can anyone suggest a simple opamp switcher?

[ulysses]

hey guys

has anyone designed or used some sort of 8pin opamp changer so you can switch between opamps for testing purposes

doesnt have to be pretty, just has to be able to change between several opamps in the "lab"

idealy it will be wired to an 8pin dummy IC and i can just slot that into a cradle in a test circuit.

cheers
ulysses

[gez]

I can't think of any easy way of doing this other than to parallel another amp and associated circuitry; then use a DPDT to switch between inputs and outputs and A/B two amps at a time.

[gez]

Hmm, possibly use three 4053s.  That would still only give you the option to switch between 2 op-amps, though.

Edit: this should give you some ideas:

http://www.geofex.com/Article_Folders/cd4053/cd4053.htm

[Ben N]

I would think that you don't have to switch all 8 pins. Power supply, for example, should be ok left on, as long as the disconnected ICs are idle. Beyond that, as Gez said, you probably need some kind of CMOS-based switching, although the on/off resistances of the CMOS switch may skew your results somewhat (sorry, too lazy to look it up).

Ben

[moro]

According to this:

http://www.beavisaudio.com/Projects/ScreamerLab/

you can have them wired in parallel and just use a DPDT to toggle the input / output of each one.

[gez]

I would think that you don't have to switch all 8 pins.

That was my initial thought too, but then I edited my post to suggest the use of 3 chips instead of 2: my concern being that the amp not in use would oscillate if the power pins were left connected, which might cause spikes/noise.

[gez]

According to this:

http://www.beavisaudio.com/Projects/ScreamerLab/

you can have them wired in parallel and just use a DPDT to toggle the input / output of each one.

Actually, considering the high input-impedance of most op-amp inputs, you'd probably only need to switch the outputs (all other pins to be tied to the respective pins of their parallel counterpart).  Good point!

Edit: though oscillation might be a problem again...

[moro]

That was my initial thought too, but then I edited my post to suggest the use of 3 chips instead of 2: my concern being that the amp not in use would oscillate if the power pins were left connected, which might cause spikes/noise.

I'm sure this is a stupid question, but why couldn't you toggle the power pins? (Instead of the in / out.)

[Ben N]

That was my initial thought too, but then I edited my post to suggest the use of 3 chips instead of 2: my concern being that the amp not in use would oscillate if the power pins were left connected, which might cause spikes/noise.

You could probably take care of that by grounding the input(s?) of the unused IC, but then, that adds a switch, too. What if only one power pin (+ or -) was switched? (I am way out of my depth here, just cogitating.)

BTW, a single CMOS throw can control a DPDT relay, so one 4053 could handle switching 6 pins two ways; any non-signal switching (like power) could be done electronically.

[Sir H C]

You can't toggle power because there is ESD protection circuits in the opamps which amount to diodes from the rails to the inputs and outputs.  Also there are wells and other structures inside the IC that are diodes so if the top rail is below the input you could get all these conducting.

[Ben N]

You can't toggle power because there is ESD protection circuits in the opamps which amount to diodes from the rails to the inputs and outputs.  Also there are wells and other structures inside the IC that are diodes so if the top rail is below the input you could get all these conducting.

Huh? 

[Sir H C]

Ok, to protect the inputs and often the outputs of op-amps, they have large diodes internally that connect to the V+ supply (usually) and ground.  So when you get an ESD zap, the current flows through the diode to the supply or ground and safely away.  So if V+ is switched off, you can get some weird paths to ground going.

Also when they make an IC they start with a wafer of doped silicon.  Usually it is doped P-type, and to make PMOS transistors (and NPN transistors) they have to put in some N doping for the connections (PMOS devices are in N type silicon and NMOS devices in P type).  So now you have an N-P junction, and therefore a diode.  Add a third terminal in there, and you have a parasitic bipolar transistor.  Have the supply wrong, you can "latch-up" the part and draw a lot of current.

[ulysses]

perhaps the easiest way would simply to use 2x 4pdt switches and switch both switches at the same time?

as i said, it doesnt have to be pretty.. just has to work im not trying to sell it.

re: two identical circuits with different opamps.. yeah i thought of that.. but it seems like a hassle to have to build two circuits to test opamps as i usually only make 1 of anything unless i really like it and make a one for a freind or two.

cheers
ulysses

[Sir H C]

If you want to test, why not just socket the opamp and then pull it out and try another?

[MikeH]

I've used the opamp switching on the beavis audio page, and it does work.  You can really tell the difference between a low noise and not-exactly-low-noise type opamp when toggling back and forth.

[Ben N]

Mike, you mean this?:
http://www.beavisaudio.com/Projects/ScreamerLab/

That would work fine IF either the opamps are high input-Z (JFET input) types, OR there is a low output-Z (BJT or opamp) input buffer feeding the opamp stage being tested. Otherwise, paralleling the inputs might lower the input imedance enough to load down the signal noticeably, I would think. In a TS it is not a problem.

Ben

[Paul Perry (Frostwave)]

If you google for 8P 2T switch, you will find that slide switches are reasonably common in this configuration.
I saw surplus ones (rotary & slide) for $12, that's without making a proper search.