|
Sagebrush posted:Everyone in tech always thinks they've discovered the one ethical use for the panopticon Im stealing this
|
#
?
Dec 28, 2020 02:12
|
|
|
|
| # ? Jun 4, 2024 15:31 |
|
|
But building it would be so cool
|
|
#
?
Dec 28, 2020 02:16
|
|
|
ante posted:But building it would be so cool -Every nuclear physicist in 1940
|
|
#
?
Dec 28, 2020 02:22
|
|
|
Didn't Oppenheimer struggle with it the rest of his life too?
|
|
#
?
Dec 28, 2020 02:26
|
|
|
Forseti posted:Didn't Oppenheimer struggle with it the rest of his life too? Quite a few of them did yeah. Well except Ed Teller, who thought nukes were fantastic and we should use them for everything.
|
|
#
?
Dec 28, 2020 02:31
|
|
|
Richard Feynman posted:[It] was Bob Wilson, who got me into it in the first place. He’s sitting there moping. I said, “What are you moping about?” He said, “It’s a terrible thing that we made.” I said, “But you started it, you got us into it.” You see, what happened to me, what happened to the rest of us is we started for a good reason but then we’re working very hard to do something, and to accomplish it, it’s a pleasure, it’s excitement. And you stop to think, you know, you just stop. After you thought at the beginning, you just stop.
|
|
#
?
Dec 28, 2020 02:35
|
|
|
Shame Boy posted:Quite a few of them did yeah. In his defense, if we'd used nukes for everything 70 years ago, the smoking crater that remained wouldn't have any of the problems we have today.
|
|
#
?
Dec 28, 2020 02:36
|
|
|
Makes sense, I guess Oppenheimer just stood out in my mind because of the Bhagavad Gita quote
|
|
#
?
Dec 28, 2020 02:41
|
|
|
I threatened to post a thread for hand-drawn or otherwise artsy PCBs if I ever got around to releasing my PNG-to-GRB converter. I'm taking a vacation this week so here's the thread: https://forums.somethingawful.com/showthread.php?threadid=3953089 and here's the software: https://github.com/stackmachineSA/emboarden/releases/ Stack Machine fucked around with this message at 03:07 on Dec 30, 2020 |
|
#
?
Dec 29, 2020 03:31
|
|
|
Nice! Windows should run it in WSL just fine too I'd imagine. When can we get a BeOS version 
|
|
#
?
Dec 29, 2020 05:45
|
|
|
Forseti posted:When can we get a BeOS version Ah that sounds like a lark! ... (hours pass) Huh. There doesn't seem to be an easy way to get "modern" C++ to build for BeOS. Have Haiku as a consolation prize. Your guess is as good as mine as to why the only program on that OS that could open the SVG files was the icon editor.  I would have just built for BeOS from Haiku, but it turns out there's this whole thing where there are 2 simultaneous GCC versions and the one based on GCC 2.95 produces backward-compatible binaries and the one based on GCC 7 doesn't!
|
|
#
?
Dec 29, 2020 22:46
|
|
|
Foxfire_ posted:These are thermometers I use at work. Depending on how big of an area you want to collect, from, you might need external optics for the IR. Even a 5 degree cone is 4" bigger at 4 feet away. Haven't tested on skillets. There are a number of variants with different temp ranges, FOVs, features I haven't deciphered. The docs imply it averages the temp of everything in FOV, so if you go too wide, the readings will be off. It lets you set the surface emissivity. I suspect it won't work at all on shiny things, but I'll test. The docs are detailed, but are a bit hard to decipher without domain specific knowledge and/or guesswork. Ie I had to look up a guide online to figure out which pin was which! I'm still not sure what it means by its interface is PWM over I2C; haven't spelunked the drivers yet. If you're going to go through with this project, the biggest challenge will be finding a suitable projector, and getting it to look even / appropriate brightness / readable etc. Not the temp taking. C++ code:Dominoes fucked around with this message at 00:44 on Dec 30, 2020 |
|
#
?
Dec 30, 2020 00:37
|
|
|
Stack Machine posted:Ah that sounds like a lark! ... (hours pass) Hahaha that's amazing, I salute your commitment to a bit! I have to admit I wanted a BeBox sooooooo bad when I saw it in an issue of boot (later renamed Maximum PC) but I've never actually bothered to install the OS even on an emulator. I still want a PC case with green-yellow-red LED bars indicating the load on each core though. My guess would be that BeOS uses SVGs for its icons. IRIX definitely uses a vector format for its icons (I have an Indy and it's awesome to run the 3d file system demo that they use in Jurassic Park on it!) so they might have aped that idea. Not sure if the ones in IRIX are SVG though.
|
|
#
?
