BS modules.

[reTrEaD]

BS = Binary Switching?
BS = Bus Structure?
BS = (something more traditional and obvious)?

You decide.

In this thread I'll present three different type of Binary Switching strategies. They apply to (but are not limited to) a floating Bus Structure which is useful for (but not limited to) those "every possible combination" schemes we occasionally see.

Pack a lunch and fill your canteen before we get started. This could be a long, meandering journey.

---

Every Possible Combination.

If we're talking about a single coil, there is only one unique combination that's worth talking about. The one that produces sound. Flipping the phase on a single coil when no other coils are present is rather meaningless.

When it comes to two coils, there are exactly six combinations (excluding those which have a cap or resistor involved) possible. We can achieve all six of those with a relatively simple rotary switch that has at least 3 poles and six positions.

When it comes to three coils, we'll need at least 5 poles and 30-ish positions. That's just not gonna happen. So where do we go from here?

One possible solution would be to employ a single pole rotary switch on each lead of each pickup and a floating Bus Structure.

---

Floating Bus Structure

This isn't a new concept.  We've seen this before in the Game Changer, although in that case it was an automated system.  Here we'll be manually switching each connection. This is a cumbersome process. It isn't something to design around for a guitar one would use on stage. But it will make it possible to audition every possible combination. This could be useful in determining what a particular person might want to include in a reduced set of combinations on a stage-worthy guitar.

We'll need two fixed buses (one hot and one ground) plus a number of floating buses.  How many floating buses? One for every possible series link. So an 8-bus structure could provide enough capability for a seven coil arrangement. That's a bit of overkill but I chose to compare methods of serving an 8-bus structure because 2³ = 8.

---

Rotary Switch

We'll need thirteen 8-position rotary switches for a seven coil guitar. (one lead of one of the pickups can be wired directly to ground).

For a HHH, we'd need eleven 7-position switches.
For a HSH, we'd need nine 6-position switches.
For a HH or HSS, we'd need seven 5-position switches.
For a SSS or HS, we'd need five 4-position switches.

Add more to the total number of switches if we want to use a capacitor or pot to bypass or couple any of the coils. But the number of buses can remain the same.


Pro:
It's relatively intuitive and if we use chicken head knobs and a numbered indications on the mounting plate, we can easily see which bus we're connecting to.

Cons:
It could get fairly pricey.
It takes a fair amount of real estate.

Let's look at other options for switching.

---

Binary Switching (single switch)

Minitoggles are small and inexpensive. A SPDT takes up a tiny bit of real estate. But what can we do with just one?  Not much.


When the lower throw is connected (0) we connect to one bus, when the upper throw is connected (1) we connect to the other bus. But two a two bus structure isn't useful.

Nothing more to see here. Let's move on.

We're gonna need more switches!

Full Binary (2-bit)

Adding a second switch, in this case a DPDT, adds some functionality.  We can have two floating buses.  That's enough for a 3 coil arrangement.  SSS or HS.  Not bad at all!


Pros:  Cost is low.  Real estate is minimal.

Cons:  It's less intuitive than a rotary since we need to start thinking in binary.  But since we're only dealing with two switches and four possible combinations it's still quite manageable. Also, the second switch (DPDT) will cost about the same as two SPDTs.

We seem to be on a good path here. Let's expand.

Full Binary (3-Bit)

Now we're talkin'! By adding just one more switch we can get as many buses as we'll need for an absurd 7-coil guitar.

Pros: Just three switches.  The 4PDT does take up more real estate but the three switches together still have a much smaller footprint than a rotary.  This still it still has a slight price advantage over a rotary and a chicken head knob.

Cons:  4PDTs tend to be a bit pricey.  Even if we compare to the price of two DPDT or four SPDTs, they cost more.  We've moved farther down the 'think in binary' path and that's not great.

Unfortunately we can't expand any farther using the same strategy. That would require a 8PDT switch. Not gonna happen.  Besides, do we really need to support more than 7 coils?

Just for fun, lets examine some different avenues as starting points.

---

Binary Tree (2-level)

Hmmm.  This looks interesting.  By using separate switches for the second level, we can avoid the need for a more complicated switch.  Two SPDTs instead of one DPDT.  That's nice.  Three switches aren't too difficult to manage.  And it's rather visual, if we arrange the switches in a triangular pattern.  Less so if we flatten them into a straight line.  Not sure where this might lead when we go to the next level though.

