Trying different Power Amp IC's in Tube Cricket???

[xixiviii]

I just finished a Tube Cricket from Beavis Audio Designs and used a standard Tayda supplied LM386N-3 for the IC in the power amp section.

Dano states in the build instructions that a LM386 doesn't work well here and to use a NJM386D (which small bear has for $0.80, Mammoth has for $0.50, etc.) 

I had about 20 LM's and nothing else so that is what I tried.  Mine seems to work fine (although I haven't tested it with a legit cabinet yet). 

Has anyone had any experiences with different 386's?  I am wondering what differences (if any) it might make?  Is it a stability related thing maybe? 

I read through a few posts regarding the tube cricket, but didn't immediately pinpoint much regarding how the IC changes might affect the circuit.

I need to order some tube sockets from small bear, so I can definitely add a couple of NJM's to my cart to try for myself - just wanted to hear form you guys what you thought.

[Mark Hammer]

As far as I know, the differences are essentially in safe heat dissipation.  The NatSem one, with the 1/3/4 suffix, vary in how much you can push them in terms of supply voltage and load, to yield different wattages.  Of course, we're talking variations under 1.5W, so its not like it makes a helluva lot of difference.

[thomasha]

I was searching the same thing.

I have build a modified tube cricket with a jfet first stage, but the lm386 is very noisy.

Without the power stage everything is nice and quiet, and the tube only introduces a light overdrive.

Using a cab the sound is nice, and the noise his quite low, but when using headphones the constant humming is very annoying, it won't change with volume , even with the lm386 input grounded, so I blame the layout, or the IC.

Since I made my own PCB, I searched for an IC with the same pinout but lower noise level, but couldn't find anything.

I guess a lm2822 or a TDA2003 would do a better job. I tested the pre-amp in a lm2822 computer speaker system and the result was better.

If you find some difference in noise using another IC let me know.

Cheers,
Thomas

[Brymus]

Add a 1n - 10n + 4k7 resistor from the spkr (-) side of the 250uf output cap to pin 1 of the 386 this will create a LPF and cut the hiss for you.
The larger the cap the more high end you cut.
I am using a 5n6 cap and can't tell any difference at guitar frequencies,but the hiss is almost completely removed.

I am doing this using the LM386's in another build I am working on.

I do prefer the NJM386D for my 386 use but they are a bit pricier now at Mouser than last time I bought some (around 1$ a piece)

[duck_arse]

in days gone by, they/we used to use a crummy op-amp that will remain numberless (it was all we had), and hung a crummy "power stage" off the end. the power stage was a pnp/npn pair of signal transistors, w/ maybe a diode or two for bias, I can't remember exactly without searching. has anyone thought of using a modern-er, more good-er-er op amp, like a tl071, and a bc548/bc558 or bc338/bc328 for lower noise and hiss?

[thomasha]

Thanks Brymus,
I saw that at http://www.transkommunikation.ch/dateien/schaltungen/space_station_receiver/informationen/08_-_reducing_hiss_in_an_lm386_audio_amplifier.pdf

but for a non-inverting circuit (pin 3 signal, pin 2 GND). As the ruby and the cricket are inverting circuits (pin 2 signal and pin 3 GND)  this increased noise. Maybe a NFB approach like the fender champ would help?

It could be the layout, since the heaters supply is very near the IC.

A picture of my build so far:



I will test it tonight, lets see

[Brymus]

I had never seen that one thanks for the link.
I am going to pin 1 though,not pin 8, maybe that makes a difference if you have a cap there for more gain as I do and is shown in the PDF you linked too.
Good luck

[PRR]

> the constant humming is very annoying, it won't change with volume , even with the lm386 input grounded, so I blame the layout, or the IC.

Does it change as you move it around the bench?

If not, then it's power supply hum.

The chip itself can only do so much. A cleaner supply helps, but only so much. But the chip can give leverage against a cap which helps significantly. The '386 has a Bypass pin, read the sheet and be generous with that cap. '386 has two input pins, the un-used one should be grounded.

Layout IS important. If high-current AC spikes from power supply run past low-level stages, even the '386's Bypass cap return, buzz is injected into the signal.

A super-clean power supply (usually regulator work better than BFC) is a no-brains path to much power PS-induced hum. (And usually you then find further hum causes.)

[thomasha]

Well, after some testing I observed that adding the resistor plus capacitor speaker and pin 1 increased volume and changed the tone, but the hiss and hum were still there. The circuit was also way more sensitive, mostly between the resistor and pin 1.

I tested a 4k7 NFB resistor between speaker and the LM386 input, and that worked, but there was also a slight decrease in volume.

Connecting the resistor to the second tube stage cathode worked the same way, with a bigger volume drop.

The real improve was by using another wall socket and a bigger capacitor at the VCC pin. Still not as quiet as I expected, but much better anyway.

Thanks