#avr Logs
Apr 02 2017
#avr Calendar
12:24 AM daey_ is now known as daey
03:45 AM Emil: polprog: did you assemble the board already?
04:28 AM polprog: Emil: i have a ton of schoolwork to do, i hope i can etch it today
04:32 AM Emil: polprog: pfft
04:32 AM Emil: disregard schoolwork, acquire electronics
04:33 AM polprog: oh id love to, but i have a 2 month backlog of homework
04:34 AM polprog: maybe ill scavenge another old laptop battery i have, some dell stuff
04:34 AM polprog: or fix the old thinkpad
04:34 AM polprog: its only 1100 here
04:34 AM polprog: a lot of time
04:36 AM polprog: also the parts are probably coming tomorrow
04:51 AM Emil: etch today, it takes an hour, assemble tomorrow, it takes 2x or more
06:17 AM Jartza: my PCBs arrive tomorrow, it seems
06:18 AM Jartza: hopefully soldering doesn't take too long
06:51 AM polprog: i've made an alternate design with 1206 components and .5 mm traces (intead of .25mm), without a ground fill, and i'll try to etch it today
06:51 AM polprog: Jartza: which project are those boards?
07:11 AM Emil: Jartza: what pcbs?
07:11 AM Emil: Share picture
07:11 AM Emil: polprog: wtf
07:11 AM Emil: polprog: _ALWAYS_ have a ground fill
07:12 AM Emil: polprog: also yeah 1206 is easier route things under
07:12 AM Emil: and 0.5mm traces are definitely much easier to make
07:12 AM Emil: polprog: tell me you take this to heart
07:13 AM Emil: akways have a ground plane
07:15 AM Emil: (except when rf design, that's different)
07:30 AM polprog: you say? ok, i'll redesign
07:30 AM polprog: always ground plane
07:31 AM polprog: semper gnd
07:31 AM Emil: always ground plane
07:32 AM Emil: semper ground :D
07:33 AM polprog: gtg
07:35 AM Emil: polprog: a) groundplanes provide better grounding b) it's easier to diy and takes less time c) they reduce spurious transmissions (at some cost with increased (minimal) capacitance) and your routing is easier.
07:41 AM polprog: oh
07:41 AM polprog: i though ground plane would be more complicated to do
07:41 AM Emil: nooo
07:41 AM Emil: mate wtf :D
07:42 AM polprog: ive never etched a pcb
07:42 AM polprog: ive only watched some tutorials on youtube and some polish ee forums
07:43 AM Emil: you just route your signals and power and then flood fill the rest for ground planes. For diy etching if you use ground planes etching a) takes less chemicals because you need to etch less away and b) it also takes a lot less time, too
07:43 AM polprog: that makes sense... :D
07:44 AM polprog: thanks. ive learned so much in this channel
07:46 AM polprog: but he actually did register
07:46 AM polprog: wrong channel
07:48 AM Emil: hmm
07:48 AM Emil: I should do a pcb etching video tutorial
07:49 AM Emil: the whole process takes ~30 minutes with fresh chemicals
07:52 AM Emil: actually ~20
08:22 AM polprog: maybe
08:23 AM polprog: if so, in english :D
08:23 AM polprog: anyway, what do you think? https://puu.sh/v6Luf/815d9fae90.png
08:23 AM Emil: polprog: finglish ;)
08:23 AM Emil: no my English is pretty good
08:23 AM polprog: some of the word in finnish i understand, i had german in school
08:23 AM Emil: polprog: what's the clearance?
08:24 AM polprog: .25 on logic and .2 on power, whoops
08:24 AM Emil: Well, we do have some German loanwords but Finnish!=German :D
08:24 AM Emil: polprog: if you can, also increase clearance to 0.5
08:24 AM Emil: it's so much easier to diy
08:24 AM Emil: but 0.25 works fine, also
08:25 AM Emil: polprog: you also want to have small holes for the drilled parhs
08:25 AM polprog: 0.5 might be problematic with 1206... 0.3 was too much
08:25 AM Emil: so that it's easier to center where to drill
08:26 AM polprog: like small holes in copper?
