News for hackers. Stuff that matters.
Re:Can you figure out how this single part boost converter w... by smoyer on Thursday June 08, 2017 @10:32PM (#14518797)
Hacks like this are a lot of fun to play with on the bench but don't design a circuit like this when you want something that "just works", has a reasonable life-time and certainly don't put a hack like this into production.

That being said, Sir Clive Sinclair made a pretty remarkable little computer (the ZX-81) but minimizing the parts count to the extreme. And there's a colloquial term for the process of minimizing your component count.

One of my favorite authors in Electronic Design was Bob Pease (now deceased). In a 1992 article [0] he recounts the process used by Earl "Madman" Muntz who was an early television pioneer. I could ruin the story but I'm going to insist you read the original instead. Enjoy!

[0] http://www.electronicdesign.com/boards/whats-all-muntzing-st...

EDIT: I also should have noted that this practice has led to me trying to create the simplest possible software algorithm that solves my problem. A good day is when the LOC count goes down (and not because I've packed more into a single line).

Re:Can you figure out how this single part boost converter w... by js2 on Friday June 09, 2017 @12:22AM (#14519161)
A good day is when the LOC count goes down (and not because I've packed more into a single line).

You can't write that and not link to http://www.folklore.org/StoryView.py?story=Negative_2000_Lin...

Re:Can you figure out how this single part boost converter w... by SAI_Peregrinus on Thursday June 08, 2017 @11:13PM (#14518929)
My favorite cute hack of a power supply is the ringing choke converter (self-oscillating flyback). It doesn't require any controller ICs, just a couple of transistors, resistors, caps, and a transformer. You get isolation, and the circuit is quite simple. With an optoisolator and zener diode you can even get regulated output!

The disadvantages are the poor efficiency, need for tuning, and need for a transformer if you don't need the isolation. Also no programming involved, so not really of as much interest to most of the HN community.

Re:Can you figure out how this single part boost converter w... by kens on Friday June 09, 2017 @12:01AM (#14519092)
If you buy a $2 phone charger, this circuit (the ringing choke converter) is what you get. It's kind of amazing how cheaply these are made. It can look just like an Apple charger on the outside, but the power quality (and safety) is much, much worse.
Re:Can you figure out how this single part boost converter w... by SAI_Peregrinus on Friday June 09, 2017 @12:23AM (#14519170)
Yep, they're used in lots of cheap power supplies, not just phone chargers.

The poor power quality and safety aren't inherent problems of the ringing choke design, those are just issues with cheap chargers. They tend to omit the output filtering and do horrible things to creepage and clearance distances.

Re:Can you figure out how this single part boost converter w... by blackguardx on Thursday June 08, 2017 @11:30PM (#14518993)
That was the first circuit I put together after cobbling together some old TV parts in 8th grade. The noise and dazzling purple glow of the high voltage arc from the flyback transformer were mesmerizing.

The urge to understand how it worked led me to study EE in college.

Re:Can you figure out how this single part boost converter w... by userbinator on Friday June 09, 2017 @08:00AM (#14520774)
All the controller ICs do is put almost everything but the coil onto one silicon die. Otherwise they work similar to the discrete version.
Re:Can you figure out how this single part boost converter w... by rhizome on Friday June 09, 2017 @12:27AM (#14519181)
Great story, I hadn't read that one before. Madman Muntz is a hero[1] of a person, love that guy, also inventing the term "teevee" (later "TV") and manufacturing and selling the Muntz Jet automobile[2]. Oh, he also invented car stereos (and bars, but watch #1).

His commercials were pretty amazing[3], too.

1. https://www.youtube.com/watch?v=69yXg9EkD-g

2. https://www.hagerty.com/apps/valuationtools/1951-muntz-jet

3. https://www.youtube.com/watch?v=FGi6VhnQkw0

Re:Can you figure out how this single part boost converter w... by phaedrus on Friday June 09, 2017 @08:51AM (#14520984)
I've had great success building digital logic with minimized part counts. When I've tried to do the same with analog circuits, I've had more mixed success. I realized that it's because digital logic is transitive (in the algebra sense) but analog is only quasitransitive. That is (assuming you're working in the realm of bits rather than abusing physical limits), in digital logic if circuit A is equivalent to circuit B and circuit B is equivalent to circuit C, you can replace A with C, and keep repeating this process until the circuit is minified. In analog electronics, you run into the sorites paradox.
Re:Can you figure out how this single part boost converter w... by th0ma5 on Thursday June 08, 2017 @11:16PM (#14518944)
Came across in the amateur radio world the "Minimal Art Session" contests where people get points based on how many radio contacts with how few parts... http://www.qrpcc.de/contestergebnisse/mas/2016/index.html
Re:Can you figure out how this single part boost converter w... by qb45 on Friday June 09, 2017 @04:51AM (#14520095)
> Hacks like this are a lot of fun to play with on the bench but don't design a circuit like this when you want something that "just works", has a reasonable life-time and certainly don't put a hack like this into production.

Dunno, do you think a constant rate PWM SMPS isn't enough to power a trivial load like LED and justifies the cost of adding a dedicated chip? LED forward voltage is well below absolute maximum for the RESET pin, peak current is limited by pulse time, looks good enough to my (amateur) eye.

Re:Can you figure out how this single part boost converter w... by smoyer on Friday June 09, 2017 @10:25AM (#14521570)
I didn't mean that you shouldn't use a boost converter - back in the RS-232 days the Maxim max232 chip saved a tremendous number of parts (otherwise you needed a +/-12VDC power supply to drive the signal wires. The idea of the IC's Vcc following the coil's transient is a bit scary and I wouldn't design that into a system.
Re:Can you figure out how this single part boost converter w... by qb45 on Saturday June 10, 2017 @03:34AM (#14526503)
I thought about using RESET of course. The Vcc thing is indeed bizarre and I wouldn't be confident that running logic circuits on a fast dV/dt power supply is a good idea.
Re:Can you figure out how this single part boost converter w... by dom0 on Friday June 09, 2017 @09:29AM (#14521183)
RSTDISABLE means that no in-circuit programming is possible any more and for re-flashing a parallel HV programmer is required; it's quite annoying for tinkering.
Re:Can you figure out how this single part boost converter w... by qb45 on Saturday June 10, 2017 @03:38AM (#14526508)
Yes, but this was about long term reliability so programming is of lesser concern.
Re:Can you figure out how this single part boost converter w... by amelius on Friday June 09, 2017 @05:29AM (#14520217)
If the objective is to use the least number of parts, then there's a simple solution: brute force.

You can of course automate this using a simple computer program and a simulator like SPICE. Slightly more intelligent solutions could use heuristic optimization techniques like genetic programming.

Re:Can you figure out how this single part boost converter w... by amelius on Friday June 09, 2017 @09:39AM (#14521244)
Not sure why this was downvoted. There's even a paper that explores this technique: [1].

[1] http://www.sciencedirect.com/science/article/pii/S1877705811...