lordpil's blog


jero32: VHB
jero32: https://www.amazon.com/Tape-RP32-width-length-Roll/dp/B007Y7E0CQ
guys do you think i could smoke the flux on some ~dpak2 size fets without burning anything?
because nothing looks burnt at all, but they havea nice flux sheen, and i swore there was smoke and def was a smell

Posted by renesis at 14:11 | permalink | 0 comments


wifi comm to pi gpio the only thing has worked seamless this project
<3 ssh ui

Posted by renesis at 13:36 | permalink | 0 comments


jero32: thats not bad for a cheap watch doesnt even have an xtal
watches have had xtals since the 70s
https://www.digikey.com/product-detail/en/ecs-inc/ECS-.327-CDX-1038/XC2121-ND/5875500
half a second a day
so couple minutes a year
sounds about right

Posted by renesis at 13:21 | permalink | 0 comments


a few seconds a dat is prob like .01%
i fucking hate this little keyboard
i should see if the mechanical one is working now that its dried out
32khz xtals are huge and more than free
0.002%

Posted by renesis at 13:08 | permalink | 0 comments


if ms resolution is good enough, sure
if you need cycle precise resolution, you can just use timers
and systick int is 1ms in HAL
HAL is slow
they can be
typically you would have a timer trigger an interupt
or you can have a timer chained to another timer, and you can check time on the fast one,, and use the slow one to see how many times it overflowed
you can do either
you can drive timer counts with overflow events from other timers
you can chain timers in stm32 but i dont know the details
i dnt think its an xtal i think its an rc, and probably yesy if you dont mine a few seconds error a day

Posted by renesis at 13:03 | permalink | 0 comments


platform screws def done
gearbox def more clicky
i guess i work on uart code

Posted by renesis at 12:30 | permalink | 0 comments


timecop: i did and it got fucked
i tried ziggler nichols method and shit changes direction so fucking fast
something slightly smoked, i think because full pwm and almost stalled
which is impressive because the h bridge is overrated as fuck
oh maybe it was ecig
anyway it can inverted pendulum for up to maybe 5 sec but its like ripping itself apart with reversals
motor gearbox is like fuck you
ssh estop is success, i stopped it before it attached my bike
aluminum on aluminum, might have been a good fight

Posted by renesis at 12:20 | permalink | 0 comments


nothing stripped, but every screw that wasnt threadlocked loosened
including magnet mount, which explains the siezure that prob wiggled motor platform screws out
shit is worse than subwoofers

Posted by renesis at 11:52 | permalink | 0 comments


k
code back to where it was before timers and alignment bullshit
increasing D seems to break shit
i think i maybe have stripped the too short PC screw holes
well, one or two of the 4, now whole platform wobbles

Posted by renesis at 11:32 | permalink | 0 comments


i am like wtf is #pragma and keil is like ya iunno either
st.com official forum rep w/ that suggestion

Posted by renesis at 10:24 | permalink | 0 comments


uint16_t tx[] = {0x90, 0, 0};
this gets cast to uint8_t* for the spi txrx function
in the function, it uses 16b xfer mode even tho set for 8b, so does casting back to uint16_t
if you trust the input variable cast, seems like it should bork
anyway looks like its transmitting right
im not sure if its using half my buffer and im just lucky because all zero after leading data
the shit is if i edit the generated code itll clobber everytime i use cubemx
mfkr

Posted by renesis at 10:04 | permalink | 0 comments


because it sometimes breaks
right and i proved it so fuck hal
the issue it is tries to use 16b transfer mode when you tell it to be 8b
how do i do that?
instead of doing 16b array and packing shit stupid
c++ not c
well fuckit lets see what its doing

