't be avoided in a particular scheme.

I think it is also helpful to bear in mind the situations where you won't be able to do so. A phase switch, for example, will always require both hot and ground to be switches for one pickup or coil. A series/parallel switch will likewise require one ground to be switched in addition to the hot. (This is called "lifting the ground", if you see that phase, you'll know a ground needs to go to a switch somewhere.)

Those are the two main examples, it also applies when several things are put in series and it is desired to turn these things on/off individually.

[stevewf]

Aug 13, 2022 0:17:36 GMT -5 MattB said:

I agree with everything posted so far, in particular the advice to study existing designs and understand how they work.

For complex designs I usually start by hand drawing a basic block diagram.  I'll draw the parts I know how to do, and then look for ways to add the rest. This is how I try to figure out how many poles I neeed for each switch, and where they need to go.

Once I have a basic layout I think might work I'll draw in individual connections, and try to add any bits I left out. If I get stuck and can't see a way round the problem I'll usually just add more poles wherever I need to. If I get everything working I can look for ways to simplify things and hopefully get rid of the extra poles.

This is the initial diagram I drew for the designs I posted in this thread: 

guitarnuts2.proboards.com/thread/9879/challenge-coil-select-phase-series?page=2

Thanks, MattB. That draft and its explanation give good excellent! insight on how you design, in particular that circuit you offered, linked just above.

I can see how it evolved! The components from the draft live on in the final, and (but) there are some differences. Thumbnails for ease:




- The series/parallel has been fleshed out (that's SW2).
- The phase swapping has been added, as well as the Vol, Tone and Blend controls. MattB has said that through experience, some controls have become so basic that a draft for the author's own use needn't mention them.
- Another item fleshed out in the final is that phase swapping is tied to coil swapping, and that is where some magic occurs. I've done phase swapping with a pair of 2P2T switches; with a 5-way, while I haven't done it, I can begin to imagine how; I've done coil swapping with various switches, but I'd never succeeded at tying the two together, and here's the but:
- The splitting of the two M coils is now being done by throws instead of a pole. The circuit still uses the 5-way superswitch (SW3), but now instead of using the pole on what was D, it uses a throw on B and another on C (for positions 2 and 4, respectively). That's how phase swapping is being tied to coil swapping: through combining a coil-swapping 2P2T switch with a phase-choosing 2P5T (in effect) switches.

This last bit is exactly the kind of decision-making that I feel like I need a breakthrough. How the heck do I decide if a pole gets the job or a throw, when I'm first laying down a draft? Rules of thumb can be time-savers, and since I'm slow at drafting, the clock goes whizzing by and many scraps of paper go into the recycle bin! What was the first inkling that led to drafting that coil split using a pole, and what revelation prompted the change toward using poles instead?

In order to gain this level of comprehension of MattB's drawing (and likewise jhng's in that same other thread), I've really had to tear it apart. I redrew subsystems, which enabled me to see how the switches' functions are distributed.

It's all taking time. More porous minds can probably do it faster, but here the brain needs to have new information kneaded into it. Sigh. Still, I do feel that I'm improving. Seems that at least I'm following the "study existing designs and understand how they work" guideline, and that it's working

[stevewf]

And here's a diagram I made while I was translating MattB's schematic to a wiring diagram for the Ash Lab

[MattB]

Ok, I'll try to explain my thought process in bit more detail.

First, here's a simplified diagram showing two ways to combine a phase reverse and a coil split. The black boxes marked phase on the diagram are there to show where the connections to the phase switch are made. Not all the connections that would be needed for the phase switch are shown. Instead the coils on the diagram have been flipped to show reverse phase.

On the left, the coil split is before the phase reverse. Flipping the coils also flips the coil split switch, because the two are hardwired together. (Note the numbering on the switch is reversed.)

This means the switch cuts the same coils, in or out of phase.

On the right, the coil split is after the phase switch. Hopefully you can see how revering phase also means the opposite coil is selected.

