CMOS 4053 circuit - please comment on biasing strategy

[tysonlt]

Hi All,

As I understand it, this would be how to use multiple 4053's to create an effects bypass looper, with one chip per loop.

(Sorry about the quality -first go with TinyCad)

Questions:

1) Assuming the R and C values are as specified in R.G's article, would this be the correct way to do it for a 9V power supply?

2) I have read some comments on the forum (which I can't find now!) saying that it is not safe to assume things about the caps inside the pedals, and for this and other reasons it would be better to use a bipolar power supply. Can anyone elaborate? Would that cut down on the components, or do you still need caps??

3) R.G mentioned elsewhere using an LM386 to generate the correct bias. Is this better than the resistor-based voltage divider shown on the 4053 bypass page? Would I still need all those Rs and Cs, or would I just set the 4.5 to Vee?

3.1) On that note, isn't there a pin on the 4053 to specify the bias voltage? Why not connect our 4.5 vref to that?

4) Would chaining these chips together result in 'tone loss', ie the multiple-boss-pedals-blanket-on-amp-effect?

5) I'm only using CMOS because I can't afford relays... unless the ones listed on tayda would work!!! Would relays make my life as an electronics amateur a bit easier? I understand flybacks and transistor drivers etc etc. Or are CMOS just as good when it comes to preserving guitar audio integrity?

5) Any other hints or words of wisdom (or parts donations ) are most welcome.

Thank you

[DavenPaget]

Of course relays would work , but being mechanical they can pop .
CMOS is alot better then JFETs of course but just don't lose power , but it any case it would revert state to being off .
TTG uses CMOS , RG uses CMOS , i haven't used CMOS switching yet but i am a official aficionado of CMOS-anything .

[tysonlt]

I was wondering about whether RG uses cmos or relays. I know he says that you can't hear the difference, but all of his looper designs still use relays. If cmos is so good and relays noisy and 'clunky', why would a black-belt like RG still use relays?

[DavenPaget]

I was wondering about whether RG uses cmos or relays. I know he says that you can't hear the difference, but all of his looper designs still use relays. If cmos is so good and relays noisy and 'clunky', why would a black-belt like RG still use relays?

It entirely depends what he does , CMOS are easier to get around since they are quieter and Relays are ... Extra parts involved .

[tysonlt]

It seems to me that cmos is more complicated... Lots of components for Biasing, whereas relays just need a transistor driver and a flyback diode...

My main goal is to know how to chain them together and get the Biasing right. Can I assume that most if not all commercial pedals will have input and output caps?

[tysonlt]

Can someone please comment on the basics idea of the schematic? I'm interested in whether the chaining is correct, and whether or not you need caps going to and from the pedals.

[PRR]

> whether or not you need caps going to and from the pedals.

You can't know WHAT will get connected.

Input and output caps are simple safe-audio precautions.

CMOS switch input pins need to be DC-biased about halfway between your supply rails, so that signal swing does not run into the supply rail and be clipped. For +9V this is +4.5V.

However this means, for the CD405x series, that your control voltage must also rise well above half of 9V, specified as 6V. If you use mechanical switches to the +9V this is fine. If you want to drive from 5V logic, it won't work. In that case the simplest plan is +/-5V, with logic and CMOS control running on zero and +5V, CMOS switches fed +/-5V, and CMOS switch ports biased to zero. With a cap because... You can't know WHAT will get connected.

[tysonlt]

OK, that sounds great...

Now excuse my ignorance, but does using a bipolar supply mean I can dispense with the Rs and Cs that RG used for biasing? And could you please show me where these safety caps should go? Sorry for the simple questions - just cos I use words like 'bipolar' doesn't mean I know what I'm talking about

Also would you mind me asking what the best way to make the +/-5v is? Is it just a 7805/7905 pair?

Thanks very much for answering. I have found the cmos switching aspect quite confusing really, I guess cos I haven't quite grasped the whats and whens of using filter caps on an audio signal.

[PRR]

> mean I can dispense with the Rs and Cs

No.

Or: only in theory in some ideal world.

It's the same Safe-Audio problem. Maybe 99% of your sources and loads leak "zero DC", but Murphy's Law ensures that you will "have" to use a DC-leaker on an important gig.

Speaking broadly: "ALL" audio box in/out jacks should be DC-blocked.
Some exceptions: loudspeaker DC-block needs BIG cap, so this is often omitted. Guitars to vacuum-tubes are traditionally not DC-blocked.



> the best way to make the +/-5v is?

It's probably about what ELSE needs power. It is not clear if you will control with simple finger-switches or simple CMOS logic or a micro-controller or some pig like a 80186.

The CMOS switch itself draws near-zero power. If the rail is only to feed the CMOS, this is one case where R-Zener, or even unregulated, is maybe "best".

[tysonlt]

Hi Paul,

Thanks so much for responding. I think you have answered a few of my beginner questions now!

