Topic for #qi-hardware is now Copyleft hardware - http://qi-hardware.com | hardware hackers join here to discuss Ben NanoNote, atben / atusb 802.15.4 wireless, and other community driven hw projects | public logging at http://en.qi-hardware.com/irclogs
<DocScrutinizer>
(t500) well it's just good enough for me
<DocScrutinizer>
better old solid stuff than the newest hottest crap
<wolfspraul>
man that's some dose of tech pessimism here
<wolfspraul>
luckily I think most new tech is actually *far* better than the old one, so I'm not worried :-)
<DocScrutinizer>
you won't get a quality like T500 for 600 bucks new
<DocScrutinizer>
and actually it's terribly hard to find any decent anti-glare non-glossy display in a laptop nowadays
<wolfspraul>
computer prices are dropping and will continue to do so, driven by semiconductor economics as you know
<wolfspraul>
I would agree there is scary little innovation in notebooks, but we all know this for 10 years, no need complaining.
<wolfspraul>
let's do better
<wolfspraul>
hey btw, since we have the pessimists meeting here - will nokia go out of business?
<wolfspraul>
:-)
<DocScrutinizer>
who's complaining?
<DocScrutinizer>
I just stated I prefer used high quality 2 year old stuff to new stuff that's inferior quality for 4 times the price
<wpwrak>
(taking diy here) ok :)
<wolfspraul>
wpwrak: but watch out we have the pessimists coming after us :-)
<wolfspraul>
for that - a sip of nice yunnan coffee...
<wolfspraul>
DocScrutinizer: I think I bought my last notebook, took it apart after buy and reassembled in new case, which was fun
<wolfspraul>
but then I switched to a screen with back-mounted thin-itx board and battery and yes, it works just fine
<DocScrutinizer>
wpwrak: wolfgang confusing sane pragmatism with pessimism
<wolfspraul>
:-)
<wolfspraul>
wpwrak: do you have any explanation why whitequark described his pcb-making in such dark words (months of fiddling, unstable, messy, eventually gave up), whereas what I saw at your place looked quite charming actually?
<wolfspraul>
what would help me if I would pick a realistic goal first - maybe atben clone?
<wolfspraul>
or you think that's too difficult
<wolfspraul>
DocScrutinizer: what's your take on the future of nokia?
<wpwrak>
whitequark: i usually get my boards right one the first try :) they're small, though. i try keep things simple ;-) it only takes a lot of attempts if i try to rush things. haste makes waste, quite literally in this case ;-)
<DocScrutinizer>
DOOOOOOM
<wolfspraul>
I somehow think in the downfall there maybe a chance of a spinoff that could continue with a Linux phone
<wpwrak>
(get right) in the sense of the PCB being what i had in the design. then i discover the flaws in the design .. ;-)
<wolfspraul>
but I'd say only 10% chance of that happening, more likely it will all be grinded down to nothing
<DocScrutinizer>
(spinoff) hard to see this perspective. I think they already fired all inhouse competence wrt linux/FOSS
<DocScrutinizer>
I expect Nokia becoming a M$-dependency
<wpwrak>
whitequark: (scaling) in fact, it scales negatively. if you make a lot of boards, the error rate goes up :)
<wolfspraul>
of course, for scaling you send to a fab
<wolfspraul>
and for cost-down
<wolfspraul>
although I also think there are a lot of interesting pcbs with unique features a standardized fab cannot easily do, but that needs to be seen case by case
<DocScrutinizer>
hm?
<DocScrutinizer>
like what?
<wolfspraul>
I remember a long discussion with Andrey the Elphel founder about how he planned to make some special cuts into pcbs to be able to perfectly aline an image sensor with the lens
<wpwrak>
(nokia) why the future tense ? ;-)
<wolfspraul>
and how he was unable to find any pcb maker willing to work with him on that, for months, until he eventually gave up
<DocScrutinizer>
wpwrak: :-P
<DocScrutinizer>
indeed
<wolfspraul>
wpwrak: please help! :-) which initial goal? atben clone?
<wolfspraul>
something else that is more interesting now?
<wolfspraul>
labsw?
