LM386, bass, sounds farty

[SirPoonga]

Ok, I was at radioshack and picked up a LM386 to make a headphone amp for the fun of it.  Just to start out I used a 10k input/volume pot going into pin 3.  I have a 220uF electrolyte on output 5 pin.  That's it right now.  It sounds fine except for low notes sound farty.  I was wondering if there is something else I could do to fix that.  I assume I am going to need something off the gain pins.

[B Tremblay]

Did you ground pin 2?

Low voltage can introduce odd distortion as well.

[SirPoonga]

You mean pin 4.  Pin 2 is + where the input goes in.  I have ground daisy chained from the 9v -, the sleeves of the input and output, and pin 4 to the lm386.  Hmmm, I don't have a ground on lug 3 of the volume pot.....

I am using a 9v battery straight into the chip.


Eidt: I just looked at the schematic for the chip.  Doh, I meant pin 3 is +.  No, I didn't ground pin 2 which is -.

Ok, I grounded the pin 2 (-).  It's a little better, however it can't get too loud before it gets really bad, especially highs now.  Do I need to put something on the gain?  Will gain amplify the signal so I don't need to turn up the volume pot tso much then?  And I noticed the lower notes aren;t amplified as much.

BTW, I am following this schematic sort of
http://www.rason.org/Projects/icamps/IMG00001.GIF
I don;t have the resistor and cap off the power though.  I do have the rest though.going to headphones instead of a speaker.

[B Tremblay]

It's always a good idea to ground the unused input.

The power supply bypass cap will probably help smooth things out a bit.

Adding the 1k/10uF network between pins 1 and 8 will add some volume, but will also allow the chip to be overdriven earlier in the Volume pot rotation.

Also, headphones have a much wider response than a speaker intended for use in a guitar amp, so they may be responsible for the harsh highs.

Measure your battery voltage.  A 386 can drain one fairly quickly.

[SirPoonga]

Brand new battery, reading 9.07V.

I've tried without anything on pin 1 and 8 also.  Same thing for the most part.

I'll have to try finding a cap to clean the power..  If you put caps in parallel you can add them, right?  I don't think I ahve 220uF caps lefts.  I have a couple of 100uF's though.

[B Tremblay]

I've found that a 100uF and 100nF in parallel work quite well.

[javacody]

There's good and bad 386 chips. I never had much luck with any other than specifically the JRC386D. I've tried the rat shack ones, they were crap, I've tried LM386-4's I got as free samples, more crap, and I've tried the JRC386BD which I also got as free samples. Again, crap. However, I noticed that my Smoky had a JRC386D chip, so I grabbed that and my amp sounds much better for it. You can get the JRC386D chips from Steve at Small Bear.

The other chips all did this gatey, crackle thing that is hard to describe. The sound just dies like a misbiased Fuzz Face.

One last thing, if you want more headroom, run the chip at 12 volts (maybe heatsink it though). This gives you more clarity.

[SirPoonga]

Or, maybe, another question to ask is there a better chip to make a headphone amp out of?  I just got the lm386 because it was available locally at radioshack.  I really don;t want to order a single chip.  Will have to put it in the next order from digikey ro mouser then.

The other quesiton is how to bass boost it?

[spongebob]

The other quesiton is how to bass boost it?

There is a schematic for a bass boost in the lm386 datasheet I think...

[SirPoonga]

I was looking at the datasheet.
Now, for a bass boost it has a resistor and cap coming from pin 1 to pin 5.  For the boost examples it has a resistor and elect cap from pin 1 to pin 8.  Can I combine the two to get bass boost and gain control?

http://www.national.com/ds/LM/LM386.pdf

[B Tremblay]

Can I combine the two to get bass boost and gain control?

Yes, as shown in this schematic:
http://headwize.com/projects/showfile.php?file=guitar_prj.htm

[SirPoonga]

I totally forgot about that circuit.  I even have it archived in my schematics folder!  I will have to give it a go.

Maybe I will start out with this circuit.  I am sure I have all the caps listed there at home.  Well, maybe not the .0047uF, but I can stop at radioshack for that.  That looks like it might work well with bass.

[Mark Hammer]

Remember that moving air at low frequencies means that the speaker cone will spend a lot of time at the forward and rear extremes of its excursion.  Remember that all a speaker is is an electromagnet that just happens to have some paper attached to it.  To hold the speaker voice coil at the forward or rear extreme of travel takes a lot of current.  If you are powering this beast with a humble 9v battery, you're not going to get a lot of current for a long time.

I would suggest listening to it with an 8-pack or 6-pack of AA's.  The resevre current capacity may allow the speaker to move without buckling under the inability of the power supply to support speaker excursions at low frequencies.  In general, one tends not to find battery-powered headphone amps aiming for low frequency content, and certainly speaker (rather than headphone) based amps don't either.  Of course for something like a Smokey, it helps that the speaker is so small no one would even dream of asking it to reproduce bass notes.

One avenue to explore is changing the 10uf cap in the feedback loop for something much smaller, such as 2.2uf or even 1uf.  You'll get your boost but it won't be as pronounced in the lower range.  You can leave the 220uf cap on the output but the amp won't take great pains (and current draw) to reproduce ultra low end.

