Clean LM386 amplifier schematic suggestions please.

[frequencycentral]

Hi, I want to build a really clean LM386 based amplifier.

I'm planning to have a tube preamp section (Subcaster which I've already built) at the front end, so I don't want the amp to have a distortion feature.

Any suggestions?

[sfx1999]

Try looking at the datasheet; there are some designs in there. The LM386 isn't really a good quality chip though.

[kurtlives]

Try looking at the datasheet; there are some designs in there. The LM386 isn't really a good quality chip though.

Why is it not a good chip? Almost all the little DIY amps I have seen use em.

[dano12]

Some things that may help with your design.

1. Supply it with 12 volts instead of 9. Get the higher rated LM386 chip. The N1 through N3 versions will go to 12 volts and the N4 will handle up to 18 volts. The NJM/JRC386 will go to 12 volts, and in my opinion, sounds better and is a bit more stable.

2. On the output of the subcaster, use a b100k pot as a voltage divider to attenuate the signal before going into the 386. You may want to find the max volume that you like, and use another fixed resistor to set a limit.

When I was designing the Tube Cricket, I had poor luck with the LM386 and went with the JRC part instead. It may work better for you.

Hope that helps.

[frequencycentral]

Some things that may help with your design.

1. Supply it with 12 volts instead of 9. Get the higher rated LM386 chip. The N1 through N3 versions will go to 12 volts and the N4 will handle up to 18 volts. The NJM/JRC386 will go to 12 volts, and in my opinion, sounds better and is a bit more stable.

2. On the output of the subcaster, use a b100k pot as a voltage divider to attenuate the signal before going into the 386. You may want to find the max volume that you like, and use another fixed resistor to set a limit.

When I was designing the Tube Cricket, I had poor luck with the LM386 and went with the JRC part instead. It may work better for you.

Hope that helps.

Thanks Dano. The chip I'm playing about with is an N1. I was planning to use 12 volts anyway as the Subcaster has that power requirement. Good idea to attenuate the volume into the 386.

[Mark Hammer]

Try this one: http://www.headwize.com/projects/guitar_prj.htm

[Ben N]

Try looking at the datasheet; there are some designs in there. The LM386 isn't really a good quality chip though.

Why is it not a good chip? Almost all the little DIY amps I have seen use em.

I think what sfx1999 meant was that it is not a really good audio chip, which is something all those DIY amps--which are designed to be easily overdriven, and to drive various speaker loads--take advantage of. If you want really clean, you will want headroom and lower distortion. Any number of chip amps (TDA xxxx) will do you better there, although the builds may be larger and more complicated due to higher power supply requirements and heatsinking.

[any]

Thanks Dano. The chip I'm playing about with is an N1. I was planning to use 12 volts anyway as the Subcaster has that power requirement. Good idea to attenuate the volume into the 386.

The N1 only puts out 325 Mw @ 8Ohm max and is the lowest power version, meaning it will also distort at lower volume levels then
the N3/N4 version. Get at least the N3 and preferably the N4 if you intend to get maximum clean volume.
(N3= approx 700Mw @ 8Ohm, N4= approx 1000Mw @ 8Ohm)

Cheers

[sfx1999]

Try looking at the datasheet; there are some designs in there. The LM386 isn't really a good quality chip though.

Why is it not a good chip? Almost all the little DIY amps I have seen use em.

I think what sfx1999 meant was that it is not a really good audio chip, which is something all those DIY amps--which are designed to be easily overdriven, and to drive various speaker loads--take advantage of. If you want really clean, you will want headroom and lower distortion. Any number of chip amps (TDA xxxx) will do you better there, although the builds may be larger and more complicated due to higher power supply requirements and heatsinking.

Yeah I wasn't referring to their reliability I was referring to their signal quality.

[frequencycentral]

Thanks Dano. The chip I'm playing about with is an N1. I was planning to use 12 volts anyway as the Subcaster has that power requirement. Good idea to attenuate the volume into the 386.

The N1 only puts out 325 Mw @ 8Ohm max and is the lowest power version, meaning it will also distort at lower volume levels then
the N3/N4 version. Get at least the N3 and preferably the N4 if you intend to get maximum clean volume.
(N3= approx 700Mw @ 8Ohm, N4= approx 1000Mw @ 8Ohm)

Cheers

Thats what I love about this forum - I didn't know there were different versions of this chip when I started this thread.

I quite like the schematic Mark Hammer suggested. I will build that one and order an N4 version of the chip in the meantime, and swap it out with my N1 when it arrives.

Thanks everyone!

[Renegadrian]

What about this?


R1= 50 k log
R2= 1,5 k
R3= 10 r

C1= 10 uF
C2= 47 uF
C3= 100 nF
C4= 10 uF
C5= 10 uF
C6= 47 nF
C7= 220 u

[Renegadrian]

what about lm384? It gives 5w output...anybody used it?