<DocScrutinizer>
yeah, sounds good
<DocScrutinizer>
not too much too fine pitched stuff I assume
<wpwrak>
wolfspraul: (atben clone) it's a little demanding on the ME side. that uSD shape ... depends a bit on how confident you are with a dremel
<wolfspraul>
I don't even have a dremel. need to get a number of tools.
<DocScrutinizer>
duh, so I mistaken sth
<wolfspraul>
my drilling machine broke yesterday because I abused it in cutting a line :-)
<wolfspraul>
well, in good chinese quality it didn't totally break, but at high speeds it sputters (stops for a few seconds). middle speeds still work :-)
<wpwrak>
wolfspraul: i would pick something simpler. than atben/atusb. why not design your own circuit ? we have plenty of parts. i'd make something with an MCU, two LEDs, and a button. add USB if you want fancy.
<wolfspraul>
nah I can only motivate myself for something I actually need
<wolfspraul>
how about labsw?
<wolfspraul>
you dismissed anything with the xc6slx9 fpga as too difficult initially
<wpwrak>
wolfspraul: RF has one nasty property: if you want it to work well, you need make sure grounding is good. now, good grounding makes the soldering harder, too, because all those large ground areas near your chips act as excellent heat traps
<wolfspraul>
an fpga/atrf board would be nice
<wolfspraul>
how about labsw?
<wolfspraul>
that could come in handy
<wpwrak>
labsw is good but also a bit large. more like a 2nd stage project :)
<wpwrak>
for the first try, the goal should be to make something where you can learn quickly. extra complexity only slows you down.
<DocScrutinizer>
:nod:
<wolfspraul>
atben clone
<wpwrak>
once you're comfortable, you can move to more interesting things
<DocScrutinizer>
more interesting like doublesided ;-)
<wpwrak>
atben = hairy ME + RF. that's not an easy project. also, the board is densely populated and your vias must be quite precise
<DocScrutinizer>
which is like one magnitude more demanding than singlesided
<wpwrak>
oh, double-sided is good. particularly if the circuit is simple. so if you screw up, you can still fix the problems with patch wire. see if the rest works.
<wolfspraul>
the bottom side of atben is just gnd?
<DocScrutinizer>
honestly, start with a single or 1.5-sided PCB (one trace side plus GND backplane, so you can't mess up the via fitting)
<wpwrak>
single-sided SMT is kinda boring. too easy ;-)
<wolfspraul>
I do see some traces there though...
<wpwrak>
wolfspraul: it's as much ground as possible :) but yes, i needed a few traces, too.
cladamw has joined #qi-hardware
<wolfspraul>
wow there is almost nothing on the board, right?
<wolfspraul>
the power only comes from ben?
<wolfspraul>
no converter, nothing? just from ben into the atrf chip?
<DocScrutinizer>
I think there's a mc as well?
<wolfspraul>
but a local crystal, I remember you tried without but the ben's clock was usable/stable enough?
<wpwrak>
power comes from the ben. and yes, very few components. very densely packed. lots of ground all over the place. a lot of vias to drill.
<wolfspraul>
was not
<DocScrutinizer>
aah that's atusb I guess
<wpwrak>
DocScrutinizer: atben and atusb have a crystal. remember the phase jitter issue ?
<DocScrutinizer>
yep
<wpwrak>
(MCU) yes, that's atusb. atben is "dumb"
<DocScrutinizer>
I recall a shematics with two chips
<DocScrutinizer>
which I think was atusb
<wolfspraul>
atrf, 8 caps (0201), 1 resistor, 1 crystal, and then one other IC - what is that one?
<wolfspraul>
I should look at the schematics, open hardware here right? :-)
<DocScrutinizer>
wpwrak: btw what fixed the jitter issue finally?
<wpwrak>
wolfspraul: the components are 0402. i didn't venture into 0201-land yet. most people find 0402 challenging enough ;-)
<DocScrutinizer>
LOL, indeed
<whitequark>
yeah, 0603 is already hard
<wolfspraul>
ah ok, yes
<DocScrutinizer>
I'm just too old for that shit
<wpwrak>
0201 may actually not that bad. i mean, already 0402 is more by smell than by sight ...
