Simple opamp problem (biasing) - can you help?

[peskywinnets]

I wish to extract the positive half of a signal & amplifiy it, then rectifiy it.

I had thought this circuit would work...



....where I reference the input to ground, signal comes in, only the +ve swing would be acted upon & amplified 50x.

It doesn't work? (no output at all)

it does work (ie I see a signal as expecte don the output) if I reference the input to half supply (the usual way of doing it).

I'm using a Microchop MCP6004 opamp.

I really wanted to save on a few components (2 resistors & a cap!) by not boiasing the input to half VCC as the final board will go inside a guitar (and I want it to be as small as possible) but not only that I want the output to be 0V quiescently (rather than half VCC)....of course I could add a blocking cap after the opamp output, but that's another component!!


So I'd really like to get the circuit above working (or an explanation why it won't can't)

Thanks for your help!

[richie00boy]

Unless you have a special op-amp, you can't give it a signal outside of its supply rails, maybe you blew the chip.

The output will never be 0V because again unless it's a special op-amp it can't swing that far.

I'm afraid your circuit needs pretty much a doubling of components to be correct, but you do seem to be aware of these just that you didn't want to use them

[peskywinnets]

to put this into context...the incoming signal is about 400mV to 800mv peak to peak ...therefore it's swinging negative by half of that - it hasn't blown the chip (like I say, when I bias it to 1/2 VCC I see an output).

i'm just wondering what's going on where I can't simply reference the input to ground.

The MCP6004 is a rail to rail opamp....why can't it swing down to 0V?

re needing a lot of work(?), one input cap, two biasing caps, one decoupling cap in the negative feedback chain...job done! (but alas, like I say, that would incur a space penalty)

[richie00boy]

There are three kind of "rail to rail" op-amp.

1. Input can be taken to either rail.
2. Output can be taken to rail.
3. Both of the above.

Not many are all three, which would be what you need.

edit: just downloaded the datasheet and yours does indeed do all three. What exactly is being connected to the non-inverting input? Just straight across a pickup?

[sofaspud]

If you only take the positive half of a signal, you've already rectified it....

[djk]

http://sound.westhost.com/appnotes/an001.htm

[CopperTop]

I'm sure your schematic is just an illustration or for simulation or whatever, but I did notice that the pin numbers shown don't seem to correspond to the MCP6004 quad op amp's pin assignments. Just thought I'd mention it in case you're not aware of it...

[djk]

His circuit can't possibly work, read the above app note for a circuit that will.

[CopperTop]

@djk

He's trying to extract a crude measure of amplitude using a minimum of components. The op amp behaves well if the input signal goes below the negative supply according to the data sheet, and he is limiting the input current. Granted it won't be 'precision' because of the diode drop, and we don't know what follows it, but it should still basically work shouldn't it?

[djk]

Why should it?

It will only pass audio peaks above .7V, and may stick to the rail from lack of input bias.