I've always wanted to build my own headphones and was curious if anyone here has any experience or tips as I am not sure if the circuit can work.
The components I will use are as follows:
Breadboard
Battery
ceramic capacitor
Resistors
polar capacitor
BC237 NPN transistor (datasheet)
BC239 NPN transistor(BC239 DATASHEET)
MIC
Audio Jack
Here is the circuit.
The components I will use are as follows:
Breadboard
Battery
ceramic capacitor
Resistors
polar capacitor
BC237 NPN transistor (datasheet)
BC239 NPN transistor(BC239 DATASHEET)
MIC
Audio Jack
Here is the circuit.
Unfortunately, that circuit has a 2.2k series resistor which will block most of the signal for the headphone output.
TIP: First build a CMOY or something well known. Choose one that has opamp and 2 batteries instead of virtual ground. Please present a scaled schematic for readability. An attached image file should fill the window when clicked.
TIP: First build a CMOY or something well known. Choose one that has opamp and 2 batteries instead of virtual ground. Please present a scaled schematic for readability. An attached image file should fill the window when clicked.
It can't, not for useful headphone level. There is no path for the DC in the output transistor except through 168k or the headphones. Also the connection of the un-labeled Q2 Q3 seems to be a multi-vibrator... not a follower as would be expected.
Copy known-good plans and learn how they work. Mark Johnson has posted one in the last couple years, although not jellybean parts or performance.
One by lineup.
Or this is known to work and easy to follow:
The "0.2A IDC" can be (was originally) a 60 Ohm 5 Watt resistor.
Last edited:
You circuit has a few problems. Mostly the output makes ~no sense. What did you draw this with? I suggest:
1. Use LTC Spice to draw and simulate your circuits, including a load resistance. 30 Ohms is a fair approximation of headphones, unless you have better information.
2. Search for Elvee's headphone amp. It uses a low voltage like you did. Others on DIYA ignore such constraints. The problem with low voltage (3 Volts) circuits is that you can't waste voltage on bias and/or transistor saturation.
I tend to dismiss "headphone amplifiers" and class-A amplifiers as playthings for kids to fiddle with, but if you are serious, (<5V) battery power and distortion can be a real challenge. Some "headphone amplifiers" have a lot of 2nd harmonic distortion, ie asymmetric waveform distortion, which many people love the sound of. Other "commercial" headphone amplifier chips like the ones from TI are for use in cell phones and not particularly hifi.
A problem with simple AB amplifiers is thermal runaway but you can simply use relatively large emitter resistors (~10 Ohms) and still get plenty of output for headphones, and less heat than a class-A amplifier.
1. Use LTC Spice to draw and simulate your circuits, including a load resistance. 30 Ohms is a fair approximation of headphones, unless you have better information.
2. Search for Elvee's headphone amp. It uses a low voltage like you did. Others on DIYA ignore such constraints. The problem with low voltage (3 Volts) circuits is that you can't waste voltage on bias and/or transistor saturation.
I tend to dismiss "headphone amplifiers" and class-A amplifiers as playthings for kids to fiddle with, but if you are serious, (<5V) battery power and distortion can be a real challenge. Some "headphone amplifiers" have a lot of 2nd harmonic distortion, ie asymmetric waveform distortion, which many people love the sound of. Other "commercial" headphone amplifier chips like the ones from TI are for use in cell phones and not particularly hifi.
A problem with simple AB amplifiers is thermal runaway but you can simply use relatively large emitter resistors (~10 Ohms) and still get plenty of output for headphones, and less heat than a class-A amplifier.
Considering some of the other possible hobbies where energy is consumed for enjoyment, I would probably have to listen to a class-AB amplifier for a whole year to redeem myself for an unnecessary drive to the shops, or for a few additional sticks thrown onto the bonfire. Even more-so with headphones.
That said, it's certainly a fun intellectual exercise if the amplifier box can be made sleeker.
On the other-other hand, I think the alleged niceness of H2 harmonic distortion is mostly a myth. IMD is co-created at the same time, so if something still sounds better with the added distortion, it's likely a 3rd factor that reduces distortion elsewhere and someone didn't bother to investigate it. A few possibilities have been covered here: https://www.diyaudio.com/community/threads/current-drive-for-loudspeakers.250272/
One thing that often seems to go with class-A-ness is open-drain / open-collector output, which often provides a high output impedance. So, if the load contains distorting semi-inductance, a transconductance amplifier can automatically ignore that because it provides a linear current instead of a linear voltage. However, there are also source-follower / emitter-follower amplifiers where those benefits are lost.