Posted by renesis at 09:57 | permalink | 0 comments


rab: fuuuuu
rab: link is true
my spi shit was working fine in main loop, then when i tried to put it into systick int and then tim7 int, shit was fucking up, 3 byte arrays declared in diff files (main vs the it.c file)
so i think im doing ints wrong but shit works fine without spi and in debug it keeps hardfaulting at this uint16 cast
no changing to uint16_t array which gets cast into uint8* worked i think
it does that
it never says why but right this is prob why all the examples are cast to 8b arrays

Posted by renesis at 09:52 | permalink | 0 comments


https://community.st.com/thread/13259
wtf

Posted by renesis at 09:38 | permalink | 0 comments


18:39:46 <+BlackMoon> renesis: milwakuee now has a cordless heat gun
with like, special battery pack?
does it just short the battery and blow air over it?

Posted by renesis at 09:12 | permalink | 0 comments


that one is important

Posted by renesis at 08:11 | permalink | 0 comments


https://stackoverflow.com/questions/18244726/why-doesnt-c-have-binary-literals
According to Rationale for International Standard - Programming Languages C §6.4.4.1 Integer constantsA proposal to add binary constants was rejected due to lack of precedent and insufficient utility.
so right theyre assholes because they took the time to learn hex so fuck everyone else
what did you do wrong
gcc implimented it
and keil does not do it

Posted by renesis at 08:06 | permalink | 0 comments


that part isnt working
heheh
didnt he have pic of the digikey bag?
jero32: bad resistor soldering
that was funny

Posted by renesis at 07:44 | permalink | 0 comments


jero32: just set it high in cubemx
or go to the init function and replace their init output
yes
lame timer int isnt broken something else is hard faulting
jero32: just set it in cubemx
if it doesnt work then your shit is prob fuckey
its just a resistor and led or there is a switch or what
go to configuration tab
its in the gpio block
you put the led backwards
]or burnt it up with the iron
those you can melt those things and they still work
*tho
you switched it with cubemx?
how do you know your psu probes werent backwards
?!
or tack a wire to it and dmm that

Posted by renesis at 07:39 | permalink | 0 comments


rab: thats almost asm
GPIOA->ODR is basically GPIOA register block, offset to data register, or with 0b00010000
i hate that you cant 0b in c
wonder how diff the hal shit compiles
what kind of complete asshole dates a single mom and doesnt want them dealing with their kid

Posted by renesis at 07:33 | permalink | 0 comments


HAL_GPIO_WritePin(GPIOA, GPIO_PIN_10, GPIO_PIN_RESET);
is how you clear
has it ever been on?

Posted by renesis at 07:17 | permalink | 0 comments


HAL_GPIO_WritePin(GPIOA, GPIO_PIN_10, GPIO_PIN_SET);
i dont have to know if its enumerated or a pin mask
i just have to make it look like that and it works
yeah so that pin define is enumerated from zero
the prof who said i dont have an engineering mind got aqll confused when i used that syntax
thats the other HAL pin diddler
yes

Posted by renesis at 07:09 | permalink | 0 comments


i dunno
it might be a position number
so it might be 0x01
or it might be an index, so actually 0, and you would have to do 1<<GPIO_PIN_0
right click, go to definition
or, you know, just do 1<<0, or 1, or 0x01
rab: man fuck i dunno if learning hal is worth it
i dont type very fast so maybe

Posted by renesis at 07:04 | permalink | 0 comments


blackmoon: oh huh
jero32: it only does that to me when i have an error and it didnt finish compiling

Posted by renesis at 06:58 | permalink | 0 comments


wtf
pointless comparison of unsigned integer with zero
fuck your opinion keil no one asked
its (uint >= 0)
its like it decomposed it to (uint > 0) || (uint == 0) and then bitched at itself for not doing !(uint)
anyway it says code has 1 warning, i am contesting my code is fucking awesome, the compiler is the problem

Posted by renesis at 04:21 | permalink | 0 comments


Top | Add to Technorati Favorites


© 2007 lordpil.   XHTML 1.0! CSS! Site design by GNAA  Blog Engine by pbx | MULTI2 | ian hanschen | lolwat