One of my classes last spring semester was on transistors, and a big part of that class also covered basic amplifiers. I already took one such class in the previous university I attended, but they never really taught us much about the practical aspects of real-world amplifiers. So I got a little excited one day after class and decided to build my own discrete headphone amplifier. Turned out much easier than I expected and worked surprisingly well too.
Attached below is a schematic of the amplifier I built, just as it was on the breadboard with all of the weird resistor values and configurations. It consists of simple, straight-up book example class A stages cascaded together. 3 transistors per channel, 6 transistors total for stereo. Simplicity at its finest! I'm sure most of you who have even the slightest of knowledge on discrete amplifiers will probably find my design laughable
schematic
Still, it goes to show how little you need to know to make one. And all of this was built from spare / junk parts I had around or collected over the years. As you can see, there are no special transistors used. Just plain old 2222 NPNs for the amplifying stage and a TIP42 PNP in a TO-220 case for the output. I probably could have built this even simpler by using only 2 transistors per channel. But I wanted to keep the gain on the amplifying transistors low for lower distortion. Not sure if that mattered or not, but regardless.
So how does it sound? Pretty
good . I would dare say even much better than many of those popular cheap-n-dirty amplifier ICs. There was virtually no hissing or humming at all. Quite surprising IMO, because I had the gain set fairly high at about 18.7 dB. Moreover, I was using a 9V switching power adapter. You'd think that would introduce some noise… but it didn't. And with this gain, it was more than enough that I didn't need to put the volume in Windows higher than the first bar (both Wave and Main Volume controls).
And now with the more technical side of things…
Frequency Response:
@ 1 KHz
@ 26 Hz
@ 212 KHz
I took the amp with me in one of the labs where I had access to a scope so I could perform some tests. As you can see from the attached scope pictures, the frequency response isn't too shabby – 26 Hz to just a little over 200 KHz at full volume right before clipping point, with a 33 Ohm resistor as the output load. While the high-end response is quite good, the shape of the low-end of the frequency response was kind of ugly. Unfortunately, halfway through the tests I noticed that I forgot to change my output coupling caps from some earlier testing. So instead of 1000 uF as I originally planned (and as the pictures show), I had 470 uF caps. CapXons at that too
. But I didn't have any other caps with me on hand, so I just left the test results alone. The low-frequency output would probably have looked better if I used 1000 uF caps. When I was listening to some music with lots of low-end bass, however, I was not able to tell a difference.
Gain:
I used the scope to determine that and compared it to my calculated values. Turns out, they weren't too far from each other, with my calculated gain of 18.7 dB (8.6 V/V) being the minimum gain for the amplifier at the above frequency response. At 1 to 10 kHz, the gain was closer to 21.7 dB (12.3 V/V).
Output:
The highest output I could get with that 9V SMPS adapter was about 3 Vp-p (peak to peak). This, with a 240-250 mVp-p input from the signal generator. Translating this into RMS values: approximately 1.06 V RMS output with about 0.08662 V RMS input.
Power:
With the above output, that means I was getting about 34.1 mW RMS into that 33 Ohm resistor. It's not a whole lot, but still plenty IMO. My headphones (cheap classic Sony MDR-023) are rated for 24 Ohms and input sensitivity of 100 dB per mW. That means I can technically push them up to 46.9 mW with this amplifier, or about 117 dB. More than enough for me.
Efficiency:
Not good. Not good…
I don't exactly remember what current it was drawing from that 9V power adapter, but I remember it was in the ballpark of 100-and-something mA and the calculated power between 1.1 to 1.4 Watts total (so about 550 to 700 mW per channel). With those numbers, that's roughly 5 to 6% efficiency with the 33 Ohm resistor and should be 7 to 9% efficiency with my headphones.
Yeah, it's pretty low
. But again, this is a class A amplifier, so this is more or less normal. I remember reading a while back somewhere that around 10% is generally to be expected. Could get it up to 50% if using impedance-matching transformer… but then there goes the simplicity!
Unfortunately, I needed my breadboard for a project in another class, so I had to take the amplifier apart. All fine though, because since then, I've built another class A amplifier – this time bigger, slightly improved, and meant to drive actual speakers. It still has similar gain and all but the output stage is slightly different (originally, I wanted to do class AB, but I still can't figure out how to do that properly without distortion).