Somewhat troubling is the 'don't care' situations represented by the grey x s.

Binary Tree (3-level)

Oh my, this got very ugly very quickly.

I'm far less concerned about the addition of four more switches than I am about the difficulty of keeping track of which bus we're connected to.  If we keep things in a triangular pattern (and waste a significant amount of real estate, it's still tough.  And all those don't care situations are really piling up.

If we flatten this physically so all the switches are in a straight line it becomes much worse.

I'm not fond of this.  Let's explore a completely different approach.

It's time to think outside the box triangle!

---

Binary Staircase (4-output)

I revisited that little corner of my mind I like to refer to as the pole vault and quickly realized I have a lot of thoughts on how best to use the pole of a multiple-throw switch but hadn't invested nearly so much thought into various ways of cascading SPDTs.

So it was time for a daydream, to allow thoughts to perform a pole dance. I think something useful might have shaken out.

Rearranging the three SPDTs we had in our 2-level binary tree, we can retain four outputs just like the tree. But now some nice things seem to happen. I like the patterns that are emerging.

This shows some promise. Let's just leap forward to the next major step.

Binary Staircase (8-Output)

As hoped, this looks just fine when expanded.

• The first thing we notice is that all the don't cares exist to the right.
• Everything that does matter exists to the left.
• The deciding point? The left-most switch which gives us a zero. (that will be dead-easy to spot)
• Putting the switches in a straight line will be a good thing.
• We can have any number of buses we desire just by adding or subtracting a step.

Pros:
The switches are cheap.
A relatively modest amount a real estate is necessary.
The switches can be arranged in a line.
The settings are visual and intuitive.
It's flexible.

Con:
The sheer number of switches required means this won't look pretty unless you're a fan of switch rash.  For a HSH, we'll need nine 5-switch arrays. 9 x 5 = 45 switches. Yikes!
5 x 3 = 15 switches for a SSS ... this is reasonable.

---

BS = Binary Staircase?
If you want to grow one of those every possible combination schemes, this seems fertile unless you have a lot of ground to till.

We could also use 4PDT switches in a binary staircase arrangement rather than a 4PxT rotary switch for other applications. But that probably falls under the heading: Just because you can, doesn't mean you should.

Which is best? Rotary, Full Binary, or Binary Staircase? That really depends on what's most important to you. But now you have some information to make a more informed decision.

Options are.

[Yogi B]

Dec 6, 2018 22:55:01 GMT -5 reTrEaD said:

BS = (something more traditional and obvious)?

# So let me introduce to you, the one and only Billy Shears #

We'll need thirteen 8-position rotary switches for a seven coil guitar. (one lead of one of the pickups can be wired directly to ground).

That might be a worthwhile optimisation in an electrically switched circuit, however I have to disagree when expected to manually perform the switching. At the stage of having teens of switches (or groups thereof), I don't think the saving of one switch is all that important -- were already past the point where most would likely be overwhelmed. And I'd say the disadvantage of adding a fourteenth switch is outweighed by the advantage that the one 'fixed ground' pickup then doesn't require special treatment.

For example, imagine setting up an essentially series combination of two parallel groups of the other six pickups (e.g. [A+B+C] × [D+E+F]) then wanting to audition the 'fixed ground' pickup in both groups ([A+B+C] × [D+E+F+G] and [D+E+F] × [A+B+C+G]) -- one would have to adjust 12 switches to rearrange the series stacking of the parallel groups, rather than just two switches required to swap any of the other pickups between groups.

Which is best? Rotary, Full Binary, or Binary Staircase?

1P8T Slide switches are also a potential alternative to the rotary switches, having the advantage of taking up less room, but the negative of requiring square rectangular holes.

---

A different approach in the case of a HHH setup, requiring only 7 busses, could be to forego binary idea altogether and instead opt for a partial ternary tree:

The second level consisting of either a single switch or a pair -- using DPDT (or 4PDT) ON/ON/ONs, or actual triple throw switches if available.

