PwrBlock is our open-source programmable power supply for test fixtures and electronics testing automation.

It is no longer just a schematic or a render. We already have the first assembled samples, and they have gone through the first real tests. This article is a short update on what PwrBlock is, what we have already confirmed on hardware, and what we are improving next.

Why We Are Building It

The idea behind PwrBlock is simple.

When building test fixtures, engineers often end up using either a full bench power supply with lots of unnecessary features, or a generic DC-DC module that works until it suddenly becomes the least predictable part of the setup.

We wanted something else: a compact, programmable, open hardware block that can be embedded directly into a test stand and controlled as part of the system.

PwrBlock is part of a larger plan to build a family of open hardware tools for electronics testing: power supplies, programmers, switch matrices, DAQ modules, and other building blocks that can be combined into complete fixtures.

What Is Inside

The current revision is built around a few key components.

Protection: TPS25983

• 2.7 V to 26 V
• 2.7 mΩ integrated 20 A FET

MCU: STM32G0C1CE

• Cortex-M0+ CPU, up to 64 MHz
• 512 KB flash
• USB Type-C Power Delivery controller

The main reason we chose this MCU family is the built-in Power Delivery controller. It lets us avoid using a dedicated USB PD IC.

Measurement: INA229

• up to 85 V
• 20-bit, ultra-precise
• SPI interface with alert output
• integrated temperature sensor, ±1 °C

This chip gives us accurate measurements of current, voltage, power, and energy at the output of the supply, without extra calibration. It also gives us temperature data.

Power stage: LM51772

• Buck-boost
• I2C configuration interface

This is the heart of the block. It can step voltage down, step it up, and be configured over I2C.

First Confirmed Power Results

One of the first things we wanted to validate was long-term operation near maximum output power.

So we ran a simple test with these conditions:

• Output: 32 V / 3 A
• Input: laboratory power supply
• Load: electronic load
• Board condition: no enclosure, no heatsink
• Duration: 3 hours under load

The results were:

• Input power: 103.64 W
• Output power: 96.17 W
• Efficiency: 92.7 > %
• Heat losses: 7.47 W

That was an important milestone for us. It means the project has already moved from concept to confirmed real-world operation close to the target power level. Although it's not correct to measure efficiency after the wiring.

Thermal Behavior

After three hours at around 96 W output power, we checked the thermal picture of the board.

The hottest point on the top side was the inductor body at 80 °C. The DC-DC converter itself stayed at 71–72 °C. On the bottom side, the hottest area was the thermal path from the inductor pads into the inner PCB layers, reaching 86 °C.

The inductor itself is rated for temperatures up to 165 °C, so we do not expect thermal problems from these results. Still, thermal chamber testing is definitely on the roadmap.

The main takeaway is clear: the block can already operate without overheating even without an enclosure. The enclosure test is still ahead, but it is already obvious where the next improvement should go. We will likely use a larger inductor, and thermal coupling into the enclosure through a thermal pad under the inductor area should work well.

Current Limiting at Low Current

Voltage regulation over I2C works exactly as expected.

We get:

• 10 mV steps in the 1–24 V range
• 20 mV steps in the 3.3–32 V range

We also added a useful extra feature: analog fine adjustment. It works well for small offsets around the digitally programmed voltage and allows us to reach about 1 mV adjustment steps.

This behavior is not explicitly described as a feature in the LM51772 documentation, but it works. We apply an offset to the ATRK input while controlling the output voltage over I2C.

On the current side, the LM51772 can limit not only peak current but also average current measured across an external shunt. That threshold can be adjusted over I2C.

This gives us a digital stepped current adjustment with roughly 50 mA resolution, but there is no practical adjustment below 500 mA.

So we tried a second approach: instead of a fixed ISET resistor, we used a transistor-based adjustable circuit driven from the MCU DAC.

That works too, but the low end of the range becomes unstable. Below 500 mA, we see output voltage ripple around 6.5 kHz.

In theory, by replacing the shunt with a larger one, we can shift the lower practical limit down to about 250 mA, while still keeping the maximum measurable current around 3.5 A.

So the practical result is:

• The internal hardware constant-current block of the LM51772 is usable for CC mode from 250 mA to 3 A
• The digital method gives about 25 mA current adjustment steps
• The analog method can give sub-1 mA adjustment
• Below 250 mA, we will need software CC mode

This is one of the most interesting findings from the first prototypes.

Enclosure

For the enclosure, we ended up with a simple design made from 1 mm sheet aluminum. It is inexpensive to manufacture even in small batches, and once assembled it turned out to be rigid enough.

The PCB is mounted to the lower half of the enclosure through standoffs soldered to the bottom side of the board. A thermal pad sits between the board and the enclosure for heat transfer. The labels are laser engraved, and the finish is anodized aluminum.

One of the early concerns was confirmed on the prototypes: anodizing isolates the enclosure electrically. The PCB standoffs did not make electrical contact with the case, so grounding did not work as intended. We will remove the anodized layer in the areas where electrical contact is needed.

After trying the block inside different fixture concepts, we also realized that we should add more mounting nuts on the lower side. That way it can be mounted almost anywhere.

We also decided to replace the current button with a similar one that has a more tactile click, and to add backlighting.

Overall, the block already feels great in hand: simple, solid, and pleasantly brutalist.

What Comes Next

The next steps are:

• thermal chamber testing
• EMC testing
• finishing the tuning of dynamic characteristics
• validating the lower current range with software CC mode
• refining the firmware
• building a real test fixture around an actual board using PwrBlock inside

Step by step, PwrBlock is turning into a practical tool for test fixtures, electronics automation, and open hardware testing systems.

Support the Project

We are building PwrBlock as an open tool for engineers who create test fixtures, small production setups, and automated test systems, or who simply want a compact programmable supply without bench-supply overhead.

If the idea of open-source tools for electronics testing resonates with you, please support the project on Crowd Supply.

https://www.crowdsupply.com/everypin/pwrblock-323

P.S. What Comes After PwrBlock

The next block in the family is CmptBlock.

It will combine a Raspberry Pi Compute Module with a 25 W power supply, analog and digital I/O, an ammeter, and an ohmmeter — basically a block that can serve as the core of a complete test fixture almost by itself.