#garfield Logs

Apr 13 2021

12:01 AM aandrew: hm, fusion 360 is saying the hole diameter is 32.05mm which is way way wrong
12:08 AM aandrew: so like you said it seems the units don't mean shit
12:08 AM rue_mohr: its been suggested that I prevent the chip from booting by removing the firmware rom
12:08 AM rue_mohr: then the jtag wont be disabled
12:10 AM rue_mohr: 25P010
12:21 AM rue_mohr: it just says rst...
12:32 AM rue_mohr: so I wonder how multi procesors work
12:32 AM rue_mohr: would they each have a set of start vectors in memory?
12:35 AM aandrew: nope
12:35 AM aandrew: they see the same memory
12:35 AM aandrew: but you program the NVIC so only the core(s) you want to see a given interrupt see it
12:36 AM rue_mohr: hmm
12:36 AM rue_mohr: one of the guys started a linux kernel on one of these controller boards
12:36 AM rue_mohr: how did the kernel know anything about that level of hardware...
12:37 AM aandrew: the kernel's extremely portable
12:38 AM rue_mohr: heh
12:38 AM rue_mohr: I'm kinda wondering if the count of vectors in that table gives anything away
12:38 AM rue_mohr: aside from knowing its an arm core, there is quite a lack of info
12:39 AM rue_mohr: my idea is this...
12:39 AM rue_mohr: IF I can get ahold of the PATA 16 bit data port
12:39 AM rue_mohr: provided it can be a real latching port
12:39 AM rue_mohr: and if I can get ahold of the dma
12:40 AM rue_mohr: then to make it stream vga from ram
12:40 AM rue_mohr: the serial port would go to the z80 for the low end interface
12:40 AM rue_mohr: and one of the cores could do the graphics operations
12:40 AM rue_mohr: draw line, circle, etc
12:41 AM rue_mohr: breaking that down into small steps could be insane
12:51 AM rue_mohr: <rue_mohr> if the dma is designed for 66Mhz
12:51 AM rue_mohr: <rue_mohr> whats pixel clock...
12:51 AM rue_mohr: <rue_mohr> must be lower than that
12:51 AM rue_mohr: <rue_mohr> from 48.6Mhz to 59.5Mhz
12:51 AM rue_mohr: <rue_mohr> oh sorry wrong channel
12:51 AM rue_mohr: so it looks like the pata interface should be able to do VGA
12:52 AM rue_mohr: 921600 pixels,
12:53 AM rue_mohr: at 8 bits per, thats 1MB of ram
12:53 AM rue_mohr: 4096k x 16 SDRAM
12:53 AM rue_mohr: so, that'll work
12:53 AM rue_mohr: I wonder
12:53 AM rue_mohr: if I took the board with 1M of ram, and changed the chip, if it would find it
12:54 AM rue_mohr: I'd like two graphic pages
04:55 AM polprog: ohh you got it working, nice
09:57 AM rue_shop1: not jtag tho
10:09 AM aandrew: yes in theory you should be able to spew that shit out parallel port if you can figur eout the peripherals
10:09 AM aandrew: personaly I'd get it working on an STM32 dev board first, prove out the design and then try to reverse engineer something
10:50 AM aandrew: hey my mechanically-inclined friends
10:50 AM aandrew: I'd like to automate roller chain blinds (basically just an opaque sheet rolled up at the top fo the window, a ball chain hanging down that you pull one way to unroll, pull the other way to roll up)
10:51 AM aandrew: lots of designs for this kind of thing on thingiverse but as far as I can see, none allow you to use the chain manually in addition to using the motor
10:52 AM aandrew: I assume you need some kind of clutch assembly so that if the force is coming from the ball chain gear side of the clutch (in either rotation) it lets it go, but if the force comes from the motor side it engages and makes the ball chain gear turn
10:52 AM aandrew: the blinds "stay" where you put them so you don't need to hold the chain when you're at the height you want
10:53 AM aandrew: is there a mechanical way to do this, or do I need to use a little magnetic clutch that engages when I want the motor to turn the chain and stays disengaged otherwise?
12:15 PM polprog: hmm, maybe a slip gear
01:20 PM aandrew: hm, https://www.thingiverse.com/thing:4704712 might work... put a spring inbetween those two inner pieces and have something that pushes them together when the motor turns
01:20 PM aandrew: motor stops turning (or backs up a little to relieve pressure) and it pops apart
01:21 PM aandrew: but getting that mechanism where the turning of the motor pushes one of the inner pieces "in" might be tricky
01:21 PM aandrew: https://lounge.mikkel.cc/uploads/8cc9a1f487007f3f/image.png hahahahaha
04:38 PM polprog: :P
06:10 PM rue_mohr: slip clutch
06:10 PM polprog: slip clutch yes
06:10 PM polprog: why did i call it a gear
06:10 PM polprog: i suck at mech jargon
06:11 PM rue_mohr: how did that poster get by everyone?