Dec 30, 2020 01:05
|
|
|
Dominoes posted:Haven't tested on skillets. There are a number of variants with different temp ranges, FOVs, features I haven't deciphered. The docs imply it averages the temp of everything in FOV, so if you go too wide, the readings will be off. It lets you set the surface emissivity. I suspect it won't work at all on shiny things, but I'll test. The lens focuses incident light of a particular wavelength onto a thin membrane inside it. That heats up the membrane and it is small enough to get to steady state quickly. It measures the temperature difference between the membrane and a reference material + the absolute temperature of the reference. Then it backcalculates how much energy has to be delivered to the membrane to produce that heat flux, then combines that with the emissivity coefficient & FOV to figure out how what temperature of object that filled the FOV would radiate that much energy into the lens. Emissivity is factory set to 1. If you are looking at a material with less it will be increasingly wrong (try the shiny side of aluminum foil, it's around 0.05). You can change the emissivity model and get better readings, but error will be larger because a hot and cold object don't radiate very different amounts. The PWM interface is a mode where it can be reconfigured to output temperature as a PWM signal instead of I2C
|
|
#
?
Dec 30, 2020 02:53
|
|
|
Thank you very much! That's a great explanation about the workings. I'd like to try this for the skillet application, sans projection; I'd use it every day, even with the caveat of it being probably unusable for steel and aluminum. For this use, temperature doesn't have to be precise. Will have to experiment with emissivity. I think the one with the smallest FOV, on a hand-pointed gooseneck swivel is the way to go. Attached to the range hood with a magnet. Probably with a laser pointer to help aim it onto the skillet.
|
|
#
?
Dec 30, 2020 03:25
|
|
|
Stack Machine posted:I threatened to post a thread for hand-drawn or otherwise artsy PCBs if I ever got around to releasing my PNG-to-GRB converter. I'm taking a vacation this week so here's the thread: this is an extremely loving cool thing to wander back into electronixthread for, hopefully i'll have something to contribute soon actually i did partially assemble a Manhattan-style crystal tester using jewellers' saw-cut Aesthetic Node Geometries, i think that counts. i posted a WIP pic at some point e: yep  guess i didnt get pics of anything post-soldering + component-adding (unrelated: i built and installed that long-delay door alarm you worked out a circuit for, i just never posted it because i became paralyzed by selecting an alarm tone that everyone else in the house didnt hate. possibly i should have assessed if other household members felt we needed a door alarm as much as i do. it does its job flawlessly and also completely silently. i take it on faith the battery hasnt run out yet  )
Ambrose Burnside fucked around with this message at 19:06 on Dec 30, 2020 |
|
#
?
Dec 30, 2020 03:37
|
|
|
Foxfire - do you know what the nominal FOV of the "standard package" is? There are (tough to find) variants with specific rated FOV, but the common one is labeled "standard package" instead of "35° FOV" etc. My best guess from piecing it together from 3p sources is it's 90 degrees, but not linear of how it balances diff parts of the FOV; the 35° version adds something physical to block part of the FOV, and the narrower ones add (price-increasing) IR optics.
Dominoes fucked around with this message at 22:21 on Dec 30, 2020 |
|
#
?
Dec 30, 2020 22:15
|
|
|
Section 10.2 of datasheet is about FOV for the 'standard package' one, which seems like no optics besides the filters probably I've only ever used the 5° one
|
|
#
?
Dec 30, 2020 23:00
|
|
|
Thanks!
|
|
#
?
Dec 30, 2020 23:51
|
|
|
1) What's a good free circuit simulating software? The circuit I'm trying to design/test is a Op-amp multivibrator into a SPDT relay to replace the analog switch in a Klomp dirty video mixer  https://www.youtube.com/watch?v=iSRWvQf3u2c&t=1440s 2) does this actually make any sense? I'm just going off some comments I got on my Stackexchange post
|
|
#
?
Dec 31, 2020 00:41
|
|
|
LTSpice https://www.analog.com/en/design-center/design-tools-and-calculators/ltspice-simulator.html
|
|
#
?
Dec 31, 2020 01:09
|
|
|
What are you trying to make happen? You have 2 toggle switches and a knob. What are they supposed to do? What are the Vin and Vout wires supposed to be carrying?
|
|
#
?
Dec 31, 2020 01:28
|
|
|
Foxfire_ posted:What are you trying to make happen? Sorry. They're carrying composite video signal. I want instead of me switching the switch real fast I have some electronics replace one of the switches. I'm trash at explaining. Edit: the circuit as drawn the pot mixes the two signals and the switches toggle what signal is on each side of the pot. Papa Was A Video Toaster fucked around with this message at 01:35 on Dec 31, 2020 |
|
#
?
