A Small General Purpose AF Power Amplifier
Vložit
- čas přidán 28. 02. 2022
- I needed an audio amplifier (AF amp) for a regenerative receiver that I am building. In 2015, Texas Instruments announced that they would discontinue the LM386 and the LM380. There are some alternate manufacturers, such as New Japan Radio Company and Unisonic Technologies, but New Japan Radio’s NJM386 is also marked as obsolete now.
The LM386 was one of the most common amps used in DIY guitar preamplifiers and sustainers due to its ability to run on a single 9V battery. It was also used in many homebrew Amateur Radio designs. In 2021, surprisingly you can still buy the TI LM386 in a DIP package, from Mouser but who knows how long that will last. You can also find them on Amazon and eBay, but these parts are often rejected or poorly performing copies.
For a one-off DIY project, it is perfectly fine to use an obsolete part, or a part you have on hand, or can get on eBay, for example. But if you have to make more than a couple, you will have to choose your parts carefully to avoid supply issues.
I wanted to design and build an audio amplifier that would perform like the LM386, use a minimal amount of parts, and use parts that will not be obsoleted very soon. I initially considered building the discrete AF amp that is in the book Experimental Methods in RD Design. However, this amplifier was designed to drive a set of walkman-like earphones, which are generally 32 ohms. It is really not suitable for driving an 8-ohm speaker. Instead, I came up with the simplest design that I could think of, which was an amp that I built way back in the late 1980s.
The amplifier consists of an op-amp voltage gain stage and an NPN/PNP output pair of transistors to drive an 8-ohm speaker. Feedback from the NPN/PNP transistor stage allows the op-amp to linearize the AF amp and overcome cross-over distortion. I used a 4558 op-amp which I had laying around and a 2N3904/2N3906 for the NPN/PNP output stage. You can pretty much use any op amp with varying performance depending on the particular op-amp you use. The best op-amp would be a low noise, rail to rail, high slew rate op-amp, but even the lowly LM358 op-amp gives satisfactory results.
In the schematic, I have the gain of the op-amp set at 10x. The second op-amp gives a gain of 2x for an overall gain of approximately 20x. You can change R1 and/or R2 to achieve different gains.
In testing, I found that using 100 ohms for both R4 and R5 gave the best slew rate performance from the 4558 op-amp and also minimum cross-over distortion. If you are using an op-amp such as an LM358 then a 330-ohm resistor at R10 gives the least crossover distortion performance. For other op-amps, you may or may not need a different value for R10 including not populating R10 in the case of the 4558 op-amp.
This circuit works well with supply voltages from 9 V up to 24 Vdc. I built the circuit up using manhattan construction. Generally, when building up a circuit manhattan style, I use a combination of round pads created from some double-sided circuit board with a hand punch, and pieces or strips of Vero stripboard. I have these pads and strips as well as various components modeled in Autodesk Fusion 360 where I can arrange the parts beforehand to optimize the layout when doing manhattan type construction. Once I have a board layout that I can visualize in 3D, I then go ahead and build it up in manhattan style.
The main advantage of this circuit is that it can be thrown together from parts that you have on hand and none of the values are critical. It can be used anywhere you would need a power amp to drive a 4 to 32-ohm speaker. It works well as a small guitar practice amp as another example.
--------------------------------------------------------
Manhattan-style construction:
www.sdmakersguild.org/the-art-...
aa7ee.wordpress.com/tag/manha...
www.unixnut.net/files/manart.pdf
Materials used in this video:
Solderless Breadboard Kit: amzn.to/3wjR2SR
Breadboard Jumper Wires: amzn.to/3wfGmF7
Jumper Wire Kit: amzn.to/3GQIBDE
Vero/Strip Board: amzn.to/3wgv3wn
Double-Sided PCB: amzn.to/3CPzaBZ
Hakko Soldering Station: amzn.to/3kavUdf
Side Cutter: amzn.to/2ZW3RHB
Rigol SD1054Z Oscilloscope: amzn.to/3o5P98T
Rigol DG4062 Arbitrary Waveform Generator: amzn.to/2ZUZLz4
Get An Amazon Prime Business Membership: amzn.to/3B8euov
Check out Audible Books: amzn.to/3yYcGfP
Amazon Prime Music: amzn.to/2VEKlNd
Check out my Amazon store: www.amazon.com/shop/philipaco...
DISCLAIMER: Links included in this description might be affiliate links. If you purchase a product or service with the links that I provide I may receive a small commission. There is no additional charge to you! Thank you for supporting my channel so I can continue to provide you with free content! - Věda a technologie
I fortunately found long ago at a recycling center more than 50 destroyed no-brand walkmans (the cheap ones that were produced in mid 2000). I managed to salvage some 40 brand new LM386 and many transistors that all were new from factory. For my needs, I think I have enough. I tested them and all were functional in the specs from datasheet.
I love that Simpson analog multimeter! I have a RadioShack analog multimeter that I still use that I made from a kit when I was a kid in the 1970s.