If you need it to produce low notes, I'd recommend going with an LM380 and an 8-pack of C's or D's.  I made a little one of these this summer, and with a 6" speaker in a modest closed cab, there is some surprisingly serious thump.

[SirPoonga]

That's interesting.  As the speakers being used are headphones.  Which, in my CD player that only uses 2 AA bats, can produce 20Hz just fine.  Bass guitar goes down to 40 Hz.  The headphones should handle it.  I will have to look at that other chip though.

Or maybe am I driving my headphones with too much power?

[spongebob]

If you really need the full low-end range you should increase the output capacitor since the output cap and the speaker form a high pass filter, the bigger the output cap the better the low-end response.

[SirPoonga]

I need a bass boost.  I added some caps to output, does nothing.

[Mark Hammer]

If you really need the full low-end range you should increase the output capacitor since the output cap and the speaker form a high pass filter, the bigger the output cap the better the low-end response.

Well, that *would* be true if the amp and speaker were capable of reproducing it.  Asking ANY amp and speaker to do the impossible, however, usually results in bass farts; that's one of the many reasons why subwoofers have gone from something that only the audiophiles at the margins used to have in 1970, to something that any little $29.95 computer speaker set has these days.  Cheap amps and cheap speakers in cheap enclosures CAN'T handle bass at the same time as other frequency content, so dedicated bass-reproduction is gonna be the only thing to permit them to do that job right (not to mention the fact that you need less power overall when bi-amping, compared to passive crossovers, but that's another matter and forum).

In some respects a paltry little amp chip like a 386, run by a battery desparately needs a much higher impedance path at the output for extreme bottom end, to save it from being asked to do the impossible.  The current needed for bass excursions simply taxes the power supply enough that a great deal of other frequency content is deleteriously affected.  Perhaps if there was a heftier power supply, but certainly not with a 9v battery, and certainly not for very long.

It IS possible to make a 386 do some nice tricks that also include bass frequencies, but if one is aiming for cleanliness in a battery-powered 386 amp, bass is not the fastest way to get there.

So, adding parallel caps at the output is a bit like giving you a steak knife to eat with.  If I give you the steak (clean, strong bass content), the knife will help, but if you're only eating pudding, maybe not so much.

[SirPoonga]

So, to make a long story short, the lm386 can't handle bass at those power levels?

Dang.  hmmmm, how can I accomplish this task.

I know my headphones can, they are headphones that I use to listen to practically everything (including jazz, blues, hip hop, rock, movies, etc...).  They have no problem producing clean bass at amplified.

Hmmm, I might just break down and buy a HA-1 (www.cafewalter.com designed for bass guitar).
Answer this, looking att eh schematic for the HA-1, why does he split the guitar input and amplify both the aux and and guitar?  Could I make a simpler version of this using just one LM833 to amplify my bass guitar then just a pass through jack for aux (which I intend onlly for a cd player).  And use the cd player volume to adjust the volume of the aux in.
I wouldn;t need the tuner part of the schematic either.  Also, if using just one LM833 could I get away with one 9v (or 12v in AAs) and a circuit to produce +9 and -9.

Edit:  I just realized there is only 1 LM833 in the schematic, doh!  It's a duel opamp, is there a single opamp out there with simular specs....

[Mark Hammer]

Actually, try paralleling the two sections of the 833.  That's right.  Solder the + inputs together, the - inputs together, and the outputs together, and then just add in the passive components that will set gain, rolloff, etc.  Should actually make a very nice mini-amp.

Again, since I may not have made the point clearly, the LM386 WILL be able to provide satisfying bass to headphones of a reasonable load and efficiency.  That is, in fact, what all those 8pin dip amplifier chips were intended for in the first place.  Once the driver cone is expected to move the substantially greater horizontal (forward/backward) distances typical of speakers to be listened to from farther away than a half-inch of ear canal, though, the capacity of that chip to move the requisite current to accomplish that task becomes severely limited.

Certainly more supply voltage will help that out in terms of providing current to the load in question, but upping the supply voltage without getting past the current thing won't improve matters drastically.  The 386 will be able to make a 12" 8-ohm speaker dance to the tune of 50hz at 500mw, but you have to be able to provide it with the current to do so.  An octet of AA's (12v) or even C-cells should be able to accomplish that.  If you have any type of compression you could apply, that would probably help out a lot too.

[spongebob]

The current needed for bass excursions simply taxes the power supply enough that a great deal of other frequency content is deleteriously affected.  Perhaps if there was a heftier power supply, but certainly not with a 9v battery, and certainly not for very long.

Yes, of course, but for headphones this shouldn't be a problem, even with batteries. We are talking about a few milliwatt (10-50?) output power maybe.

SirPoonga, did you have a look at the runoffgroove.com Ruby amp?
It's a proven and working design and a good starting point for a headphone amp:
http://www.runoffgroove.com/ruby.html

And I wouldn't be concerned about not having enough gain, unless you are deaf! The LM386 has more than enough power to drive most headphones to ear-piercing volumes.