08:26 AM Emil: polprog: 0.25 clearance is fine, as long as traces are thicker
08:26 AM Emil: polprog: yeah holes
08:26 AM Emil: a small hole on the copper where to drill
08:26 AM polprog: im not sure if kicad can do this
08:26 AM Emil: or no copper at all, preferably, where you drill
08:26 AM Emil: polprog: it can
08:27 AM Emil: without any issues
08:27 AM polprog: how :D ?
08:27 AM Emil: also mate wtf, extend those mosfet pads over the fins :D
08:27 AM Emil: dont leave them under the part
08:28 AM polprog: i dont get it
08:28 AM Emil: secondly no mounting hole, so add one if you feel like it
08:28 AM polprog: i'll add the mounting hole after the process
08:28 AM polprog: you mean the g1, g2 etc traces should not go under chips?
08:29 AM Emil: polprog: ypu'll want to really etch the copper away from your holes, first
08:29 AM polprog: ok, ill put a hole there
08:29 AM Emil: polprog: yeah, the traces should not be left under the component
08:30 AM Emil: you can route them under of course
08:30 AM Emil: but have big enough pads that you can soldee on them
08:30 AM Emil: so that the fins dont cover the whole pad
08:31 AM polprog: oh, you mean those "fins" :P
08:31 AM Emil: also turn on the pad filling so that you can see the actual pads, also, not just the outline
08:32 AM Emil: nice use of the 0 ohm resistor, also
08:32 AM polprog: but the footprint should be ok, i managed to hand solder tqfp to a regular footprint
08:32 AM Emil: alrighty
08:33 AM Emil: then it's fine
08:33 AM polprog: i couldn't get the ground to the logic part of the pin header, then the 0 resistor came to my mind :)
08:35 AM Emil: it seems to me your ground plane clearance is higher than your trace clearance
08:35 AM polprog: let me check
08:36 AM polprog: it's .5mm
08:36 AM polprog: the think traces have .2
08:36 AM polprog: thin*
08:36 AM Emil: polprog: if your overall clearance is 0.25 just use that everywhere
08:36 AM Emil: polprog: use 0.25
08:37 AM Emil: with home etching it makes a difference
08:37 AM polprog: looks better now
08:37 AM polprog: so the general rule of home etch design is to remove as small copper as possible?
08:38 AM Emil: as little as possible, yes
08:38 AM Emil: it also doesn't hurt even if you have them fabficated for you in a shop
08:38 AM polprog: yeah, probably
08:39 AM polprog: what is the probablility of a total fail when etching? like, what's the most difficult part of the process?
08:39 AM specing: not scraping off the mask
08:40 AM specing: with your fingers
08:40 AM specing: happened to me when making tqfp board
08:40 AM Emil: polprog: it's super easy. If your printer is running out of ink that's pretty much the only part that hurts
08:40 AM Emil: specing: eh
08:40 AM Emil: use acetone :D
08:40 AM polprog: i'll be using the toner transfer method
08:40 AM specing: I was using laser printer -> transparent sheet -> UV -> etching
08:41 AM Emil: toner transfer is nice because you can just use pure copper sheets
08:41 AM Emil: but uv is superior
08:42 AM Lambda_Aurigae: uv is a bit pricier
08:42 AM polprog: i think that as long as it's not in the chemical bath i can wash the toner off and try again
08:42 AM Emil: Lambda_Aurigae: true
08:42 AM Lambda_Aurigae: I use press-n-peel blue sheets on straight copper..toner transfer.
08:42 AM Lambda_Aurigae: different brands of printers have different fill properties too.
08:42 AM Tom_L: i don't like those as well as photo paper
08:43 AM Tom_L: but it takes a while to find the 'right' photo paper
08:43 AM Lambda_Aurigae: yup.
08:44 AM Lambda_Aurigae: had almost as good luck with old glossy magazine
08:44 AM polprog: yeah, i need to get glossy paper
08:44 AM Lambda_Aurigae: never really had good luck with photo paper.
08:44 AM Tom_L: best i found was from Austrailia
08:45 AM Lambda_Aurigae: the press-n-peel gets a bit pricy too but I bought a couple hundred of sheets some years back on sale for 0.10 USD per sheet
08:45 AM Emil: Print transparent sheet (or actually even normal paper works but with longer activation time), activate the protection chemical under uv light, wash it away in naoh and etch in natrium sulfate in a bubble machine
08:45 AM Lambda_Aurigae: have more press-n-peel than I do copper clad.
