arm-netbook October 2015

arm-netbook@lists.phcomp.co.uk

8 participants
9 discussions

adafruit spi lcd hx8357d and arduino uno
by Luke Kenneth Casson Leighton 29 Oct '15

29 Oct '15

.... arrived today, and were up and running, working perfectly, in under 15 minutes: http://git.rhombus-tech.net/?p=eoma-firmware.git;a=tree;f=eoma68/arduino_un… i used a makefile that was easily found with a google search "debian arduino uno makefile", followed the instructions _in_ the makefile, created a blink test, ran it, ran "make upload" and it worked. total time: well under 3 minutes. next was the SPI-based LCD, soldering took a few minutes, copying the library files into the main directory a few more - done. i've now modified the adafruit source code so that it can do 9-bit SPI bit-banging, by setting "IM1=0" (shorting that to ground) and i'm now in a position to begin two-way testing of the adafruit LCD with an STM32F072, and the yunlea HX8357D LCD with the arduino uno. this is a heck of a lot quicker than messing about with both unknown hardware and unknown software. l.

2 2

laptop 15in pcb3
by Luke Kenneth Casson Leighton 27 Oct '15

27 Oct '15

i'm on to assembly of pcb3, now, which is the power / charger board. pcb assembly whilst listening to ozric tentacles is a thoroughly recommended experience btw :) this one is slightly fiddly because it's not a "prototype" PCB, so is not covered with as much solder (on the pads). things aren't properly sticking unless i "pre-tin" them. the LTC4155 QFN i wicked its pads across a soldering iron and that has gone down fine: 0402 resistors however are _really_ acting up. next time i will go round every single pad with a soldering iron, blobbing them before even considering putting components on. regarding PCB2: the SPI interface to the LCD, i've given up on the current approach, because of the two "unknowns" - unknown hardware and unknown software - and have asked the sponsor to obtain an adafruit HX8357D LCD and an arduino. i will then be able to work from two "known-goods", as a way to triage which (or if both) of the two "unknowns" is "not working". there are slight differences between the yunlea LCD and the adafruid LCD, which i will be able to compensate for by: (a) getting the adafruit LCD working with the STM32F072 and (b) getting the arduino to talk to the yunlea LCD. having these successfully cross-talk i will then be in a position to finally put the yunlea LCD onto the STM32F072. that's the plan, anyway. l.

1 0

SPI stress!
by Luke Kenneth Casson Leighton 20 Oct '15

20 Oct '15

RRRRRR :) i'm attempting to get the yunlea spi lcd with an hx8357 up and running. i'm resorting to bit-banging the spi at the moment because it needs 9-bit initial "control" signal followed by 8-bit "parameters".... but once again just like with the I2C i have "unknown hardware" combined with "unknown software". sooo.... i think what i will do is try a loop-back SPI software test. connect SPI1 to SPI2 on the STM32F072, and put SPI2 into slave mode and SPI1 into master.... then see if i can get them to talk to each other. *sigh*... :) l.

2 3

"recycling not the answer"
by Luke Kenneth Casson Leighton 19 Oct '15

19 Oct '15

http://www.lowtechmagazine.com/2009/06/embodied-energy-of-digital-technolog… if the energy cost of the plastics injection molding is enough to power a 30 watt laptop for 1,000 days (3 years), then clearly, recycling of the plastic, doing exactly the same thing, is *not* doing the environment any favours. the answer is, (a) instead, to keep that plastic casework in use for as long as possible (for example by allowing the main processor and memory to be continuously upgraded on a rolling basis) and/or (b) develop different casework production techniques that are far more energy-efficient. l.