<whitequark>
LOL smell
<DocScrutinizer>
I heard 0201 is oxidizing away from mere plain air, too terrible mass/surface ratio ;-P
<wpwrak>
for experts: 0201, with a trace crossing under the component, and have a via in one of the pads. if you can pull that off with DIY, you rock ;-)
<whitequark>
DocScrutinizer: actually, I have some problems placing SMT components because my hands are shaking. not much, but enough to shift a 0603 horizontally by one position
<DocScrutinizer>
sure
<whitequark>
and I'm 19 and I never drank alcohol, nor do I commonly drink coffee
<DocScrutinizer>
and I'm not focusing anymore on things closer than 40cm
<whitequark>
so I guess that's implicit limitations of human beings
<DocScrutinizer>
a nogo for unarmed eye
<DocScrutinizer>
SMT soldering
<wolfspraul>
wpwrak: I still think atben clone is a good first target
<wolfspraul>
I need something I can use myself
<wolfspraul>
that or labsw
<whitequark>
uh, I'm somewhat lucky here because I don't focus on things further than 40cm :)
<wolfspraul>
anything fpga you outlawed, for now. ok. accepted.
<DocScrutinizer>
hihi
<wolfspraul>
with atben it's just a few parts and I focus on the pcb and see how I can get it done and documented, should be neat.
<whitequark>
wolfspraul: actually, just try making an atben. it's not like you are commiting to a longterm relationship with it or something like that
<wpwrak>
whitequark: maybe you're not placing your arm/hand well. or maybe yuo have a small defect in your motoric system. try to rest hand and arm as much as possible. use small angled tweezers. if you notice your hand is shaking, do something else for a while. then 0603 should be no problem.
<whitequark>
wolfspraul: if it will work, you're two orders of magnitude better at making pcbs than me
<whitequark>
wolfspraul: if it won't, try something else
<wolfspraul>
there is no way it will not work
<whitequark>
wpwrak: I don't really have problems placing 0603's, it's perfectly manageable but requires some concentration, accuracy and preciseness
<whitequark>
wpwrak: whereas for 0805's or even 1206's it's mostly a no-brainer
<wolfspraul>
wpwrak: what tools do I need? I start with a 0.8mm fr4, dual 1oz copper
<wpwrak>
whitequark: (0603) ah okay. that it does :)
<wolfspraul>
do you etch nowadays, or drill?
<wolfspraul>
I mean the copper
<wolfspraul>
for etching, the process is still what I saw with laser printer and toner transfer?
<wpwrak>
1) i cnc-drill all the via holes
<wpwrak>
2) i cnc-mill any larger holes and the board outline
<whitequark>
wpwrak: ahh, so that's how you managed to diy atbens. cnc explains everything.
<wpwrak>
3) i clean the board (steel wool and alcohol)
<wpwrak>
4) i toner transfer the layout. this is a bit tricky. if you can find that HP paper, that would be best. else, you need to search around. may taken 5-10 attempts before you find a paper that's good.
<wolfspraul>
HP will pay you to claim that they are still related to anything innovative :-)
<wpwrak>
whitequark: yeah. atben is a bit tight for manual work. particularly the uSD card shape. of course, it can be done, but ...
<wolfspraul>
you have this Roland some-15 machine, right?
<wpwrak>
wolfspraul: yeah, and even with such an old paper ;-)
<wpwrak>
mdx-15, yes
<DocScrutinizer>
the engineer's 3D-printer
<wolfspraul>
any recommendations what is cheap and good nowadays, and easy to program/control?
<DocScrutinizer>
;-D
<wpwrak>
not sure if it makes sense to get such a critter, though. i mean the mdx-15. it's a nice package, much like an office printer, but it also has its drawbacks
<whitequark>
oh mdx-15. I WANT IT
<DocScrutinizer>
you'll not get a decent CNC for <1000 bucks
<wpwrak>
drawbacks being the weird language, the weak spindle, and the small work area
<whitequark>
but $3.5k isn't really manageable for me
<whitequark>
DocScrutinizer: any advice on a good CNC in $1k-$1.5k range?