So when I said I knew the phase switch would work without drawing it I only really needed to see two things.

1)The coils are connected in series and in phase.

2)There is nothing that needs to go before the phase switch.

As long as those two things are true, the version on the right should always work.

Your questions about poles and throws doesn't really fit with how I think about circuits. I mostly think in terms of modules and how they fit together. It's only when I run into trouble that I start thinking about individual lugs and wires. I realise this isn't very helpful advice for a relative beginner, and I'm not sure it's the best way, but it's what works for me. I don't think there are any easy shortcuts to understanding. When I started learning about guitar wiring I struggled with lots of things, but I kept working at it and I started recognising parts of diagrams without having to trace every connection. From there it got a lot easier.

For this design I ended up having to trace some wires, because even though I drew it, I didn't really fully understand the first schematic I posted. This is one drawback of my way of thinking- it can easily lead to blind spots and oversights.

Here is the initial rough sketch I posted, with the middle pickup edited out.

The middle pickup has to connect to the parallel/series switch at some point. The two 5-way poles on either side are the obvious way to do that.

So I looked for every possible way I could connect them, then figured out what would happen.

I looked at three ways.

1)Split the middle pickup in half and put one either side. Like this:

Splitting a humbucker in half and treating it like two different pickups is a trick i use quite a bit. It's very handy for combining a split humbucker with other pickups, like we want to do here.

The problem is, I couldn't make the phase switch work.

The middle pickup coils aren't connected in series, so the standard method I described above won't work.

I need to switch both ends of both pickups. That means four poles. I want to put both phase controls on the same switch, so I don't want to use more than two.

Ok, looks like I need to keep the middle unsplit so I can phase reverse it with two poles.

2) Connect the middle pickup centre join to both sides of the parallel/series switch.

I think this one would have worked. The only issue is, it leaves a pickup hanging from the centre join in two positions. That's best avoided where possible, becausee it could be a source of noise.

3) The same thing, but inside out. What if I connect the middle where the ends are and the ends where the middle is?

This is the version I went with, because it only leaves a pickup hanging in one position.

When I was drawing out the schematic I realised that in series mode there would be no sound in positions 1 and 5 because there was no connection to one half of the parallel/series switch. So I added wires to fix that. Then i realised I'd just shorted out positions 1 and 5 in parallel mode, so I added extra poles to SW2 to fix that.

That was the first schematic I posted in the other thread.

I had a feeling I was missing something, so I traced the connections in positions 1 and 5 and wrote them down.

In parallel mode, position one, the neck + goes through SW4D and SW2C to "hot". (I usually draw the volume and tone pots last, because that's the easy part, and they get in the way otherwise. So when I'm working on a design I think of hot and ground as the two end points of the circuit)

In series mode, position 1, the neck + goes through SW4D, SW2C, SW2B, SW4B, SW2A and SW4A to hot.

So it ends up going to hot however the parallel/series switch is set. That should have been obvious, but apparently I needed to remind myself.

So if it always goes to the same place in position one, it doesn't need to go through all those switches. I can connect the neck + directly to SW4A lug 1, and that means I don't need SW3C.

Can I do the same thing in position 5?

I need a connection to ground in position 5 that's always on. I don't have a way to do that right now. Can I free up a pole anywhere?

SW4C only has two connections, and one of them is to ground. Can I move the other?

It goes to SW4A. It doesn't need to go through both. I can connect the middle pickup centre join directly to SW4A lug 2.

Now SW4C can be used to connect the bridge pickup to ground in position 5, and the middle pickup centre join to ground in position 4.

That's the second schematic I posted in the other thread.

In the rough sketch I drew, the pole that would become SW4A was there to stop things hanging from hot. So on that sketch I ran everything that was going to hot through that switch, because I knew that would work and it meant I didn't have to think about it. Then when I was drawing the parallel/series switching in the schematic I wasn't thinking about SW4A because it isn't part of my mental map of how parallel/series switches work.