06:11 PM polprog: did i tell you..? the english class im taking this semester is english in car mechanics
06:11 PM polprog: our handbook is based off a toyota service manual
06:11 PM rue_mohr: odd
06:12 PM polprog: toyota learning materials, not one particular manual
06:12 PM rue_mohr: k
06:12 PM polprog: i have to take an english class every semester for some reason, i wish they just left me alone
06:12 PM polprog: but that class is very neat
06:13 PM polprog: back from work?
06:19 PM polprog: nighters
06:27 PM rue_mohr: hah
06:27 PM rue_mohr: we are supposed to hack hard drives now
06:28 PM rue_mohr: I'm going to try resistors between magic pads and lifting chip select lines and stuff
06:30 PM rue_mohr: https://twitter.com/4jawjan/status/1382098856922521605/photo/1
06:30 PM rue_mohr: uh
06:31 PM rue_mohr: why would you not cnc that.
06:52 PM aandrew: why not cnc what, what looks like a threading bit?
06:52 PM rue_mohr: but
06:52 PM rue_mohr: thats got to be a $3000 tap
06:53 PM aandrew: yeah it odesn't look cheap
06:55 PM Tom_L: 4" 5 npt
06:57 PM Tom_L: ~ $650
06:57 PM Tom_L: https://drillsandcutters.com/4-8-npt-carbon-steel-pipe-tap/
06:58 PM Tom_L: not the same thread
06:58 PM Tom_L: err yeay it is, i read the side of the tap wrong the first time
07:01 PM aandrew: 4" 8 NPT I think
07:01 PM aandrew: that seems ... cheap?
07:10 PM aandrew: slip clutch eh
07:10 PM aandrew: interesting
08:39 PM rue_mohr: Tom_L, you had a 5x7 font around
08:40 PM rue_mohr: I thought I saved it but I dont know where I would have put it
08:41 PM Tom_L: bit file?
08:41 PM rue_mohr: dunno
08:41 PM rue_mohr: anything I can use
08:41 PM rue_mohr: :)
08:41 PM Tom_L: well i'm not sure either
08:41 PM Tom_L: was it cad or for lcd?
08:42 PM rue_mohr: I think you used it for an lcd
08:42 PM rue_mohr: for a font on a graphic?
08:43 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/rue/LCD/Fonts/
08:43 PM Tom_L: maybe that's it
08:47 PM aandrew: rue_mohr: I think I have one handy if you need one
08:47 PM rue_mohr: I cant work out how wide the chars are
08:48 PM rue_mohr: Tom_L, yours might be 8x8
08:48 PM rue_mohr: aandrew, 5x7?
08:48 PM rue_mohr: 0x00, 0x3c, 0xfc, 0x78, 0xfc, 0xf0, 0x00, 0x00, // '#'
08:48 PM rue_mohr: yes, tom I think you have 8x8 there
08:49 PM Tom_L: that's likely it then
08:49 PM rue_mohr: ok
08:49 PM rue_mohr: close!
08:49 PM aandrew: rue_mohr: I think it is yes
08:50 PM aandrew: www.mixdown.ca/dump/oled.c oled.h and tahomabd.h
08:50 PM rue_mohr: hook me up
08:50 PM aandrew: the last is the font
08:50 PM aandrew: pretty sure it's 5x7
08:50 PM rue_mohr: www.mixdown.ca/dump/oled.h
08:51 PM rue_mohr: www.mixdown.ca/dump/tahomabd.h
08:51 PM aandrew: nope it's variable width
08:51 PM rue_mohr: drat
08:52 PM aandrew: sorry
08:52 PM rue_mohr: wtf, twitter is muting my thread
08:54 PM rue_mohr: https://github.com/adafruit/monochron/blob/master/firmware/font5x7.h
08:54 PM rue_mohr: ok
08:54 PM rue_mohr: less the WTF
08:55 PM rue_mohr: but its for arduino, so ok
08:55 PM rue_mohr: oh ffs, twitter switched me to top tweets again
08:56 PM Tom_L: is that like king of the hill?
08:56 PM rue_mohr: yea I will never make top tweets
09:24 PM rue_shop1: 274002
09:24 PM rue_shop1: yes 16 bit
09:44 PM rue_mohr: its not blank rue
09:51 PM Tom_L: question is, is it worth keeping
09:52 PM rue_mohr: yup
11:35 PM rue_mohr: aaha if I switch the 74138 for a 4028 its fixed