#garfield Logs
Apr 14 2025
#garfield Calendar
12:11 AM aandrew: rue_mohr: did you take a look at that data aquisition over HDMI capture dongle link I pasted a few days ago?
12:26 AM rue_mohr: uh?
12:26 AM rue_mohr: I think I missed that one
12:26 AM rue_mohr: I didn't know you could override the protocol on the chips
12:26 AM aandrew: https://github.com/steve-m/hsdaoh?tab=readme-ov-file
12:26 AM aandrew: you can't, but you can override the encoding and get raw frames, so you encode the data you want to capture at high speed into HDMI frames
12:26 AM aandrew: this is another one of those "dammit why did I think of that" ideas
12:26 AM rue_mohr: ah 'use fpga to frame your data'
12:26 AM rue_mohr: mmm
02:06 AM rifraf: ok thanks Tom_L can anyone else confirm same behaviur and will try make the apk again
08:30 AM polprog: rue_mohr: is that by chance a circuit from the ARVID system>
08:30 AM polprog: ?
08:32 AM rue_mohr: polprog, er? which circuit?
08:35 AM rue_mohr: bbl
08:41 AM polprog: 07:04 < rue_mohr> I have an old russian circuit for storing serial data on VHS tapes
02:24 PM rifraf17: tom i made new apk, can give you link if you are able to try it
03:00 PM polprog: rifraf17: what does the apk do?
03:01 PM rifraf17: detects RPM of a spinning top and other cool data
03:01 PM rifraf17: wanna try?
03:02 PM polprog: oh neat
03:02 PM polprog: sure
03:02 PM rifraf17: you need andriod, something to spin, something to spin it on so it won't fall of phone, and to protect screen a bit, like a very small saucer
03:03 PM rifraf17: but fist need to see if it installs ok, i will make a new apk now, still debugging some maths
03:03 PM polprog: sure
03:05 PM rifraf17: ok is still unsigned apk, http://desire3d.com/images/public/vibesensing.apk
03:05 PM polprog: allright
03:05 PM polprog: ill test it when im done with erranda
03:05 PM polprog: errands*
03:05 PM polprog: just happen to have a spinning top here
03:06 PM rifraf17: cool :)
03:14 PM Tom_L: rifraf,
03:14 PM rifraf: yes
03:14 PM Tom_L: oh i see the new link now
03:16 PM rifraf: screenshot of what it should look like https://desire3d.com/images/public/vibesense1.jpg
03:17 PM rifraf: also you will likely need to isolate your phone from external vibrations, i sit it on a piece of foam or something soft and flat
03:19 PM Tom_L: ok it installed ok and is open
03:19 PM Tom_L: but i don't have a top
03:19 PM rifraf: make one, you have enough machines
03:19 PM Tom_L: i'm tired
03:19 PM rifraf: 5 minutes you could hack a basic one up on the lathe
03:20 PM rifraf: Tom_L the the graph will be all over the place trying to pickup vibrations?
03:20 PM Tom_L: i had a shitttttty one that will spin like 10 sec and got a 1560 rpm from it
03:20 PM rifraf: if so then its seeing your accel ok, and vibelevel should go close to 0 when phone is flat
03:21 PM rifraf: ok, default fft rate needs to be tweek, settings down the bottom, if you see FFT rate change to 1024, the default is broken at moment
03:21 PM Tom_L: you should slow the sample rate down on leveling
03:22 PM Tom_L: it's sitting on my desk at -112ish
03:23 PM Tom_L: with nothing spinning just sitting there i get rpm readings
03:23 PM rifraf: ok, try something soft under phone, 10 second spin should still hopefully give you a few similar rpm
03:23 PM Tom_L: the room is not a super quiet room either
03:23 PM rifraf: yes we are picking up minute vibrations
03:24 PM Tom_L: what about ppl talking?