On the first schematic, I made both parts work individually without seeing that they could work better together.

So in the end my way of thinking led me to making the design more complicated than it needed to be, and I had to go back and find my mistakes.

I still think my approach is worthwhile. Without a rough sketch I'm not sure I would have seen a way to start.

And finally, here's something that helped me a lot when I starting to learn.

When you study a diagram and figure out how it works, wait a few days and try to draw it from first principles. Not by remembering the whole thing, but by applying the same methods from scratch. Sometimes it can take a while to see things in practice even when you understand the theory.

Here's a good one to get you started. It's very simple in the sense that it uses very few parts, but it's also very clever and kind of unconventional. It gets a lot out of a regular 5-way switch by connecting three pickups in series. It took me quite a few attempts before I could reproduce it without a lot of headscratching.

Jul 7, 2021 23:15:00 GMT -5 Yogi B said:

Jul 6, 2021 20:57:00 GMT -5 Yogi B said:

I think the thing that bugs me is that even with a standard Strat switch or 8-terminal import switch, it's still possible to have a wiring scheme that's substantially more May-esque...

Scratch that, we can do better. With a standard Strat switch, two phase switches, and a DPDT otherwise going spare, it's possible to get ALL the selections that are available on the RS.

SPOILER: Click to show

[stevewf]

Wow, MattB, that's a lot of time you've given. Thank you! You've done a great job of showing your method by example. And of course it doesn't bother me that the example is related to a request made earlier by Mr Muggins here (me).

Also thanks for the exercise "Brian May with a 5-way." Looking it over, yes, it's clever and simple. And yes, I'll see if I can redraw it by concept, because that's a learning method that I now see I've already been applying (e.g. when I redraw "subsystems").

But again, wow! I'm grateful for the input, and I'll venture that it makes this forum richer.

[stevewf]

As an exercise of making myself understand better, here's an attempt at echoing some of MattB's thought process, with an example set f drawings.



Splitter and PhaseSwapper: We can consider each of those functions as a separate "subsystem".

There are four drawings here: Nested-Overview, #A=Splitter, #B=PhaseSwap and #C=Nested-Expanded

In this example, a humbucker splitter is nested inside a phase swapper. Accordingly, the orange circuitry (Splitter) is enveloped by the blue circuitry (PhaseSwapper).

CoilSplitter (orange): The control is a 1P2T switch. One way to consider this subsystem is: "Switch position#1 shorts the South coil and pos#2 shorts the North coil; we don't (yet) know/care which end of the pair of coils gets assigned (later) the Hot and which gets the Ground."

PhaseSwapper (blue): The control is a 2P2T that dictates "#Down assigns the Hot to the left side of the orange subsystem and Ground to the right, and #Up assigns Hot-right, Ground-Left; we don't know/care what's inside the orange subsystem."

Critical to this way of thinking, #B shows the detail of the PhaseSwapper mixed with an overview of the Splitter. Verbally: "PhaseSwap whatever is in the orange subsystem." Analogous to Blue = PhaseSwap(Orange)" in algebra. The word "whatever" shows that we don't need to consider what's inside the orange circuitry at this point of the design process. That "whatever" idea can be translated to a pair of parentheses.

Working backwards from there, we could say the first drawing amounts to "Overview = Blue(Orange)."

Then the final drawing may be obtained by first defining the details of each of those separate subsystems (#A and #B) and then putting them together with their details expressed (#C).

---

Note: this little exercise treats the left side of MattB's helpful schematics (earlier in this thread). This happens to be the "wrong" set of drawings for my original request, but it's all I have time for right now! Without having fleshed it out yet, I believe the same principle of nested subsystems will apply.