A couple of diodes between the bases of the output transistors, with a trimmer resistor across them to set idle current, would eliminate most all of your crossover distortion. Keep in mind that the fast slews of the op-amp trying to eliminate the crossover distortion (caused by running the output transistors in pure class B) might well create noise in your receiver. This would disappear with my suggested change, which converts the output stage to class AB.
But that would reduce efficiency and power handling. For voice reproduction class B with NFB is good enough and I don't think it creates much interference. 0
@@westelaudio943 Well, that of course falls under the aegis of one of my favorite adages: "If it works, don't fix it." 🙂
Nice use of cad to design manhatten boards
Thank U for this video.
By the way there is a lot of IC's which considered very good with less components as well as it has a wider operation voltage rang Such as TBA820m , TDA7052 , TDA2822 TDA7231 , TA7331 , TA7368 , CD7368 , LM4871 , FM8002a and the powerful TA7252AP
The LM4871 'boomer' (or clones like the 8002) have taken the place of the LM386 in my uses. Cheap, super easy to set up, flexible, and bridged. something like 3W from 5V supply.
That said, a buffered op amp like this is pretty easy to set up too, and permits surprising power. I modified your circuit a tiny bit, and am getting about 2W output (on an +/-9V supply), with a (very) small amount of crossover distortion, using an NE5532 op amp and some random TO-126 transistors. Component count is minimal, especially if used on a split power supply.
Thank you forn the information on the ICs. I'm back after a few decades off the key and needing to find more modern amp chips.
I've seen videos telling / showing how to tame / perfect the LM386 by using .27 capacitors on 3 of the pins. That's done to make the IC quieter (noise) and to eliminate some BCI. I'm wondering if there's the need for that sort of mod for the 8002 etc.
Thank you.
Thanks! Nice project.
Surprised how clean that output is. That said, most of the circuits I've seen for this include two diodes to bias the transistors apart from each other.
Yes, I experimented with a few such circuits and the crossover distortion was too high for my liking.
I remember using a 741 as a simple amp some 50 years ago.
Привітання,як вам гарно жилося, років 50 назад у нас не було такого для дітей 10-15 років,бо не було такого різноманіття у радіо деталях, дякую вам,але я всерівно мав успіх
Back in the early 1980s the 741 was my first IC. I remember studying it's schematic and trying to work out how they got all those components into such a tiny plastic box.
@@user-rq9po2zv4k- Yes, I had electronics as a kid but I was very poor and had to save up to buy anything so it was all very precious to me.
Very interesting video. Thank you. (I just subscribed.)
I really like the look of that construction. what pcb material did you use for the positive bus bar? It looks a lot thinner than normal clad pcb.
thanks
nice
One diode between the transistor bases halves the slewrate requirements.
Nice informative video. Can you give a link to the schematic?
It's at 4.19 in video
Це є великий клас
I think R6 and C2 are redundant. The output of the first stage is at the correct DC bias to connect directly to the second stage.
Pardon me for asking, but what are the purposes of R7 and R8 on the +V rail. Also is it suitable as an audio amplifier, as I can't get any sound out to a speaker. Thanks
Voltage divider to reference U1A and U1B at 1/2 Vcc.
I don't think you need R9. Unless I'm missing something, I don't see it doing anything useful. I'm also worried that U1B is going to have stability problems. My quick suggestion would be to have a series resistor between its output and the bases of the transistors. (My gut says 47R to start.) I'd also add a small value capacitor as direct feedback from the output to the inverting input. It may be fine now, but I think you'll get oscillations if you use an opamp with a higher GBP. When using op-amps with output boost stages in the feedback loop, you need to keep the GPB of the op-amp lower than that of the boost stage.
However, it's working for you, so my suggestions could be unnecessary. :D
It's very easy to build a 3 transistor circuit which is functionally near equal to the LM386.
I do think the LM386 will continue to be fabricated into the future, but it is a bit under engineered, as it prefers 16 ohm loads at higher power supply voltages.
Your circuit is nice, though.
Buf634 saves everybody
Thanks for the video, it didn't occur to me that the ubiquitous 386 would be put out to grass but it happens to us all in the end. Newbie question, what would be the nearest drop in replacement/alternative when a schematic calls for an LM 386?
Tda2822. 8 pins IC. 2channel 2x1watt or. 1x3watt. Work from 3 to 18 volt. Cost. 3 US dollar. ..
sorry for a stupid question, but what is the purpose of R10?
R10 is in the circuit to provide a load on the output transistors at high frequencies. The reactance of C4 is high at audio frequencies, which essentially takes R10 out of the circuit. But at much higher frequencies, the reactance of C4 is reduced and R10 becomes an increasingly heavy load on the output. This helps prevent high frequency oscillation.
tda2822M is used very often has 2 channels which can be bridged. If you want to do discrete transistor 8050 8550 is a great pair. also used a thermally compensated bias with a pot and a sense transistor. 3-5 ma bias is enough. Opamp is not needed at all. just 1 single transistor can drive it . May want to bootstrap load resistor for driver transistor.
Быстрее будет вырезать кружочки или прямоугольники на плате
But more prone to short circuits to ground.
Интересный материал, но не на русском. И переводчик технический не осилит.