<wpwrak>
plus, inconvenient mounting (you can't clamp/screw things down. the only option is double-sided tape)
<DocScrutinizer>
not at all
<DocScrutinizer>
ask roh - he seems more up to date with such stuff
<wpwrak>
i don't know how whose low-cost chinese machines compare, though. maybe they have their own set of issues
<wolfspraul>
I will probably just buy something cheap here in beijing
<wolfspraul>
that's the great thing about all the chinese "crap", it's wonderful for experimenting and learning, and then knowing what is really worth spending money for
<wpwrak>
hehe ;-)
<wolfspraul>
since they will challenge every screw - "not needed"
<DocScrutinizer>
just to find you got a reproducable precicion of >1mm ;-D
<wolfspraul>
that's certainly up to me to find out :-)
<wolfspraul>
both words "reproducible" and "precision" are not translatable into Chinese...
<wolfspraul>
:-)
<wolfspraul>
no need
<DocScrutinizer>
indeed
<wolfspraul>
they make their billions without them
<whitequark>
rofl
<whitequark>
so true
<wolfspraul>
yes once you get the hang it's so empowering really
<DocScrutinizer>
yeah, for PCB on CNC they simply produce 50 boards and select the 2 that incidentally were good
<wolfspraul>
you say pcb is hard
<wolfspraul>
ha ha
<wolfspraul>
I can do a brain surgery tomorrow!
<wpwrak>
yeah. the mdx-15 is pretty accurate in X/Y. the table moves quite a lot in the Z direction, though. maybe 0.5 mm if you push hard. now, normally there shouldn't be much vertical pressure, but you can still sometimes run into problems.
<wolfspraul>
definitely
<wolfspraul>
chinese docs can do it too, so can I
<wolfspraul>
just give me some tools, and I will start
<wolfspraul>
first couple patients will probably die, but after some hundred it will get better
<wolfspraul>
a friend here has one of those hyped makerbots
<wolfspraul>
let me see what that is
<wpwrak>
happy ? :)
<wolfspraul>
except for unpacking and showing people how cool they are they will have no use for it
<wpwrak>
;-)
<wpwrak>
makerbots and the like are additive. so you need to apply the material you want in the end. alas, they don't print FR4, copper, and such
<wolfspraul>
ah ok, that probably won't help then
<wpwrak>
so for pcbs, you want a mill
<wolfspraul>
I just want a simple drill/mill like you have
<wolfspraul>
ideally fully controllable from standard linux tools etc
<wpwrak>
you'll need a more or less dedicated PC for the parallel port
<wolfspraul>
dremel? you mean a hand dremel? you need that for what?
<wpwrak>
all the cheap ones leave the control work to the PC. and the PC is bit-banging the motor signals. as low-level as it gets :)
<wpwrak>
(dremel) that would be the tool if you don't have CNC. or for the occasional retouching.
<wolfspraul>
I fail to see the application for the additive 3d printers I've seen so far
<wolfspraul>
they are insanely slow, large tolerances, just a joke really
<wolfspraul>
maybe I could make doorstops or so
<wolfspraul>
or candle holders
<wolfspraul>
that would fit aesthetically
<wpwrak>
3d printers are less messy than CNC mills. and some can also print overhangs. that's something the basic CNC mills can't make
<DocScrutinizer>
proper 3D printers start at nn kBucks
<DocScrutinizer>
and weigh tons
<wolfspraul>
yes sure, but we try to be very practical here
<DocScrutinizer>
wpwrak: yup
<wolfspraul>
wpwrak: it seems you are happy with your mdx-15 and will continue to use it... so that's a good guidance
<wolfspraul>
and you built a neat noise-cancelation box around it :-)
<wolfspraul>
what exactly is that HP paper?
<wpwrak>
perhaps an in-house 3d printer can be good for very rough prototypes. so that you can hold the thing in your hands, turn it around, feel it.
<DocScrutinizer>
goldelico has printed GTA02 cases now, in translucent white
<wpwrak>
the mdx-15 is all i have :) if i had a lot of money, i'd probably try to find a 4 axis machine. but we're quickly talking USD 10k+
<wpwrak>
i may need a bigger apartment as well ;-)
<wolfspraul>
still need more data
<wolfspraul>
what HP paper you recommend?
<wolfspraul>
how did that transfer work again?