[Edit:] Now I see that I could tie the coil shorting function with the phase swapper, meaning when one's changed, so the other is. This could be done by using a 3P2T switch; all poles are still present and wired the same, but you change both functions together, no choice. Further, now my brain wants nest all this between a 5-waySS. I wonder if I'll end up with what MattB posted after all. That'll be later - it's only lunchtime now. [/Edit:]

[stevewf]

Next stage:



(D) "Ties" the coil splitting choice with the phase selection, which, in circuit vocabulary, means unifying two separate switches into a single switch. In this case, they were the 1P2T and the 2P2T and now it's a single 3P2T.

(E) Encapsulates all the preceding development into a module (green box)

(F) Translates "use a 5-way to select when the humbucker is active, and choose which phase" into a pair of poles on a 5-way.

(G) Expands the contents of the green module within the new 5-way setup. [Edit:] I chose not to define this as a finished "module because there are other throws on these poles that will probably get used in separate "peer-level" circuitry; this choice might get reversed later.

(H) shows the whole nest as a module. The unused poles and throws on the 5-way have been greyed out; they remain available for other parts of the circuitry, but the two poles of the 5-way that are used now bear labels. I foresee that there'll be peer-level circuits being added soon, and that phasing will need to be consistent. In case I employ those unused throws that will share the poles I've already assigned, I'm hoping the labels will help me keep the phasing straight when the time comes.[/Edit:]

We now have, if I can continue this arithmetic analogy: CoilActivator_PhaseChoice(PhaseSwap(Splitter)). I don't know how long this analogy can continue to make sense as the development continues, but here we are so far.

[Yogi B]

Aug 10, 2022 12:15:55 GMT -5 stevewf said:

Given specific wants/needs for a circuit, what's your mentality on selecting switches when you start designing?

Worry about it later.

Often when first 'sketching' out an idea I'll allow myself extra poles than usually exist on common switches with the hope that I can rearrange and simplify once I have something that functions correctly. By "sketching" I don't mean in the sense of drawing, I do this textually: switches with one pole per line and (usually) the common in the first column. (I almost always break that second rule with double throw switches.) That allows me to focus on the logic and allows a much quicker pace of revisions rather than worrying about visual layout and legibility — plus if it's written in a consistent enough format I can double check the switching logic programmatically. (There's a reason why the below is almost exactly valid python syntax )

I'll try my best to illustrate this with a recreation of a real life example. The scheme in question was to have two modes: the first is regular Strat switching (albeit with a master tone, so no switching of tone pots); the second, alternate mode, is a bit Red Special & a tiny bit Tele — more specifically it gives B×M, B×−M, B+N, N×−M, N×M.

Firstly I'll look at permanent connections to the pickups, we want series and OoP combos so we can't permanently ground all three pickups. But, looking at these combos they all contain the middle pickup. The bridge & neck are only both selected when in parallel, therefore it's easier to have B & N permanently connected to ground & leave the middle 'floating' (for now). Written out, that gives something this:

pickup_b(pos='b+', neg=GRND),
pickup_m(pos='m+', neg='m-'),
pickup_n(pos='n+', neg=GRND),

Also, since we are doing regular Strat switching with the first mode (throughout which the middle pickup can be grounded) we can fill out one pole of our mode toggle switch:

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
),

In the above '--' is a currently unused net which I'll come back to later, but I've included it now because we know 'm-' will need to be connected to *something* in order to get the series combos. It's a separate net because if we used 'm-' directly that would cause problems since in the parallel mode 'm-' is joined with ground.

Next I'll start filling out the selector switch. Purely from intuition it's very unlikely that a regular Strat switch will do, so I'll use a switch with 5 independent throws.

In positions 1 & 2 we want the bridge connected to the hot output but only in Strat mode. In position 3 we want the opposite: the bridge connected to output only in the alternate mode. That would look like this:

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'A3', None, None,
),

Both a help and a hindrance of doing this textually means that I need to name the nets of the circuit, this makes the result somewhat self documenting — however meaningful names are hard, and harder when I'm restricting myself to just two characters for the sake of easy column alignment. In the above 'OS' stands for "Output: Strat mode" — i.e. connected to hot output in the Strat mode, disconnected in the alternate mode.