2 4

stm32f072 usb firmware
by Luke Kenneth Casson Leighton 18 Oct '15

18 Oct '15

hi peter, et al, i managed to find a variant of libopencm3 which has usb support for the stm32f072, i compiled the cdcacm example and it worked. i will write up some notes later but here's a start: https://github.com/kuldeepdhaka/libopencm3/tree/usb-rewrite https://github.com/kuldeepdhaka/libopencm3-examples/tree/usb-rewrite http://hands.com/~lkcl/usb_req_done.patch peter i sent phil the ssh key, please do consider joining the mailing list so that things can be coordinated without cc. the above branch is... well... it's literally one of the first times that libopencm3 has run usb code on an stm32f072 nucleo board. it means that certain things aren't going to be there. with two of us, peter, doing testing and development, the libopencm3 team on irc.freenode.net will have the encouragement they need to move things forward: we are, however, going to have to help them out, they're pretty busy. the similarity to the stm32f1 series is sufficiently strong that i have confidence in their work, which means we can both press ahead with firmware development whilst the usb code is moved mainline and stabilised with unit testing. l.

3 7

Open sourced A13 Kicad design mail server
by joem 18 Oct '15

18 Oct '15

Open sourced A13 Kicad design mail server https://www.own-mailbox.com/libre.html Found them on kickstarter: https://www.kickstarter.com/projects/1547898916/own-mailbox-the-first-100-c… Seems these guys have done some real hard work. As well release their design on KiCAD, they also made a distro that runs on A13 which has a modified mail server that accepts mail replies and posts it at secure https links to standard email addresses which don't offer security built in. The KiCAD designs are missing some libs due to the way the paths have been hard coded into the project files. I'm sure they will get around to repairing it. Its a more powerful design than a raspberry pi, but many times smaller.

3 7

linus torvalds says "2016 is the year of the arm laptop"
by Luke Kenneth Casson Leighton 15 Oct '15

15 Oct '15

http://linux.slashdot.org/story/15/10/09/1838249/linus-2016-will-be-the-yea… surpriiiise!

3 12

laptop_15in working PCB2!
by Luke Kenneth Casson Leighton 05 Oct '15

05 Oct '15

yippee, i tested the PCB2 i put together yesterday, and successfully uploaded a test program to it, turning it into an example USB-HID "mouse" that generates events, wobbling the cursor back and forwards. over the past few days i found someone who coincidentally maintains a branch with USB support for the exact STM32F model ARM Cortex M0 that i chose, and i got a working "usb hid" example up and running on a demo board i have, here ($EUR 10 retail. amazing). yesterday i populated PCB2 with components, i had got the STM32F072 the wrong way round, burned one out! used the IR solderer to remove it and stuck a new one down, this time the right way round :) then, with a lot of fiddling i managed to get the right connections to reset and put it into "DFU" - device firmware upload - mode. then i just uploaded the firmware tested previously on the demo board, and yay it worked. so i will do a bit more firmware programming, there is the keyboard matrix to do, I2C interface for the touchpanel (converting to real USB-HID events), and also there is the SPI interface for the 480x320 LCD. testing the LCD and capacitive panel for the first time will be interesting. l.

1 1

laptop and microdesktop pcbs
by Luke Kenneth Casson Leighton 03 Oct '15

03 Oct '15

ok so: after finding that there's someone who is maintaining an STM32F072 USB branch, and testing it successfully to create both a usb-hid and usb-cdcacm example i put the majority of the components on the laptop STM32F072 PCB, with the exception of the 3.3v regulator which isn't here yet. i have a source of 3.3v power so that can be tested later. assembly of the much smaller STM32F072 PCB was only a day, with two uses of the IRDA lamp. i did however have to go round all of the 0.5mm pitch QFP and FFC pins individually with a soldering iron. surprisingly, QFNs tend to stick first time. i've also ordered the laptop PCB3 which is the power board, that one is going to be fun and games, especially with the EMF from the rather large (4A) regulator. i've put in a huge number of GND vias surrounding the high-frequency power output which goes to the inductor, we'll just have to see what happens. also i had redesigned the microdesktop board some weeks back, and will be de-populating one of the existing boards and transferring all the components. this i anticipate to be pretty straightforward, although perplexingly the microdesktop board has a staggering 120 components. i don't quite understand why, given that it's 4in x 3.5in and that the CPU Card itself is around 3.5in x 2in! still, there we go... :) l.

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

arm-netbook October 2015