<wpwrak>
C6039A: HP Premium Photo Paper, glossy
<DocScrutinizer>
o.O
<wolfspraul>
first print on the laser, then use your stepper-motor modified whatever-it-was to transfer/burn the toner onto the copper?
<wpwrak>
or the same kind of paper but in a different size. the product code changes with size and number of sheets in the package.
<wolfspraul>
how about just using an iron to transfer it?
<wpwrak>
the laminator, yes
<wolfspraul>
you need heat & pressure?
<DocScrutinizer>
yup, just like inside laserprinter
<wpwrak>
you can use a clothes iron instead. need a bit of figuring out the technique, though. ironically, i found that women seem to have a lot less difficulties with it than men ;-) (a few friends have built circuits using this process)
<wpwrak>
heat and pressure, yes
<wolfspraul>
ok, sounds like iron first
<wolfspraul>
what chemicals do you need for the etching?
<DocScrutinizer>
umm, now it gets funny
<wpwrak>
yeah. also modifying the laminator can be a bit messy. i made a very simple circuit that works but poorly
<wolfspraul>
anything to watch out for in the laser printer?
<wolfspraul>
sure I will try to skip that
<wpwrak>
i use muriatic acid and peroxide
<DocScrutinizer>
yeah, you might mess it up
<wolfspraul>
I will go out and buy cheap chinese tools, whatever I can find, and then reverse from there. rather than following a perfect spec.
<wolfspraul>
it was a 2-step process, right?
<wolfspraul>
you have the etching bath, put it in there for xx seconds, then out and into a second cleaning bath?
<DocScrutinizer>
also I think tones can differ vastly
<DocScrutinizer>
toners*
xwalk has quit [Ping timeout: 248 seconds]
<wolfspraul>
maybe but no point to over-plan and assume the worst
<DocScrutinizer>
make sure you're printing as 'fat' as possible
<wpwrak>
(laser printer) the straighter the paper path, the better. some have a backdoor you can open and then the paper doesn't get turned around
<wolfspraul>
like I said, when the experiments are cheap and fast, they are fun (and valuable) no matter what the outcome
<wolfspraul>
straight - excellent, makes sense
xiangfu has joined #qi-hardware
<wpwrak>
but in general, it seems that any printer will do. you may have to vary the transfer paper, though
<wolfspraul>
can you prepare the copper in some way for better transfer?
<wolfspraul>
I would think there are many ways to transfer it over
<DocScrutinizer>
I think wpwrak already mentioned copper conditioning
<DocScrutinizer>
cleaning and polishing
<wpwrak>
you must clean the copper before the transfer. so you scrub it with the steel wool until it's shiny. then you wash/rub off the steel wool with alcohol.
<wolfspraul>
sandpaper also ok?
<DocScrutinizer>
eeeek
<DocScrutinizer>
probably not
<wpwrak>
then quickly, before it oxidizes again, do the transfer. if it fails, remove the toner with paint thinner or acetone, then repeat the cleaning process
<wpwrak>
maybe if you can find very very fine sandpaper. but steel wool should be easier. any supermarket has that ;-)
<wolfspraul>
sandpaper vs steel wool?
<wolfspraul>
very fine, hmm
<wolfspraul>
corning 500 is no problem here
<wolfspraul>
they have sandpaper everywhere, from about 100 to 500 I think, have never tried to look for extremes though
<DocScrutinizer>
more like 1400
<wolfspraul>
and the steel wool is less intrusive? I would have never thought
<DocScrutinizer>
well, steel wool is less hard than sand paper
<wolfspraul>
I would think that the steel wool creates deep scratches, whereas say a corning 500 sandpaper feels quite smooth
<wolfspraul>
ok, thanks
<wolfspraul>
I will try
<DocScrutinizer>
and copper is harder than most other materials you usually apply steel wool to
<wpwrak>
you want the fine steel wool. with almost hair-thin steel. not the coarse with "noodles"
<wolfspraul>
depends on applied pressure too, of course
<wolfspraul>
ok [hair thin]
<wolfspraul>
the etching...
<wolfspraul>
need more data
<wolfspraul>
2 step?
<wpwrak>
you can get scratches, yes. you can also experiment a bit with what works best.
