The excercise used a lousy old 741 opamp because the excercise was written by a very old teacher in 1967 when the 741 was the only opamp available.

The lowpass filter was used because in 1967 an old AM radio did not produce any high audio frequencies and the 741 opamp produced lots of hiss.

 

OK I started fresh with a TL071CP and tested it out and getting the gain I should ~10. Swapped the 10kΩ for a 4.7kΩ and increased the gain to 21.8 measured. Added an LM380N-8 and the gain was too high so I put the 10kΩ resistor back in. Still too much gain so I replaced the 380 with a LM386 and things kinda worked but was getting oscillation. So I'm gonna break out some perf board and solder this thing up and see if I can make it work. Very frustrating.

 

The lm386 has a gain of 20 without adding any additional components. Why are you replacing opamps with power amps?

 

An audio product is supposed to have too much gain then you can turn up the volume control when the signal is weak or turn down the volume control when the signal is too loud.

The datasheet for an LM380 says that its gain is 34dB which is 50 times and for an LM396 with no added resistor its gain is 26 dB which is 20 times.

 

The lm386 has a gain of 20 without adding any additional components. Why are you replacing opamps with power amps?

I'm not. The 386 is the follow-on power amp for the TL071 pre-amp.

 

An audio product is supposed to have too much gain then you can turn up the volume control when the signal is weak or turn down the volume control when the signal is too loud.

The datasheet for an LM380 says that its gain is 34dB which is 50 times and for an LM396 with no added resistor its gain is 26 dB which is 20 times.

The next part of the exercise is the volume control. There was so much gain with the LM380 it was picking up a radio station without the audio input attached. I really need to use the Faraday Box I just built but I had to order some more patch cords as I don't have enough now. I thought I'd go big and try the LM380 as I've never used it before and wanted to see how it compared to the LM386 without the 10uF capacitor. This supposedly simple little lab exercise is turning into an "interesting" experience as the old chinese curse goes. But I am learning from it.

 

Any amplifier will pickup interference if it is built on a breadboard with fairly long wires all over the place. The wires act like antennas. The many strips of contacts also act like antennas.

 

And that is exactly what I am encountering. Never really worked with designing and building amplifiers before and no real experience with transistors or op amps so this is new to me. I was a Project Engineer specializing in Process Control and Electrical Controls and DCS interfaces so I was more of a user than an electronic designer. And a Ham and did some basic radio troubleshooting and alignment on tube Xcvrs and designed and built some dipole antennas and other Ham gear. This solid-state stuff is new territory for me. I did do some logic chip stuff back in the 70s and 80s building Iambic Memory Keyers and their power supplies and regulators for DC logic but never got into transistors. Making up for lost time.

 

Having a few inexpensive 12-0-12 transformers the AC should measure 17 VAC should be useful.
Comparing op amp datasheets "Recommended operating conditions"
https://www.ti.com/lit/ds/symlink/lm833.pdf
https://www.ti.com/lit/ds/symlink/lm741.pdf

 

On the LM386 and LM380 is a bypass pin shown with a capacitor to activate it. Should it be grounded if not used? What exactly does this bypass pin do?

 

The bypass pin has resistors to bias the amplifier so that its output voltage swing is maximum and is symmetrical. It is powered from the power supply that might have mains hum or battery voltage fluctuations that get into the output. The bypass pin can have a bypass capacitor to ground which filters out the disturbances. The LM386 datasheet shows 10uF.

 

Grounding the pin(s) would make the output go to ground.

 

Preamp and opamp stages done and tested.

Socketed for R1 & R2 to adjust gain. Ground rail along bottom and a few extra pin sockets for attaching instrument. 100nF caps on +/-15V supply.


Noisy @ 50mV input but getting the Av=10

EDIT: Using TL071!

 

bypass pin can have a bypass capacitor to ground which filters out the disturbances.

Sounds like it wouldn't hurt to have it then. Thx.

 

Is all that noise on the input? Is the noise from a signal generator or is it stray pickup?
You must use a shielded audio cable at the input. Audio cables made with RCA plugs on the ends always use shielded cable.
The shield keeps out the interference.

 

Cranked the voulume up on the Echo/Alexa and looks from the output I'm good to be under the 400mV input for the LM386 so I'll start there on the power amp section.

 

Is all that noise on the input? Is the noise from a signal generator or is it stray pickup?
You must use a shielded audio cable at the input. Audio cables made with RCA plugs on the ends always use shielded cable.
The shield keeps out the interference.

The sig gen output is on BNC coax with just a couple of inches to a mini grabber. I can connect BNC coax directly to the scope input and there is no noise. This is why I built the Faraday box but my extra patch cables ordered that I need won't be here until Friday at the earliest. Anytime I get into mV or lower inputs I have to fight noise. I thought it was because the sig gen is in an unground plastic case but I do get a clean signal from it surprisingly.

 

Once I put the power amp section together my next hurdle will be coupling the speaker to it. The LM386 and LM380 show different values for their coupling circuits. I understand using a coupling capacitor but they also use a smaller value decpoupling in the circuit? And the coupling cap sizing I assume is based on the output voltage or wattage? The LM380 typicals are all for the 14 pin even though the PDF is also for the 8 pin.

 

I do not have an Alexa, instead mine is a Google version. For the first time I asked it to play a sinewave of 1kHz and it did. It was at a fixed level with no distortion.

Yours has its level varying all over the place. The many waves are crammed together horizontally. Reduce the time per horizontal graticule to see fewer waves so that you san see their shape. I do not know why the level keeps changing.

When I spread out your crammed together waves then the "sinewave" looks like digital steps. Is your Siglent 'scope a digital one?

 

Didn't ask for a sine wave, it was streaming WWOZ out of New Orleans if you want to hear some really good heritage and Jazz music. They say it's like Gumbo, a little bit of Dis and a little bit of Dat. Yes, the digital scope is a Siglent 1102CML+ 100Mhz that when I updated the firmware magically became a 1202CML+ 200Mhz.