Power Boosted LM386?

[blackcorvo]

Here's yet another crazy idea I've had. Something similar to this:



BUT! Using a LM386 instead of the op-amp listed in that schematic. The idea would be to "boost" the output current of the 386 to get more power from it. Perhaps using a pair of TIPs you could get some more serious power, but I have a couple pairs of BD135/136 transistors I could test with (maybe I could get like 3w from that? I dunno).

I know it would be perhaps more viable and stable to just use a more powerful chip, but I like the overdrive on the 386, and it would be cool (and kinda funny) to try and make a "high power" amp out of that tiny little sand bug.
Also, I have been playing around with a Ruby-ish circuit to get a "voxy" sound and I think I'm getting close. For now I have a single MPF102 booster direct coupled to the buffer, and I'm using this neat trick to get a Top Boost type of tone and I'm liking it so far. I'll post a schematic with the proper values once I get home.

There's also the possibility of using two Rubys in bridge mode, each with it's own "power booster" like in this other schematic:



But at that point I think I might be going too far deep into wonderland. Thoughts?

[tca]

You could try something like this:



maintaining the above schematic for the power source.

Or using six 1.5V volts to get +- 4.5V and use a single LM386  (you'llhave to hack some other resistor values for the tips).

P.S.
And a buffer.

P.P.S.
But I do find the LM386 noisy.

[Seljer]

I've built the bottom circuit shown here: http://www.runoffgroove.com/littlegem.html

Works great, much louder than just one LM386, though theres a bit of crossover distortion/asymmetry when you overdrive it (but it's not like a lone LM386 is any better in this regard)


You don't need any extra transistors on the output, the chip itself has plenty of oomph to drive an 8ohm speaker to the maximum voltage the supply allows.

[teemuk]

Unless you design the thing to drive lower output impedances the added output transistors are just robbing away the max. voltage swing and output power. I think LM386 is already pretty capable within its limits.

Speaking of those, the biggest one is not current drive capability but limited voltage swing. With 9 - 12 V peak-to-peak you can do only so much. Bridging, IMO, is an easier choice than trying to pump up the design to withstand lower load impedances, but if we're going to start adding output transistors then why not use an output stage with voltage gain and higher voltages than those powering the LM386? At that point one probably should concluce that it's probably even better idea to abandon the use of LM386 as "voltage amp" altogether and resort into using integrated circuits more ideal for such tasks. i.e. even as a basic driver a generic opamp has far greater voltage excursion capabilities than a LM386.

[blackcorvo]

Well, alright. I'll try bridging later, then.

Anyways, the vox-sounding circuit I have is this one:
http://blackcorvo.tumblr.com/image/104446683865

And here's a sound sample:
http://youtu.be/VrluUl1yL_w

I just kinda wanted more juice from the 386. I guess the "push a louder chip" idea works better for that.

[CLAAS]

Very nice sounding!!!

[blackcorvo]

Apparently someone else already had this idea before I did. Look what I found on Google:



I just quickly tested it with BD135/BD136 for the transistors and... no smoke so far! Didn't play too loud because of the time now.
Only thing different is that I'm using 1Rs instead of 2R2s because it's what I have.

Anybody knows if it will live for long or not by keeping that value?

(Oh and by the way, I tried bridging. Sounded terrible. It was like using a bag of chips as an amp from how crackly it sounded.)

[PRR]

> "boost" the output current of the 386

The LM380 _IS_ the LM386 before they clobbered it. Much more power output, same topology. More gain, more legs. More rated voltage and more useful current.

Using the 6 Volt supply you show....

The peak voltage is 3 Volts (really less).

3 Volts across 8 Ohms is 0.375 Amps.

The '386 can reasonably deliver 0.5 Amps.

So the '386 is already ample; and as teemuk says, with the booster you are losing 0.6V each side.

Taking 0.5 Amps as the '386 limit, and 8 Ohm load, you can swing 4V peak already. That's 8V peak-to-peak. There's at least a volt of loss in the '386, so on 9V supply it is as good as it gets.

You could raise the supply voltage but at 12V minus the ~~1.4V loss of the added transistors the difference from 9V is hardly audible, certainly not "WOW!".

Going over 12V is risky on most '386 parts.

A possible drawback with boosted '386 is that the boosters are outside the NFB loop built into the '386. Power boosters are traditionally nasty, and "always" used inside a larger NFB loop.

The real answer is lower load impedance. Stick with 4 Volt peak swing but take speaker impedance down to 2 Ohms. Now we must deliver 4V/2 = 2 Amps to the load. The '386 won't do that, but it can order the buffers to do it. This gives about 4 Watts.

Two Ohm speakers are rare in guitar-land, but four 8-Ohm speakers parallel is right there. AND using multiple speakers is near as good as a bigger speaker: better grasp on the air, more air-shake per electrical Watt. As an extreme example, four light-cone Tens with 4 Watts will be mighty loud. Blow the crap out of the usual '386 and 4-inch speaker.

The bridge allows such powers with higher impedance, but as you note, it does NOT clip the same. (But many of the car-sound chips are bridged amps, tuned to "pleasing" overload, and easily dump 8 Watts in 8 Ohms at 12 Volt supply.)

There is yet another trick for boosting but if it goes wrong you could smoke a buck of parts in an instant.

I do think that LM380 is the short answer and car-sound chips are the heavy answer.

[bool]

TDA2003. Cheap (costs close to a single LM386 or less if you look around but there may be counterfeits), sturdy, durable, almost badass.

If you're into minimalistic stuff, look up for TDA7056.

[blackcorvo]

I do think that LM380 is the short answer and car-sound chips are the heavy answer.

TDA2003. Cheap (costs close to a single LM386 or less if you look around but there may be counterfeits), sturdy, durable, almost badass.

If you're into minimalistic stuff, look up for TDA7056.

What you're not understanding is that I want the LM386's distorted sound, but louder. The preamp I put before it is clean, It simply boosts the signal.

These other options you mentioned I've tried before (except for the TDA7056, I have used the TDA7240 tho), but they always sounded terrible. Way too mid-heavy, no bass nor treble, and they don't take preamps well at all. And the TDA2003 oscillates and blows up way too easily.

I'm starting to consider trying something more outside of the box: transformer couple the LM386's output to a couple of power tubes in push-pull.
I can very easily power both off the same power supply, I just have no idea where to find driver transformers in Brazil. I don't even think they have those here, to be honest. I'm probably gonna use a small P.T. for that instead.

[bool]

If that's the case, I'd try the LM380 first - or for a "true" LM386 tone, something like a "power-boosted opamp" with a faux current-dumping BJT pair. Like here (scroll down)

http://www.tubecad.com/2009/02/blog0157.htm

BD135/6 would most probably "just work". TBH, I only used similar circuit with an ordinary opamp. A a sidenote you would be left with no shortcircuit protection, so you'd have to carefully dimension the feed-through resistor (14 Ohm in the tubecad schem) to trip the chip protection, because the transistors would just pass what's thrown at them.


Ps. If you managed to kill a TDA2003, you surely did something weird. The ones I've got and used, survived a hell lot of experimentation, miswiring and/or decades of regular use. Are you sure it wasn't a 2030 - which by my account is a rather "nervous" chip.