Dec 31, 2020 01:32
|
|
|
If you have a 10AWG cable supplying 50 amps for 10 feet, but right in the middle you replace one inch of it with a thinner gauge, is that the same amount of resistance and voltage drop as if you replaced the entire cable with that same gauge? I guess what I'm wondering is, does "a chain is only as strong as its weakest link" apply directly to cabling resistance?
|
|
#
?
Dec 31, 2020 03:13
|
|
|
TVsVeryOwn posted:Sorry. They're carrying composite video signal. I want instead of me switching the switch real fast I have some electronics replace one of the switches. I'm trash at explaining. Zero VGS posted:If you have a 10AWG cable supplying 50 amps for 10 feet, but right in the middle you replace one inch of it with a thinner gauge, is that the same amount of resistance and voltage drop as if you replaced the entire cable with that same gauge? Now replace one with a 100Ω resistor. End-to-end will be 1090Ω and measurements inside will change depending on if they cross the bigger resistor. Wire with sections of different thickness is the same thing, just smaller resistances and continuous instead of lumped
|
|
#
?
Dec 31, 2020 03:34
|
|
|
Foxfire_ posted:I'd expect that doing any analog add/mean/whatever to two unsyncronized composite video signals will just give you garbage by wrecking the vsync and hsync information. Sure does, that's where the Klomp dirty video mixer gets it's name. It makes no attempt to sync the signals. Hence why I want to switch the signals digitally over time.
|
|
#
?
Dec 31, 2020 03:53
|
|
|
Zero VGS posted:If you have a 10AWG cable supplying 50 amps for 10 feet, but right in the middle you replace one inch of it with a thinner gauge, is that the same amount of resistance and voltage drop as if you replaced the entire cable with that same gauge? https://en.wikipedia.org/wiki/Kirchhoff%27s_circuit_laws say that the resistance of any closed single circuit is the sum of all its individual resistances, the current measured at any point is the same everywhere in the loop, and the voltage must sum to zero (where the source is a positive value and each drop is negative). So this means that in the trivial case the resistance of the whole cable will be slightly increased by the thinner segment, but not as much as if you replaced the whole cable with a wire of the thinner diameter. The thinner part doesn't act like a "roadblock" or something; the current is decreased (given a constant voltage) everywhere in the wire simultaneously. This is one of the ways the hydraulic analogy breaks down. However, since the current is constant in the entire wire but the thinner section does have higher resistance than the rest of the wire, that means you'll have a greater voltage drop over that particular section, and more power dissipation, and a smaller amount of metal to handle the heating, and the thin part will get hotter and burn out before the rest. This is the concept of a fuse. But the fuse passes just as much current as the rest of the cable (ohmic heating aside...) until the point where it melts down.
|
|
#
?
Dec 31, 2020 04:01
|
|
|
Zero VGS posted:If you have a 10AWG cable supplying 50 amps for 10 feet, but right in the middle you replace one inch of it with a thinner gauge, is that the same amount of resistance and voltage drop as if you replaced the entire cable with that same gauge? Edit: beaten, but I'll keep this here anyway: So there are two things going on here. There's the resistance and voltage drop, which foxfire covered, and then there's the fusing current, the amount of current the wire can pass before it melts (or gets to some arbitrarily chosen "bad" temperature). Because both the power dissipation and thermal conductivity to the environment are proportional to length for a given diameter, the temperature of a long-enough run of wire will only depend on its diameter and the current through it. All of which is to say that, if your concern is thermal your run is only as good as its narrowest segment, but if it's electrical you can just add up the segment resistances. Stack Machine fucked around with this message at 04:17 on Dec 31, 2020 |
|
#
?
Dec 31, 2020 04:13
|
|
|
Is there a reason to get the raspberry pi 4 over the 3? I want to do this using his software as well. I think the 3 will be fine, don't see a reason to pay extra for the 4. These were made before version 2 even. https://woodgears.ca/tech/pi_holder.html His Divine Shadow fucked around with this message at 13:10 on Dec 31, 2020 |
|
#
?
Dec 31, 2020 13:05
|
|
|
Isn't the 1GB Pi 4 the same price as the 3B? The Pi 4 is much faster and has better I/O (I think notable the ethernet port is on PCIe instead of sharing the USB bus with the USB ports) but the Pi 3 is plenty fast. IMO, the Pi 4 is faster but still not fast enough over the 3 to do anything different with it really. You can get more RAM but that makes it go up in price pretty quickly. The Pi 3's drivers are also more mature unless the 4 has caught up recently. For video, they both offload to the GPU anyway so from a quick skim of that it looks like it's just a video camera with simple motion detect? I think the 3 should do that easily.
|
|
#
?
