Ben Keppers’ Work Log

  • Post author:
  • Post last modified:April 6, 2024
  • Reading time:12 mins read
  • Post category:Work Logs

4/6/2024 – ATX Power Supply | Finishing Layout

This week I finally got around to finishing my layout. Because I’m mostly copying the evaluation board for the initial revision of the 5V standby flyback I followed the layout of the evaluation board closely. I did wind up making a few tweaks to accommodate some different sized components that I have. My final layout is below.

Going into next week we’ll order boards and components. Once they come in it’ll be crunch time to try to get them assembled and working before design expo.

3/30/2024 – ATX Power Supply | Still Layout

I didn’t put as much time into layout this week as I should have, so I don’t have my portion done yet. unfortunately this means next week will also be layout and there will be very little time for troubleshooting between when the boards come in and the end of the semester.

3/23/2024 – ATX Power Supply | Layout

This week I got started with layout for the 5V standby section. I hope to be finished sometime next week.

3/16/2024 – ATX Power Supply | Design Review Preparation

This week was mostly dedicated to cleaning up schematics and making a document with all of our design decisions to send to our sponsor for review.

3/9/2024 – ATX Power Supply | Component Selection

This week I got through compiling a BOM and assigning footprints for the 5V standby section. This has mostly been a matter of going through the BOM for the evaluation board I’m basing my design on, searching the components on Digikey, and putting them into my own BOM. Next week my team will be doing a schematic review with Plexus, who’s sponsoring this project, and I’ll get started on layout for the 5V Standby section.

2/24/2024 – ATX Power Supply | Transformers and EMC

This week I got word from Wurth Elektronik that they shipped samples of the transformers for the 5V standby flyback. I’ve also spent some time on the input EMI filter design based on a YouTube video I found. I’ve chosen initial values for X and Y capacitors based on the video’s recommendations. This week if I get some time I’ll spend some time on Digikey looking for a common mode choke. From the research I’ve done, I need to find one that has a decent amount of attenuation at our switching frequencies (60 kHz and 100 kHz) and their first few harmonics.

2/18/2024 – ATX Power Supply | Finishing 5V Standby Design

When I was doing my portion of project planning I said that by the end of this week I’d be done with the 5V standby design and have everything put into KiCad. I’ve tried to work through the design guide in the UCC38C42 datasheet for a while, but I’ve never been able to figure out a way to make sure my design will actually work when I get done with it. I did some more searching, and I stumbled across the UCC28704 and its evaluation board. It happens to take line voltage and step it down to 5V at 2A . . . exactly what I need. The one issue with it is the transformer it uses is only available directly from Würth Elektronik. The core appears to be available on DigiKey though, so I can always wind my own. Long story short, I migrated the evaluation board schematic into KiCad, and it will be the first iteration of our 5V standby supply.

2/10/2024 – ATX Power Supply | Transformer Selection

This week I found a transformer that should work with the controller IC I selected. This turned out to be a bit of a process, and the transformer I found appears to be the only transformer on DigiKey or Mouser that meets the specs I need. I hoped to run through some calculations and select a mosfet to use, but I ran out of time in the week. Hopefully I’ll get a chance to get some more work done tomorrow.

2/3/2024 – ATX Power Supply | 5V Standby Stuff

This week I read through the datasheet for the UCC38C42 PWM controller that I’m planning to use for the 5V standby. I have a better understanding of how it works, and I found design equations from the datasheet that I can use. Next week my goal will be to find a transformer and to select a MOSFET.

1/27/2024 – ATX Power Supply | The plan and my plan

To start off the semester we’re doing a thing where we each come up with a drawing of what our project is. Then we’ll compare them to make sure we’re all on the same page. I came up with a block diagram of our project in KiCad of how our project’s circuitry will be laid out. Going from the mains input to the main outputs, we have some EMI filtering going into an active power factor correction circuit, followed by an LLC converter that will produce a 12 V output. This 12 V output is then converted to 5 V, 3.3 V, and -12 V. Finally, a 5 V standby output is also required, so we’ll have a second converter coming out of the EMI filtering to produce the 5 V standby rail. This is a slightly atypical organization, but we think it will be simpler to design and get working. Typically there’s one main converter that produces all four main rails with a smaller 5 V standby converter.