Similarly, 'A3' is "Alternate position 3 output", it could be overall "Output: Alternative mode" (i.e. 'OA') but that could be getting ahead of ourselves — later potentially finding out we've painted ourselves into a corner — so at this stage, whilst just trying to get something functional without worrying about efficiency, 'A3' it remains.

Despite naming the nets, we've yet to actually implement their functionality, for that we'll need to add another pole to the mode toggle switch:

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
    'OS', OUTP, 'A3',
),

And we can sort out the neck pickup via a second pole on the selector switch:

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'A3', None, None,
    'n+', None, None, 'A3', 'OS', 'OS',
),

Now we just have to figure out what to do with the middle pickup. If the selector switch were a megaswitch we'd have a further two poles, we can use one to switch 'm+' and the other to switch '--' from earlier. I'll start with the Strat position 3 first, all that requires is that 'm+' is connected to hot, but only in the Strat mode — we already have a net that does this, 'OS'. For now I'll leave all the other throws disconnected:

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'A3', None, None,
    'n+', None, None, 'A3', 'OS', 'OS',
    'm+', None, None, 'OS', None, None,
    '--', None, None, None, None, None,
),

For the remaining 1,2,4,5 series positions we basically need a phase switch that swaps the selection between 'm+' and '--', for now I'll call these 'xx' & 'yy'. Since we want the both outer positions to be in phase and the notch positions out of phase, these must be arranged thusly:

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'A3', None, None,
    'n+', None, None, 'A3', 'OS', 'OS',
    'm+', 'xx', 'yy', 'OS', 'yy', 'xx',
    '--', 'yy', 'xx', None, 'xx', 'yy',
),

Again, these new nets don't actually do anything useful yet, so let's add another two poles to the mode toggle:

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
    'OS', OUTP, 'A3',
    None, 'xx', None,
    None, 'yy', None,
),

First, let's concentrate on Strat mode. In positions 1 & 5, 'm+' is connected to 'xx', since in Strat mode 'm-' is grounded 'm+' and thus 'xx' should remain disconnected from any other net. In positions 2 & 4, 'm+' is connected to 'yy', thus for the middle pickup to be selected as needed 'yy' must connect to the output — so let's do that:

(The connections from '--' to 'xx' or 'yy' matter not, since in Strat mode '--' is disconnected.)

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
    'OS', OUTP, 'A3',
    None, 'xx', None,
    OUTP, 'yy', None,
),

Since the bridge & neck are the ground-most in the series combos, one end of the middle pickup must connect to the output — in order to get the specified phasing behaviour that end is: 'm+' for positions 1 & 5; and 'm-' (via '--') in positions 2 & 4. Thus in the alt mode we need 'xx' connected to the output.

Now we only have to figure out what to do with 'yy': in positions 1 & 2 it should connect to 'b+' and in positions 4 & 5 connect to 'n+'. We could establish a new net & add a fifth pole to the megaswitch to accomplish this, however we already have a net that seems to fit the bill: 'OS'. There just two questions we have to answer before we can be confident of that choice. Does 'OS' connect to anything else in those positions that could cause issues? Remembering that we're currently considering alternate mode, the answer to that is no. Secondly, what about position 3: here we have 'OS' connected to 'm+' — is adding 'yy' to that a problem? Again the answer is no.