My initial thinking was to maximise output impedance across the frequency range, but looking at the different elements that make up the total impedance of a magnetic speaker, the bass resonance probably benefits from low output impedance instead. And that's a pain. Going beyond 'DC servo' style negative feedback, which gives a high damping factor only in the bass, is fiendishly difficult, but doable.
Most class-AB designs seem to jump the gun, trying to create the perfect voltage amplifier, which is kind-of useless in the real world which doesn't have perfect speakers.
That said, it's certainly a fun intellectual exercise if the amplifier box can be made sleeker.
On the other-other hand, I think the alleged niceness of H2 harmonic distortion is mostly a myth. IMD is co-created at the same time, so if something still sounds better with the added distortion, it's likely a 3rd factor that reduces distortion elsewhere and someone didn't bother to investigate it. A few possibilities have been covered here: https://www.diyaudio.com/community/threads/current-drive-for-loudspeakers.250272/
One thing that often seems to go with class-A-ness is open-drain / open-collector output, which often provides a high output impedance. So, if the load contains distorting semi-inductance, a transconductance amplifier can automatically ignore that because it provides a linear current instead of a linear voltage. However, there are also source-follower / emitter-follower amplifiers where those benefits are lost.
My initial thinking was to maximise output impedance across the frequency range, but looking at the different elements that make up the total impedance of a magnetic speaker, the bass resonance probably benefits from low output impedance instead. And that's a pain. Going beyond 'DC servo' style negative feedback, which gives a high damping factor only in the bass, is fiendishly difficult, but doable.
Most class-AB designs seem to jump the gun, trying to create the perfect voltage amplifier, which is kind-of useless in the real world which doesn't have perfect speakers.
First - design. What is the output (load R and wattage) and what is the input (ie is it line level and what's the maximum voltage peaks?) - then what is the gain required to go from the output to the input. All the bells and whistles on top of that concept can be added in after.
You can then evaluate designs and ideas against your requirements but also borrow concepts accordily.
Next comes - how do you want to evaluate it? Fun with a spectrum analyser (ie a sound card will do) but also getting the bode plot out really helps you start to understand how the design is working.
^^^ sorry input output gain!
After working that out your world is your oyster.
For example - low impedance headphones need only minimal voltage gain (the same voltage in will often be more than enough for the voltage peak for the headphone) but instead the headphone power (P=I*I*R) means you’re looking for say 30mW into 55R means mAs only for
current but more than your line in - the design becomes more of a “buffer”, where the buffer supplies the current required to support the power requirements but doesn’t amplify voltage.
Higher impedance headphones (600R) means higher peak voltage than the line input peaks but less current - so a voltage amp design fits the bill.
There’s lots more, but that’s the fun in learning and experimenting.
Bigger speaker amps will need more voltage and current delivery, but the principle of gain (even gain of frequency spectrum - hence bode plots) remains the same.
From that you’ll learn about phase, about filtering and frequency response, noise, distortion… but those build on core understanding. You’ll find new designs that solve multiple problems in unique ways. Switch to digital and class D and that knowledge will follow you..
I would advise a pen and pencil, then work through the operating points, maths of NPN/PNP etc.
After working that out your world is your oyster.
For example - low impedance headphones need only minimal voltage gain (the same voltage in will often be more than enough for the voltage peak for the headphone) but instead the headphone power (P=I*I*R) means you’re looking for say 30mW into 55R means mAs only for
current but more than your line in - the design becomes more of a “buffer”, where the buffer supplies the current required to support the power requirements but doesn’t amplify voltage.
Higher impedance headphones (600R) means higher peak voltage than the line input peaks but less current - so a voltage amp design fits the bill.
There’s lots more, but that’s the fun in learning and experimenting.
Bigger speaker amps will need more voltage and current delivery, but the principle of gain (even gain of frequency spectrum - hence bode plots) remains the same.
From that you’ll learn about phase, about filtering and frequency response, noise, distortion… but those build on core understanding. You’ll find new designs that solve multiple problems in unique ways. Switch to digital and class D and that knowledge will follow you..
I would advise a pen and pencil, then work through the operating points, maths of NPN/PNP etc.
Last edited:
My headphones are 250 ohms, and the last time I built an amplifier for them, they only needed SOT-23 sized output transistors to supply 7-10mA of idle current. Maybe a slightly larger size like a SOIC-8 or SOT-223 or even a D-PAK could have provided a bit more thermal overhead. Never mind the actual design which was flawed anyway, but please, for all the TO-220 or bigger designs you've gotta ask yourself is it really necessary to put an 8 cylinder truck motor onto a lawnmower?
Gate charge should be kept to a minimum, but a downside of big transistors are those parasistic capacitances between G and D and so on.
Gate charge should be kept to a minimum, but a downside of big transistors are those parasistic capacitances between G and D and so on.