valvestate the lm386

[soren_jepsen]

I'm fiddling around with the lm386, seemingly learning a lot  :lmao:

I stumble across this thread about valvestating the lm386 (one of ardrics posts, it has a link) , but i don't really get it. Could anyone explaine?

http://www.diystompboxes.com/smfforum/index.php?topic=45678.msg334715

Thanks 

[teemuk]

I think this page should explain it in quite detail:
http://sound.westhost.com/project56.htm
I had a thread here that introduced the thing but the image host has lost the images so I'm afraid it will not be very helpful anymore:
http://www.ssguitar.com/index.php?topic=75.msg842#msg842

I'm also afraid that this thing won't work that well with LM386 as it has somewhat an internally fixed gain setting. You should try it with an opamp that has it's output buffered with push-pull emitter follower or something.

Edit:
As you might have figured out the feedback configuration affects the frequency response of the power amplifier. How? This article should give a nice presentation:
http://milbert.com/articles/TvsT/tvtiega.html
(See figure 1.)

[blindsjc]

Hi friends,
Can someone help me to implement this variable impedance
trick on a LM3886 amp? Maybe this can make it sounds better
with a guitar.

The amps is a
http://sound.westhost.com/project19.htm
built and working...

can I add the info of
http://sound.westhost.com/project56.htm
with this amp just adding the feedback resistor?

Thanks for any info
Blindsjc

[teemuk]

The Rod Elliott article already explains most of it. That configuration ("mixed mode feedback") is pretty typical nowadays so you can find further info from pretty much any schematic of a modern guitar amp with a chip output stage (i.e. low power Marshall Valvestates, Peaveys, Fenders etc.)

Here's one example from Dean Markley:
http://www.deanmarkley.com/Info/LegacyAmps/Schematics/D1515.pdf

Nevermind that the chip and component values are different, the basic configuration is the same either way: You can see that the output stage is just a basic non-inverting (power) opamp. (This is true even in discrete designs). Now, the additional current feedback path (formed by current sensing resistor R16 and feedback resistor R14) is just connected to the inverting input as well. Well, not directly since the capacitor of the RC circuit is conveniently used to isolate the DC levels. And there you have it: It´s an extremely simple configuration.

By the way, the particular Dean Markley output stage was discussed here some time ago in a thread about Dean Markely amps (can´t remember the exact title, though).

Of course some variations of the configuration exist but this, I believe, is the most basic and easiest to implement to an existing design.

About component values... Well, you got to experiment: The current sense resistor is typically either 0.1 or 0.22 ohms so there´s not much room for choice in that. (You want to keep it low for efficiency´s sake) The feedback resistor defines the output impedance (and "damping") and this is basically a compromise between two configurations:

a) Low R: Lower overall gain but an increased output impedance. This results to lower damping factor and more response to changes in speaker impedance. "Tube-like".
b) High R: More overall gain but an increased damping. This provides a more flatter frequency response. Basically it´s closer to the existing setup of the amp design in project 19.

Rod Elliott covers the advantages and disadvantages of both setups pretty well in his article. I have also written a lot about the subject on this forum so run a search (try "mixed mode" as the keyword).

So... there are no rules for selecting "proper" component values. I guess about 1K is a pretty good starting point for the value of that feedback resistor: Not too high and not too low either. Experimenting is the keyword here. If you are handy with SPICE simulation it can save a lot of time. As a side note, you can make the value of that feedback resistor tweakable with a potentiometer and you have yourself a "damping control".

[blindsjc]

Thanks a lot Teemuk,
I always learn a lot with your replies.
I was thinking that this was a lot more
difficult, but, killed doubts. Thanks.

Blindsjc