[reTrEaD]

You brought up some excellent points, Yogi B

Dec 7, 2018 15:44:34 GMT -5 Yogi B said:

We'll need thirteen 8-position rotary switches for a seven coil guitar. (one lead of one of the pickups can be wired directly to ground).

That might be a worthwhile optimisation in an electrically switched circuit, however I have to disagree when expected to manually perform the switching. At the stage of having teens of switches (or groups thereof), I don't think the saving of one switch is all that important -- were already past the point where most would likely be overwhelmed. And I'd say the disadvantage of adding a fourteenth switch is outweighed by the advantage that the one 'fixed ground' pickup then doesn't require special treatment.

For example, imagine setting up an essentially series combination of two parallel groups of the other six pickups (e.g. [A+B+C] × [D+E+F]) then wanting to audition the 'fixed ground' pickup in both groups ([A+B+C] × [D+E+F+G] and [D+E+F] × [A+B+C+G]) -- one would have to adjust 12 switches to rearrange the series stacking of the parallel groups, rather than just two switches required to swap any of the other pickups between groups.

I agree. There's a false economy I suggested there. Save a bit on switches, forfeit ease of use.

---

Which is best? Rotary, Full Binary, or Binary Staircase?

1P8T Slide switches are also a potential alternative to the rotary switches, having the advantage of taking up less room, but the negative of requiring square rectangular holes.

That's a somewhat attractive choice. Although slide switches with more than two positions tend to be a bit fiddly, if the panel graphics are right, they could be very visual and intuitive. Plus, there would be exactly one switch to operate for each node. There's a lot to like there.

A ganged pushbutton array, where when one switch is pressed, releases the previous selection, would be very easy to operate. But those things take up way too much room.

---

A different approach in the case of a HHH setup, requiring only 7 busses, could be to forego binary idea altogether and instead opt for a partial ternary tree:

The second level consisting of either a single switch or a pair -- using DPDT (or 4PDT) ON/ON/ONs, or actual triple throw switches if available.

If synthesizing a 2P3T switch for single switch to serve the second level, what would we call that?  A truncated full ternary?  idk.

I think it might be preferable to opt for a tree instead (with three separate switches).  And if we twist things about, it would be somewhat visual and intuitive if we put the three switches in a row, with the switch for the first level between the other two.

It would still be a partial tree ...




Waddaya think?

[Yogi B]

Dec 8, 2018 13:20:40 GMT -5 reTrEaD said:

And if we twist things about, it would be somewhat visual and intuitive if we put the three switches in a row, with the switch for the first level between the other two.

I did consider that arrangement, symmetry is a plus, but my concern would be over how one assigns the busses to the switch positions.

In the asymmetrical version it seems most logical to me that the least branching path (0X) be connected to the ground bus, thereby requiring at most two switches to flip in order to guarantee deselection a pickup by switching both ends of the pickup to the ground bus. On that basis, the 22 path is probably the most sensible place for the hot bus.


For the symmetrical arrangement I see two obvious options:

• keep the ground bus as the least switched branch (i.e. in the middle), but then where do we put the hot bus? The choice is arbitrary and ruins the symmetry;
• keep the hot and ground busses as the outermost branches, this keeps the symmetry but makes one of the floating busses seem undeservedly special.

[JohnH]

This is great Blue-sky thinking, and deserves its space without too much constraint by practical issues. In due course, out of this 'Apollo space program', may come the better 'non-stick frying pan' that we can all use.

Towards that end, would any of the design emphasis change if you considered arrangements with say no more than 6 coils, and no need for more than say 4 at a time? Within those ranges, I would expect to find virtually every combo that has ever been used in practice.



….or, consider that there are only ever three pickup positions, B, M and N, and each of those can only ever have 1 or 2 coils (ok so set aside triple-buckers). So an almost fully versatile scheme might involve a 6x rotary for each pickup position (if it be a 2-coil one), followed by whatever is the best fully comprehensive wiring for three pickups. Though there's a few that this couldn't achieve, such as a neck coil in parallel with one coil of a bridge HB, with the other B coil in series.

[reTrEaD]

Dec 8, 2018 19:05:33 GMT -5 JohnH said:

This is great Blue-sky thinking, and deserves its space without too much constraint by practical issues. In due course, out of this 'Apollo space program', may come the better 'non-stick frying pan' that we can all use.

