Author Topic: Diodes in ground leg?  (Read 2487 times)

PBE6

Diodes in ground leg?
« on: July 14, 2015, 02:47:25 AM »
Just had a wacky idea about putting diodes in the ground leg of a non-inverting opamp and wondering if it serves any useful purpose. I was going to breadboard it, but it's too late for me to be making noise at the moment so it'll have to wait until tomorrow.

The starting point for my idea was a circuit like this:

With one diode in the ground leg only half the waveform would be affected, so it's probably best done with a diode pair instead.

The total current through the forward-biased ground leg diode will be:

iD = vin / (RD + Rg) = vin / (vD/iD + Rg) = (vin - vD) / Rg

For small signals the diode doesn't conduct much and the diode resistance is very high, so the circuit acts like a buffer. As the signal gets bigger, the diode begins to conduct more and the diode resistance decreases dramatically. This has the effect of ramping up the gain very high for signals above a certain threshold. In some sense this is akin to an expander because it increases the dynamic range of the signal. It will actually increase the signal a bit too much with certain resistor pairs, so I'm thinking that clipping diodes (either in the feedback loop a la the Tubescreamer or straight to ground a la the MXR Dist+) are a must.

Anyhoo, I tried simulating the Tubescreamer-esque variant and it makes a neat shape:

As you can see it follows the larger signal for a portion of the cycle before shooting up past a certain threshold. It looks like a Tubescreamer waveform (shown below) except more pinched:

As I said I haven't tried this out in real life yet so I don't know how it sounds, but it looked cool so I thought I'd share.

Anyone else tried diodes in the ground leg before? Useful or useless?




teemuk

Re: Diodes in ground leg?
« Reply #1 on: July 14, 2015, 04:35:15 AM »

Rob Strand

Re: Diodes in ground leg?
« Reply #2 on: July 14, 2015, 05:12:13 AM »
You need to add a discharge path on the cap using say a resistor, otherwise it will just stay charged.
Send:     . .- .-. - .... / - --- / --. --- .-. -

duck_arse

Re: Diodes in ground leg?
« Reply #3 on: July 14, 2015, 09:03:59 AM »
what waveform if you add a normal R//C leg in parallel to the diode leg?
by request

PBE6

Re: Diodes in ground leg?
« Reply #4 on: July 14, 2015, 11:05:31 AM »
@duck_arse it depends a bit on the resistor values, for Rg2 close to Rg1 it looks pretty much like a Tubescreamer:

For Rg2 much smaller, it gets a bit mangled and shifted (though I'm sure a large part is due to filtering alone):

PBE6

Re: Diodes in ground leg?
« Reply #5 on: July 14, 2015, 11:07:32 AM »
Thanks teemuk, I will take a closer look tonight! Dumb question, what's a good DIY Stompboxes analogue for a current source?

samhay

Re: Diodes in ground leg?
« Reply #6 on: July 14, 2015, 12:54:27 PM »
In my experience, it is best to put a resistor in parallel with the diodes. Also, you need to hit the input of the op-amp with a fairly hot signal to get any effect. In the end, I didn't find it very musical, but best results were with Schottky diodes and a boost in front.
I'm a refugee of the great dropbox purge of '17.
Project details (schematics, layouts, etc) are slowly being added here: http://samdump.wordpress.com