View on GitHub

Quorten Blog 1

First blog for all Quorten's blog-like writings

How much current can different AWG wire gauges handle? 26 AWG is fairly common in basic electronics, and it is rated for about 2.5 A of current. This means that that you should limit your circuit design to put a maximum of 1 A of current through 26 AWG wire, i.e. at least half the current rating of the wire. And this nicely coincides with the 1 A power consumption limit of the Raspberry Pi Zero computer board.

This also means you should make sure your modular connectors you use are rated for 2 A of current if you are using them for the main power connector to the Raspberry Pi Zero. All other lower power connectors, e.g. for connected peripherals, can go just fine with 1 A modular connectors, i.e. Molex PicoBlade.

20191109/DuckDuckGo 26 awg wire current
20191109/https://www.engineeringtoolbox.com/wire-gauges-d_419.html

Please see my next article on the subject for important further discussion.

What’s the difference between a BJT and a MOSFET transistor? Wikipedia has lots of information on the individual devices per se, but not much in the way of how they differ. So I go searching the larger Internet for an answer, and indeed a good answer I have found.

  • BJT transistors are current-controlled devices, but MOSFET transistors are voltage-controlled devices. Thus a MOSFET only consumes gate power during switching, but a BJT transistor base power continuously.

    Thus, if you have a circuit that requires current to flow freely through the base of a BJT transistor, say in the cases of capacitor timers/oscillators and low-pass filters, a swap-out to a MOSFET transistor will not fully work as intended. This is also the reason why MOSFETs are used primarily in digital switching circuits for power or logic rather than as amplifiers per se.

    However, if you convert your analog signal into a digital signal for the MOSFET amplification stage and keep the low-pass resistor-capacitor network outside of the MOSFET stage, then you can construct an amplifier using MOSFETs. The Class-D amplifier is an excellent example.

    BJT transistors are sometimes useful when your primary goal is to minimize the cost of a circuit. An audio DAC and amplifier can be built more cheaply using a BJT transistor as it allows you to use a smaller capacitor on the input-side of the amplifier rather than on the output-side as with a Class-D amplifier.

Read on →

So, I’ve been just about fed up with my cheap Foto&Tech Pentax IR remote release trigger. It’s not working at the moment, and I can’t quite tell if it is because my button hack needs to be adjusted, the battery needs to be replaced, or the entire thing has just failed due to too much hackery. The battery voltage looks good, but who knows, maybe it is just extra picky.

In any case, now I know that if I am doing just internal studio remote release shoots in a studio environment, there is no need to go cableless. A cable trigger, by virtue of not having a battery, is one less part to fail that could get in the way of my productivity.

So, shopping around for the cable release trigger, unfortunately it appears that there are also complaints with the quality of the official cable release. Really? Not only do their remotes have their downsides, but even a simple cable release comes with downsides too? Now I’ve already bought the IR remote from Foto&Tech and its button was too cheap, now I don’t know if I want to bother with buying their cable release.

But, here’s the really good news. The Pentax CS-205 cable release is simply a standard 3.5mm stereo audio jack. That’s easy enough to build my own, wire one push button from left channel to ground for half-push, another push button from right channel to ground for pull-push, and you’re in business!

Read on →

Again, I reiterate, because this is important! Previously I’ve linked to an article about Linus Torvalds agreeing that Linux is bloated, but that was a few years ago. Well, we’re still at the progressive slowing down of the Linux kernel. Here is yet another newer article on the same subject. But, some of the specifics rae different. In light of the Meltdown and Spectre bugs, some of the mitigations have resulted in notable slowdowns, but these are not the result of the largest slowdowns. The largest slowdowns were due to configuration errors, some introduced by newer changes, others introduced simply by inability to auto-configure with newer hardware.

Companies like Red Hat and Google spend 6-12 months tuning the kernel, but Linux kernel releases come out every 2-3 months. Google uses a team of 100 engineers to do this.