My goal for this week is to finalize the 5 V standby circuit. Right now I have a circuit from TI’s webench power circuit generator, but I’m hesitant to blindly trust it. Throughout this week I’ll go through the datasheet for the flyback chip I’m using and verify that the circuit I have looks like it should work.

11/12/2023 – ATX Power Supply | 5V Standby

We’ve began selecting transformers and MOSFETs for the PFC and converter. This week I worked on finding a controller for the 5V standby rail. The target for this rail is 2 or 3 A at 5V, and it needs to be isolated. Searching TI’s WEBENCH design tool led me to the UCC28742. This chip looks promising. It requires relatively few external components, it’s one of the cheaper options, and it’s in stock on DigiKey. Next week my goal will be to come up with a rough design for this chip that will work well next to our main power conversion stages.

11/4/2023 – ATX Power Supply | Development Boards

In order to do some amount of testing of our design before we order parts we ordered a couple evaluation boards this week. For our LLC chip we had two options. The first has a single 120Vac input and included PFC correction and everything else that’s necessary. The second requires a ~400Vdc input and a 120Vac input, and it doesn’t have PFC. We chose to go with this board along with an evaluation board for the PFC chip Ben Abel has been working on a design for. Hopefully some hands on experience with these evaluation boards will give us a better understanding of how the chips we’re using work, and they’ll help us improve our design.

10/28/2023 – ATX Power Supply | Switching to an LLC for Main Power Conversion

This week and last week we made the decision to use an LLC converter for our initial conversion rather than a flyback. In our research it seemed like flybacks are typically only used for lower power applications. We plan to use the UCC256404 controller from TI. I found a Simplis simulation model and a spreadsheet to assist with calculating component values for the LLC circuit. We also found a series of SMPS transformers on DigiKey that we plan to use.

10/6/2023 – ATX Power Supply | Finding a Flyback Controller

My goal for this week and last week was to find a potential flyback controller IC and develop a spice simulation for it. From what my team has found so far a flyback controller may not be able to produce enough current, but the relative simplicity of a flyback makes it a very appealing option. After some searching I found the UCC28C52 from TI. This chip looks somewhat promising. Section 9.2 of its datasheet gives a schematic for a isolated 12V power supply rated for 4A. That’s a lot less than what we need, but hopefully with a different mosfet we can get closer to out 50A target.

There’s a PSpice model available for the UCC28C52, and I’ve been working on converting it to a QSpice model so we can use it with the rest of our simulations. I haven’t gotten it working yet, but I think I’ve gotten it down to one major error. After I work through this error it will hopefully work. Otherwise we may need to look at simulating the flyback in PSpice.

9/23/2023 – ATX Power Supply | Getting Started with EMI Filtering

This week’s goal was to learn what conducted EMC requirements we need to meet and how to design an EMI input filter to meet them. Ideally I’d be able to choose components for the input filter as well, but I thought we may need to know more about our switching architecture first. To begin, I looked for the conducted emissions limits we’ll need to meet. According to the Intel ATX Version 3.0 Design Guide we must meet Class B for both conducted and radiated emissions for FCC Part 15, EN55023, and CISPR 22. Being in the U.S. I decided FCC Part 15 would be a good starting point, and luckily these limits are readily available through the National Archives eCFR system. Conducted emissions are in a table under §15.107.

Frequency of Emission (MHz)Conducted Limit (dBμV)
Quasi-PeakAverage
0.15 – 0.566 to 55*56 to 46*
0.5 – 55646
5 – 306050
*Decreases with the logarithm of frequency Citation: 47 CFR 15.107

It’s nice to know what the limits are, but the limits themselves don’t say much about how to meet them. Fortunately the internet exists, and I was able to find this YouTube video that discusses choosing initial values for EMI input filter components. With this video and the above limits, I believe I’ve reached my goal this week, and in the future it should be relatively simple to choose components. Next week we’ll need to put a lot of work into ironing out our DC-DC conversion and case design.