#GreatScott #video on #FOC #control #electronics for #brushless #motors #toread
on 02026-08-25#Great-Scott #video on soft-start #electronics design to eliminate inrush current spikes. He rigs up an impressive alumina-bodied-resistor current shunt of 200mΩ and 50W, which should be good up to 15 amps continuous duty. He finds a 10Ω green disc “ZNR1” NTC thermistor labeled “SCK 103” used in another power supply to solve this problem; it drops from 10Ω down to some 297mΩ when it warms up. But to solve the NTC thermistor cooldown problem he rigs up a relay circuit which shorts out a 3.3Ω series resistor when its IRFZ44N coil MOSFET is triggered by an RC time delay (τ = 50kΩ·10μF = 500ms). That might potentially have the same problem if the inrush-current capacitor’s discharge has a shorter time constant than the delay capacitor, but that seems unlikely.
on 02025-12-17#GreatScott #video on controlling #motors with encoders
on 02025-11-11#GreatScott #video #toread on the #TL431 #electronics #hardware
on 02025-01-08#video by #GreatScott from two years ago endorsing the #Raspberry-Pi RP2040 Pico #microcontrollers #hardware #electronics
on 02024-08-27#video #GreatScott on tiny power #electronics #hardware, building PCBs around a TI TPS63020 4-amp buck-boost converter, which takes 1.8–5.5V and provides 1.2–5.5V (programmed with a voltage divider) with 25μA of quiescent current, using a 2.4MHz switching frequency so it can use tiny external inductors and capacitors, with only 8 passives in all. So he builds the new PCB with 0805s (because he find 0604s hard to hand-solder). The dadashit has a very nice suggested layout. Despite worrying about hand-soldering, he paste-soldered the components with a stencil and reflowed them on a tiny 3-cm-square hotplate he evidently had, labeled MHP30. The 5V output was 5.115V (2.3% high, a big sloppy), while the 3.3V output was 3.393V (2.8% high). He got 2.8A out at 3.3V and 2.04A at 5V, with 71%–91% efficiency, the higher numbers being at lower loads. He did two more tighter layouts where he had to hand-solder some caps on the back after reflow; the current capability decreased to only 2.0A at 3.3V but the efficiency was the same. Comments suggest TPS631000 and TPS63900 as even lower-power alternatives.
on 02024-05-21#video on #electronics #hardware of a synchronous dc-dc converter by #GreatScott. First, for comparison, he rigs up a simple buck converter, whose switch is an IRF5305 p-MOSFET with its gate pulled up to Vcc (12V?) by a 100Ω resistor or down to ground by a BC547 with a 220Ω base resistor; the freewheel diode is a rather large 1N5822 Schottky (½W?), and the output capacitor after the hand-wound 100μH inductor is a 330μF electrolytic. The BC547’s base is driven by a knockoff Arduino Teensy generating an 80kHz PWM signal. So he replaces the freewheel diode with an IRFB3077 p-MOSFET to cut the power losses from 40mW at 400mA to about 0.4mW. To ensure dead time and avoid shoot-through, he drives the MOSFETs with an IR2184 half-bridge MOSFET driver chip and nine passives it apparently requires for its gate-drive network for some reason. He got 92% efficiency in CCM.
on 02024-05-15#GreatScott cheapest Arduino alternative 20¢ #video #hardware He’s recommending the Holtek HT66F002, apparently unaware of the Padauk chips, even though he posted this only three weeks ago. It’s an 8-SOP with 1kB, an 8MHz internal oscillator, 6 GPIOs, a 12-bit ADC, and a 10-bit timer, but no I²C. He needs I²C, PWM output, GPIOs, an ADC, and low-power sleep. He ended up installing some proprietary Holtek IDE and buying a Holtek programmer, ugh, and then spending a couple of days debugging Blink before figuring out that he needed to disable the watchdog. It slept at 11μA, considerably better than the datasheet number. The instruction set is unfamiliar: one accumulator, conditional skip instructions, Harvard architecture, no multiplies. They don’t document the instruction encoding. He upgraded to the 32¢ HT66F0176 to get I²C instead of trying to bitbang it, then discovered that Holtek microcontrollers only support I²C in slave mode, so he’s going to try the 50¢ ATtiny402. Some Holtek μCs are under 10¢. (Oh, and apparently JLCPCB will be offering multicolored PCB silkscreen now.)
on 02024-04-05#GreatScott build an HV generator. His HV transformer has a low-voltage feedback winding on the primary side which he uses to turn on an NPN BJT to ground, via a 100Ω resistor and a UF4007 diode on the base, which turns on the other (113μH) primary coil; apparently this makes a 19kHz blocking oscillator like a Joule Thief, but this time with a high-voltage secondary. He burned up his first transistor pretty soon by dissipating 2.8 watts. An IRF740 N-MOSFET died even sooner. So he switched to a flyback converter built around an IRF840 driven by an external square wave and a DRC snubber. He reports that above 200kHz he could no longer feel the shock, but did burn his skin. But it still wasn’t as efficient as he wanted, so he drove it with the ZVS oscillator he’d built previously for induction heating. #electronics #hardware #video
on 02024-04-04