Since this power supply is just a fun design for an upcoming Nixie tube clock project of mine, I have the time to achieve ESE. While in Part 1 I described the equations and simulations, in this Part 2, I collected experimental results to complete the design. In the process of finalizing the design, I was able to discover a couple of key design improvements and I’ll share these changes with you. The updated schematic, BOM, Kicad Layout, and design files are located at Github.
In AC/DC power converters beyond a few watts, during the initial application of power an excessive inrush current will flow when the input capacitors are suddenly charged. If unhindered the inrush current can easily exceed 50 A at the peak of the AC cycle and severely stress the converter’s fuse and input rectifiers, thereby significantly reducing the reliability and life expectancy of the modules. Universal power supplies (supplies which accept a wide range of input voltages) are particularly susceptible to high inrush current since their input capacitors must be large enough to handle line voltages as low as 110 VAC, as well as voltages as high as 305 VAC at start-up. In these environments, a power-supply failure or a tripped circuit breaker can be inconvenient at best, and expensive or dangerous at worst.
Designing a power supply for FPGA includes multiple voltage, ripple management and power sequencing, here’s an app note from Maxim Integrated. Link here (PDF)
Field-programmable gate arrays (FPGAs) and complex programmable logic devices (CPLDs) require 3 to 15, or even more, voltage rails. The logic fabric is usually at the latest process technology node that determines the core supply voltage. Configuration, housekeeping circuitry, various I/Os, serializer/deserializer (SerDes) transceivers, clock managers, and other functions all have differing requirements for voltage rails, sequencing/tracking, and voltage ripple limits. An engineer must consider all of these issues when designing a power supply for an FPGA.
My original plan was to find a replacement LCD and restore the unit to its original full functionality. But the LCD used in this unit is likely specifically made for the 169X series of power supplies and through some initial research I realized it would be extremely difficult to get hold of unless I could find a donor unit with a functional LCD inside. After I received the power supply, I realized that it had more issues than just the broken LCD itself. During my initial testing, I found that the output would not reach higher than 10 to 11 volts even with the over voltage protection set to the maximum value (20.5V). So clearly I have more homework to do, and for the time being let’s simply strip it down and see what’s inside.
Teardown and repair of an Agilent E3632A DC power supply from The Signal Path:
In this episode Shahriar & Rosanah investigate an Agilent power supply which does not appear to power on. It can be quickly observed that the fuse has failed on the unit. Using an isolation transformer a small amount of AC voltage is applied to the unit after the fuse replacement. It is clear that a short is present somewhere in the instrument since even at 10V AC the instrument consumes more than 1A.
Nicu Florica has been working on a power supply project, inspired by Albasete’s power supply unit with LM723 and Arduino volt and ampermeter:
It use an Arduino nano board with i2c LCD1602 display, a active buzzer for indicate shortcircuit case. Also, I put DS18B20 temperature sensor and relay for power a cooler when tenmperature is bigger than a threshold level.
For albasete version, I write psu_reber_ver3ro.ino sketch. In this sketch I put value for albasete case (R1 = 1k put between GND and A1 port, R2 = 39k put between +OUT and A1) and value for threshold (temax) and hysteresis value
What does one do when designing a power supply? Well, build a power supply tester, of course. One of the simplest things to build is a constant current load. This will allow for testing of the endurance of the power supply, as most of the designs out there are using slow components.
However, I wanted to make a better one: one that I could hook up to my Analog Discovery and generate a test waveform to be able to connect and disconnect the load fast. This is a weekend project, so all parts are not the best for the purpose, just what I had around.
I decided that the best way to exploit the results of the reverse engineering effort was to design a controller board that would host the NXA66 and expose its functionality via a front panel. I’d throw in a few simple extras myself such as current monitoring and data logging and finally I’d implement it as a through-hole design so that it could be implemented by people of all skill and equipment levels.
The end result will be a bench-power supply that’s cheap to build and has a current supply level greater than that of most supplies priced at hobbyist levels.
[Wolf] came into possession of an Extech power supply that wasn’t quite in working order. It has been used in battery manufacturing and was fairly corroded. He was able to fix it but found there was an issue with the power supply that wasn’t a defect. By design when you turn off the outputs, the voltmeters read zero. That means you can’t adjust the voltage to a known value without turning on the outputs. Sure, you ought to disconnect things before you adjust, but you can only hope you’ll remember.
At first, he tried to use the existing output control switch, but that really cut power. Instead, he turned to a small microcontroller board usually used for servo control. He added a few nice looking pushbuttons to the front panel. There was plenty of room in the enclosure to mount the controller board and four relays. You can see the final result in the video below.
You can guess the rest. The micro is able to read the controls, set the power supply, and switch the outputs off without killing the metering. This required some major mechanical surgery on the output terminals, by the way. In addition, the micro monitors the voltage output with an analog to digital converter and stores state when the power is dropping out. That way it can restore things on the next power event.
[Wolf] did eight videos covering each step of the process. You can find the result in the video below, but be sure to watch the ones that lead up to it as well.