|
huh. it literally is just three potentiometers in the same package. looks like they do it for redundancy, which makes sense if it's controlling a vehicle. with three of them you can vote on the outputs. weird
|
#
?
May 15, 2020 07:31
|
|
|
|
| # ? Apr 23, 2024 15:35 |
|
|
maybe the three encoders are slightly offset to give greater precision?
|
|
#
?
May 15, 2020 09:02
|
|
|
I used to work on ECU tuning/reverse engineering, and I found it kind of interesting that the accelerator pedal on some of them actually logged two values, in opposite directions (like 0 to 5V, and 5 to 0V as pedal is depressed), maybe two opposing pots can help account for certain types of faults better than two completely same phase  So I was thinking maybe it could be some kind of "3 phase" reading, but I'm not really sure how that makes sense with pots, since I don't think pots are made with sinusoidal "taper". Or like they could still be linear taper and be a 3 phase sawtooth? But then at least one or two of them would have to wrap around from hi to lo at some midpoint, which seems weird. Or they could be 2 in phase with each other, and just one going the opposite direction. No idea, just rambling at this point... probably easier to just measure what they do than guess much more about it. Or try to request datasheet.
|
|
#
?
May 15, 2020 10:34
|
|
|
peepsalot posted:I used to work on ECU tuning/reverse engineering, and I found it kind of interesting that the accelerator pedal on some of them actually logged two values, in opposite directions (like 0 to 5V, and 5 to 0V as pedal is depressed), maybe two opposing pots can help account for certain types of faults better than two completely same phase I think it's a safety issue. Say there is a fault where both pots are shorted to the 5v reference (or 12v for that matter). Instead of the ECU thinking the driver just floored the pedal, the discrepancy is seen by the ECU.
|
|
#
?
May 15, 2020 11:47
|
|
|
Glorgnole posted:ok i found the actual device. you probably already have this info since you have the actual hardware but i feel proud of my detective work. peepsalot posted:I would have expected some kind of encoder too, but looks like its actually just 3 potentiometers in a housing. glorgnole, I have no idea how you figured it out but you nailed it! nice work! thank you both, I ran into the same problem because of that financial group having way better seo I’ll read the manuals for the details but if they really are just stacked pots, I should be able to feed that into an arduino and make it show up as a joystick in windows. I just have to figure out how to define the values for throttle/neutral/brake, but some experimentation should get me mostly there. holy poo poo thank you again!
|
|
#
?
May 15, 2020 13:57
|
|
|
Neat. I'm going to guess that reading the pots is one thing but getting the right "feel" from the values/movement will end up the bigger fish to fry. Proudly show him an early prototype that generates only fart noises.
|
|
#
?
May 15, 2020 16:30
|
|
|
The Eyes Have It posted:Proudly show him an early prototype that generates only fart noises. only once the pitch has been mapped to the throttle
|
|
#
?
May 15, 2020 16:39
|
|
|
yeah I might have to try to map an acceleration curve to the throttle. the manual for the system even mentions that, so it’s probably not just linear then, yeah I’ll probably have to make a separate curve to map the fart pitches to.
|
|
#
?
May 15, 2020 17:28
|
|
|
did another doom inspired composition https://www.youtube.com/watch?v=WG9M67-aj2c
|
|
#
?
May 17, 2020 04:32
|
|
|
current project status: im sprucing up our back yard. we're gonna have to have our wedding ceremony back there with like 10 people max all distanced properly from each other
|
|
#
?
May 18, 2020 19:22
|
|
|
I’ve been learning how to write C programs for the original Mac, but then as most of the docs are in Pascal I’ve been learning that too. Think Pascal is pretty cool and let’s you just write some QuickDraw code and have it appear in a window without having to mess with any of the setup stuff.  This is a very modified version of the Tiled Lines example from https://generativeartistry.com/tutorials/tiled-lines/ Also I found a great repository of PDFs of old Mac programming books at https://vintageapple.org/macprogramming/
|
|
#
?
May 19, 2020 02:06
|
|
|
quickdraw owns
|
|
#
?
May 19, 2020 05:07
|
|
|
https://i.imgur.com/sqBZLwA.mp4 I think I'm getting close on the sim racing throttle and brake controller for my friend! I've gotten it to the point where the arduino shows up in Windows as a gamepad (hence the two way axis for the throttle and the single axis for the brake, although in practice this shouldnt matter) and it reacts the way I would expect it to when I give it input Weirdly, despite being designed to control a multi-ton steel box at highway speeds, the potentiometers are surprisingly sloppy in their zero readings. I wrote a serial monitor to get the raw pot values and I wound up having to write a deadzone into the software to account for the fact that when they return to zero, its never quite the same zero. The actual module used in cars has to be using a combined deadzone and also comparing the outputs of the three pots to divine where the stick actually is
|
|
#
?
