Kat Mitchell

Hello there, long time no see. Here are some floppy disks I finished today. #pixelart #vaporwave #lofi #retro

Plans for 2026: Music Cassettes, Floppy Disks, Retrofuturism, and more! | Northernpaws Get more from Northernpaws on Patreon

Made a post on my Patreon giving a small summary of the first couple months of 2026, and a peak at my project plans for the year moving forward!

www.patreon.com/posts/plans-...

フットプリント起こすとき #shapoart

sleep well!

They’re always issues that support clearly knows come up, since they have prerecorded help material, but it’s all left out of the manuals!

I can only assume it’s purposefully left out to force businesses into ongoing support contracts they (in this case) don’t even need after it runs.

I can’t believe how often i’m running into these problems with tech support companies for my commercial client - they’ve repeatedly got hardware failing within warranty periods, or not working on delivery, and tech support tries to get large payouts to give one or two photos showing what to do.

It had a lovely mess of issues, including a dead hidden lipo battery for a sensor that tech support refused to tell us the location of and was left out of the manual. Eventually after a month they decided to waive the $500/m support fee and help get it to run.. by sending me a pre-recorded video

Big win today! After months of dealing with unhelpful tech support, coordinating with multiple people, and waiting for parts.. We finally got an automated cotton candy machine running I’ve been helping a client with since November!

I feel like I need more info on why you have an x-ray interlock key

Still roughing in the controls, but the general idea is:
- one column of navigation/menu buttons down the left
- contextual buttons along the bottom of the display
- numpad and some navigation keys on the right side for entering coordinates
- rotary encoder for menu navigation and value entry

The plan is to have a 7" LCD driven by an STM32H7 that does most of the processing and USB-C communication with the game. A 100BASE-TX ethernet PHY (or optional Wi-Fi AT-modem) for downloading and querying community star charts.

Schematic half finished, routing & layout only roughed in.

I started roughing in an idea I've had for awhile for a controller for Elite:Dangerous - a fairly basic navigation/systems panel that connects to the game's local API to display ship status and navigation data, but can also load advanced navigation data from community start charts and plan routes.

why we stop making computers look fresh

I drove past a local bank last night with a yellow LED bar sign under their logo that flashed a few times and then printed “EFI SHELL VERSION xxxx (xxxx)” across the display in huge letters.. makes me wonder if that is suppose to happen, and what the process is that triggered it - an update maybe?

I just hope you don’t get surprised by a bill from the courier a month or two down the road - I’ve had that happen to me before with FedEx after 3 months 😕

Lesson learned to not trust the values recommended by the manufacture - especially when they don't document why those values were selected for the recommended circuit.

My guess is the circuit was tuned for telephony use, they seem to regularly cut out bass, which supposedly helps voice clarity?

So I took my tweezers and tried briefly shorting one of the DC blocking caps, and suddenly my audio signal had bass!

I went and ran the calculations myself, and yeah, assuming an 600 ohm line impedance it should have been at least 47uF. Bodging in some 47uF caps has solved it for now.

The manufacture of the audio codec I used recommended 1uF DC blocking capacitors on the line inputs and outputs. I noticed looking at schematics for other audio line-level drivers, the DC blocking caps tended to be between 47-100uF - that's quite a large difference!

I finally got around to removing the bead and bridging the pads.. and the bass was still missing! I was rather confused and started looking over the schematics again in detail, but everything seemed correct.

So I started comparing my audio path to other schematics.. and oh

Last week I got all the code working to start playing audio samples on the board, and realized that all of them were also missing their bass! My first suspect was a ferrite board that was in the audio ground path, a leftover from testing an old FM antenna circuit that coupled to the aux cable.

Story time:

I was having several issues on the music sequencer boards a couple of months ago where it seemed like my sine wave oscillators I was testing were missing most of their bass.

At the time I chalked it up to possibly a problem with my oscillator code, even though it seemed fine.

A dual-board option makes future revisions much more cost effective, as I can make changes to things like the input layout without needing to also buy and assemble an entirely new audio processing circuit unnecessarily when all I needed to change was a few buttons!

I'm also going with a dual-board design for this next revision. The main board with the audio processor will sit behind the display and do most of the heavy lifting, while input (buttons, lights, sensors, etc.) will be handled by a secondary board.

Using a BGA footprint for the STM32 also gave me a lot more options for peripherals - notably now have enough pins to use the STM32's LTDC driver to drive the LCD over an RGB interface. This simplified routing because I no longer have the memory bus routed to both the LCD and the SDRAM!

It's amazing seeing how far I've been able to condense down the board compared to the first revision from July.

Using a few BGA components and via-in-pad has helped greatly, but a large part of it can be attributed to the learning I was able to do on the first board and apply to this one.

A screenshot of a circuit board layout, with a web colourful traces flowing out from a microcontroller to various components on the board.

The schematic and routing for the next revision of my music sequencer hardware is well underway! I decided to swap to using BMA footprints for the MCU and SDRAM, which has ended up saving a significant amount of space and made routing much easier.

STM32 FMC SDRAM Overhaul and NOR/PSRAM/SRAM Support by northernpaws · Pull Request #5293 · embassy-rs/embassy This PR intends to overhaul the STM32 FMC driver to resolve several issues with the current driver, and to fill it out with a variety of missing features. In particular, the FMC driver in it’s curr...

Additionally, I've started building out a re-write of FMC driver in Embassy. Currently it's got a lot of problems that stem from it's reliance on an unmaintained external crate to provide most of the driver functionality, so I'm stripping that out with a native driver.

github.com/embassy-rs/e...

Extend STM32 DMA interface with memory-to-memory transfer functionality by northernpaws · Pull Request #5204 · embassy-rs/embassy This PR enables the use of DMA memory-to-memory transfers, a core feature of DMA that's missing in the STM32 HAL and required to make the DMA interface feature complete. Memory-to-memory transf...

We also added support for memory-to-memory DMA transfers which was missing from the STM32 DMA driver. We've been using these to drive our sequencer display via an 8080-parallel interface with the FMC, without DMA mem-to-mem the display framerate was too low.

github.com/embassy-rs/e...

Add support for fractional PLL multipliers on STMH743ZI by northernpaws · Pull Request #5155 · embassy-rs/embassy The STM32H743zi has an option for fractional PLL multipliers that's currently not exposed by the PLL configs or used by the VCO clock calculation. This commit adds support for the fractional mu...

We added missing support for the fractional PLL multiplier on the STM32H7 line. This was particularly important for my projects, because the fractional multiplier allows tuning the SAI clock to be significantly closer to the desired sample rate than is otherwise possible.

github.com/embassy-rs/e...

Been busy the last couple of weeks - got a couple of PRs merged into Embassy, been cleaning up my sequencer code in various areas to start building the foundation for writing the user interface, and started work on the next revision of the sequencer hardware!