As an expercience desinger on power electronics I know buck converters can be really tricky with several strings aatached, so I recommend two things:
1. rated power shall meet used power, and input & output voltage shall not be to far apart to reduce efficieny losses (heat -> cooler)
2. don't start to build one on your own on opamps, unless you're an experienced electronics engineer. get a preconfigured designed one. nearly every semicondutor company provides those.
I had own a desktop, analog PSU built on transformer with voltage regulation based on multiple transformer taps selected by a set of relays with a current limiting made on op-amps. It had 2x150W output power and a significant weight that I didn't enjoy to carry at all
Then I decided to go for a SMPS unit with an adequate power and weight of 1/10 of the previous one.
From my experience I can tell, that analog design reacts incomparably faster in overcurrent situations, when the PSU is switching from constant voltage to constant current mode. I've been thinking to record the response of each one with an oscilloscope to have unbeatable proof and gut the SMPS one, as the delay in reaction could be related to improperly designed feedback signal.
Of course, the SMPS are excellent choice if You're looking forward compactness, efficacy of energy conversion and a wide range of input voltage, and You can fore-tell the power consumption to calculate adequate inductors and so on. For a lab purpose, I'd rather stick to analog one.
It seems like a reasonable option if you need to do something custom (e.g., for high currents or weird voltages) and low-BOM. They run much faster than typical switching speeds, tolerate a good range of voltages, and don't need a whole lot of power.
When building hobby boards, building in a buck/boost always feels like a big challenge. I've so far opted to buy small boards from Ali express, and have footprint that matches them on my project, and solder them directly on top. Choosing correct ICs and inductors feels too daunting.
As a hobbyist, I'm usually fine with what USB provides. Either I have a microcontroller board that connects to a computer's USB port or I can can plug a spare USB power supply into an outlet, so I don't need to think much about this stuff.
So it seems like this sort of thing will be more useful in battery-powered circuits? What's a good way to regulate voltage in that situation?
Even for battery operation, a USB battery pack might make sense? Looks like there are decent ones for $20 and some of the devices here are in the $2-10 range, without the battery.
Regarding your use case, I think it really depends. Most 32-bit MCUs want 3.3 V. Some chips may want 1.8 V. But if you're using these premade "prototyping boards", they usually come with all the necessary regulation. You might still need some way of powering accessories, though - LCD, OLED, motors, etc. The onboard regulator on the prototyping board might be not enough for that. Some will come with their own prototyping boards with their own regulators, but some won't. Plus, the costs of these boards add up.
Battery power - you can certainly run most 8-bit MCUs off of 2-3 AA/AAA cells, or a single Li-poly cell, without regulation. Peripherals might be more hit-and-miss. 32-bit MCUs generally need the supply voltage to stay within narrower bounds, so some regulation is necessary, and linear regulators tend to be wasteful. These plug-and-play switching modules I mention in the text - stuff like CUI VXO7803-500 - are probably the simplest hobbyist-friendly choice. That CUI module costs $2.40 or something like that, and you don't need external inductors to make it work.
It is probably worth noting that in practical buck and boost converters, the diodes are often replaced with synchronously operated MOSFET switches. This causes no functional difference, but avoids the inherent p-n voltage drop.
As an expercience desinger on power electronics I know buck converters can be really tricky with several strings aatached, so I recommend two things:
1. rated power shall meet used power, and input & output voltage shall not be to far apart to reduce efficieny losses (heat -> cooler)
2. don't start to build one on your own on opamps, unless you're an experienced electronics engineer. get a preconfigured designed one. nearly every semicondutor company provides those.
I wonder how well a boost or buck converter could be implemented with an AVR-DB controller. I can imagine the internal op-amps could come in handy.
I had own a desktop, analog PSU built on transformer with voltage regulation based on multiple transformer taps selected by a set of relays with a current limiting made on op-amps. It had 2x150W output power and a significant weight that I didn't enjoy to carry at all
Then I decided to go for a SMPS unit with an adequate power and weight of 1/10 of the previous one.
From my experience I can tell, that analog design reacts incomparably faster in overcurrent situations, when the PSU is switching from constant voltage to constant current mode. I've been thinking to record the response of each one with an oscilloscope to have unbeatable proof and gut the SMPS one, as the delay in reaction could be related to improperly designed feedback signal.
Of course, the SMPS are excellent choice if You're looking forward compactness, efficacy of energy conversion and a wide range of input voltage, and You can fore-tell the power consumption to calculate adequate inductors and so on. For a lab purpose, I'd rather stick to analog one.
It seems like a reasonable option if you need to do something custom (e.g., for high currents or weird voltages) and low-BOM. They run much faster than typical switching speeds, tolerate a good range of voltages, and don't need a whole lot of power.
When building hobby boards, building in a buck/boost always feels like a big challenge. I've so far opted to buy small boards from Ali express, and have footprint that matches them on my project, and solder them directly on top. Choosing correct ICs and inductors feels too daunting.
Look into the plug-and-play modules I mention briefly in the text - CUI VXO7803-500 and the like. They're pretty inexpensive.
As a hobbyist, I'm usually fine with what USB provides. Either I have a microcontroller board that connects to a computer's USB port or I can can plug a spare USB power supply into an outlet, so I don't need to think much about this stuff.
So it seems like this sort of thing will be more useful in battery-powered circuits? What's a good way to regulate voltage in that situation?
Even for battery operation, a USB battery pack might make sense? Looks like there are decent ones for $20 and some of the devices here are in the $2-10 range, without the battery.
Regarding your use case, I think it really depends. Most 32-bit MCUs want 3.3 V. Some chips may want 1.8 V. But if you're using these premade "prototyping boards", they usually come with all the necessary regulation. You might still need some way of powering accessories, though - LCD, OLED, motors, etc. The onboard regulator on the prototyping board might be not enough for that. Some will come with their own prototyping boards with their own regulators, but some won't. Plus, the costs of these boards add up.
Battery power - you can certainly run most 8-bit MCUs off of 2-3 AA/AAA cells, or a single Li-poly cell, without regulation. Peripherals might be more hit-and-miss. 32-bit MCUs generally need the supply voltage to stay within narrower bounds, so some regulation is necessary, and linear regulators tend to be wasteful. These plug-and-play switching modules I mention in the text - stuff like CUI VXO7803-500 - are probably the simplest hobbyist-friendly choice. That CUI module costs $2.40 or something like that, and you don't need external inductors to make it work.
It is probably worth noting that in practical buck and boost converters, the diodes are often replaced with synchronously operated MOSFET switches. This causes no functional difference, but avoids the inherent p-n voltage drop.