Quick - and clean! - bias voltage

[R.G.]

Use an LM386. The LM386 self biases to half the power supply voltage from 4 to 15V supplies, and already has its inputs pulled to ground by internal resistors. You only need to hook up power, ground and the output pin becomes 1/2 the power supply.

It's a much more solid reference voltage than the resistor/resistor/cap in many effects.

[Tim Escobedo]

Great tip, R.G. It never ocurred to me!

[puretube]

great idea! especially since it`s very low impedance...

[Paul Marossy]

That's useful information. Thanks.

[downweverything]

how much extra current does it use?

[R.G.]

Ah - a sharp one!

The static current is 4ma.

You can get lower static current with resistors. The trade off is the low impedance. Some circuits really, really need the low impedance on a faked ground, which is what most reference voltage circuits are. I have seen setups that cannot be stabilized right with resistor/cap references. Buffering it with an opamp or other active chip fixes them right up.

[downweverything]

yeah prob an awesome idea for stompboxes because they dont require tons of current .  ive thought about using those in amplifiers before but for something with high current draws those devices have limits.  never thought about using them for stompboxes though.  awesome idea. thanks.  i think im going to do that on my next one.

[Peter Snowberg]

Great idea R.G.!

Take care,
-Peter

[Tim Escobedo]

In my tests, the low power JRC versions seem to use less current. As little as half what the Nat Semi versions do. May be worth exploiting.

[will]

Hi,

Do you think an unused 1/2 of other dual opamps like TL072 or a 4558 would work as well?

Regards,
Will

[puretube]

...if you hook up a resistive voltage-divider in front of it, and make it a voltage-follower: yes, but not such a low output-impedance as the 386.

R.G.: how about shorting the input-side of the 386 to ground with a cap, so as not to have it floating AC-wise, and thus picking up "noise" ?

[ExpAnonColin]

Nice... in all of my 386-playing-with, I didn't notice.  Thanks for the tip!

-Colin

[R.G.]

how about shorting the input-side of the 386 to ground with a cap, so as not to have it floating AC-wise, and thus picking up "noise" ?

I worried about that before I posted. It's one of those ideas that may be good, may not offer many advantages except in unusual circumstances. Here's why:

The inputs of the 386 are already tied to ground through internal 50K resistors, so the input is not exactly floating. The resistors are not that big, so noise pickup is likely to be small. Certainly noise pickup is going to be as small or smaller than the 386 used as a gain stage with no input, which is acceptable.

The small resistance value also means you need a biggish cap to shunt the input to ground, which negates some of the advantage of the one-chip-solution.

I suspect that you would get almost the same result if you grounded the input, just hard tied it to ground. I *think* that it would work just fine, but I don't have one on a breadboard to try it with right now.

[petemoore]

solid and easy...
 I always end up using board space or having the divider [say two 10k's and and two 10uf's] creating a little over the board sculpted 'motif'...
 A quick tally of parts count/pricing...the two caps alone cost...the low profile and EZ hookup make this the divider I'm trying out on the next one.

[David]

This bias trick is cool!  I can't wait to try it out on my Flatline.  Setting up the bias network on the breadboard was driving me nuts (a short trip in my case!   :mrgreen: )  Will this even work with the lowly 386-1s from Rat Shack, or do we need to use 386-3s?

[smoguzbenjamin]

the n1 worked for me. Nice trick RG!

[javacody]

David,
   those lm386 1's at Radio Shack are actually lm386 2's. They'll take 12 volts just fine with no problems.

[Phorhas]

The LM386 self biases to half the power supply voltage from 4 to 15V supplies

what about power supplies of the 18v to 40v?

is there a simple, clean and stable way of doing that ?

[Peter Snowberg]

The best way there would probably be to use an opamp that has low impedance and can take the voltage. An NE5534 will run up to 44V and has a very low output impedance with a current drain of typically 4ma or so. Add a pair of 100K resistors as a divider with a filter cap at the input and there you go.

[R.G.]

what about power supplies of the 18v to 40v?

You could use the same trick. All of National's power amp chips self bias to half the supply, I think. They certainly could be biased there, anyway. The problem is that for the higher voltages, it's no longer cheap to do this.

The LM1875 (http://www.national.com/ds/LM/LM1875.pdf) will go to 50V total, but I think you have to use a couple of resistors and a cap to bias it at half. It's a TO-220 package.