***NEW*** LM386 practice amp with FET preamp for more gain.

[cab42]

For a bridged 386, see ROG's Little Gem MKII: http://www.runoffgroove.com/littlegem.html
EZ

I knew I have seen a bridged 386 somewhere... lol

mac

I have a TDA2822, I have thought about using for at practice amp. It's stereo, but the channels can be bridged in a Little GEM II type configuration. My hope is that it should give more headroom than the 386.

Regards

Carsten

[andronico]

Hi Mac.

Congratulations !    Your practice amp sounds very good, I´ve made a ruby to try and yours will be in my future list.  I´m from Argentina too, I live in Buenos Aires but have family living in MDQ !

Un abrazo.

julián

[mac]

Could you explain what de diode's in the pre-amp section do (i know what the do but i haven't seen diode's used like you use em in your amp) and which 'path' the sound signal follows?

The ac signal passes through the diodes, which are turned on by r5,r6 and r7. At first sight it seems that they are blocking the signal, but this is ac... I guess lol

Hi Mac.
Congratulations !    Your practice amp sounds very good, I´ve made a ruby to try and yours will be in my future list.  I´m from Argentina too, I live in Buenos Aires but have family living in MDQ !
Un abrazo.

Voy seguido, nos juntamos y hacemos un poco de ruido jaja

mac

[mac]

A new clean demo clip, copy the link and paste it on a new browser window since tripod.com does not allow remote hosting:

http://guitarfxs.tripod.com/mp3/jade_clean.mp3

http://guitarfxs.tripod.com/mp3/jade_clean.mp3

mac

[yertle]

I still don't get the Led part in your circuit, you said AC will pass trough the diodes, I don't see why. I thought that the positive part of the AC signal will pass through a diode that's 'pointing the same way' as the signal (if it is over the threshold of the diode but in this circuit that's already taken care of by the V+ and teh resistors) and that the negative part will go through a 'reversed' diode (this is the way diode's work in distortion circuits with anti-parallel diodes to ground right?). I would really like to understand whats going on in this circuit, I hope that someone could explain this.

Btw, if the ac signal is completely passing through the diode's, what's the effect on it to the sound?

Clean sounds nice, although not completely clean, what settings did you use for that clip?

[Eb7+9]

nice sounds mac ... 

... the diodes are being used for their non-linear resistance ... if you look at both diodes as a combined resistance you'll see it's part of a voltage divider acting between the Drain and the equivalent resistance seen at the volume pot ... the diodes are biased weakly so that minute variations in current through them cause a relative increase or decrease around their idle resistance value ...

you have to argue on a half-cycle basis, take a sine wave input for argument's sake ... when the Drain voltage goes high the first diode gets slightly less biased and the second one gets slightly more biased ... the incremental resistance, what the signal sees, is increased in the first and reduced in the second ... during the second half cycle the reverse happens ...

because of the logarithmic current-to-voltage nature of diodes the increase in resistance on the weaker biased diode is greater than the decrease in resistance on the stronger biased diode ... this means that the incremental sum resistance of both is greater than the idling sum value, and the increase is monotonic with signal amplitude ... to a point at least ... if the diodes and their biasing are well enough matched then the two half cycles will be attenuated through the divider in an equal way ...

at least this is what happens when the stage is operating in class-A ... with stronger signal levels you get to onset of clipping, and then into class-B fuzz and hard limiting ...

Marty/John - maybe something to explore in the Tornado dividers ...

[mac]

.. the diodes are being used for their non-linear resistance ... if you look at both diodes as a combined resistance you'll see it's part of a voltage divider acting between the Drain and the equivalent resistance seen at the volume pot ... the diodes are biased weakly so that minute variations in current through them cause a relative increase or decrease around their idle resistance value ...
you have to argue on a half-cycle basis, take a sine wave input for argument's sake ... when the Drain voltage goes high the first diode gets slightly less biased and the second one gets slightly more biased ... the incremental resistance, what the signal sees, is increased in the first and reduced in the second ... during the second half cycle the reverse happens ...
because of the logarithmic current-to-voltage nature of diodes the increase in resistance on the weaker biased diode is greater than the decrease in resistance on the stronger biased diode ... this means that the incremental sum resistance of both is greater than the idling sum value, and the increase is monotonic with signal amplitude ... to a point at least ... if the diodes and their biasing are well enough matched then the two half cycles will be attenuated through the divider in an equal way ...
at least this is what happens when the stage is operating in class-A ... with stronger signal levels you get to onset of clipping, and then into class-B fuzz and hard limiting ...
Marty/John - maybe something to explore in the Tornado dividers ...

The LEDs I used when I was testing were different, both red but one with the red case and the other transparent. 1.65V and 1.5V resp. Since the 1.65V were conducting less I replaced the 10K to gnd with a 8.2K. Also I used a 100K-B pot instead of the fixed 47K (R6). I'm using 100K now. As I wrote in the schem, play with the diode network resistors.

To be honest, I included the diode network because I had oscillations problems above certain gain, and a big resistor or shunting signal to gnd was not atractive and sounded horrible. When I read about the diode network (Joe Davisson maybe?) I said 'why not?'. It reduced the oscillations a lot but not completely. Then I realized that I was tweaking ROG version so the fet drain was going to the 386 (-) input, pin 2. I changed to (+), pin 3 and eureka! No more oscillations. I was too lazy to remove the diodes... 

mac

[mac]

I have a TDA2822, I have thought about using for at practice amp. It's stereo, but the channels can be bridged in a Little GEM II type configuration. My hope is that it should give more headroom than the 386.
Regards
Carsten

Success! I used two bridged LM386-N, this ones have more power the the one I used first. I connected the 386s almost exactly as in the Little GEM MKII, after the 0.22uf. I only added a 0.022uf cap from (+) pin 3 to gnd, and a dual 10K pot to control both chips gain at the same time. And the caps are not 10uf but 2.2uf polyester. The bad news that the 386s, the 2.2uf and the 10 resistor get hot. Maybe a fan or small heatsink for the chips.
As soon as I draw the new schem I'll post it. MAybe a new thread.

mac