03:24 PM rifraf: is all fine when you have a top spinning properly
03:24 PM Tom_L: ok well it appears to work
03:24 PM rifraf: and have a phone on foam or similar
03:24 PM rifraf: cool thanks, thats great initial feedback
03:24 PM Tom_L: stack of napkins
03:25 PM Tom_L: send me a top :)
03:25 PM Tom_L: and i'll test it out good
03:25 PM rifraf: you can reset and use pause resume to clear screen
03:25 PM Tom_L: i did
03:25 PM rifraf: ok will do soon
03:25 PM Tom_L: i'm halfway kidding on that
03:26 PM rifraf: well i have so many, if you are happy with not perfect have alot
03:26 PM Tom_L: my lathe is in sad shape and i doubt i could get a decent top from it
03:27 PM Tom_L: i've tested other products out
03:28 PM rifraf: yes this can do things like work out car rpm, or vibrations of moters etc, but made for tops
03:28 PM rifraf: use it to debug machinery vibrations etc
03:28 PM Tom_L: several of those mesa electronics boards and of recent i was the primary debugger of this: https://www.gnipsel.com/linuxcnc/flexgui/index.html
03:28 PM Tom_L: https://github.com/jethornton/flexgui
03:29 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/single_point_threadmilling.png
03:29 PM rifraf: cool, well have a few bugs left if you want something to do
03:30 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/touch-probe3.png
03:30 PM Tom_L: i did that gui up to test with
03:30 PM Tom_L: and around 40+ subroutines for various operations
03:30 PM rifraf: but you need to understand vibration and FFT / PSD a bit
03:30 PM Tom_L: we'll see how good a teacher you are :)
03:31 PM rifraf: heh you can ask chatgtp, it can teach you faster
03:31 PM Tom_L: what are they?
03:44 PM rifraf: who?
03:44 PM Tom_L: FFT / PSD
03:45 PM rifraf: fast fourier transforms and power spectral density, fun stuffs
03:45 PM Tom_L: Fast Fourier Transform
03:45 PM Tom_L: if you understand it i suppose
03:45 PM Tom_L: above my pay grade
03:45 PM rifraf: yep, cannot work out rpm without them
03:45 PM rifraf: i guess you will not be debugging then
03:45 PM polprog: fast foyer transform
03:46 PM polprog: when you do all 4 walls in a day
03:46 PM Tom_L: nope
03:46 PM Tom_L: pol can
03:46 PM polprog: yeah i could try though i never did any real android app work
03:47 PM polprog: youre doing it in Java?
03:47 PM Tom_L: are you using android studio?
03:47 PM rifraf: kotlin
03:47 PM rifraf: yes studio
03:47 PM rifraf: debugging ok, is just stuff i need time for
03:47 PM Tom_L: https://kotlinlang.org/
03:47 PM polprog: thats what the cool kids use now
03:48 PM polprog: its not bad
03:48 PM Tom_L: rifraf was cool long before
03:48 PM rifraf: i just want to get the world of spinning tops to test this app, and then they will see they need my tops
03:48 PM Tom_L: create a need
03:48 PM rifraf: but would be awesome to see famous tops compared in time
03:49 PM rifraf: well is one of a few apps that interact with a physical item
03:49 PM rifraf: sure there will be a few cracked phone screens, but thats not my problem
03:50 PM rifraf: if you had one of my concave lenses the phone will be fine
03:50 PM Tom_L: spinning the top on the phone?
03:51 PM Tom_L: a makeup mirror would do ehh?
03:52 PM Tom_L: https://www.amazon.com/CRILSTYLEO-Magnified-Bathroom-Application-Single-Sided/dp/B0DP1GBQHX?gQT=1
03:53 PM Tom_L: maybe smaller
03:54 PM Tom_L: or a Hubble mirror
03:54 PM Tom_L: :)
03:55 PM rifraf: makeup mirroes are great
03:55 PM rifraf: the concave ones
03:55 PM Tom_L: right
03:55 PM Tom_L: do you get those in bulk?
