#avr Logs
Aug 10 2017
#avr Calendar
03:40 AM MrFahrenheit: anyone use avrdude to read fuse sections from an elf file?
03:42 AM MrFahrenheit: is it just fuse:e:file.elf or?
03:42 AM MrFahrenheit: or rather fuse:w:file.elf:e
04:00 AM Emil: MrFahrenheit: Fuses are burtn separately
04:00 AM Emil: burnt*
04:00 AM dunz0r: MrFahrenheit: You need to read them off the chip
04:00 AM MrFahrenheit: does it recognize the location of each fuse byte in the .fuse elf section?
04:00 AM MrFahrenheit: because there's only one .fuse section
04:02 AM MrFahrenheit: never mind, I just located the elf handling part in the avrdude code and it does do that
04:03 AM MrFahrenheit: line 780 https://svn.savannah.gnu.org/viewvc/avrdude/trunk/avrdude/fileio.c?view=markup
04:03 AM MrFahrenheit: *784
04:04 AM MrFahrenheit: so I guess the command is then just -U lfuse:w:file.elf:e -U hfuse:w:file.elf:e
04:08 AM MrFahrenheit: I read somewhere that it used to expect separate .hfuse and .lfuse sections which is why I was asking, but looking at the code I don't think that was ever the case
05:14 AM Lambda_Aurigae: make sure your efuse unused bits are set to 1
05:14 AM Lambda_Aurigae: otherwise you will get verify errors.
05:15 AM MrFahrenheit: I'm gonna make two make rules, program-fuses and program-efuse
05:16 AM Lambda_Aurigae: I just do them all together myself.
05:16 AM Lambda_Aurigae: it's just that unused bits in fuses read as 1 but people often write them as 0 and get verify errors.
05:18 AM MrFahrenheit: ah, I see
05:19 AM Lambda_Aurigae: atmega328 is bad for that...I've seen many people come in here from ardweeny land and complain that they can't set fuses for just that reason.
05:20 AM Lambda_Aurigae: avrdude will reset fuses back if it can't verify,,,by default anyhow.
05:22 AM Lambda_Aurigae: "Friends are gods ways of apologizing for our families"
05:41 AM NoHitWonder__: its a shame you cant get atmega328PB anywhere
05:41 AM NoHitWonder__: so i switched to this http://www.mouser.fi/ProductDetail/STMicroelectronics/STM32F070RBT6TR/?qs=sGAEpiMZZMu3xu3GWjvQiE0ppv5LuDUhwKFrPRY5IGs%3d
05:41 AM NoHitWonder__: 4 uarts and usb
05:42 AM MrFahrenheit: that pin pitch
05:44 AM NoHitWonder__: its nice that those have 5v tolerant IOs
05:54 AM MrFahrenheit: I just finished writing my makefile, supports c, asm, simavr+gdb, avrdude, micronucleus, removes unused code and data, generates dependency files and is pretty clear and shortish
06:02 AM polprog: wow
06:02 AM polprog: nice
06:03 AM polprog: my makefile has just "all", "flash", and "debug" procedures defined :P
06:05 AM MrFahrenheit: took me a while to get code and data removal (--gc-sections) to work
06:05 AM MrFahrenheit: I spent a long time writing my own linker script before I realized I could just include the default one and it worked fine
06:10 AM polprog: heh
07:49 AM rebecc_ is now known as rebecca
09:59 AM enh: morning
10:47 AM polprog: waiting for my logic analyzer
10:47 AM polprog: saelae clone
10:48 AM MrFahrenheit: the $5 one?
10:49 AM polprog: yeah
10:49 AM MrFahrenheit: works great with sigrok's pulseview
10:50 AM polprog: that's why i bought it
10:51 AM MrFahrenheit: although mine glitches out some 20% of the time
11:29 AM polprog: well see
11:29 AM polprog: i use my scope mostly anyway, and if only pulseview could decode from analog traces, i wouldn't bother with that LA
11:29 AM polprog: ive got a 4 ch scope
11:29 AM polprog: so it's like half og the analyzer xD
11:53 AM arti_ is now known as arti
12:09 PM enh: Anybody know a way of depositing copper over graphite?
12:09 PM cehteh: electrolysis
12:10 PM enh: electroplating, you mean
12:10 PM enh: dunno if it adheres
12:10 PM cehteh: whatever the english word is yes
12:10 PM enh: electrolysis is separation
12:10 PM cehteh: nothing adheres well to graphite, but when the plating is thick enouh it shouldnt matter
12:11 PM enh: i can laser cut white paper if it is covered with graphite. My laser is very weak (808n, 500mW)
12:11 PM cehteh: matter much .. i mean do you need physical strength or whats the application`
12:12 PM enh: I 've been wondering for some time if it would be better to prepare a bed for PCB tracks andthen deposit copper over the tracks, instead of removing it
12:13 PM cehteh: thats how 'professional' pcb makers do it
12:13 PM enh: really?
