LM380N output resistor and capacitor needed?

Thread Starter

portreathbeach

Joined Mar 7, 2010
143
I am using an 8 pin LM380N as a small amplifier.

From the datasheet it shows a 0.1uF cap and 2.7R connected in series from the output pin to ground.

There is also a capacitor (4.7uF) connected from the bypass pin to ground.


I have tried the chip with and without all of the extra components. I hear no difference with the bypass cap in circuit or out. If I insert the 2.7R and 0.1uF cap, the sound is worse than without it.

Do these components have to be used? If I don't use them, will the chip get damaged?
 

flat5

Joined Nov 13, 2008
403
Do you have an oscilloscope? A good AC voltmeter?
Check the output pin for oscillation with and without a square wave signal.
If the output is clean you don't need the insurance these components provide...for now.
You should not be able to hear a difference with/without them.

Without the stated test equipment you could monitor the current drawn by the circuit with and without the network and also with and without a square wave signal.
If the current is the same it's not oscillating.
 
Last edited:

MikeML

Joined Oct 2, 2009
5,444
What is your power supply? If pure dc (battery), then you can operate without the bypass on pin 8. The bypass reduces hum if the amp is operated on a non-regulated, inadequately filtered AC to DC supply, like a wall-wart.

I would not operate without the anti-oscillation network (100nF, 2.7Ω) on pin 6. Every circuit suggested by National (now TI) shows it. Without it, the amp is likely to break into oscillation at ~100kHz, overheating the chip, and discharging the battery much faster.
 

Thread Starter

portreathbeach

Joined Mar 7, 2010
143
I have a scope and have scoped the output. There is noise without the resistor/cap network, 0.5v pk-pk. With the network inserted, the noise is lessened. I have noticed that with the network installed, 100mA is being drawn from the supply all the time, without it, only 9mA is drawn!

The circuit will be powered from a 12v car battery, so I guess pin 8 can be left free.

If I use the 2.7R and cap on the output, I get a slight hiss all of the time. Without it, I don't.
 

MikeML

Joined Oct 2, 2009
5,444
Which means that your layout must be so bad that the thing is oscillating with or without the anti-oscillating network :(

Do you have a 470uF power supply bypass cap wired from pin 7 to pins 4&5 as shown here?

 

Thread Starter

portreathbeach

Joined Mar 7, 2010
143
Yes. It is currently on breadboard. I may transfer to veroboard later for better connections. I didn't have a 470uF handy, so I'm using a 1000uF.
 

MikeML

Joined Oct 2, 2009
5,444
I doubt you can stabilize it on a breadboard. I would build it "dead-bug" style on a piece of copper-clad PCB material. That way, the ground plane is low-inductance to prevent it taking off at RF frequencies, and will help get heat out of the dip. Bend the grounded pins down onto the copper plane, let all other pins stick up to create isolated terminals to solder the other things to.
 

MikeML

Joined Oct 2, 2009
5,444
If it prevents the device from oscillating, then it will be dissipating almost zero power because there is almost zero power in the real audio signal at the frequencies where the network has a low impedance.
 

Thread Starter

portreathbeach

Joined Mar 7, 2010
143
OK. I've built it up on veroboard. There is a faint buzz from the speaker, but way better than when it was on breadboard. It has minimal static sound when nothing is being sent into it and sits at 9mA draw.

Does this sound OK?
 

MikeML

Joined Oct 2, 2009
5,444
I doubt that the buzz is because the device is oscillating, especially if the Iq is only 9mA. Hang the scope on it, and see if you can determine the frequency of the buzz?
 

flat5

Joined Nov 13, 2008
403
http://www.ti.com/lit/ds/symlink/lm380.pdf
7-25 ma. You are ok.
Short the input and see if the hum goes away.
We don't know if you are using a volume control or what.
You have not provided a schematic.

"The circuit will be powered from a 12v car battery, so I guess pin 8 can be left free."
Will the battery be in a running car? There may be more to think about.
 
Last edited:

Roderick Young

Joined Feb 22, 2015
408
I fixed a production design (a video game) almost 40 years ago, where the original designer thought they could get away without that 0.1 uF snubber on the output of the LM380N. Sound came through, but it was very feeble. Possibly the designer didn't count on digitally-produced square waves, which contain lots of high frequency components, on the input. 0.1 uF and a 2.7 ohm resistor are cheap insurance, if you ask me.
 
Last edited:

Thread Starter

portreathbeach

Joined Mar 7, 2010
143
Hi,

I currently make a product for BMX starting gates. The PCB I designed a few years back simply has a PIC which outputs to 4 bright LEDs for the countdown lights and a PWM that switches a small transistor to drive a speaker to get the tones. It also has a MOSFET to switch an electro magnet to release the start gate. The system is usually connected to a car battery, not connected to the car. I supose some buyers may run it from the aux socket in the car, I'm not sure.

I decided improve the design and incorporate speech like the professional versions do, so I picked up a voice recorder module to have a play with. It uses an ISD1820 chip on it, pretty nice little IC actually.

So I wanted to get the PWM and the ISD1820 to output through the speaker. I have just assembled a very crude setup. I have the PWM pin from the PIC through a 0.1uF cap, then 10K resistor. The output from the ISD1820 board goes through a 0.1uF cap and a 1K resistor. They are then both connected together and to ground through a 100K resistor. The point above the 100K resistor is the input into the LM380.

When the ISD1820 is not outputting, the speaker output pin goes high resistance, I also change the PWM output pin of my PIC to an input to make it high resistance when not being used also.

Seems to be working OK, the recording isn't the best though.

To record onto the chip, I used the 3.5mm jack on the back of my Mac. Ground to ground and the left and right channels are connected to Mic and Mic Ref via 0.1uF caps. The left channel plays the inverse of the right channel otherwise you would hear nothing.

I may see if I can get an IC that can be written too directly from the computer, rather than having to record each chip manually.
 

flat5

Joined Nov 13, 2008
403
Not my place to say so but you should start your own thread as this has nothing to do with the LM380.
I'll delete this comment later, most likely.
 

ian field

Joined Oct 27, 2012
6,536
Do you have an oscilloscope? A good AC voltmeter?
Check the output pin for oscillation with and without a square wave signal.
If the output is clean you don't need the insurance these components provide...for now.
You should not be able to hear a difference with/without them.

Without the stated test equipment you could monitor the current drawn by the circuit with and without the network and also with and without a square wave signal.
If the current is the same it's not oscillating.
Its usually "parasitic oscillation" - just a blip of oscillation close to the peaks. That probably isn't going to do much to to the current draw unless it goes full blown.
 

Thread Starter

portreathbeach

Joined Mar 7, 2010
143
I scoped the buzz, and it seems like static, there isn't any frequencies I can actually identify. Anyway, I'll keep playing about with it and see what I can come up with.

Cheers for all the comments :)
 
Top