20191109/https://blog.acolyer.org/2019/11/04/an-analysis-of-performance-evolution-of-linuxs-core-operations/

`systemd` replacement for rc.local

2019-11-09

Categories: unix  
Tags: unix  

One of the big problems in systemd for many simple use cases is the claimed obsolescence of rc.local, yet many times there is a real need to just run some simple shell commands at startup. So, how do you do it in the modern world? You write an additional systemd unit file that will call your own rc.local script, or in the case of a single command, call your command directly.

Place in /etc/systemd/system/rclocal.service:

[Unit]
Description=Run rc.local shell commands
[Service]
ExecStart=/usr/local/bin/rclocal.sh
[Install]
WantedBy=multi-user.target

20191108/DuckDuckGo systemd run shell commands at boot
20191108/DuckDuckGo systemd run shell commands rc.local
20191108/https://www.redhat.com/sysadmin/replacing-rclocal-systemd

I have an LED display circuit I am wiring up to a Raspberry Pi. It has a number of display cells that would each require a number of protective resistors, but I don’t have a particularly large number of spare resistors on hand. Can I gang together multiple LED cells onto a single resistor? The short answer is no, a single resistor only has enough power rating to support a single LED. Such a wiring would only work if you could guarantee that only a single LED was powered on at a time, such as a hardware multiplexer circuit. But failing that, say you have a shift register instead, you should use one resistor per LED.

Here is why. The typical power rating on resistors is 1/8 Watt or 1/4 Watt. A typical resistor with specifications around 2 V 50 mA will consume about 1/10 Watt power. With ceiling room allowance for double the power capacity to be on the same side, that really doesn’t leave you with any room at all for multiple LEDs on a single resistor. Definitely nothing on the order of 10, but maybe something on the order of 2 if you are careful.

20191106/DuckDuckGo resistor current rating
20191106/https://www.dummies.com/programming/electronics/components/electronics-components-resistor-power-ratings/

This site has some good info on how to calculate the required resistance nad power values for a LED’s protective resistor.

20191106/DuckDuckGo led current
20191106/https://www.sparkfun.com/tutorials/219

For many decades, standards in 3D models was troubled area, but with the rise in popularity of 3D printing, it feels like de facto standards are finally emerging. Printers are strong forces in standardizing data formats.

One notable artifact of injection-molded plastic parts is the conspicuous inclusion of part numbers and technical markings. Many 3D printed designs are devoid of such details since they need not pass through as many “middlemen.”

How is EOMA68 doing? Any updates? Nope, none since my last notes. However, Libre RISC-V M-Class is making some interesting progress. More EU NLNet grants coming, and building interest over the Libre RISC-V M-Class filling a niche market of more programmable GPGPUs that has not been able to open up due to the inertia of the incumbent players.

20191104/https://www.crowdsupply.com/libre-risc-v/m-class/updates/intriguing-ideas

All electronic parts on a printed circuit board are designed to be able to withstand the temperatures of a reflow soldering oven. Wait, what about the black plastic integrated circuit packaging? Most plastics have a lower melting point than solder. The “plastic” IC packages clearly can’t be plastic since they would melt when exposed to the heat of reflow soldering.

So, what material are they actually made out of? It is epoxy, a thermoset plastic, rather than a thermoplastic. Therefore, it will not melt when it is overheated. More specifically, it is formulated so that high temperatures decomposition will result in charring.

20191105/DuckDuckGo what kind of plastic is used for ic packages
20191105/https://electronics.stackexchange.com/questions/217423/ics-chips-are-typically-packaged-in-what-material
broken/http://www.ti.com/en/download/qlty/SEMICONDUCTOR_PACKAGING_ASSEMBLY_TECHNOLOGY-MISC.pdf 20191105/DuckDuckGo semiconductor packaging technology misc
20191105/http://educypedia.karadimov.info/library/SEMICONDUCTOR_PACKAGING_ASSEMBLY_TECHNOLOGY-MISC.pdf

Here is some more great information on the plastic packaging and considerations for reflow soldering.

20191105/DuckDuckGo temperature range of plastic ic packaging
20191105/DuckDuckGo plastic ic package max temperature
20191105/https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/an/an113.pdf

Read on →