Help with replacement for a rotary switch

[Vegard Prøysen]

Hello.

I hope you can help me out with this one. I am looking to build a multi path signal switch for my guitar gear (rack).
I was going to use two simple rotary switches, one 5 pole and one 3 pole, both 3 or 4 throw. I really only need three positions but 4 is nice to have in case I add a unit later on.
I would also add a 9v circuit for some leds to indicate what I have chosen.

Do any of you know of any "chips" that can replace 5 pole 3 or 4 throw rotary switches and not degrade the audio sound?

My initial plan was to do the following with the switches. Let me use 5 pole 3 position and 3 pole 3 position as an example here:
Switch #1
Pole 1: Guitar -> 3 x pre-amps
Pole 2: 3 x pre-amps -> Power amp
Pole 3: Leds
Pole 4: 3 x pre-amp effects send - > Pole 1 Switch #2
Pole 5: Effects signal -> 3 x pre-amp effects return

Switch #2
Pole 1: Effects send from pre-amp -> 3 x effect units
Pole 2: 3 x signal from effects units -> Pole 5 Switch #1
Pole 3: Leds

A pretty simple circuit, but it is actually not that easy to find these switches. When I do find them they cost like 3 bucks and are some old military stuff from Soviet or cheap looking ones from China so I am not sure they're durable or any good for audio circuits. Also, it's cool to have one of those fancy pantsy small smooth turning knobs that just controls a "chip".

I have an AVR dragon if I need to program anything, but I know too little about the different "chips" to go out and just pick one up.
Any help here is appreciated.
Thank you.

Vegard

 

[KrisBlueNZ]

Hello Vegard and welcome to Electronics Point

Digikey is a great resource for finding out what components are available in a given category. There are two relevant sections on the Digikey web site:

1. Integrated Circuits (ICs) > Logic - Signal Switches, Multiplexers, Decoders: http://www.digikey.com/product-search/en?FV=fff40027,fff802b3,2dc09bf,2dc1141,2dc1316&stock=1&quantity=1&pageSize=250

2. Integrated Circuits (ICs) > Interface - Analog Switches, Multiplexers, Demultiplexers: http://www.digikey.com/product-search/en?FV=fff40027,fff802ff,1b80139,1b801cc,1b801cd,1b801d4,1b801d6,1b801d7,1b80222,1b8028b,1b8029c&stock=1&quantity=1&pageSize=250

The classic analogue switch is the CD4066, part of the old CMOS 4000 series. (Other manufacturers use different naming schemes: HEF4066, HCF4066, MC14066.) There's also a 4016 which has poorer specifications, so ignore that.

The CMOS 4000 family also includes the 4051, 4052, 4053 and 4067, which provide analogue multiplexing/demultiplexing with different numbers of poles and positions.

All CMOS 4000 series analog switches perform better at higher supply voltages. I would suggest a 12V supply. Signals also need to be biased at around half the supply voltage, and should have high levels, for best signal quality.

There are also many and varied audio switching devices from Analog Devices, Fairchild, Linear Technology, Maxim, and others. If you really want to waste an afternoon, have a thorough look through those selection tables I linked above. There are links to manufacturers' data sheets for almost all the devices listed.

You can reduce the number of devices in the table by narrowing down your selection parameters. For example, select only through-hole devices, or devices in leaded packages (i.e. not "no-lead" packages that are impossible to hand-solder), or devices with a single supply, etc.

 

[Vegard Prøysen]

Hello KrisBlue.

Thanks a lot for the info.
I've decided to go for the CD4066. I assume it doesn't matter much which brand I go for, or do you have any bad experience with any of them?
One thing though, I was looking at the data sheet from Texas Instruments and it says that on-state resistance decreases with the voltage. If I understand it correctly this is the resistance between my input signal and output. I want that resistance as small as possible since these are audio signals I am processing. You mentioned biasing a 12V down to 5-6V, but wouldn't it be better to run a stable 12V into the chip if the on-state resistance increases with lower voltages?

Anyway, I'll go for 8xCD4066's, two rotary switches and a circuit to lower whatever input voltage I feed into it.
When everyone finds out they really need a rack unit like this and I have to start mass producing, I will be sure to put a picture of a kiwi as a logo and name it KB100 or something like that

Vegard

 

[Vegard Prøysen]

After reaing your reply again it seems you're saying I should power the switch with 12V and bias the control voltage.
That's the plan, thanks again.

 

[KrisBlueNZ]

Yes, power the 4066es with 12V (or 15V if it's well-regulated and will never go much higher) and bias the signals on both sides of the switches at half supply voltage. This gives you the highest allowable signal amplitude, and having all points biased from a single half-supply rail means you will avoid pops at switching times.

The control voltages need to swing all the way between VSS and VDD. You may find that slowing down the transitions at the control inputs with an R-C network will reduce pops at switching times too.

All CMOS switches have lower ON-resistance at higher supply voltages, not just the TI one. But some brands to have lower typical and maximum ON-resistance so check out as many brands as you can find data sheets for on Digikey.

 

[Vegard Prøysen]

Thanks again.

I am not sure I want to bias my signals on both sides though, as that may alter my sound (?). If my signal drops I might have to buffer it and then I'll do it, but I am not interested in altering my signal too much (or at all if possible).
I have ordered some parts now so I will be putting together a prototype once they arrive. I ordered some DG411DJZ's too. They had very low on-state resistance so I'll build two circuits and compare the 4066 to this one. I'll also try biased and non-biased signals.(I do however prefer minor pops to a change in my signal, lower and higher are bot equally bad).

Once I've decided on the layout I will make a PCB and install everything in a 1 unit rack case.

 

[KrisBlueNZ]

I am not sure I want to bias my signals on both sides though, as that may alter my sound (?).

No it won't; if anything, it will prevent changes to the sound.

The inputs and outputs of analogue switches will always have some DC level.

For best performance with the 4066, this DC level should be half way between VDD and VSS (unless the data sheet specifies otherwise). You can do this by using VSS as your signal ground and shifting inputs and outputs up to a half supply voltage using capacitor coupling, or you can actually run the switch ICs from split supplies and bias the inputs and outputs at 0V. In the latter case, the 4066 will need level shifters to generate the full VSS-to-VDD swing for the control inputs.

The DG411 does not need level shifters and is intended to run from split supplies, for convenience.

Feel free to post your circuit designs here for comment.

 

[Vegard Prøysen]

Ok, maybe I am thinking of biasing as a buffer. Thanks for your help so far.

I am waiting for the parts to arrive now and I will try a few circuits when they do. I will post the design I am happiest with. It can probably be improved, so comments will be appreciated.

 

[KrisBlueNZ]

Cool