Збирач потоків

Hi everyone!, after 10 months of working and improving on my accelerator, its finally complete! This device accelerates a magnet in circles using 4 electromagnets and hall effect sensors (I've tried IR sensors but failed😔). Those sensors detect the magnet and then a N-MOSFET switches the coil on and off at the right moment, which leads to acceleration of the magnet. I've also used a 12v--> 5v voltage regulator and for one reason or another I've put a quick ignition and fire hazard or whatever you call it on the voltage regulator.

If you wanna know more, or just wanna see the accelerator in action you find the youtube video at the KIWIvolt youtube channel.

I'm thinking to make a part 2 in which the magnet is a sphere and thinking of replacing the breadboard with a PCB. If you have any other ideas or wishes please let me know so i can adjust it, to perfect my accelerator even further.

So i used HEF4094BP, i did the same circuit in this video 4094 shift register long time ago, then in 2022 i bought raspberry pi pico, and in this year i write a long code with MicroPython to count from 1 to 9 and repeat the loop, but i need to optimise it next time.

Over the last few weeks I’ve worked on an Arduino board connected through an ADC converter into 3 magnetometers. They are set orthogonally to one another (around the clear box) so that the magnetic field strength and direction at a given point can be found. The whole lot gets power through a USB cable that allows you to model the direction and strength in python. It’s been an absolute blast building it :)

Left: 1974 (Matsushita Electric)

Right: 2021 (Rubycon)

Both 16V 1,000μF.

Same voltage rating and capacitance, but shrunk this much in about 50 years.

For a biochemical project of mine I needed a very precise scale. The ones I bought were underwhelming, so I decided to just solder one myself.

The sensitivity is kind of ridiculous. Sitting near the scale, I can see my heartbeat in the signal when streamed to a PC. Someone walking on a different floor makes the reading jump — and I live in a concrete building. The coil can lift about 20 g. With different coils, you could trade off dynamic range vs. precision. For my purposes, the precision is already overkill.

Components were about $100 total. The most expensive part was the neodymium magnet.

The principle is electromagnetic force restoration. A 110 Ω coil suspended on a lever lever sits above a neodymium ring magnet. The lever height is held constant by a feedback loop that uses an IR photointerrupter. The current required to hold the weight is directly proportional to the mass.

For current sensing I used a 10 Ω shunt resistor (RJ711, 5 ppm/°C TCR) and a 24-bit ADC (ADS1232). The signal is read by an Arduino Nano and displayed on a small LCD (SLC0801B).

The photointerrupter is built from a generic IR LED and IR photodiode. The LED is driven with a constant current source (using a 2N7000 MOSFET), while the photodiode is reverse-biased for fast response.

The circuit runs from a low-drift 2.0 V reference (REF5020), which provides a stable reference for the ADC. After dividing it to 0.5 V, it also biases the photodiode stage and provides the ADC’s negative input.

The coil current is controlled with an N-channel power MOSFET (IRF540N) acting as a low-side driver, operated in its ohmic region. Its gate is driven by the photointerrupter circuit.

Zero-drift op-amps (OPA187) buffer the reference voltages, drive the photointerrupter, and control the coil current.

I also added a capacitive touch button for tare, so you don’t have to touch the scale directly — that’s surprisingly important at this sensitivity.

The schematic looks a bit op-amp heavy, but it’s actually pretty straightforward.

Challenges and possible improvements - The lever tends to oscillate, so the feedback loop has to be very fast. A lighter lever with a higher resonant frequency would help, and might require a lower-gate-capacitance MOSFET. - All components in the feedback path need low temperature coefficients to minimize drift. - To fully eliminate drift, one would need to monitor and compensate for coil temperature, photointerrupter temperature, as well as ambient air temperature, humidity, and pressure (for buoyancy effects). - A parallel guide system will eventually be needed so measurements are independent of where the weight is placed on the lever.

This build definitely requires some electronics background, so it’s not a first-project type of thing. But if you’re comfortable with soldering and op-amps, it’s very doable.

Hope you like it 🙂

This was a brain fart moment upon finding out they were .25 watt, we needed 9 watt capable. This is a lovely bundle of 36 that has next to no resistance now 🤦 .... 20ohm

Miss the old micro SD upgrade days

I've had an obsession with rockets/flight controllers and decided to make an open source flight controller from scratch (nicknamed Athena). I've added the Github repo/design files if anyone wants to take a closer look.

👉Github repo / Design files

• Triple MCU: STM32H753VIT6 (MPU), STM32H743VIT6 (TPU), STM32G474RET6 (SPU)
• 6 Pyro Channels: Direct 12V battery connection with fuse protection
• 6 PWM Channels: 2 for TVC (Thrust Vector Control), 4 for fin control
• Sensors: Triple ICM-45686 IMUs, LIS2MDLTR magnetometer, ICP-20100 & BMP388 barometers
• GNSS & Communication: NEO-M8U-06B GPS, LoRa RA-02 telemetry, Bluetooth DA14531MOD
• Storage: SD Card + Winbond W25Q256JV flash memory
• Power Management: 7.4-12V LiPo battery with BQ25703ARSNR charger, USB-C PD support
• 6-Layer PCB: Signal/GND/Power/Signal/GND/Signal