12:13 PM enh: I thought they milled the PCB
12:13 PM cehteh: but doing that well needs a lot nasty chemicals and few more process steps
12:13 PM cehteh: milling? takes waay to long
12:14 PM Emil: milling is kawaii, though
12:14 PM cehteh: dunno whats state of the art, but when i learned about it they uses photoresist for positive masking, some baths to prime the raw material with a thin copper layer (no graphite, that doesnt adhere)
12:15 PM cehteh: and then electroplating up to the desired thickness
12:15 PM day_ is now known as daey
12:15 PM enh: another stuff I've been thinking of is to laser cut brown paper and then wet it with epoxi to make it harder
12:15 PM cehteh: and later maybe tinning
12:15 PM cehteh: hardpaper/pertinax :D
12:16 PM cehteh: but that uses phenol/formaldehyde resins instead epoxy
12:16 PM enh: sintering (pressing) copper powder into the track beds would also be possible. With Sn/Pb powder. Then warm all up and wait until solder fuses.
12:17 PM cehteh: now guess why no one does that?
12:17 PM cehteh: i mean the etching method is still state of the art for hobbyists for some reason
12:17 PM enh: no one does new stuff, until it is not new anymore
12:17 PM enh: i agree
12:18 PM enh: I was just searching for a simple way to make PCBs with my CNC and a small laser
12:18 PM cehteh: i dont mind the etching .. but some easy way to mask the pcb would be nice
12:18 PM cehteh: as in directly printing a mask on the pcb
12:18 PM enh: If the laser could drill the material it would be wonderfull.
12:19 PM cehteh: no toner transfer or UV exposure
12:19 PM enh: printing a mask over the pcb is easy. Mask all PCB, then remove the mask with the laser.
12:19 PM enh: I do this with black ink
12:19 PM enh: spray
12:19 PM enh: Then I can etch
12:19 PM cehteh: xerox has these wax printers would be nice if it would be possible to print masks with those
12:20 PM enh: You probably can remove wax from the top of a PCB with laser too, but it should be black wax
12:20 PM cehteh: heh you can wipe it off with some solvent
12:20 PM enh: http://gabuleu.blogspot.com.br
12:21 PM enh: I mean using the alser to draw on the mask
12:22 PM enh: Finding a diode laser with a round spot is not easy.
12:22 PM enh: it needs lenses with a fancy name
12:22 PM tpw_rules: fresnel?
12:23 PM enh: nope
12:25 PM enh: They are present into DVD heads
12:30 PM enh: Sorry, I forgot the name. did you see the laser etching results?
12:33 PM day_ is now known as daey
12:36 PM cehteh: i wonder if one could make a stamp from some conductive material (graphite block?) and then use EDM to erode a PCB
12:37 PM cehteh: https://en.wikipedia.org/wiki/File:EDMWorkpiece.jpg
12:37 PM enh: anamorphic lenses...
12:38 PM enh: cehteh: graphit is brittle. Your idea is good, even on copper.
12:38 PM enh: There are CNCs with EDMs
12:38 PM enh: but the wire etches fast too
12:39 PM cehteh: since the stamp has no contact to the PCB brittle wont matter
12:39 PM enh: The arc erodes the materials. Both sides.
12:39 PM cehteh: but ah noticed my error since the inner traces become insulated it wont work well
12:39 PM JanC is now known as Guest25300
12:39 PM JanC_ is now known as JanC
12:40 PM enh: I tried to make a machine to etch a 50um hole in a metal gasket.
12:40 PM cehteh: i'd really prefer the wax ink printing method .. if such a printer would be available (would need a modifed printer, only one color, no silicone fixating step)
12:41 PM enh: The hole gets bigger than the wire, or bents, or many other problems appear
12:41 PM cehteh: yes
12:41 PM cehteh: EDM can be very precise when you have the necessary experience and right parameters
12:41 PM cehteh: but shooting blindly will be bad
12:41 PM HighInBC: EDM all sounds the same
12:42 PM HighInBC: (Electronic Dance Music) :)
12:42 PM enh: I can try to remove wax ink from the PCB using the CNC. Hard will be to place a uniform layer of wax over it
12:42 PM cehteh: https://en.wikipedia.org/wiki/Electrical_discharge_machining
12:42 PM cehteh: enh: i meant Print the mask in the first step
12:42 PM cehteh: nothing to remove
12:43 PM cehteh: just finished mask, etching, wipe wax off
12:43 PM cehteh: print on the copper layer directly
12:43 PM enh: I believe it can be done. Something like 3d printing
12:43 PM cehteh: https://en.wikipedia.org/wiki/Solid_ink
12:44 PM cehteh: the printers exists and printouts are really nice
12:44 PM enh: Maybe could even print the mask later
12:44 PM cehteh: but i dont know if you can feed pcbs into them, they are expensive, you dont need 4 colors and you must disable the final fixating (silicone oil) step
12:45 PM enh: You will be limited to 2 layers
12:45 PM cehteh: https://www.youtube.com/watch?v=-9Vna-IQUiQ
12:45 PM cehteh: hah
12:47 PM enh: nice!