May 19, 2020 15:39
|
|
|
I swear I'll get back to playing with antideluvian macs soon too but recently I've been obsessed with this ill-advised piece of kit: Its entire purpose is producing 180V for driving vacuum tube plates at approximately rectified-US-line-voltage from 12V input (e: using a separate 5V USB supply for the control electronics because why not gently caress everybody up with some weird USB poo poo?). The high voltage side is isolated from the input power brick which is ideally also isolated from the mains. It also has a little buck regulator for driving 6.3V filaments from the same input. It's all built around a single quad comparator IC, using one for an oscillator, one each for the high voltage and low voltage power supplies and one is tied off now but I think I'll use it as a current limit to protect the high voltage side. There are plenty of better ways to do this; plenty of all-in-one controller ICs, but my hope is for this to have some pedagogical value as well. Power electronics is unforgiving. When poo poo goes awry the first indication is frequently crackling followed by the smell of smoke. This can deliver watts at 180V. I have not yet been shocked by it and work with one hand behind my back. The custom proto board this is built on has its own story but I'll have to wait and share that if they're ever used for their original purpose. Stack Machine fucked around with this message at 04:43 on May 20, 2020 |
|
#
?
May 20, 2020 04:40
|
|
|
i made a script a few years ago that runs several png optimization programs in parallel, then chooses the smallest of the resulting files and feeds it back to all of the optimizers until the file stops becoming smaller. like really bruteforce that png. anyway, turns out, there either have been some advancements in png optimizers, or i just somehow managed to choose 4 of the worst ones to use: https://css-ig.net/png-tools-overview i used advpng, advdef, OptiPNG and pngout, you can find them near the bottom of the "savings" table  anyway, i'm redoing the script as a java program to try out CompletableFuture and parallel processing. also i added pingo to the 4 listed above and shockingly so far the results have been: just use pingo lol
|
|
#
?
May 20, 2020 14:47
|
|
|
thanks for sharing this! ive been using oxipng and thinking it was the best of the batch and now im gonna check out pingo too
|
|
#
?
May 20, 2020 14:58
|
|
|
Wheany posted:anyway, i'm redoing the script as a java program to try out CompletableFuture and parallel processing. also i added pingo to the 4 listed above and shockingly so far the results have been: just use pingo lol so far it seems that for "wallpaper like" images pingo always wins, but for icons some of the other tools manage to get a few more bytes squeezed off
|
|
#
?
May 23, 2020 08:53
|
|
|
Wheany posted:i made a script a few years ago that runs several png optimization programs in parallel, then chooses the smallest of the resulting files and feeds it back to all of the optimizers until the file stops becoming smaller. like really bruteforce that png. how common was it to ever see any benefit from a second pass, seems like that is the classic case where it's most likely neutral or worse
|
|
#
?
May 23, 2020 20:45
|
|
|
in something like imageoptim on macos would normally shave a few more % off with zopfli on a second pass in that app itll run several different png optimization libraries in combinations and pick the best but i think the second pass helps partly due to a practical limit of how many times itll iterate overall and that itll keep a local log of the tests tried so it may be going further into those slow combinations
|
|
#
?
May 24, 2020 05:11
|
|
|
i enjoy gardening now
|
|
#
?
May 24, 2020 05:31
|
|
|
Stack Machine posted:I swear I'll get back to playing with antideluvian macs soon too but recently I've been obsessed with this ill-advised piece of kit: What kind of transformer is that and where did you get it? I’ve been playing around with driving an array of nixie bar graph tubes, but they need a little more current than the boost converter with cheap inductor power supplies I’ve experimented with so far.
|
|
#
?
May 24, 2020 06:54
|
|
|
It's hand-wound on an RM-8 pot core (that's the form factor. These come in a few different shapes and sizes.). If you want to get it from digi-key or whatever, you go to the magnetics section and buy a "bobbin", the actual plastic bit with the pins, a ferrite core and some clips to hold the core halves together and then wind some magnet wire on the bobbin, then stick the core in when it's wound. Don't worry if you don't get the right clips because you can always tape or superglue the core pieces together. These are ferrite cores only useful for switching power supplies btw. If you want to run something straight off the mains down at 50Hz or 60Hz you'll need to buy some fancy laminated iron core thing. There are a few different circuit topologies for power supplies using these kinds of ferrite core transformers, and old EE neckbeards will chastise you for calling the ones with gapped cores like this one "transformers" instead of "coupled inductors" despite the fact that nobody who grew up in the CRT era ever heard of a TV "flyback inductor". If you want to discuss switching regulators we should probably take it over to the learning electronics thread but I'll drop a link here to a guy who's done what you're describing for nixie tubes: https://www.dos4ever.com/flyback/flyback.html It probably goes without saying but please be careful. Flyback converters can crank out moderate power at lethal voltages. This is an idiot spare time project for more than one reason.
|
|
#
?
May 24, 2020 17:33
|
|
|
BobHoward posted:how common was it to ever see any benefit from a second pass, seems like that is the classic case where it's most likely neutral or worse not very common at all. but every now and then a second pass helps a little bit and even more rarely a third pass also helps. anyway, i optimized a about 10k png files with pingo and then compared the results to the original file with imagemagick and... there was 0 difference between the original and optimized files (as expected, but pingo's web page warns that you should make backups, so i did). using pingo -sa <filename> (-sa being the "brute force" lossless optimization), the time taken was about 6 seconds for 1080p files on a i7-6700, which is very reasonable imo. also the space saved was about 40% compared to irfanview's lowest png compression, 20% for files that were probably saved at the default level. Wheany fucked around with this message at 21:35 on May 24, 2020 |
|
#
?
