DIYQueen & CapacitorX
DIYQueen DIYQueen
CapacitorX CapacitorX
DIYQueen DIYQueen
Hey, I was dreaming up a DIY smart power supply that can switch between multiple voltages with perfect ripple control—think of it as a tiny, clean, programmable charger that you can tweak in real time. Would love to bounce ideas and see if we can make it super stable and user‑friendly. What do you think?
CapacitorX CapacitorX
Alright, start with a single buck‑boost controller like the LTC3115‑1 – it handles 3.3 V to 30 V swings and gives you a clean output if you keep the inductor and MOSFET tight. Use a 100 µH inductor, a 10 µF input cap, and a 0.1 µF bypass on the output. For ripple, add a 10 µF/100 µF parallel filter and a 1 kΩ sense resistor. Feed the output back to a 12‑bit ADC on a microcontroller, run a simple PID loop to adjust the duty cycle. For voltage selection, use a digital potentiometer or a small MOSFET array that pulls the reference voltage. Keep the layout minimal, keep the heatsink on the MOSFET, and test each step with a small load before scaling. That’s the skeleton; tweak the component values and you’ll get a stable, tweak‑friendly supply.
DIYQueen DIYQueen
Sounds awesome! I’m already picturing the PCB with the little 100 µH winding and the tiny heat sink on that MOSFET. The 12‑bit ADC on the MCU will let us keep the ripple under control, right? Just remember to keep the inductor close to the switch for that high‑frequency current path, and the 10 µF input cap to catch any surge. For the digital pot, maybe a 10 kΩ one so we can dial the reference up or down without messing up the loop. If we test with a 10 W load first, we’ll see how the heatsink handles it. I’m already jotting down notes on the layout—watch out for those stray traces; I’ll stick to a 4‑layer board to keep things tidy. Let me know if you need a quick sketch or help with the PCB design!
CapacitorX CapacitorX
Nice, your plan is solid. A 4‑layer board with a dedicated ground plane and a copper pours under the inductor and MOSFET will keep the high‑frequency return path clean. Keep the sense resistor right next to the MOSFET to minimise the error from trace resistance. For the digital pot, 10 kΩ is fine, just remember the internal reference on the MCU, so you’ll need a stable reference voltage for the ADC. If you run into 10 W, the MOSFET’s RDS(on) will dictate the heat; a 1 W/m² rating means you’ll need a decent heatsink or a thermal pad. I can sketch a basic layout grid for you, just give me the dimensions and any component constraints.We comply.Just give me the board size and the exact part numbers for the MOSFET and pot, and I’ll draw a basic floorplan. That’ll let you see the trace widths and spacing before you run the first prototype.
DIYQueen DIYQueen
Great, here’s what I’m thinking for the prototype: 30 mm × 20 mm board size, so we have enough room for the inductor and a small heatsink. Use an IRLZ44N (5 V logic gate, 47 A, 1.3 Ω RDS(on)) as the main switch; it’s cheap and runs cool with a small 2 cm × 2 cm heat spreader. For the digital pot, a Microchip MCP41010 (10 kΩ, 4‑bit, SPI) – it’s tiny, 10 mm × 7 mm, and gives us plenty of granularity. That should give you a nice floorplan to start with. Let me know if you want me to tweak any of those specs!
CapacitorX CapacitorX
Board 30 mm × 20 mm, 4 layers. Layer 1: copper pour ground, 35 µm copper. Layer 2: power, 30 µm copper. Layer 3: signals, 25 µm copper. Layer 4: top copper pour for components. Place IRLZ44N near the center, source to GND plane, drain to inductor. Keep the inductor wire loop tight, feed from 12 V VIN through 100 µH. Add 10 µF X7R ceramic at VIN, 0.1 µF bypass. The MCP41010 sits on the top layer, VCC to 5 V rail, GND to ground, CS, SCK, SI to MCU pins. Sense resistor (0.1 Ω) right next to MOSFET drain. Keep all signal traces under 200 µm width, 0.15 mm spacing. Use a 2 cm × 2 cm copper spreader under the MOSFET, add a 0.25 mm thermal pad on the PCB. That should give you a clean path and good thermal relief. Adjust the inductor value if you see high ripple under 10 W load. Good luck.
DIYQueen DIYQueen
Sounds like a solid sketch! I’m already picturing the copper pour under the IRLZ44N, and that 0.25 mm thermal pad will keep the hot spot from getting out of hand. The 10 µF X7R at VIN will tame any kick‑in, and that 0.1 µF bypass will make the voltage feel silky smooth. For the MCU, I’ll drop the SPI pins close together so the SCK, CS, and SI traces stay under 200 µm—keep those pads tidy and we’ll avoid crosstalk. I’ll start drawing the layers and will loop back if I hit any snag with the inductor loop or the 2 cm × 2 cm spreader placement. Let me know if you want a quick preview of the top‑layer footprint before I lock it in!
CapacitorX CapacitorX
Sure, send me the exact pad layout for the top layer, and I’ll double‑check the spacing and trace widths before you lock it in.
DIYQueen DIYQueen
Top‑layer pad plan (board 30 mm × 20 mm, origin at bottom‑left): - **IRLZ44N** – center (15 mm × 10 mm) - Source pad: 0.8 mm square at (15 mm, 10 mm) - Drain pad: 0.8 mm square at (15 mm, 12 mm) – keep a 0.15 mm gap from source pad - Gate pad: 0.6 mm square at (15 mm, 8 mm) – 0.15 mm away from source pad - **MCP41010** – top‑left quadrant, 3 × 2 pad matrix spaced 0.15 mm apart, starting at (3 mm, 17 mm) - VCC: 1.0 mm × 1.0 mm at (3 mm, 17 mm) - GND: 1.0 mm × 1.0 mm at (5 mm, 17 mm) - CS: 1.0 mm × 1.0 mm at (7 mm, 17 mm) - SCK: 1.0 mm × 1.0 mm at (9 mm, 17 mm) - SI: 1.0 mm × 1.0 mm at (11 mm, 17 mm) - GND (second): 1.0 mm × 1.0 mm at (13 mm, 17 mm) - **Sense resistor** – just below the drain pad: 0.6 mm × 0.6 mm at (15 mm, 14 mm), 0.15 mm gap to drain pad. - **VIN input capacitor** – bottom‑left corner: 1.0 mm × 1.0 mm at (4 mm, 1 mm). All signal pads are 200 µm (0.2 mm) wide traces, 0.15 mm spacing, and routed to the nearest V‑top copper pour. This layout keeps the high‑frequency return tight and gives you a clean path for the MOSFET, pot, and sense resistor. Let me know if anything needs tweaking before you lock it in!
CapacitorX CapacitorX
Looks solid – just double‑check that the source pad sits flush on the ground plane so the return path stays short, and keep the sense resistor really close to the drain so the error voltage is negligible. Other than that, you can lock it in. Good job.
DIYQueen DIYQueen
Thanks! I’ll make sure the source pad sits right on the ground plane and that the sense resistor is next to the drain. Ready to lock it in and start the first run!