this is all spi1, spi2 has a nrf24l01 wired in, got a working setup with a maple.
cycled through the displays, 1 a bit light on the contrast, but quite enough for test.
all displays functional
tried another maple – nope
removed bread board to safe distance, but duplicated it, checked ‘safe mini and nokia’ ok
then went through the maples, none
created a red pill version on another breadboard, wow, winning streak… … …
3 of 4
went back to try the maple version on the copied ‘maple’, now more are working.
went on to blue pills, wow good, good, good
pleasant 2-3 hrs though, 4 if wiring rebuilds & scrutiny included,
sssssoooooo
what i did do was change the board type.
maples – all bootloader v2
pills – st-link, always do.
multiple re-compiles and downloads, couple of the maples got the updated boot loader
what i didn’t do was restart the ide?
is it possible something is getting set and not being cleared or not being flushed on a board change.
anyone else had the it wasn’t working, but all of a sudden or i came back in the morning and it worked moments?
why oh why does this happen to me … … …
the last 52yrs if i include senior school, 44 if not; and no it’s not me being paranoid.
why is the feline so sure i won’t mind if she just rolls over on anything in the way, under my hand for a head scratch, 3 times so far.
probably more a case of i don’t care, but she is grounded!
the ultimate i suppose would be to tgz everything ~/ard165 and ./.arduino15 and sketchbook.
cut sketchbook right down to the 2 ino’s and required libraries. i’ll even cut my use of soft links to zero,
tgz that setup, flush, populate & try for maple and pill types in turn.
thoughts? chapter 2?
stephen
A: Gremlins
- download.jpg (5.64 KiB) Viewed 448 times
once upon a time a while back, I actually named a kitten Gimzo and it stuck. No one renamed it!
you hear stories about red setters being daft, they’d be totally outclassed by Gizmo.
currently copying directories in preparation … … …
stephen
into each, created sketchbook, copied as required from ~/sketchbook to sketchbook/hardware and sketchbook/libraries, only libraries required.
created copy of ard165 in each
removed ~/.arduino15
started arduino, set preferences to ~/test_maple/sketchbook
restarted arduino, install avr boards @1.6.5 and cortex-m3 1.6.5
restarted arduino, selected maple mini
quite happily compiled and programmed 7, all working
repeat the above under test_pill and select generic c8t6 / stlink, 2 red and 2 blue pills programmed and working
anyone care to try the next step as well as me?
off to create test_both
stephen
yes, i am trying a pill(stlink) and a maple(bl2) at the same time.
edit source to define LED for each and // the one i want
pill-> ok 1hz
maple-> ok 1 hz
set times to 100/1000
pill(stlink setting remembered)->ok 100/1000
maple(also remembered bl2/acm0)-> ok 100/1000
pill is doing 100/1000
maple is actually doing 1000/100
the /tmp/build* is re-used, so i copied that dir, first change has 22 files different changing to maple, reverting to pill it was 6
repeated switching for 100/200 maple, pill, then 200/600 maple ok,
pill hung up? nope, i forgot to switch the led.
i ‘ll set it up again and then try the 5110 sketches.
stephen
On the Blue and Red Pill, the LED is between Vcc and the pin, rather than the pin and GND.
I forget If I put a define in the bootloader to compensate for this… I know I did for the “button” in the booloader, as some boards have an active High “button” and some are active low.
But if course the Red and Blue Pill don’t have a user “button”
Re:Gremlins
They must have come out if hibernation, as all of my nRF51822 boards have packed up.
Perhaps the Arduino IDE doesn’t like the year 2016 or something ( as I’m using the IDE to compile for nRF51822), but I have got all sorts of strange things going on.
The nRF51822 is a bluetooth low energy board, and it basically transmits its identity about once a second ( if you put it in the right mode).
One of my gremlins is that if i read back the flash from one module and flash into another module, the data it is sending is different.
One module sends its name as “SimpleChat” which is the name of the sketch but on another module its sending nRF51 as its name…
Very odd indeed.
Actually I suspect the Jlink programmer I’m using, ( which was working in 2015) so I will go back to the STM32 based BlackMagic Probe.
But I know for the moment the gremlins are still in my system 
and of course, all of st annes is built on a sand, although some bits are firmer than others.
stephen
and of course, all of st annes is built on a sand, although some bits are firmer than others.
stephen
did dig out a trench for a spoil pipe, sides stayed up, but all sand.
i’ve not yet got around to trying my nokia sketches alternating maple and pill.
the more i think on it, it would fit in with the all woes from those spi displays and sd cards issues.
there is the common factor of spi as well.
if i try i2c sketches with some bit of hardware, i always start with the i2c scanner so i know it’s there!
is there a spi scanner?:D is there a ‘standard’ for id codes?
stephen
Devices may be write only, in which case even if you read from the SPI you may get either 0xFF or 0x00 all the time.
multiple bytes and bits to be looked for.
thought about and looked up a few, but the ‘one library fits all displays’ libraries want you to explicitly declare
the chip, having an initialising sequence for each, not look for which the chip is.
didn’t think to look if the read the id as a check.
stephen