So, finally, putting everything together we have a functional scheme:

pickup_b('b+', GRND),
pickup_m('m+', 'm-'),
pickup_n('n+', GRND),

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
    'OS', OUTP, 'A3',
    None, 'xx', OUTP,
    OUTP, 'yy', 'OS',
),

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'A3', None, None,
    'n+', None, None, 'A3', 'OS', 'OS',
    'm+', 'xx', 'yy', 'OS', 'yy', 'xx',
    '--', 'yy', 'xx', None, 'xx', 'yy',
),

If we had no further constraints we could stop there. However, whilst a megaswitch can be fairly easily obtained and substituted in place of an original Strat switch, the 4PDT for the mode toggle may not as be feasible due to considerations of space or aesthetics (4PDT push/pulls are extremely rare, so you're pretty much limited to a Fender S-1 switch if you want a stealth installation). If we could reduce the requirement to only a DPDT it'd mean we could use a regular push-pull or push-push.

The most obvious starting point is the net 'yy', in the Strat position it's connected to OUTP and therefore connected to 'OS', thus actually connected to 'OS' in both positions. As such we can permanently merge 'yy' into 'OS', freeing up one pole:

pickup_b('b+', GRND),
pickup_m('m+', 'm-'),
pickup_n('n+', GRND),

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
    'OS', OUTP, 'A3',
    None, 'xx', OUTP,
),

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'A3', None, None,
    'n+', None, None, 'A3', 'OS', 'OS',
    'm+', 'xx', 'OS', 'OS', 'OS', 'xx',
    '--', 'OS', 'xx', None, 'xx', 'OS',
),

We can do similar with 'xx', that can be safely merged with 'A3'. Both are connected to the output in the alt mode and disconnected in the Strat mode, and there'll be no interaction connecting the two since 'xx' is only connected to other nets in positions 1, 2, 4 & 5 and 'A3' only in position 3. For symmetry I'll name this new net 'OA' for 'alternate output' — we've now completed the objective for the simplified wiring:

pickup_b('b+', GRND),
pickup_m('m+', 'm-'),
pickup_n('n+', GRND),

toggle_mode(
#   thrw1 commn thrw2
    GRND, 'm-', '--',
    'OS', OUTP, 'OA',
),

selector_switch(
#   com    t1    t2    t3    t4    t5
    'b+', 'OS', 'OS', 'OA', None, None,
    'n+', None, None, 'OA', 'OS', 'OS',
    'm+', 'OA', 'OS', 'OS', 'OS', 'OA',
    '--', 'OS', 'OA', None, 'OA', 'OS',
),

---

Something I set out to show was swapping nets between the commons & throws of a switch for our advantage, however the above failed at that. It could be applied here, but is unnecessary. For example you can rearrange the selector switch such that the two 'throw nets' are exchanged with the four 'common nets', like so:

selector_switch(
#   com    t1    t2    t3    t4    t5
    'OA', None, None, 'b+', None, None,
    'OA', 'm+', '--', 'n+', '--', 'm+',
    'OS', 'b+', 'b+', None, 'n+', 'n+',
    'OS', '--', 'm+', 'm+', 'm+', '--',
),

I know I have applied this to my own wiring schemes, but I can't remember a specific instance. And with the midstage virtual circuits being ephemeral, it's difficult to search for after the fact.

What I did find is an example from a circuit by alexis, namely this one: New SSS all combinations mod - invisible - original 5 way.

In particular I'm looking at switch 2, the rotary 4P6T 'mode switch', over on the right-hand side. Here's my textual representation thereof:

rotary_4P6T(
    #  c     1     2     3     4     5     6
    'gr', OUTP, OUTP, 'pk', 'li', OUTP, 'li',
    'li', None, None, None, None, GRND, 'pk',
    'pp', None, None, None, None, None, OUTP,
    'pk', None, GRND, None, 'pp', 'pp', None,
),

(Here the net names are purely based upon the wire colours from the diagram: 'gr' — green; 'pk' — pink; 'li' — lilac; 'pp' — purple.)

With the first pole being full but the rest being pretty sparse, this is a good candidate for modification. So, looking at the first pole's third throw this connects 'pk' to 'gr' we can move that connection such that it occurs on the pole with 'pk' connected to the common.

