DIY Stompboxes => Building your own stompbox => Topic started by: RDV on March 31, 2004, 05:38:22 PM
I love 386 based amps but the freq. response leaves something to be desired. I realize you're only going to get so much low-end from a 1/2 watt amp, but that's not what my problem is. I've always dug the high-end part of guitar. Stuff like The Beatles "And your Bird Can Sing", The Stooges "Search To Destroy" & U2's first couple of records. Sizzle is my thing. 386 based amps have a deficit of high-end in my opinion, especially when arpeggiating chords and such. I like the "Ruby" at Runoffgroove, but it suffers from the same malady. My solution has been to use a .015uF in the place of the .047 after the JFet, but limiting the low-end going into the circuit greatly reduces the apparent gain. I've got pins 1 & 8 wired together and it still will barely break up with the volume wide open. I know this because with the .047 there's plenty of apparent gain.
I'd like to get some ideas on how to put a gain stage of some sort there instead of the buffer to make up for the loss of gain due to the higher cap. I thought about a Mosfet gain stage like Gez's that's in R.O.G.'s version of the EA Trem, or an LPB Transistor stage, or just converting the JFet buffer into a gain stage.
Thanks to all who reply.
I've got pins 1 & 8 wired together and it still will barely break up with the volume wide open.
I believe that increasing the resistance between these two will increase gain. That's how I understood it after looking at this datasheet : http://www.nari.ee.ethz.ch/teaching/PPS/PPS02/doc/LM386.pdf.
"To make the LM386 a more versatile amplifier, two pins (1
and 8) are provided for gain control. With pins 1 and 8 open
the [internal] 1.35 kW resistor sets the gain at 20 (26 dB). If a capacitor
is put from pin 1 to 8, bypassing the 1.35 kW resistor, the
gain will go up to 200 (46 dB). If a resistor is placed in series
with the capacitor, the gain can be set to any value from 20
to 200. Gain control can also be done by capacitively coupling
a resistor (or FET) from pin 1 to ground."
Whereit says a "resistor placed in series with the capacitor" across pins 1 and 8, this can be a pot for your gain control. The last part, "capacitively couping a resistor to ground" is what is done in the Smash Drive, I believe. http://www.diystompboxes.com/pedals/schems/smashdrive.gif
Hope that helps, and that I know what I'm talking about. :)
sorry alex,, acutally the gain is set via lowering the resistor,
u can put a cap adn resistor there as it states. u can also put a cap across there to get it to amplify certain freq and not others.
but since pins one and eight are tied with the resistor addign more resistance there effectively does nothing since it will follow the least resitant path
however you can bridge the pins adn then add a reistor cap to ground to get a little mroe gain out of it, and some freq shaping like we do in normal opamps..
as far as the .015uf having less gain, i am quite unsure of that.it will lessen the bass freq yes but it should retain the gain with the high end
i have used the lm386 chip for years and the only time inoticed the problem you are having is when the chip was dying..
My Ruby sounded dark and muffled until I put a small resistor and cap at the output (I think it's called a Zobel network?), it is used in the Little Gem and is also proposed in the datasheet. You could try this, at least it won't hurt.
Also, the output cap (220uF electrolytic) seems to have quite some impact on the sound, just try a few different ones here.
BTW, what impact does the size of the output cap have? Bigger = more bass?
the mentioned network prevents the amp from oscillating,
which may be ultrasonic, but can cause the sound to be dull...
What sort of speaker are you using?
I tried to be clear, but failed as usual. The gain does not go down(I have pins 1 & 8 tied together for a gain of 200) because of the .015uF in cap, but filters more lows therefore having the effect of "seeming to lower the gain. If you don't believe me, then socket this cap(.047uF & work your way up from there) & you'll see for yourself.
My whole aim was to get suggestions for a front-end gain stage, to make up for the apparent loss caused by filtering the lows at the input.
I probably will try the Little Gem style filter at the end to see if that would clear up the sound with the stock in-cap value(.047uF)
What sort of speaker are you using? A variety, but usually an Eminence 10" which has plleennttyy of treble normally. It sounds best through my Peavey 412m.
You could add the input gain stage of the Big Daddy circuit and tweak the coupling cap for an on-board treble booster.
Also, try using a 25k volume pot. This will help with gain when using a smaller input cap. Or, you could remove the input volume pot entirely and use the Little Gem's 25 ohm volume control at the output.
Grace/Big Daddy: http://runoffgroove.com/grace.html
Little Gem: http://runoffgroove.com/littlegem.html
I'll try the larger value input volume pot first. I really like having it because I can just jumper pins 1 & 8 and it's kinda like a non-master volume amp.
Sorry, I should make sure I know what I'm talking about before jumping in. :oops: At least I learned something though.
I added a 50KB pot for the volume, and the Zobel network from the Little Gem, and retained my .015 in-cap and it improved the apparent gain and cleared up the sound even more till it sounds kinda like a Hiwatt. I'm happy! With a boost in front it's Pete Townshend city!
Thanks to all who replied
Long Live Rock(I need it every night)
There was a project in Nuts and Volts last year for a TV sound enhancer for those with age-related upper-spectrum hearing loss. Essentially what it was was a 386-based mini-amp for either headphones or extension speakers. The "trick" it employed was to shape the frequency response of the 386 with the compensation loop.
Normally one uses this to set the overall gain, and employs a large value cap to let the full usable bandwidth pass through the fixed resistor in the loop. In the project they used a small value cap to restrict feedback for lower mids and bass and apply greater gain to the sibilance part of the signal.
Applying this logic to the miniamp, you could have two parallel paths in the compensation loop: one path with a large-value cap (10uf) and medium value resistor (e.g., 3k3) and another "secret passage" for high end with a 1k resistor and, say, a 470nf cap. This will provide flat response below some treble cutoff and additional boost above that cutoff.