For hams who build their own radios, mastering the black art of radio frequency electronics is a necessary first step to getting on the air. But if voice transmissions are a goal, some level of mastery of the audio frequency side of the equation is needed as well. If your signal is clipped and distorted, the ham on the other side will have trouble hearing you, and if your receive audio is poor, good luck digging a weak signal out of the weeds.
Hams often give short shrift to the audio in their homebrew transceivers, and [Vasily Ivanenko] wants to change that with this comprehensive guide to audio amplifiers for the ham. He knows whereof he speaks; one of his other hobbies is jazz guitar and amplifiers, and it really shows in the variety of amps he discusses and the theory behind them. He describes a number of amps that perform well and are easy to build. Most of them are based on discrete transistors — many, many transistors — but he does provide some op amp designs and even a design for the venerable LM386, which he generally decries as the easy way out unless it’s optimized. He also goes into a great deal of detail on building AF oscillators and good filters with low harmonics for testing amps. We especially like the tip about using the FFT function of an oscilloscope and a signal generator to estimate total harmonic distortion.
The whole article is really worth a read, and applying some of these tips will help everyone do a better job designing audio amps, not just the hams. And if building amps from discrete transistors has you baffled, start with the basics: [Jenny]’s excellent Biasing That Transistor series.
[via Dangerous Prototypes]
Some day I want to make a circuit like that but also put a bunch of random entomology specimens pinned to the ground plane here and there, maybe a few rhino beetles and a butterfly.
I’d like to see that!
I’ve never had to deliberately add bugs to any of my circuits.
From the link…
> go ahead & use abundant NFB, but, of course, not to the point where you create an oscillator.
Really. How does one make an amplifier oscillate with negative feedback.
Urg…
Just have enough phase delay in the circuit – all you need for oscillation is gain and > 90 degree phase shift at any frequency, and the circuit can oscillate at that frequency. I think. I’ve been wrong about so many things lately, I hesitate to commit. I do know that op-amps need to have compensation to avoid oscillating when operated at unity gain, i.e., maximum negative feedback.
But then it’s known as positive feedback ….
I guess I wasn’t clear about this: the phase shift is not in the feedback loop; it’s inherent in the amplifier. So if you treat the amplifier as a black box (and who doesn’t?), then the loop overall can have positive feedback at some frequencies even though you’re using nothing but negative feedback outside the amplifier. Op-amps with internal “compensation” just roll off the high frequency response so that there is no frequency where you get better-than-unity gain and > 90 degree phase shift in the amplifier itself, which could cause oscillation in the worst case, which is -1 feedback, or unity loop gain. These days, almost all amplifiers designed for use as op-amps have internal compensation. But back in the day, designers didn’t like that – they preferred to have the full bandwidth of the amplifier for circuits that have greater than unity gain, i.e., less 1 negative feedback, so there were both compensated and uncompensated op-amps in the catalogs. The uncompensated amps usually specified the minimum loop gain for guaranteed stable operation.
Negative feedback is 180 degrees out of phase. Negative feedback lowers gain. And amplifier with 100% negative feedback has unity (voltage) gain. It is actually a pretty common topology, and it will not oscillate.
On the flip side, any layout with anything that has a lot of gain and that dead bug layout is prone to oscillate, but that is poor layout.
And it actually only works that way because op-amp designers have added compensation to guarantee that there is no frequency at which there is 180 degree phase shift and more than unity gain within the amplifier. Unity gain is the worst-case scenario for op-amps. If you use 100% negative feedback in an amplifier without any high-frequency rolloff, chances are, it WILL oscillate, because that negative feedback will become positive feedback at some frequency, and by using a lot of negative feedback, you’re actually introducing a lot of positive feedback at this frequency.
There’s a great application note from Texas Instruments that goes into great detail about this: http://www.ti.com/lit/an/sloa020a/sloa020a.pdf.
If the amplifier is not unity gain stable, it can oscillate with 100% feedback.
1/8 watt resistors are cheap these days. The wire is thinner and you can tighten up the low power sections very nicely.
I saw a lecture by Bob Heil (of Heil microphone fame) about the importance of good audio in ham radio. It was very eye-opening. Fixing poor input audio adds much greater intelligibility to the received audio and is even better than more power. His demonstrations were simple and profound.
Great article and encouragement regarding soldering skills. ;-)
I need to read into and study amp circuits more… though wondering how do we get a better SNR, lower noise floor and more dynamic range cost effective? I guess the Burr Brown and certain TI components are the best for lowest noise, highest SNR & dynamic range. Wondering what the Rohde and Schwartz systems are using to get like -160db. I have to read into amp theory more.
I just picked up some Goodwill and online computer speaker system THX related stuff for the circuits to use/study (Klipsch ProMedia 2.1, Altec Lansing ADA995, and Sound Blaster Titanium HD) … though wondering what the latest better investment may be that isn’t out of my price point. Looks like for cards the Sound Blaster ZxR. Thinking I’m going to invest in a Behringer U-Phoria UMC202HD Audiophile 2×2 USB Audio Interface and a Creative 70SB109500000 Sound Blaster X-Fi USB Sound Card for the laptop driven mobile rig next. Shielding the laptop is a pain though… need like optocouplers for the cables seems and the laptop in a TEMPEST box with a way dimmer screen RFI/EMI window.