We can also do similar for the connections between 'gr' & 'li' at positions 4 & 6. Position 4 is easy, but at position 6 'li' is already connected to 'pk'. Thankfully there's a pole with 'pk' at its common and throw 6 open, therefore we can connect 'gr' there. Alternatively we can swap the connection between 'li' & 'pk' to there, then use throw 6 of the 'li' pole to connect to 'gr' — this is what I've done in the below:

rotary_4P6T(
    #  c     1     2     3     4     5     6
    'gr', OUTP, OUTP, None, None, OUTP, None,
    'li', None, None, None, 'gr', GRND, 'gr',
    'pp', None, None, None, None, None, OUTP,
    'pk', None, GRND, 'gr', 'pp', 'pp', 'li',
),

Now we've done that, the first pole only makes connections between 'gr' & OUTP, thus we can swap the two — i.e. have OUTP on the common and 'gr' on the throws. We can also do the same with the third pole that connects 'pp' & OUTP. After doing such, we get the following:

rotary_4P6T(
    #  c     1     2     3     4     5     6
    OUTP, 'gr', 'gr', None, None, 'gr', None,
    'li', None, None, None, 'gr', GRND, 'gr',
    OUTP, None, None, None, None, None, 'pp',
    'pk', None, GRND, 'gr', 'pp', 'pp', 'li',
),

Finally, since there's now two poles that share the same common and have non-overlapping throw connections we can merge the two poles into one, giving us a spare pole or enabling the use of a switch with one less pole:

rotary_4P6T(
    #  c     1     2     3     4     5     6
    OUTP, 'gr', 'gr', None, None, 'gr', 'pp',
    'li', None, None, None, 'gr', GRND, 'gr',
    'pk', None, GRND, 'gr', 'pp', 'pp', 'li',
    None, None, None, None, None, None, None,
),

---

Anyway, I hope the above has been insightful, even if long-winded and probably hard to follow if you're not accustomed to reading circuits as text.

[stevewf]

Aug 20, 2022 4:16:10 GMT -5 Yogi B said:

By "sketching" I don't mean in the sense of drawing, I do this textually

Wow. That is a method that I hadn't seen and had never occurred to me! It's gonna take some time for me to generate a response that is worthy!

Some aspects of the process are reverse order from drawings (like identifying/naming sub-networks upon their creation, rather than drawing and then identifying). Perhaps being versed in both would give one a richer understanding overall (read: make one more facile at creating designs).

One thing I'm toying with is to attempt to translate this progression into drawings, stage by stage. That's kind of counter-productive, but it might help me transition to this textual quasi-python mode of design. There's much to translate: Naming parts of the network; spotting poles that can be consolidated; swapping pole networks and throw networks (indeed, considering them as separate types of networks to begin with).

But that'll be for when I have a larger chunk of time.

Anyway, I hope the above has been insightful, even if long-winded and probably hard to follow if you're not accustomed to reading circuits as text.

it may seem long-winded, but the detail is much (, much!) appreciated, and exactly the kind of response I was hoping for, a real eye-opener. Thank you, Yogi B!

[Yogi B]

Aug 20, 2022 4:16:10 GMT -5 Yogi B said:

(There's a reason why the below is almost exactly valid python syntax )

I've realized it is actually still valid python: I was thinking it wasn't because of the extra commas after the closing parentheses of the function calls, but that would just make the expressions into tuples. (Pointless tuples, but nonetheless still syntactically valid.)

In truth, the reason why the commas are there at all is because the results of those functions are passed as arguments to a netlist class. Since the main purpose is illustration of the idea I could've just omitted them, along with being generally much less syntax heavy (both tables or spreadsheet cells would work as equally well), but decided against it.

Aug 20, 2022 10:45:47 GMT -5 stevewf said:

One thing I'm toying with is to attempt to translate this progression into drawings, stage by stage. That's kind of counter-productive, but it might help me transition to this textual quasi-python mode of design.

