The opacity glass panel needed 24V AC to switch states. The only available power source on the product was a USB-C port running at 5V. My first instinct was to look for a glass supplier with a lower drive voltage. None of the alternatives fit the spec. So we boosted.
That project clarified something I’d been taking for granted: most hardware engineers reach for step-down regulators by default and treat boost converters as the exception. It’s backwards. If your power supply doesn’t match what your load needs, you go up or down. The direction shouldn’t change how hard you think about it.
The topology is simpler than the reputation
A boost converter uses the same basic building blocks as a buck: a switching regulator IC, an inductor, a diode, and a handful of capacitors. The MT3608 is a classic example. Under a dollar in quantity. Takes 2V in, delivers up to 28V out. The design process is close enough to a standard step-down that if you’ve built a buck converter before, you can get a boost working on a first prototype without a lot of drama.
What changes is how you think about the power budget.
P = V x I. If you’re boosting 5V at 1A (5W) up to 25V, you’re not getting 1A at 25V. With a good layout and 85% to 90% efficiency (realistic for a modern boost IC), you land around 180mA to 200mA at the output. That’s not a surprise if you think it through first. But I’ve seen people size a boost converter by output voltage alone and then wonder why their load drops out under current.
Efficiency matters more in a boost than in a typical LDO. Every percent you lose becomes heat at the switch and inductor. For small loads like driving an opacity glass panel, that’s manageable. For anything over a few watts, component selection, switching frequency, and PCB layout start having real consequences on thermal performance.
When the voltage goes up, the safety story changes
Here’s the part that trips people up on first contact with higher voltages: creepage and clearance.
These are the minimum distances you need to maintain between conductors at different potentials. The required distances scale with voltage. At 5V, standard PCB design rules cover you. At 50V, you start consulting IPC-2221 tables. At 100V or above, your PCB house needs to know what you’re building, and your design review should include a trace-by-trace voltage audit.
Medical devices raise the stakes further. IEC 60601 places strict limits on voltages and currents in patient-applicable parts. What’s acceptable in an industrial display controller could be a certification blocker in a device that makes physical contact with a patient. The definition of “high voltage” shifts depending on the standard and the application. 24V is low voltage by industrial standards. It’s not low voltage everywhere.
None of this is a reason to avoid boosting. It’s a reason to know you’re doing it and design accordingly from the start, not as a retrofit.
How I typically approach it on a prototype
I start from a pre-certified AC/DC module taking 110V or 220V and outputting 5V or 12V. Mean Well makes good ones. The isolation and mains safety certification are already handled. The regulatory exposure for that stage is minimal, and it gives me a clean DC rail to work from.
From there, if I need higher voltages, I boost. I build close to the reference design in the IC datasheet, then test under real load conditions early. Not in simulation. Real current draw, real thermal measurements, real efficiency numbers. I’ve switched topology more than once after measured efficiency fell well below the datasheet typical curve under the actual operating conditions.
Noise is the other thing I check early. Boost converters switch fast and inject ripple onto the output and back onto the input. If you have sensitive analog circuits nearby, that ripple shows up where you don’t want it. Filtering and layout discipline matter more than most datasheets suggest for this.
The decision to boost is a system question: what input is available, what does the load need, and at what power level? That question belongs in the architecture phase, not after the schematic is done.
One number that consistently surprises people
You can generate thousands of volts from a 9V battery. A flyback converter or a dedicated high-voltage boost IC can push 9V up to 1000V or more. The output current will be tiny, maybe microamps, but the voltage is real. An ESD discharge from a fingertip reaches several thousand volts. The energy is low, but it’s enough to destroy unprotected silicon.
Voltage alone is not the right unit for reasoning about whether something is safe or useful. Power (watts) and energy (joules, or amp-hours for batteries) are what tell you whether a circuit can actually do what you’re asking of it. Get those numbers on paper before you build. The voltage you need is a consequence of the application, not the starting point.