12:47 PM enh: This can be multi layered
12:48 PM cehteh: thats the idea, you get outstanding resolution for a hobbyist (no resolution loss on transfer) and it would be easy and fast
12:49 PM cehteh: but the printers are expensive
12:49 PM cehteh: and these are rather printers which want to be uses regulary, wax heatup time is as much as 30 minutes, usually you leave them on
12:50 PM cehteh: but the molten wax ages over time when you dont print much
12:50 PM enh: hum
12:50 PM cehteh: make only sense when you print a lot
12:50 PM cehteh: well some people use laser printers to print toner on PCB directly
12:51 PM cehteh: but toner has often these porous characteristic and not so clean edges
12:51 PM enh: What do you think of the way I tried? Too bad?
12:51 PM cehteh: results will never be as good as wax
12:51 PM cehteh: try it
12:51 PM cehteh: i bet it will be slow, expensive, unreliable
12:51 PM cehteh: *but* maybe you work something out
12:52 PM enh: I could make a small PCB with it
12:52 PM enh: wax would be better, indeed
12:52 PM enh: the link above shows the results
12:52 PM cehteh: and could be extremely fast, no matter of how much PCBs you want to produce
12:53 PM enh: I did this about two years ago
12:53 PM cehteh: easy for a single prototype, and still economic for 1000's of pcb's
12:53 PM enh: the spray+laser test
12:53 PM cehteh: yes works somewhat
12:53 PM cehteh: or toner transfer
12:54 PM enh: There still exists the other problems, hole metal plating, drilling, layer aligning
12:54 PM cehteh: but it involves too much steps, loosing resolution and the process is not completely reliable
12:54 PM cehteh: you can get good results one day and crap another day
12:54 PM enh: yep
12:54 PM cehteh: such a wax thing could be much much better .. would be my idea
12:54 PM cehteh: but i dunno :)
12:55 PM enh: A good device for a mekr lab
12:55 PM enh: maker
12:55 PM enh: Probably there it would have enough demand
12:55 PM enh: How would you insulate the layers?
12:57 PM enh: This is quite a nice picture: https://lh3.googleusercontent.com/-Kv7Bp8IJuvE/VXnTTQ2-dcI/AAAAAAAAvXo/duF0LLWcu7Y/Photoon6-8-15at21.21-2015-06-11-15-18.jpg
01:00 PM enh: I like the wax printer idea. Would be indeed very practical
01:05 PM cehteh: https://www.youtube.com/watch?v=aUsNaK867cE
01:11 PM enh: And this? https://www.youtube.com/watch?v=ZHjqAbnmEvw
01:11 PM enh: Both are nice. Drilling and hole plating is another problem. Bigger, I believe
01:12 PM enh: If I could machine the substrate, with holes and everything, than deposit copper, I would even be able to make 3D circuits
01:12 PM enh: and hole plating would not be a problem
01:13 PM enh: miling a 3d circuit in fiberglass layers can be done
01:14 PM enh: but how to deposit copper over it?
01:15 PM cehteh: router or anything alike is just too slow
01:15 PM cehteh: would work for prototypes
01:15 PM cehteh: but then you could mill it directly
01:20 PM enh: milling is slow, cehteh. And strongly sensitive to board flatness
01:21 PM enh: swithcing the mill for a laser is a lot faster
02:04 PM MrFahrenheit: to expose the pcb or to evaporate the copper?
05:06 PM ohsix: sup, need a not so super important 'reset'; periphs will get reinitialized anyways, just worrying about heap, i think jumping to _start is sufficient but i thought i'd get some opinions
05:11 PM ohsix: or 0x0000 which is what the bootloader seems to jump to. all i'm avoiding is the bootloader really
05:18 PM tpw_rules: yeah that's probably good enough
06:28 PM ohsix: turns out the bootloader skips the bootloader if the reset was from the watchdog so i can just do the normal thing \m/
07:29 PM Lambda_Aurigae: https://blog.heypete.com/2017/07/29/a-look-inside-the-ds3231-real-time-clock/
07:29 PM Lambda_Aurigae: interesting clock chip.
08:52 PM enh: hi
09:59 PM Emil_ is now known as Guest11632