Towards that end, would any of the design emphasis change if you considered arrangements with say no more than 6 coils, and no need for more than say 4 at a time? Within those ranges, I would expect to find virtually every combo that has ever been used in practice.

John, if we're still talking about a mule designed around a multiple bus structure, I don't think there's much economy to be gained by limiting the number of coils chosen at one time. Since each end of each coil is managed by a switch (or group of switches) the total number of coils dictates the number of groups. With a six coil complement, you'll want 12 switches (or groups). We can limit that to no more than 3 series links (a total of 5 buses) and that will limit the complexity of a group to some degree. But that really doesn't limit the number of coils selected at one time to 4. It just means that there can't be more than 4 series elements. We could use either or both of the remaining 2 coils in parallel with the entire circuit or any one series element.

[reTrEaD]

Dec 8, 2018 17:57:53 GMT -5 Yogi B said:

Dec 8, 2018 13:20:40 GMT -5 reTrEaD said:

And if we twist things about, it would be somewhat visual and intuitive if we put the three switches in a row, with the switch for the first level between the other two.

I did consider that arrangement, symmetry is a plus, but my concern would be over how one assigns the busses to the switch positions.

In the asymmetrical version it seems most logical to me that the least branching path (0X) be connected to the ground bus, thereby requiring at most two switches to flip in order to guarantee deselection a pickup by switching both ends of the pickup to the ground bus. On that basis, the 22 path is probably the most sensible place for the hot bus.


For the symmetrical arrangement I see two obvious options:

• keep the ground bus as the least switched branch (i.e. in the middle), but then where do we put the hot bus? The choice is arbitrary and ruins the symmetry;
• keep the hot and ground busses as the outermost branches, this keeps the symmetry but makes one of the floating busses seem undeservedly special.

Before I continue rambling, I should pause for a moment to explain how I'll describe the perspective of the physical arrangement of the switches.

We'll view the guitar from the front side with the long axis as vertical, with the head at the top.

Each group will have it's switches arranged in a row, across the short axis of the guitar.
The row for (+) of the neck-most coil will be closest to the neck (top).
As we move successively down the long axis of the guitar, when we get to the last row, the row for the (-) of the bridge-most coil will be closest to the (bottom).

The direction of travel the toggle of each switch will be in the short axis of the guitar (left-to-right).

---

I like the asymmetrical arrangement you first proposed for a 2-switch row.
Let's say the first level switch is on the left, the second level switch is on the right.

With the toggle of first level switch in the left position, the position of the second level switch doesn't matter.
Make that the connection to the ground bus? I'm all for that.
And with the first level switch in the right position and the second level switch in the right position, making that the hot bus is also logical.

---

For a 3-switch row (symmetrical), the middle switch in the center position does imply some degree of 'undeserved specialness' to the middle one of the floating buses (because the other switch are now both in 'don't care' situations, but center positions are rather un-special because of the extra care it takes to get to that position. We can't just slap the toggle one way or the other.

So I'd opt for having the hot and ground as outer-most branches in the 3-switch symmetrical group.

---

When we reduce the number of buses by one (for a HSH), the most logical way seems to be having a binary in the first level and a ternary (or ternaries) in the second level. In that case, there are no 'don't cares' if using a single 2P3T for the second level.

And if we do a symmetrical 3-switch arrangement, there are equal number of 'don't cares' on both sides.

[Deleted]

You do know there is hexadecimal switches (Basically 4 x SPST Switches all in one package making 16 combinations)

I've been using then as selector for pickups and capacitors
10nF .. 1) 4n7F 2) 12nF 4) 22nF .. gives me 10nF, 14n7F, 22nF, 27n7F, 32nF, 36n7F, 44nF and 48n7F .. as capacitors sum up.
Used 8) as a splitter , the tone pot is 250k lined making it 125k but when 8bit is selected it becomes back to 250k.

This can be used to select pick ups or bring in effects 1,2 and 4bits for pickups and 8 for preset filter or in my case (design not made it yet) for a bass tone pot, as I don't have C1M pots .. I'm using 500K resister with a 500K potentiometer, the 8Bit from the selector is another 500k that links to the 500K potentiometer, so making it 500k or 250k before a 500k potentiometer with a 2n2F linked around that .