[frequencycentral]

Ok, I built this http://www.headwize.com/projects/guitar_prj.htm onto my Subcaster board. It's the schematic Mark suggested.

It's worked out pretty good. I didn't use C5, C6 and R5 from the schematic, wanting the distortion to happen in the tube not the chip. Also didn't use C3 - didn't seem nessesary. I changed the Subcaster 50k gain pot for a 100k, which I thought would enable a cleaner sound at lower gain - it seem to work. With the gain cranked up its nice and dirty! I've still only got an N1, but it's loud enough for home through the speaker in my VJ clone. I'll pick up an N4 when I can. Just for fun I took it to band practice tonight, played it  through a Crate 4x12, then a Marshall 2x15. Obviously it didn't fill the room, but the tone was excellent. Its not even boxed yet - my band think I'm nuts but they were fascinated! Tried it driven by my LPB-1 - more intense tone.

So now I have to decide - head or combo? And build an LPB-1 in for drive maybe? Also thinking of an effects loop between the Subcaster and the 386......hmm?

It's been christened 'Doctor Analogue's Nasty Little B@$t@rd!'

[sfx1999]

You could try the Omega from ROG as a booster. You could also try one of the amp simulating projects, like The English Channel (AC-30 brilliant channel), Thor (Plexi), Professor Tweed (Tweed Princeton), Eighteen (18W). Another option would be to combine the Omega with the Thor.

[Renegadrian]

Also, what about TDA2005 and TDA2003?
I believe they would be clean, they are used in small car systems if I ain't wrong...and use 9-12V
Seem quite easy to build one...

[frequencycentral]

Also, what about TDA2005 and TDA2003?
I believe they would be clean, they are used in small car systems if I ain't wrong...and use 9-12V
Seem quite easy to build one...

Nice ideas Adriano, but this was a 'use up spare parts/circuits' project. I originally built the Subcaster for a pedal - but it works out that it makes more sense to use a 12au7 as it's half the current draw of a 6111. So I thought 'what shall I do with the Subcaster board?' Answer - use it as a preamp in a little amp, as I had a spare 386 knocking about that seemed be the logical way to go.

The TDA2003/TDA2005 idea is tempting - But now I have four guitar amps: Roland Bolt 60 (80's - their only tube amp), Harley Benton GA5 (Thomann VJ clone), Vox Pathfinder 15R (Used mostly with my Rhodes piano at home - the onboard tremolo and spring reverb suit it down to the ground) - and now my new Doctor Analogue Nasty Little B@$t@rd.

Can a man have too many guitar amps? I have a sh*tload of keyboards too, including a grand piano - you have to stop somewhere!

[frequencycentral]

Double post - sorry!

[Renegadrian]

I think I'll try the TDA2005 in stereo to connect my ZOOMs (I love them...) or to connect my mp3 player for my mum while she's ironing...

[Renegadrian]

Made a small 386 ckt - N1 - it distorts yeah...gotta try a N4...
I used THIS LAYOUT HERE


edit
nah it distorts badly even with a N4 - no clean amp here...

[R.G.]

Let's do some thinking here.

The LM386 is a solid state amp with a decent feedback factor. When it is fed a signal that does not drive it beyond its power supply limits, it does not distort in a gross fashion. What distortions is has under these conditions are the small distortions of any Class AB amp - crossover and very low THD. The datasheet says it will be under 0.2%.

What makes an LM386 distort is feeding it a signal large enough to drive its output beyond its power supply limits. If you have an LM386 powered from 9V, It cannot swing its output more than 0V to 9V. In fact, it can't go that far; it can only get to 6V peak to peak with a 9V power supply and an 8 ohm speaker load. If  you load it down with a 4 ohm load, it can only get to about 3.5V peak to peak before distorting.

The LM386 has a variable gain, set by the impedance between pins 1 and 8. If you leave 1 and 8 open, the gain is 20. If you bypass 1 to 8 with a capacitor, the gain goes up to 200. You can set any value between there with a resistor in series with the capacitor.

Here's how you make it not distort, as mentioned in an earlier post: don't feed it an input signal so big that it's driven to distortion, considering its gain. If you have it powering an 8 ohm speaker and running from 9V, it can't swing more than 6V peak to peak, 3V peak or 2.12V rms. The lowest gain you can make it have is 20, so the biggest input it can take without distorting is 2.12V/20 = 0.1Vrms.

That's about the output of a single coil pickup. You can't have *any* gain before the LM386 and have it not distort with this power supply and loading. Putting a 4 ohm speaker on it would require you to pad a raw guitar output down to keep it from distorting.

This gets better (a little) when using higher power supplies. Even better when you change to a different chip to get much bigger power supplies.

No amp will be clean if you feed it an input signal bigger than its maximum output swing divided by the gain it has. Period.