Hey all. I'm a hobbyist tinkerer, and for my next project I'm wanting to build a USB-powered audio amplifier for some headphones. The headphones aren't anything great, and I'll find some generic Burr-Brown op-amps (suggestions?) to fire them with. USB is for power-only.
I know I could buy some powered headphones, or a premade amp, but that defeats the whole purpose of me building it.
I'd like guidance on building a small SMPS, even as far as winding my own inductors if there's nothing off-the-shelf (I have plenty of scavenged toroid cores). I'd like to use a boost converter (I'm still not very versed with inductors), and some mechanism to invert the voltage.
I am fairly familiar with driving N-channel equipment, and prefer FETs. I'd like for the supply to be self-regulating and hang around +- 9v regardless of output load (as I understand it, minimum load is quite the pain for a dumb SMPS).
Suggestions and such welcome!
I know I could buy some powered headphones, or a premade amp, but that defeats the whole purpose of me building it.
I'd like guidance on building a small SMPS, even as far as winding my own inductors if there's nothing off-the-shelf (I have plenty of scavenged toroid cores). I'd like to use a boost converter (I'm still not very versed with inductors), and some mechanism to invert the voltage.
I am fairly familiar with driving N-channel equipment, and prefer FETs. I'd like for the supply to be self-regulating and hang around +- 9v regardless of output load (as I understand it, minimum load is quite the pain for a dumb SMPS).
Suggestions and such welcome!
You're picking a particularly challenging combination of components:
SMPS, USB powered & headphones.
The headphones will be ruthless in their discrimination of PSU noise.
Depending on your PC, the USB power could be quite contaminated with all sorts of hash starting at the 1KHz USB frame rate and going on up to the UHF band (and beyond
I can't measure)
Not to say it can't be done well but you'll need to consider DC to light in your filter assessment.
Keep us posted and we'll help however we can
SMPS, USB powered & headphones.
The headphones will be ruthless in their discrimination of PSU noise.
Depending on your PC, the USB power could be quite contaminated with all sorts of hash starting at the 1KHz USB frame rate and going on up to the UHF band (and beyond
I can't measure)Not to say it can't be done well but you'll need to consider DC to light in your filter assessment.
Keep us posted and we'll help however we can
Do you really need that much voltage for the headphones? Rail-to-rail op-amps fed from filtered 5V may produce enough output. Another option is to produce a -5V rail and use the same R2R approach (doubles the output voltage).
If you want +/-9V you have to go for complex capacitive charge pumps, or a boost for +9V and an inductor based flyback for -9V, or a transformer based flyback for both rails.
A single rail and AC coupling may be somewhat easier (single charge pump or boost converter).
BTW: I have seen really simple self-oscillating charge pumps done with 74HC or CD4000 logic gates (buffers). One or two gates are used as an oscillator, and the rest of sections are paralleled to to drive the capacitors and diodes from a low impedance.
If you want +/-9V you have to go for complex capacitive charge pumps, or a boost for +9V and an inductor based flyback for -9V, or a transformer based flyback for both rails.
A single rail and AC coupling may be somewhat easier (single charge pump or boost converter).
BTW: I have seen really simple self-oscillating charge pumps done with 74HC or CD4000 logic gates (buffers). One or two gates are used as an oscillator, and the rest of sections are paralleled to to drive the capacitors and diodes from a low impedance.
One thing I was thinking of doing was getting something like this little Mouser job and building a push-pull driver for it. If I drive between the center-tap and one of the leads, I can use the other lead as a feedback to switch the transistor off. It's a common trick used on HV stuff, and will make the transformer always run at a resonant frequency (google 3055 flyback)
Since it's a 1:1 winding, I should be able to get 5v from each pair on the output, giving me 5v 0v -5v (in theory). Pump the output from that through a voltage doubler on the top and bottom, then an LCL filter, then some linear regulators, and I should be able to get a nice clean +-9v supply. Thoughts?
(google 3055 flyback)Since it's a 1:1 winding, I should be able to get 5v from each pair on the output, giving me 5v 0v -5v (in theory). Pump the output from that through a voltage doubler on the top and bottom, then an LCL filter, then some linear regulators, and I should be able to get a nice clean +-9v supply. Thoughts?
All-in-one isn't a criteria. I'd prefer stuff I can build on a breadboard, but it's not a requirement -- something I can build on a breadboard and debug via trial-and-error is simpler to me than printing a SMT board and hoping it works. I'd rather spend $10 on parts and have a discrete headphone amp I (mostly) designed and built myself, rather than a $1 amp on a chip that you solder a headphone jack to and call it done.
I'd also like to do this more DIY as a way to learn interactions between various components -- I'm still not very comfortable with my knowledge of transistors and op-amps operating in anything but full saturation mode.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.