08:45 AM Emil: whole process takes 20 minutes and produces good quality diy pcbs
08:45 AM Tom_L: some transparencies tend to let go of the toner too soon
08:45 AM Emil: making two sided boards that way is also easy
08:46 AM Lambda_Aurigae: Emil, mine takes about the same with ferric chloride or ammonium nitrate etchant.
08:46 AM Emil: Lambda_Aurigae: yeah you can change the etchant of course
08:46 AM Lambda_Aurigae: and I have a big t-shirt press for doing the iron on process...it heats from both sides so it works for 2-sided boards.
08:46 AM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/etching/etch6.jpg
08:46 AM Tom_L: i think those are pretty clean
08:46 AM polprog: they look retro
08:47 AM polprog: :)
08:47 AM Lambda_Aurigae: I just print 2 sheets with alignment marks, tape them together, then slide the board in the middle and iron.
08:47 AM polprog: nice
08:47 AM Emil: Tom_L: not very good quality ;)
08:47 AM Emil: also
08:47 AM Emil: >no ground plane
08:47 AM polprog: boo
08:47 AM Emil: into the trash it goes
08:47 AM Emil: ;D
08:48 AM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/temp/tiny/tinyboard1.jpg
08:48 AM Tom_L: probably most proud of thsoe
08:48 AM Tom_L: tiny 10 with 3 traces under it
08:49 AM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/temp/tiny/tinyTPI1.jpg
08:49 AM polprog: wow
08:50 AM polprog: which pb software allows you to use custom fonts?
08:50 AM polprog: kicad has one single font
08:50 AM Emil: polprog: kicad also allows that imho
08:50 AM Tom_L: dunno, i've always used eagle
08:50 AM Emil: eagle is cancer
08:50 AM Tom_L: nowdays it is
08:50 AM polprog: i tried eagle but i got so pissed at the size limit back then
08:50 AM polprog: i closed eagle tutorial playlist and opened kicad
08:50 AM Tom_L: mine has no limits :D
08:51 AM polprog: student, bought licence or torrent?
08:51 AM Tom_L: yes
08:51 AM Tom_L: i purchased an older version
08:52 AM Tom_L: i also sold products with it
08:53 AM polprog: i think i might design something to sell one day
08:53 AM Tom_L: i try to make all my hobbies cost effective
08:54 AM polprog: i like this phrase
08:59 AM Emil: polprog: stick with kicad
08:59 AM Emil: https://goo.gl/photos/45ta2bBBXmPCrGsU6
08:59 AM Emil: here's a diu one with 0.25mm clearance and trace width
08:59 AM Emil: https://goo.gl/photos/hnthX64313ydxAiR7
09:00 AM polprog: sure i will stick with kicad, absolutely not because im a cheapskate
09:00 AM polprog: ;)
09:00 AM Emil: You can probably tell I had some bright ideas while making it :D
09:00 AM polprog: what's on that board?
09:00 AM Emil: "hey if I flip this pinout I can route it easier!"
09:00 AM polprog: what's it for*
09:00 AM Emil: "oh yeah oops"
09:01 AM Tom_L: those look pretty good except you didn't use enough flux
09:01 AM Emil: It's atmega328pb, nrf24l01+ and some other things
09:01 AM Tom_L: solder blobbed up on ya
09:02 AM Emil: Tom_L: I dont usually use any extra flux but what is on the solder itself, yeah it blobbs if you dont solder fast enough :/
09:02 AM polprog: Tom_L: i'm lurking through your site, i have the same tiny atom PC somewhere ;P
09:02 AM Emil: Tom_L: but thanks
09:02 AM Tom_L: i always use a flux pen on smt
09:02 AM Tom_L: it helps
09:02 AM Emil: It was designed and fabricated in a single night :D
09:02 AM Emil: but then I didnt have the time to program it ":D"
09:03 AM polprog: ive walked the whole Warsaw looking for a flux pen
09:03 AM Tom_L: polprog it's in a box now
09:03 AM polprog: it was way before i started ordering stuff
09:04 AM Emil: https://goo.gl/photos/kpAB4nkUpu32CKqPA
09:04 AM Emil: here's a 3d view
09:04 AM Emil: You dont need flux if you are good enough
09:04 AM Emil: per se
09:04 AM Emil: extra flux*
09:05 AM Emil: There's flux on the solder itself
09:05 AM Emil: but then you have to solder well and quickly
09:07 AM Emil: But on the pics you can see I opted to tin some of the traces because 0.25mm is at the limit of how big your traces can be
09:07 AM polprog: here's my last big project, it probably has some problems like trace width and stuff, but it works and im pround of it https://puu.sh/v6Nqp/9d38d34693.png
09:07 AM Emil: For clearance it's fine
09:07 AM polprog: the only assembled version is outside the house now
09:08 AM Emil: Lambda_Aurigae: mate push your updates to the avrdude and libc upstream
09:08 AM polprog: huge fail afterwards, optocouplers cant take much juice through
09:08 AM Emil: it's annoying not having the support for pb automatically
09:08 AM Lambda_Aurigae: Emil, I don't have updates for avrdude or libc.