I am building a MIDI + analog switching system. It will be housed in two units. The first unit, a gutted 1RU router box, will contain a bunch of footswitches and a PIC. The PIC will do two things: 1) send MIDI commands, and 2) send serial data to the second box. In the second box will be the 4053s, controlled by some 74HC595 latches. I plan to send the pulse signals straight over a standard network cable between the two boxes.

So the main box will just be a standard, 5v PIC system. For that I will use a normal 7805 setup. I also need to settle on a ADC multiplexing solution such as R-2R, as I will be using about 30 footswitches.

For the actual switching box, I plan to chain a couple of 74HC595 latches that will drive the control pins of the 4053s. I *think* that the 595 can accept TTL level signals from the PIC, and latch CMOS level outputs. Have to investigate this.

As for the connection, I was just going to connect the PIC output to the 595 input directly over some network cable. This seems to be how RG does it in his remote indicating switcher.

So in short, it is really one circuit, with a PIC at one end and CMOS at the other. My theory level is at the 'shifting ping-pong balls through pipes' stage at the moment, hence the frequent questions about biasing... (BTW I think while reading your last post I finally understood what biasing does... is provides a middle point like a taught string in the centre of a pipe, and the actual guitar signal only 'vibrates' the string by a certain small amount, but being biased it will never hit the edges of the pipe, ie the power rails, and clip. Correct?)

Regarding the resistors, RG's diagram at http://www.geofex.com/article_folders/cd4053/cd4053.htm only uses one resistor tied to Vref, but you are also using one tied to ground. Could you please explain this? Also, I really don't get what a series cap does. I thought that would eventually stop signal flowing once it was fully charged... or is the point that it will only fully charge if a certain amount of power comes through... kinda like a limiter?

Thanks for drawing a picture. It is my poor understanding of biasing that is generating so many questions about how to chain these CMOS chips together.

Apologies for the extremely long post  

[PRR]

> a taught string in the centre of a pipe

"taut"... or loose like a Slinky.... but exactly right. If the string lays on the bottom of the pipe, it can vibrate "up" but not "down", which spoils the effect. If it comes through the center of the pipe we have maximum possible vibration room for that size of pipe. If the string height is not for-sure known, we want some gizmo to "level shift" it from the as-arrived level to mid-pipe level.

If we did know the string came at "ground", we could add a 4.5V battery in series to shift it up halfway in a 9V pipe. However if we only want the vibration not the initial height, a capacitor will eventually (split-second) charge-up to the difference between the as-arrived DC level and our internal DC level.

[tysonlt]

Lol, 'taught' - I had changed it from 'tight'!

OK, so the way I understand it is that the 4.5v charges the cap via the 1meg resistor, until the voltage drop across the cap is 4.5v. (at this point would the flow to ground stop? Is that why it's called dc blocking?)

When a feeble little guitar signal comes along, the new electrons coming into the capacitor force an equal number of electrons to flow out the other side, into the chip. So a crude analogy is that the cap is acting as a sort of microphone...

I imagine caps as described here: http://amasci.com/emotor/cap1.html

So in this case, if my above understanding is correct the bias voltage coming through the resistor pushes the rubber diaphragm all the way to the left, until the voltages are equal and the 'water' stops flowing. Then the guitar signal pushes back on the diaphragm from the other direction, which results in the chip getting 4.5v, plus or minus our small guitar signal.

PHEW!!! Am I right? Do I get a biscuit??

[tysonlt]

If I have understood it, then that is very very clever, and I'm glad someone else thought of it as I never would have. This forum totally gets the gold medal for having lots of giants' shoulders to stand on.

I can see now why this would not affect 'tone' at all. It's just moving the bit of string as you say!

[alparent]

WOW! this is very interesting.

I'm also building something using uC and CMOS but needed a bunch of transistors to feed the CMOS trigers. ('cause 5v from the uC wasn't enaught)
But if I use bi-poloar ........ I can triger the CMOS directly from the uC. (Now 5v IS enaught!)

OK....now how do I get +5v and -5v from a 9v battery?

Thanks Paul

[tysonlt]

Bonjour alparent,

There is an article at geofex about getting bipolar supply from a 9v battery: http://www.geofex.com/circuits/+9_to_-9.htm

iPhone users: "There's an app for that". Audio hackers: "There's a chip for that".

I have also been looking at the maxim-ic range of chips. They are very interesting. They are designed specifically for audio switching and claim to be 'popless' - I assume without the need for biasing... ~alph on this forum uses the MAX395 successfully. I am looking at the MAX4573: serially controlled, 11xSPST! So you only need 3 uC pins to shift in the serial data, and then tell the chip to flick all of the switches instantly. Very cool. Also, they seem to be able to accept TTL-level switching signals directly from a uC, as they all run on 5v single, just like the PIC.

Ordered a couple of samples, see how I go

[alparent]

So I'll end up needing more stuff ($$$) then just using 0-9v and using uC to tansistors to CMOS.   

[tysonlt]

Yep, but I think there are some advantages to using bipolar............ don't fully understand it myself though.... there are some other posts about it somewhere...