Dec 31, 2020 15:32
|
|
|
The USB bandwidth is also much higher on the 4, since it uses a USB 3 controller on that same PCIe bus instead of a single USB 2 port on the SoC itself.
|
|
#
?
Dec 31, 2020 15:58
|
|
|
Wrong thread.
|
|
#
?
Dec 31, 2020 19:42
|
|
|
bobua - I'm working through code for your proj now since it seems super useful, but am not attempting the projection. Should be able to test it out on a dev board Mon; waiting on the display. Basic ideas: - Magnetic base with hole in middle for bolt - Project box that screws into base. Small PCB in project box, with 1" OLED display mounted on it. (how?) - Batteries in sep box, or in main proj box. - Gooseneck or swivel that screws into project box, also with a bolt. Has sensor on end. - Maybe additional optics to restrict sensor FOV - STM32F1 or Lx series MCU, powered by 3AAA batteries. - Device is in a low power mode by default. It occasionally wakes up to check temp using a timer or RTC. If temp is above a thresh, it shows on display - Displayed temp is color-mapped to the reading. (eg really hot might be red or white) - Two buttons: One for skillet surface type (3 broad categories that map to emissivity: teflon, cast iron, steel/aluminum), and C/F temp unit - The buttons are mapped directly to interrupts which change the setting, and wake the display for a second to show you it's set. - The main loop wakes from low power, checks temp, displays if above the thresh, then sleeps and disables the display if temp falls below thresh I figure the OLED might be suitable for batteries, since it's small, will be off when not cooking, and most of the display will be black. Could be wrong though! I've written most of the code and it compiles, but can't ops test until I get the display. I'll show you the code if you're still pursuing this; you can probably translate the ideas, even if you're using a different platform. If you pursue the projector, this will get more complicated, especially in the physical assembly and additional considerations. I'm attempting to learn how to use Mutexes and ref-counted smart pointers to deal with sharing state and sensor info across interrupts, as opposed to unsafe mem reads and writes... Although the latter should probably be fine too. The mutex approach is more verbose, but should prevent race conditions. Dominoes fucked around with this message at 20:02 on Dec 31, 2020 |
|
#
?
Dec 31, 2020 19:49
|
|
|
I got myself an hp EliteDesk off of eBay as an auxiliary workstation for Christmas. Got it today and though it's an AMD AM4 platform, it has an RS232 port for some reason  Is this useful other than for hooking up to my 20 year old Agilent 54622D oscilloscope with an actual serial cable and convenience vs a USB UART? I guess probably mostly if you have stuff that runs at the full voltage allowed on RS-232 spec? Wonder what it was ordered for, I'm guessing it was for use in an industrial control system.
|
|
#
?
Dec 31, 2020 23:00
|
|
|
Lots of those micro computers are used for PoS systems, which still use serial for barcode scanners and the like. Every computer in my lab had a serial port, because they're still pretty common for lots of scientific and industrial gear. A USB->serial adapter gets you 80% of the functionality, but there's plenty of stuff that needs an actual hardware UART.
|
|
#
?
Jan 1, 2021 17:23
|
|
|
sharkytm posted:Lots of those micro computers are used for PoS systems, which still use serial for barcode scanners and the like. Hopefully things have changed but back when I needed to do serial stuff with a PC sometimes the buffering of the USB always hosed up what I was trying to do (usually control some sort of simple motion device or mechanical thing)
|
|
#
?
Jan 1, 2021 17:31
|
|
|
CarForumPoster posted:Hopefully things have changed but back when I needed to do serial stuff with a PC sometimes the buffering of the USB always hosed up what I was trying to do (usually control some sort of simple motion device or mechanical thing) Nope. Usb to serial still sucks. There's variable latency, fifo buffers, and inconsistent voltages to contend with.
|
|
#
?
Jan 1, 2021 17:58
|
|
|
Also counterfeit chips in lots of USB adapters. Although I'm in Linux and it doesn't seem to be as big an issue since I don't have to run the manufacturer's driver anyway so counterfeit chips aren't as likely to give me an issue I think. I have a USB one that I've used before to pull waveform data from my scope (4MB at 57600bps  )Wish I could find a GPIB adapter for not a zillion dollars.
|
|
#
?
Jan 1, 2021 18:08
|
|
|
Forseti posted:
UART is still a very common protocol for debug ports on microcontrollers, and when it comes to debug, the fewer layers of translation, the better.
|
|
#
?
Jan 1, 2021 18:10
|
|
|
|
| # ? Jun 4, 2024 15:31 |
|
|
You need a transceiver (eg ST3232B) on the MCU side to safely plug into an actual RS-232 port though don't you? Or is RS-232 a misnomer for most PC serial ports? I figured you'd have to be able to tolerate the +/- 15V to be safe.
|
|
#
?
Jan 1, 2021 18:20
|
|