09:09 AM Emil: Lambda_Aurigae: you have the configs for pb, though
09:09 AM Lambda_Aurigae: and why would you support Lead(pb)?
09:09 AM Emil: Lambda_Aurigae: atmega328pb
09:09 AM Lambda_Aurigae: I don't have configs for that
09:09 AM Lambda_Aurigae: never had one of those
09:09 AM Lambda_Aurigae: don't use the 328
09:09 AM Lambda_Aurigae: of any variant
09:09 AM Emil: oh was it LeoNerd then
09:09 AM Emil: interesting
09:10 AM Emil: polprog: what is that project?
09:10 AM Emil: polprog: also opto couplers should really be only used as signaling
09:10 AM Emil: when you need isolation
09:16 AM Emil: polprog: oh and that project is an rc controller
09:16 AM Emil: polprog: for robowars
09:22 AM Lambda_Aurigae: robowars need to be autonomous dangit!
09:23 AM Emil: Lambda_Aurigae: if only :D
09:23 AM Lambda_Aurigae: that's just the way it should be
09:23 AM Lambda_Aurigae: otherwise it's waldo-wars
09:26 AM Emil: where does that come from? :D
09:28 AM Lambda_Aurigae: humor, attempt at, one each.
09:28 AM Lambda_Aurigae: but a robot,,by definition,,,a machine capable of carrying out a complex series of actions automatically, especially one programmable by a computer.
09:29 AM Emil: no but explain the waldo-war psrt
09:29 AM Lambda_Aurigae: waldo is a remotely operated device
09:29 AM Emil: never heard of such definition
09:29 AM Lambda_Aurigae: a remote manipulator, as for puppets, operated either mechanically or electronically.
09:29 AM Emil: waldo is the red white striped goofy guy you try to find in books
09:30 AM Lambda_Aurigae: https://en.wikipedia.org/wiki/Remote_manipulator
09:30 AM Lambda_Aurigae: the name was coined by Robert Heinlein
09:31 AM Lambda_Aurigae: most of what people call robots are actually remote manipulators or "waldos"
09:31 AM Lambda_Aurigae: a robot is supposed to be autonomous.
09:32 AM Lambda_Aurigae: like the difference between missile and rocket...many of what we call rockets are actually missiles.
09:33 AM Lambda_Aurigae: but that depends on which era of definitions one uses.
09:36 AM polprog: Emil: yeah, i designed that a bit of time ago
09:37 AM polprog: i'd have used mosfets today but this is good for switching relays
09:37 AM Emil: polprog: it's a relay control box?
09:37 AM Emil: polprog: you used opto couplers to drive relays directly? :D
09:39 AM polprog: i didn't know stuff i do now
09:39 AM polprog: it worked
09:39 AM polprog: it can surely be done better
11:28 AM rue_bed: Lambda_Aurigae, unfortunatly, 'waldo' isn't an internet-searchable term
11:32 AM JanC is now known as Guest10783
11:32 AM JanC_ is now known as JanC
11:34 AM Lambda_Aurigae: well, it is searchable
11:35 AM Lambda_Aurigae: but, you get where's waldo crap
11:37 AM Chillum: like trying to pirate that movie called "xxx", brilliant anti-piracy method'
11:37 AM Chillum: good thing it was a terrible movie
11:37 AM Lambda_Aurigae: the first one was ok.
