Reddit:Electronics

This is the next update on the programmable power supply project (you can find previous posts and more details in its umbrella repository condevtion/i2c-pps, while schematics itself is in condevtion/i2c-pps-hw).

During the past week I managed to select exact market available components for the device and create detailed BOM. I need parts for 3 copies of the power supply - two sets to actually build devices and one on standby just in case. Honestly, I expected the BOM to be 3 times cheaper (or at least 2) but costs for hundred components quickly add up. In the first picture above (left chart) you can see average unit price of a part per its type with quite expected the BQ25758S controller being the most expensive thing. However, as the right chart shows total amount of capacitors easily gives them the lead in final cost, which is $108.88 (or $36.29 per set). For just one set the total is $48.99 making it almost buy two get one for free.

The next picture shows quantities of parts per device and totals for 3 devices with additional components (marked green) per part type (total here is 393 for all 3 sets). The latest allows to reduce cost even further due to substantially lower prices on bigger quantities.

Now, knowing exact parts and their footprints I can start designing PCB itself.

The first started as a way to test ADCs and parallel I/O, and I turned it into a toy oscilloscope using some software I wrote for my Raspberry Pi. I didn't really understand op-amp input bias current and so it doesn't really work properly with the probe in 10x mode. The offset is huge, but I now understand the mistake. I also used one more op-amp than I really needed, and could've gotten away with cheaper ones, but it works up to 50MS/s!

The second board is a buffered variable-gain amplifier test with voltage-variable gain and bias. I fell down a rabbit hole w/oscilloscopes and am working on making an improved 2-channel one with modern components, so I broke out some of the front end into a test board and just finished building it. It's a miracle the QFN op-amp works, I was sure I'd bridge something underneath it.

There's a subtle crucial mistake in the second design, all you need to know to spot it is that the second amp is an LMH6505. It somehow does partially function still!

Demo at https://www.youtube.com/watch?v=sKtSpykOBXY

This is the Morse code trainer I designed. It runs on an AVR128DA48 microcontroller with a 2.42 inch 128x64 OLED and a custom-designed capacitive touch sensor PCB straight key. It also includes an NRF24L01+ radio module to allow 2-way send and receive of Morse code between nearby devices. The whole thing is powered by a rechargeable 3.7V 800mAh LiPo battery. I also designed the enclosure and 3D print it out of PET-G filament.

Happy to answer any questions!

Helloo

I have been building my setup for a year or two i participated in isef last year. as a Iraqis second year in isef i really like my setup. any suggestions

After about 3 months and a lot of dedication, I successfully completed my project.

It's almost exactly Grant's project, the only modification is that the SRAM has 8KB, a 32KB one will arrive soon and, since the wiring was already done with it in mind, the change will be easy.

As you can see i have gone completely my own way to make this oscillator, it uses a 25KHz xtal and a 2n3904 transistor, 1M ohm pot and one 5k pot, the power supply comes from 15Vscaled down to 9V using 100k pot + 2n3904 + 1k resistor, i know the picture shows 10k but that didn't give me full voltage range so use 100k instead. I have no idea how it got this working and i am somewhat suprised that 2n3904 can oscillate at 20MHz+.