CMOS switch diode query

[Mark Hammer]

I'd like to try and draw a more compact PCB layout for the electronic switch circuit that is project #15 in Anderton's EPFM. (https://msu.edu/~dougl126/Electronic%20Projects%20for%20Musicians.pdf)

The circuit uses 10 different diodes to both drop a +/-9V supply down to +/-7.5 (or thereabouts) as well as shape the making and breaking of switch action to eliminate audible clicks and pops.

Although the bill of materials and schematic stipulate 1N4001 diodes throughout, I was wondering if use of 1N914/4148 would have any bearing whatsoever on the functioning of the circuit.  Certainly use of 1/8W resistors, radial caps for the electros, and small silicon diodes would allow for the circuit layout to be much more compact than it presently is.

I thought that the larger 1N4001s would be appropriate for supply drop, especially since I would be likely powering multiple switch circuits from the same common supply voltages.  On the other hand, if 1N914s can be safely used for all 10 diodes, that'd be nice.

I know there are plenty of other great electronic switch circuits out there, and that there are other sorts of switches, be they relays, JFETs, or even LDRs.  What I like about the EPFM circuit is that it provides easy flexibility to use latching or momentary switching of the same circuit, without any changes, simply be running parallel lines out to a latching and non-latching SPST footswitch.

[Digital Larry]

I think your biggest concern would be the forward current.  I think 1N914's are rated for 100 mA and 1N4001 10x at 1 amp.  I doubt this circuit uses even 100 mA based on a quick look see.

[Mark Hammer]

Yep.  The book says it draws +18ma/-8ma.  So let's say I'm good for four 4016 chips drawing from the same supply.  And if one assumes that every board fully replicates the EPFM circuit (i.e., includes the supply drop-down diodes), then each board is well within the operating limits of the smaller diodes.

I'm hoping to use this for a floor multi-FX unit.  Funny how little things can invite more extensive obligations.  My late mother would buy a table lamp she liked, but the lamp would require new end-tables to go with it, and the end tables beget other living-room changes.  In my case, I bought a large used Hammond sloped-front chassis for $10 that someone had obviously used to rehouse a whole bunch of effects; many of them likely Boss pedals.  I thought "Bargain!", and wanted to do the same thing, except with my own choice of FX and circuits, and some external loops and options for external switching.  And that begets developing switching boards. 

[amz-fx]

Yep.  The book says it draws +18ma/-8ma. 

That seems like a lot of current, but even so, you could use red LEDs instead of the 1N4001 diodes for a single switch. The drop across the LEDs will put you in the right voltage range and they can handle 18ma.

If you want to power multiple copies of the circuit, then you would be better off going to a regulated supply. 

regards, Jack

[alparent]

If you pick the right CD4016 (I would use CD4066 instead) you can skip 6 of those diodes.
This switching is very interesting. We could work together on this.
I could cook up a board in eagle and make this with SMD parts. Even smaller.

[R.G.]

Use 1N4001s for the power supply droppers. They'll wipe off the requisite volt and a half with no worries about current capability, even for many circuits. One set of 1N400x diodes will suffice for many electronic switches. The others willl be OK as 1N4148/1N914.

(1) how small do you want to go
and
(2) are you allergic to SMD?

[armdnrdy]

Remember what year this book was published?....and which parts were easily sourced....

Alain hit it on the head.

Take note of project section "modifications".

"A 4016 with a "B" suffix (such as CD4016B) indicates that the part can withstand ±9V. If you use this type of IC you may replace D6-D11 with wire jumpers."

Look Ma....no diodes!

[R.G.]

Mark is one of the few people I'd trust to actually go and find "B" series chips for his work.

Any time I put things up on the net, I assume that beginners who can barely read the part/type numbers will be building it and get everything possible wrong - not because they're not motivated, just because they don't have an abundant pad of mistakes they've already made.

@Mark: if it's for your use only, you can use the "B" chips. If anyone else ever might put in "A" chips, stick in places where either diodes or wires could go.

Actually, one way to go would be to run these from +8V ( i.e. 9V minus a diode) and -7V from a charge pump running from the +8. Actually, it's a target-rich environment.   

[alparent]

If you pick the right CD4016 (I would use CD4066 instead) you can skip 6 of those diodes.

Got this from the book:
"Additionally, you cannot substitute for the 4016-
type IC specified in this project. While some switching
ICs are considered as identical plug-in replacements for
the 4016 (4066 and 4116, among others), they will not
work in our circuit."

But doesn't say why? Can anybody educate me?

And this circuit uses a latching switch (I don't like the click of a latching switch)
I would use this http://diy.thcustom.com/switching-relay-true-bypass-circuits-not-using-a-microcontroller/ The NE555 one.
Then a silent non-latching switch could be used.

I've already done this to activate relays. But using the 4016 instead of a relay is even more appealing to me.
I could even design the board so you can leave the 555 parts out and use as a mom. switch.

I think I know what I'm breadboarding tonight!

[R.G.]

IIRC, Anderton used the CD4016 because it had a direct path from one of the enable pins to the gates of one of the transmission gate switches. This let hims slow down the sudden transition on the gates and thereby eliminate the feedthrough click of the control gate signal into the audio path. The 4066 and other chips didn't have the unbuffered path from a pin to the control gates inside the chip, so slowing down the pin switch voltage didn't slow down the control signal at the actual MOSFET gates inside the chip.

I messed with this circuit back when, and sure enough it can be tinkered in. I moved on to the CD4053 for signal switching, and scratched out ways to minimize the clicks with it. Different top level solution.

[alparent]

Thanks for that info 

[Mark Hammer]

Thank you, gentlemen.  This is all very useful information.

And yes, it is for my own use (though I suppose others would be welcome to make use as well).  I have a bunch of 4016 and 4066 chips at home (as well as 4053s), but never bothered to look at the suffix prior to today.  If I'm lucky, it's possible some of them are B-type.  I'll cross my fingers until I get home from work.

And while we're all here, let's show some collective love towards Commander Keen, who has been through some surgical aggravation of late, and is too much of a soldier to complain.  A speedy and complete recovery, brother.  May your tap-dancing career resume soon.   

[PRR]

> seems like a lot of current

Looks like 12mA just for the LED. With modern LEDs you could probably run a lot less.

I'd watch for cap discharge paths through diodes, but I see only 7mA so '914 diodes should be over-ample.

I think he got a huge bag of '4001 diodes. Also note that the front and back jacks way out-size the circuit even with the old-school parts.

[Mark Hammer]

As luck would have it, it seems like all the 4016 chips I had bought some time back were of the B (BCP) type.  Huzzah!