03:56 PM rifraf: i use lenses, and use
03:56 PM rifraf: yes*
03:57 PM Tom_L: well i probably helped all i can on this as far as debugging then
03:58 PM Tom_L: the apk behaves better
03:58 PM rifraf: yep, you can also tell me things that don't work as expected, alot of the settings are just there for debugging, but great if you want to test different types of vibration, come up with some handy use cases
03:58 PM Tom_L: spindle rpm
03:58 PM rifraf: yeah, i made that one with studio, before i just used an extract apk tool to use on existing app
03:59 PM Tom_L: engine rpm
03:59 PM Tom_L: heartbeats
03:59 PM Tom_L: for wearables
03:59 PM rifraf: yep, tape phone to spindle,, sit flat on top, adjust max freq for the graph and try it out, maybe lower sample rate and fft
04:00 PM rifraf: it will work
04:00 PM Tom_L: you might get someone in #linuxcnc to help.... dunno
04:01 PM rifraf: i did all thins before with my top testing machine, arduiono and adxl345
04:01 PM rifraf: so i knew what i wanted from the phone app
04:01 PM Tom_L: nice
04:01 PM rifraf: i don't need help, just people to try it out and get feedback to improve before i put on play store
04:02 PM rifraf: AI has been a bitg yet frustrating help, but now the code too big for it, and it gets lost very easy if i am not careful
04:02 PM Tom_L: is there a criteria to meet to qualify for the playstore?
04:04 PM rifraf: less the iuphone apps, need developer account and a few things that not too hard
04:06 PM rifraf: rue_mohr i am making a second app already, notepad list app for phone, just for fun but can add features your need
04:07 PM rifraf: am using it daily already, just to make a quick text reminder and edit or delete them as done
04:09 PM rifraf: polprog did you get chance to install yet?
04:11 PM polprog: no, im a bit busy
04:11 PM rifraf: no worries, sound like it should work when you ready
04:11 PM polprog: https://polprog.net/rozne1/ircjunk/philips_la/front_panel.png
04:11 PM polprog: just tested this, turns on
04:11 PM polprog: i will have to make a logic pod for it
04:13 PM rifraf: wow, from 1985
04:13 PM polprog: yeah
04:13 PM polprog: the logic pod should be easy, its just a few TTL register chips
04:18 PM polprog: the service manual has schematics inside
04:18 PM polprog: of everything
04:27 PM Tom_L: rifraf, i use advanced editor
04:42 PM polprog: nigt
05:06 PM Tom_L: rifraf, doing pause/resume you lose the donut and get a color changing circle if that's the desired effect
06:16 PM aandrew: Very weird failure of this computer. Won’t power on but there’s an internal switch when I press the fans come on for as long as the button’s pressed
06:28 PM rue_mohr: brain dead
06:29 PM rue_mohr: managment processor sounds dead
06:30 PM rue_mohr: maybe try a cmos clear, but I doubt it
06:37 PM aandrew: Power supply appears fine, so yeah something on the mb… gotta check the drive and pcie card I had in there
06:40 PM rue_mohr: I wonder if it opened some ground connections
06:40 PM rue_mohr: were you able to see the bottom of the board?
06:40 PM rue_mohr: are there any burned traces on the top?
06:40 PM rue_mohr: the ground plane shoudl be pretty solid
06:45 PM aandrew: Nothing obvious yet, haven’t removed the motherboard
06:46 PM aandrew: Have verified that 5V, 12V and 3.3V rails do come up on the drive connector, but none of the not populated caps show any significant voltage when it is forced on
06:46 PM aandrew: Resetting cmos didn’t do anything as predicted
06:55 PM aandrew: Found it
06:58 PM aandrew: https://ibb.co/LzVRXDWb
07:05 PM aandrew: Looks like a TVS which makes sense but removing it didn’t fix anything
07:06 PM aandrew: I’m not going to spend much time on this, but I do hope the ssd and pcie are fine. If the ram and cpu are still good that’d also be a bonus
07:13 PM rue_mohr: the fuses would be on +v tho
07:13 PM rue_mohr: you suspected a 24V thru the ground bus
07:13 PM rue_mohr: how much current is the power supply that did it?
07:16 PM rue_shop4: well that really sucks, how did the +24 get to ground again?
07:16 PM rue_shop4: did the drive have anything important on it?