11:37 AM Lambda_Aurigae: the second one was not so much
11:37 AM Lambda_Aurigae: it wasn't James Bond with Connery, but,,,
11:38 AM Lambda_Aurigae: for a Vin movie it was ok.
11:40 AM Chillum: I agree with that last statement
11:40 AM Chillum: by that standard it was ok
11:41 AM Lambda_Aurigae: not something I would want to pay to watch
11:41 AM Chillum: his movies tend to be overly fast and/or furious for my taste
11:41 AM Lambda_Aurigae: which, I didn't
02:30 PM polprog: not tching today :( no toner in printer, i'll have to get it printed tomorrow
02:51 PM specing: polprog: acqiure more toner by sacrificing a goat to the printer gods
02:51 PM Emil: polprog: invest in uv
02:52 PM specing: Get an UV flashlight?
02:53 PM polprog: i think if this bare pcb etching is a success i might do uv
02:54 PM polprog: but on the other hand i don't feel like going out to get prints on foil
02:54 PM Emil: eh?
02:54 PM Emil: prints on foil wat?
02:55 PM polprog: for uv you need to have your stuff imprinted on transparent foil, like the slided back then
02:55 PM polprog: right?
02:55 PM Emil: wat
02:55 PM Emil: wat
02:55 PM Emil: wat
02:55 PM polprog: at least that's how ive seen it done
02:55 PM Emil: wat
02:55 PM Emil: wat
02:55 PM Emil: wat
02:57 PM specing: Emil: 6 wats? That is not very bright!
02:58 PM Emil: 1) Print your circuit on a normal transparent laser sheet 2) Actiave the pcb under uv light 3) Wash away activated chemical in NaOH 4) Etch in your solution of choice, I use Natrium Sulphate 5) ... 6) Profit
02:59 PM Emil: s/chemical/protective film
02:59 PM polprog: transparetn laser sheet? so not regular paper
02:59 PM polprog: right?
02:59 PM Emil: You can also use regular paper
02:59 PM Emil: Just takes longer to activate
02:59 PM polprog: i mean the material on this photo
02:59 PM polprog: https://technick.net/img/guide_pcb/guide_pcb_artwork.jpg
02:59 PM specing: TIL transparent laser sheet is not "foil"
03:00 PM Emil: this is foil: https://www.google.fi/search?q=foil&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjZzIeTwYbTAhWBE5oKHb-bDigQ_AUICCgB&biw=1440&bih=831
03:00 PM specing: polprog: also an UV flashlight might be $25, but having the ability to see human waste everywhere is priceless!
03:00 PM Emil: http://crisissurvivortips.com/wp-content/uploads/2014/03/foil.jpg
03:00 PM specing: thats an alu foil
03:01 PM Emil: sheet!=foil
03:01 PM Emil: film you could use
03:01 PM polprog: well, that means i used a wrong term
03:01 PM polprog: i means a film
03:01 PM specing: damn
03:01 PM * specing updates vocabulary
03:01 PM polprog: s/means/meant/
03:02 PM Emil: polprog: it's easily acquired and cheap
03:02 PM Emil: You can print them at your own home
03:02 PM specing: Ok, I need an UV flashlight.
03:03 PM polprog: ill try the toner method first
03:03 PM specing: s/need/want/
03:03 PM polprog: i don't have space in my room for a lightbox
03:03 PM polprog: and i probably cant be bothered to make a uv flashlight
03:10 PM specing: put it on your ceiling
03:10 PM specing: you can buy an UV flashlight
03:59 PM Jan-: what are the DDBx registers for?
03:59 PM Jan-: er, not registers.
03:59 PM Jan-: defines. values. whatever.
03:59 PM Emil: eh?
04:00 PM * Jan- is confusing herself
04:00 PM Jan-: for some reason I put #define NRF_CSN_PIN DDB3
04:00 PM Jan-: I can't even remember what a DDB3 is
04:00 PM Emil: No such thing
04:00 PM Jan-: PB3 will always be the third bit of PORTB, right
04:01 PM Emil: fourth
04:01 PM Emil: and not necessarily, but pretty much _always_
04:01 PM Jan-: well okay, so I may as well define NRF_CSN_PIN 3
04:01 PM Emil: Jan-: bits are zero indexed, remember that
04:02 PM Jan-: yeah.