Anyways, when I get around to it, I'll draw and post a schematic of this “new” design. It's nothing to brag about either (only capable of about 600 mW into 4 to 8 Ohms), but still quite fun to build and have around nonetheless.
Attached below is a schematic of the amplifier I built, just as it was on the breadboard with all of the weird resistor values and configurations. It consists of simple, straight-up book example class A stages cascaded together. 3 transistors per channel, 6 transistors total for stereo. Simplicity at its finest! I'm sure most of you who have even the slightest of knowledge on discrete amplifiers will probably find my design laughable
schematic
Still, it goes to show how little you need to know to make one. And all of this was built from spare / junk parts I had around or collected over the years. As you can see, there are no special transistors used. Just plain old 2222 NPNs for the amplifying stage and a TIP42 PNP in a TO-220 case for the output. I probably could have built this even simpler by using only 2 transistors per channel. But I wanted to keep the gain on the amplifying transistors low for lower distortion. Not sure if that mattered or not, but regardless.
So how does it sound? Pretty
good . I would dare say even much better than many of those popular cheap-n-dirty amplifier ICs. There was virtually no hissing or humming at all. Quite surprising IMO, because I had the gain set fairly high at about 18.7 dB. Moreover, I was using a 9V switching power adapter. You'd think that would introduce some noise… but it didn't. And with this gain, it was more than enough that I didn't need to put the volume in Windows higher than the first bar (both Wave and Main Volume controls).And now with the more technical side of things…
Frequency Response:
@ 1 KHz
@ 26 Hz
@ 212 KHz
I took the amp with me in one of the labs where I had access to a scope so I could perform some tests. As you can see from the attached scope pictures, the frequency response isn't too shabby – 26 Hz to just a little over 200 KHz at full volume right before clipping point, with a 33 Ohm resistor as the output load. While the high-end response is quite good, the shape of the low-end of the frequency response was kind of ugly. Unfortunately, halfway through the tests I noticed that I forgot to change my output coupling caps from some earlier testing. So instead of 1000 uF as I originally planned (and as the pictures show), I had 470 uF caps. CapXons at that too
. But I didn't have any other caps with me on hand, so I just left the test results alone. The low-frequency output would probably have looked better if I used 1000 uF caps. When I was listening to some music with lots of low-end bass, however, I was not able to tell a difference.Gain:
I used the scope to determine that and compared it to my calculated values. Turns out, they weren't too far from each other, with my calculated gain of 18.7 dB (8.6 V/V) being the minimum gain for the amplifier at the above frequency response. At 1 to 10 kHz, the gain was closer to 21.7 dB (12.3 V/V).
Output:
The highest output I could get with that 9V SMPS adapter was about 3 Vp-p (peak to peak). This, with a 240-250 mVp-p input from the signal generator. Translating this into RMS values: approximately 1.06 V RMS output with about 0.08662 V RMS input.
Power:
With the above output, that means I was getting about 34.1 mW RMS into that 33 Ohm resistor. It's not a whole lot, but still plenty IMO. My headphones (cheap classic Sony MDR-023) are rated for 24 Ohms and input sensitivity of 100 dB per mW. That means I can technically push them up to 46.9 mW with this amplifier, or about 117 dB. More than enough for me.
Efficiency:
Not good. Not good…
I don't exactly remember what current it was drawing from that 9V power adapter, but I remember it was in the ballpark of 100-and-something mA and the calculated power between 1.1 to 1.4 Watts total (so about 550 to 700 mW per channel). With those numbers, that's roughly 5 to 6% efficiency with the 33 Ohm resistor and should be 7 to 9% efficiency with my headphones.
Yeah, it's pretty low
. But again, this is a class A amplifier, so this is more or less normal. I remember reading a while back somewhere that around 10% is generally to be expected. Could get it up to 50% if using impedance-matching transformer… but then there goes the simplicity! Unfortunately, I needed my breadboard for a project in another class, so I had to take the amplifier apart. All fine though, because since then, I've built another class A amplifier – this time bigger, slightly improved, and meant to drive actual speakers. It still has similar gain and all but the output stage is slightly different (originally, I wanted to do class AB, but I still can't figure out how to do that properly without distortion).
Anyways, when I get around to it, I'll draw and post a schematic of this “new” design. It's nothing to brag about either (only capable of about 600 mW into 4 to 8 Ohms), but still quite fun to build and have around nonetheless.
Attached Files
.
Comment