07:16 PM rue_shop4: I'm waiting for a robot to charge so I can try some new code
07:18 PM aandrew: The power supply that did it is 800W and was happy to keep putting 24V out during all of this
07:18 PM aandrew: Nothing on the drive is irreplaceable but just annoying
07:19 PM aandrew: Yeah I’m gonna call this cooked. It’s not a hard failure but that just makes it more difficult to diagnose. Not worth my time
07:20 PM rue_mohr: so I presume it also cooked the arduino on the printer
07:20 PM aandrew: I could buy the schematic for this dell motherboard (E93839) but I’ll just grab another sff from work insyead
07:21 PM aandrew: Nope. The printer is working 100%
07:21 PM rue_mohr: what, how?
07:21 PM aandrew: I told you, it’s weird
07:22 PM rue_mohr: so a 24loop from usb ground to mains ground?
07:23 PM rue_mohr: maybe find a usb isolator for this printer
07:24 PM aandrew: the bed heater connector was in contact with the metal chassis. That chassis was sitting at 24V which is odd because I fully expected the chassis to be connected to the earth/green on the power cable
07:24 PM aandrew: if I touched the USB shield and a random chassis screw I was measuring 24V
07:25 PM rue_mohr: you know
07:25 PM aandrew: the USB shield is not touching the chassis (it should be) and I would have assumed that the screws connecting the printer motherboard to the chassis would have been metal and an earth/shield connection
07:25 PM rue_mohr: the disasters are ALWAYS the bed heater
07:25 PM rue_mohr: I dont have bed heaters on any of my printers and I cant imagine the trouble its saved me
07:26 PM rue_mohr: some power supply outputs are floating
07:27 PM rue_mohr: I chained 3 power supplies once for 36V
07:27 PM aandrew: yeah I didn't take this printer completely apart to see
07:27 PM rue_mohr: two of them I didn't ahve to modify and the other was so ugly to mod I made it the bottom supply
07:28 PM rue_mohr: it doesn't make sense that a usb cable could carry enough current to do that much damage
07:28 PM rue_mohr: hey, did you open the hub it killeD?
07:30 PM Tom_L: i parallel'd 3 for more current
07:30 PM rue_mohr: were any of them upset about that?
07:30 PM Tom_L: probably didn't need the current
07:30 PM Tom_L: no i'm still running it that way
07:31 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/rue/Control2.jpg
07:32 PM rue_mohr: oh but not switching
07:32 PM rue_mohr: heh
07:32 PM rue_mohr: those wont care
07:36 PM aandrew: nope I haaven't opened that yet -- it acutally killed two hubs
07:36 PM rue_mohr: maybe see what it did to them
07:36 PM aandrew: a real cheap usb2 one I don't care about and a nice usb3 one I rather liked, had it for years
07:38 PM aandrew: yeah I am, just updating the firmware on the expensive webcam that was hooked up and didn't get fried
07:38 PM aandrew: it's not the resolution so much as the fact that it has either 8x or 10x *optical* zoom along with PTZ
07:38 PM aandrew: makes it nice to point at things around my garage while I'm upstairs
08:53 PM rue_shop4: hahah, my scrolling code isn't working, and the bug I didn't successfully find had me fixing bugs I didn't know about
08:58 PM rue_shop4: wtf, is something doing cr-lf converstion???
08:58 PM rue_shop4: 0x0a should be \n
09:02 PM rue_shop4: at drat
09:02 PM rue_shop4: and now the extra time to scroll the screen causes it to drop serial characters
09:05 PM Tom_L: java?
09:05 PM Tom_L: i had to do something special for java when i exported a file to it
09:05 PM Tom_L: it didn't like a few characters
09:06 PM rue_shop4: C, 8051
09:09 PM rue_shop4: well thats awkward
09:10 PM rue_shop4: something isn't fast enough
09:10 PM rue_shop4: I'm not sure if its the 8051 or the pcd
09:10 PM rue_shop4: maybe the lcd cant keep up with 9600 baud data?
09:10 PM rue_shop4: frustrating
09:13 PM rue_shop4: well, I could pin the performance a bit by iplementing the busy flag check
09:13 PM rue_shop4: ditch the static delays
09:13 PM rue_shop4: ugh