04:02 PM Jan-: Basically I have my chip-enable line plugged into D13 on an arduino nano, which is PB5
04:03 PM Jan-: so if I set NRF_CE_PIN to 5 and NRF_PORT to PORTB, I can then enable it with NRF_PORT |= 1<<NRF_CE_PIN
04:03 PM Jan-: ...I think.
04:03 PM * Jan- is not an expert
04:03 PM Emil: yeah
04:03 PM Emil: Exactly so : )
04:04 PM Jan-: also, the nrf24l01 chip is a nightmare :(
04:04 PM * Jan- sobs
04:04 PM Emil: nrf24l01+?
04:04 PM Emil: But yeah it's quite the pickle to get started wit
04:04 PM Jan-: I'm not sure. I think these are the basic ones.
04:04 PM Emil: probably the + one
04:04 PM Emil: they don't really make the older one anymore
04:04 PM Jan-: well I've had these ages.
04:04 PM Emil: ah
04:05 PM Emil: well the registers are the same iirc, just some changes to the plus one (some improved features)
04:05 PM Jan-: almost exactly the same and not so you would get major incompatibility as far as I can tell
04:08 PM Jan-: I'm starting to doubt the hardware, but it's *never* the hardware :(
04:09 PM Jan-: if anyone is clueful about the nrf24l01, I'd be grateful for any help
04:09 PM Jan-: I'm on evening 4 and I'm getting nowhere.
04:41 PM Jan-: ooh.
04:42 PM Jan-: I think I fixed it.
05:01 PM * Tom_L smacks Jan- with a clue bat
05:02 PM Jan-: what?!
05:02 PM Jan-: this is all your fault, anyway
05:02 PM Jan-: you sold me that AVR programmer
05:02 PM Jan-: (whichiamnotusing) but anyway it STILL IS!
05:06 PM Emil: wat
05:10 PM Jan-: I own one of Tom_L's AVR programmers.
05:10 PM Jan-: He's an expert, he makes tools :D
05:47 PM polprog: weird, i did some testing of that old thinkpad and it seems that the display works from time to time only, like 1 in 20 boots it kicks in...
06:40 PM T_X: hi! I'm currently trying to program an ATTiny861A via avrdude. unfortunately, unsuccessful so far, avrdude says:
06:41 PM T_X: "error: selected device is not supported by programmer" -> Good morning ladies and gentleman! and anyone in between or outside!
06:41 PM T_X: The truth is out there. And within you!
06:41 PM T_X: ups
06:42 PM T_X: (that copy & pasted something weird from the paste buffer - sorry)
06:42 PM T_X: "error: selected device is not supported by programmer" -> https://paste.debian.net/925648/
06:43 PM T_X: I get the same error message with those two programmers (mySmartUSB MK2 and ICprog-AVR 2.0)
06:48 PM Lambda_Aurigae: so,
06:48 PM Lambda_Aurigae: attiny841...what programmer does it require.
06:51 PM Lambda_Aurigae: looks like any simple ISP programmer should work.
06:51 PM Lambda_Aurigae: built an stk200 clone.
06:58 PM Lambda_Aurigae: T_X, looks like you need to get one of the general purpose programmers rather than the ones that have chip specific programming information on the programmer.
07:16 PM Emil: Jan-: wait, you are programming what with wat
07:54 PM T_X: Lambda_Aurigae: hm, okay. I think a friend had used at least the ICprog-AVR 2.0 with the exact same ATTiny861A. I think he had just used a command with avrdude to pretend that its a different Atmel chip
07:54 PM Lambda_Aurigae: oh.
07:54 PM Lambda_Aurigae: have fun.
07:54 PM T_X: but I currently don't know/rememeber the according avrdude command :D. ok, I'll ask him again then when I next see him :)
07:55 PM Lambda_Aurigae: no clue as that's not any kind of "standard" programmer
07:55 PM T_X: anyway, thanks Lambda_Aurigae!
07:55 PM T_X: ok