May 24, 2020 21:33
|
|
|
Wheany posted:not very common at all. but every now and then a second pass helps a little bit and even more rarely a third pass also helps. this was puzzling me so i looked poo poo up and (i think) am more enlightened now. i was aware that png used some kind of LZW adjacent dictionary compression so it didn't make sense that recompression could ever help (recompressing the output of one dictionary encoder with another rarely does anything helpful) apparently prior to dictionary compression, png goes through an encoding step which transforms raw image data into delta coded, since that results in datasets much more suited to dictionary compression. it's sort of like a more flexible / generalized notion of run length encoding, as far as i can tell. since there's a bunch of different encoding options in this first phase and the choices made can produce very different compression ratios after dictionary encoding, analyzing the image and making better delta coding choices must be the space where png optimizers do their thing idk why an optimizer would do better or worse based on the encoding of the input image though. it should be losslessly extracting pixel values from the original and then running its implementation of an encoder on the fully uncompressed dataset
|
|
#
?
May 25, 2020 10:39
|
|
|
a key factor is that the choice of data encoding is made per row of the image. it's really slow to try all permutations so normal tools often use a heuristic. some just use a fixed encoding choice for all rows, usually paeth since it results in the best average compression rate. good old DEFLATE is used for compression which has an effort setting of 0-9. saving PNGs is very slow even now so lots of tools compromise on that especially older ones, which is another avenue for improvement. i've written a parallel PNG encoder before that massively speeds up encoding because i needed to deal with very large images. it compromises compression rate by <1% as a tradeoff because it has to split the data into multiple DEFLATE streams. i don't know how common this is then there's a bunch of redundant headers tools can add. that probably only matters for tiny images.
|
|
#
?
May 25, 2020 11:44
|
|
|
gently caress yeah, i posted some game on newgrounds. if anyone was there during its golden years, this should throw you back https://www.newgrounds.com/portal/view/756186 it's a portal simulator where you play as the last member of the lock legion, trying to revive it and take over the world. it's basically cookie clicker meets plage inc meets newgrounds. it generates fun names automatically and has a pretty accurate portal model in it. i made it all in react, and i'm loving glad i did. this type of game definitely is a good fit for that.
|
|
#
?
May 27, 2020 01:11
|
|
|
God drat that does take me back. I had completely forgotten about the whole locks thing. Like wasn't there a strawberry lock or something dumb like that haha
|
|
#
?
May 27, 2020 07:42
|
|
|
spankmeister posted:God drat that does take me back. I had completely forgotten about the whole locks thing. Like wasn't there a strawberry lock or something dumb like that haha yes it was absolutely dumb and I was a part of it. ran the website for the lock legion back in the day and all. we were a spin-off from the clock crew, and actually for a while briefly surpassed them in productivity
|
|
#
?
May 27, 2020 08:21
|
|
|
lol rad
|
|
#
?
May 27, 2020 08:26
|
|
|
i'm thinking right now, this game engine is universal enough that i could convert this game to a Goon Gobbler type of thing, where instead of the NG portal you have the sa forums. with auto-generated thread titles and trying to get as many goons possible. yay/nay?
|
|
#
?
May 27, 2020 15:55
|
|
|
:yay:
|
|
#
?
May 27, 2020 16:04
|
|
|
you need a random event where you lose another vertebra once every few years and have to figure out how to fund a new one via patreon
|
|
#
?
May 28, 2020 10:16
|
|
|
a wide array of bone craftspeople, bone factories, bone mines, bone synthesizers
|
|
#
?
May 28, 2020 21:27
|
|
|
and maybe a coconut nintendo
|
|
#
?
May 28, 2020 21:33
|
|
|
i got adobe creative suite for work and i've been learning what some of the animation tools do, just for the heck of it today i made a cardboard dude and made him sing one of my songs https://www.youtube.com/watch?v=wqYcfd814e0
|
|
#
?
May 30, 2020 16:47
|
|
|
its wonderful
|
|
#
?
May 30, 2020 20:01
|
|
|
did another DOOM inspired composition, arranged the fft height field in polar coords https://www.youtube.com/watch?v=mqmW32KV97s
|
|
#
?
Jun 4, 2020 06:41
|
|
|
orly posted:did another DOOM inspired composition, arranged the fft height field in polar coords neat
|
|
#
?
Jun 4, 2020 15:56
|
|
|
PIZZA.BAT posted:neat thx  Did another recently with sheet music viz:https://www.youtube.com/watch?v=3rUEawH_g4Q
|
|
#
?
Jun 10, 2020 20:52
|
|
|
|
| # ? Apr 23, 2024 15:35 |
|
|
i've been making a web template converter https://springload.github.io/metacomponent/
|
|
#
?
Jun 14, 2020 09:06
|
|
