exxos's Atari, Amiga & retro forum

After making the onboard LED of my fpga board blink, it's time to address some other pins that can be accessed through the expansion board.

I did a little assembly of leds on a single strip of 15 pins :


Then I wrote some code to actually setup a given pin and make it blink

And then I test (some pins of the connectors seems to be wrong on the second row and thus do not blink)

Today is about programming an ESP32 with the IDF toolchain (not arduino), I dived into the rabbit hole of addressing a LCD TFT controller (the ST7789) through SPI, without using existing libraries (I can always find something that don't please me to justify a 'not done by me' syndrome, like most of them seems to be singleton designed -no more than one screen out of the box-, and messes directly with setting SPI on the host mcu -I may want to manage SPI myself-). Anyway, to make a bridge with retrocomputing, my plan is to implement the frame buffer model and console of a Thomson MO5.

Yet another project in the air, maybe I will actually complete that one next year... https://github.com/sporniket/dop-module ... own-xl4005


The idea is to convert around 15V/20V into 5V up to 4/5 Amps using one of those Chinese specialized IC. The most difficult part to chose is the inductor, as it has to support quite high current. I finally selected a through-hole one from CoilCraft.

A cheap module based on this chip already exists, and there is even a video from Robert Feranec with guest Min Zhang that demonstrate how noisy it is and how to improve that :



So I hope to do a little better :D

Before this iteration, I did a version with more CMS (notably the inductor) and some fuses (PTC fuses). Finally I ruled out the fuses (less components, and I remember some post mortem from Great Scott on its LIPO charger published on Elektor, that it was causing malfunctions because of the inrush current at startup to fill the capacitors).

I did my first order of assembled pcb + CMS components...


JLCPCB spotted an orientation error of the diode and fixed it without fail (I was too eager to complete this first order). On the other hand, they prevented me to select the variable resistor that I needed (they "discontinued" it), so I will have to order them at mouser (with the other through-hole required components). So next years, I have blasted my allowance for november/december :D

As usual, all is open sourced, and the gerbers, bom and cpl (with orientation fixed) are here : https://github.com/sporniket/dop-module ... tag/v1.0.0

Time to mess with my TFT panel and my ESP32. This time trying to work with DMA capable memory allocation to serve as a framebuffer to be send via SPI.

I finally had to mount the panel on a protoboard of a stalled project to get something again. It never worked on breadboard, and realized that a good connection was required ; so last time I made connection using crafted wires with dupont connectors, but when going back after some months, it does not work anymore. I hope that a soldered socket will be the silver bullet.


Anyway it works, so now I will be able to create a lib that allocate big chunks of memory as specified by a set of symbols defined using the configuration script of the toolchain. Meaning that after separating the problems of device independent SPI basic configuration, ESP32 specific SPI Host driver configuration, and low level addressing of some commands of the tft controller ST7789, I will deal with separating the problem of memory allocation.

But before that, it seems that some of my libraries do not compile anymore after some update of the esp32 libs, so I made a demo that also use the libraries through git submodule, in order to have a working example as well as a workspace to maintains said libraries : https://github.com/sporniket/demo-task-gpio-button-led

After a long waiting after ordering (not in stock at the time of ordering), I finally got my scope. Between this Hantek DS02D10 and the Owon SDS1102, I choose the former because of the builtin decoder for some serial protocols (i2c, spi, uart). But I had to buy a supplemental set of high bandwidth probes, as it comes with only one.


I will take some time to get used to it, then I will try to assess the ripples that I get using the USB-PD adapter that powers my STe

I have the Owon SD1002, it's not bad for the price, though does have limitations. The Hantek looks nice.

A project for which I wish to have a working firmware this summer : a wifi connected clock, i.e. the hour use the network time to set up (no more manual time settings when switching between winter time and summer time), brightness adaptable to the surrounding (no blinding in the middle of the night), and possibly one alarm clock that can be set up for school days. So I use a cheap Esp32-based dev board (lolin32 lite), with the native toolkit, not the Arduino one.


At first I planned to use existing libraries for my display module (based on a TM1637 chip), but the ones I found for the native toolkit are only adaptation from Arduino ones, meaning manual bit banging the two wires interface to implement a malformed i2c protocol where the target peripheral address is not sent -meaning in effect that this controller cannot share the two wires with any other devices-.

So I attempted and quickly succeeded to use the i2c driver of the ESP32 to send commands to the display, very satisfying. But now I am going into the rabbit hole of turning my code into a proper library.

Those last weeks, I was working on HDL stuff, and I succeeded to output something on screen through the HDMI port (technically I output DVI frames, not HDMI frames).

All the code (tmds encoding and all) was written with Amaranth-hdl.


I will tidy up my code a little before switching back to retro stuff.

Took this morning to assemble this sample breadboard-friendly speaker driver circuit I made.


It's not ready for prime time, though. Firstly, the footprint for the transistor sure did not require to fold pins, but it was a PITA at soldering time : the pad are small and packed, so after trimming the legs, I had to use flux and solder braid to rework it (remove bloby solder bridges). Secondly, my mounting holes for the speaker are a bit off, fortunately I could use two holes.

About the circuit in itself, it is a simple NPN transistor in emitter-follower circuit to amplify the sound for a little 3W, 4Ohms speaker.


For the sound, I use a simple NE555 circuit. And this ought to be a first step toward a little toy synth project