I should say this isn't the only method I use, and drawing schematics is still very useful and swapping between the two is often useful in its own right, by giving a fresh perspective to work from. I likely would have realized the 4PDT from my example could be simplified much earlier if I'd done that.

Instead I went straight to a wiring diagram where I'm more concerned about correctly copying the connections across, whilst also deciding which 'virtual' pole maps to which 'real' pole in order to make the layout & subsequent wiring as tidy as possible. Ideally I don't want to get side-tracked during that process by wondering if the wiring logic could be improved, so I try to put that to the back of my mind. As such it wasn't until about three years later when revisiting the original text based design, due to its similarity to someone's request here, that I thought "hang on a minute..." and realised the simplifications were possible.

[sumgai]

Aug 16, 2022 11:29:39 GMT -5 stevewf said:

I don't know how long this analogy can continue to make sense as the development continues, but here we are so far.

Most of the time, the analogy should carry you all the way through the design process. There are times when it will break down, or seem to anyways, but if you stay focused, you should be able to achieve your goals using this method.

Aug 20, 2022 4:16:10 GMT -5 Yogi B said:

Feb 11, 54578 4:16:40 GMT -5 stevewf said:

Given specific wants/needs for a circuit, what's your mentality on selecting switches when you start designing?

Worry about it later.

Exactly.

steve, the point of the design process is simple: solve a puzzle in logic.  Implementation is where we make the design concept into a reality, meaning we now have to consider which physical switch to use, and as we've said, a good understanding of what switch can perform which function is necessary to making good "physical construction" choices.  But of course, during that construction, we have other considerations as well as the design itself, things like user desires for operability purposes, etc. - that's when we reach the "worry about it later" stage that Yogi mentions.

Now, it could be argued that  a designer works out the whole thing from start to finished product, But separating the two processes we've described makes things easier to observe the K.I.S.S. principle.  You know the one, right?

HTH


sumgai

[jhng]

What a fascinating thread!

stevewf -- if you want work on your guitar circuit design skills, here are a couple of problems to try. The first is quite easy, but illuminating once you see it. The second is a bit of a rabbit hole, but still possible.


1. Wire a standard 5-way strat switch to give:

N
N+M+B
M
M+B
B


2. Rewire the stock parts of a 1954 strat to get all 'standard' series and parallel combinations of the three pickups. By 'standard' I mean: three single coils, three parallel pairs, three series pairs, all in parallel and all in series. i.e. 11 combinations.

P.S. Extra kudos if you do the second one on an actual 1954 strat.

[stevewf]

Aug 24, 2022 11:26:13 GMT -5 jhng said:

What a fascinating thread!

stevewf -- if you want work on your guitar circuit design skills, here are a couple of problems to try. The first is quite easy, but illuminating once you see it. The second is a bit of a rabbit hole, but still possible.


1. Wire a standard 5-way strat switch to give:

N
N+M+B
M
M+B
B


2. Rewire the stock parts of a 1954 strat to get all 'standard' series and parallel combinations of the three pickups. By 'standard' I mean: three single coils, three parallel pairs, three series pairs, all in parallel and all in series. i.e. 11 combinations.

P.S. Extra kudos if you do the second one on an actual 1954 strat.

Thanks for the exercises. Maybe I'll dabble while I wait for parts to arrive. Does "standard strat switch" mean 2P5T or 2P3T with notch positions? Meanwhile, lemme see if there happens to be a '54 strat lying around in my parts boxes...

[jhng]

Aug 30, 2022 19:52:08 GMT -5 stevewf said:

Thanks for the exercises. Maybe I'll dabble while I wait for parts to arrive. Does "standard strat switch" mean 2P5T or 2P3T with notch positions? Meanwhile, lemme see if there happens to be a '54 strat lying around in my parts boxes...

It's the 2P3T with notch positions for the first problem. The second problem is really tricky to achieve, but you can learn a lot along the way.