<wolfspraul>
you buy the chemicals or make yourself? which chemicals do you buy?
<wolfspraul>
I recently discovered a "home improvement" market nearby
<wolfspraul>
actually VERY NEARBY
<wolfspraul>
a bit scary
<DocScrutinizer>
something you go to prison when buying in USA ;-D
<wolfspraul>
they have little shops who are piling up chemicals in their little barracks until under the roof
<wolfspraul>
kids playing there, on the barrels etc.
<wolfspraul>
urgh
<wolfspraul>
and I am *sure* those people know *nothing* about chemistry
<wolfspraul>
nothing, zero
<DocScrutinizer>
yeah
<wpwrak>
etching: mix the etchant, toss the pcb inside, move it around until it's finished, remove, rinse, check visually. if you find insufficiently etched bits, toss it back in. otherwise, wash carefully, then proceed with tinning.
<DocScrutinizer>
that's China how we love and know it
<wolfspraul>
I will keep my shopping there very focused
<wolfspraul>
in, want this, that and that, pay, out
<wpwrak>
how nearby ? downstairs neighbour ? ;-)
<wolfspraul>
with all my wikipedia research about chemicals, I know 100 times more about those things than them who are storing 100 times more right where they sleep with their whole family...
<wolfspraul>
scratch that, 10,000 times more than me
<wpwrak>
always keep the muriatic acid next to the caustic soda and nothing can go wrong ;-)
<DocScrutinizer>
if you can get ammoniumperclorate, I think it's a good etching
<wpwrak>
maybe also keep a few canisters of 50% peroxide around, just in case :)
<wolfspraul>
ok one by one
<DocScrutinizer>
IIRC
<wolfspraul>
so it's only 1 bath, the rinse with normal water?
<DocScrutinizer>
wtf is muriatic acid?
<wpwrak>
yes. it's only 1 ml or so that you need to get rid of
<wolfspraul>
tap water here probably has a few hundred unknown chemicals in it anyway
<wpwrak>
DocScrutinizer: HCl
<DocScrutinizer>
aah
<wolfspraul>
wait, I need this in the right order
<wolfspraul>
1. what do you buy
<wolfspraul>
2. how to you mix/prepare
<wpwrak>
wolfspraul: there are several chemical processes what people use for etching pcbs
<wolfspraul>
3. the bath you described and remember from my visit
<wolfspraul>
4. rinse with tap water
<wolfspraul>
5. how to dispose of the rest
<wolfspraul>
sure I can imagine, but let's just focus on yours
<wolfspraul>
and watch out joking with me, I may not get it
<wolfspraul>
:-)
<wpwrak>
there if Fe(III)Cl (or some such), there is ammonium chloride, there is natrium chloride, and there's HCl+Peroxide. plus a few even more exotic ones.
<wolfspraul>
then I end up with some bomb, since the people I buy from certainly know even less than me
<wolfspraul>
yes but let's focus on your process
<wpwrak>
ah, let's talk about the fertilizer then .. ;-)
<wolfspraul>
can you walk me through?
<wolfspraul>
just what you do
<wolfspraul>
what do you buy
<wolfspraul>
how do you prepare/mix
<wpwrak>
FeCl = traditional, needs heating, considered pretty "safe" to handle.
<wolfspraul>
the application I think I'm clear (as described above), then disposal
<DocScrutinizer>
slow and messy
<DocScrutinizer>
much dirt
<wpwrak>
ammonium chloride and natrium chloride (actually, it's not NaCl ... must be something slightly different) = cleaner than FeCl, still considered "safe", work at room temperature
<DocScrutinizer>
natrium chlorite maybe?
<wpwrak>
FeCl drawbacks: it's a ugly brew, rost-brown to nearly black. stains everything it touches. loses transparency after a bit of use. will not evaporate.
<wolfspraul>
what is your known best and latest practice?
<DocScrutinizer>
riuns your cloths and skin
<wolfspraul>
I will focus on that
<wpwrak>
ammonium chloride and such drawbacks = harder to obtain, slow
<whitequark>
there's also ammonium persulphate
<DocScrutinizer>
yeah
<whitequark>
it's pretty awesome: easy to obtain, clean
<DocScrutinizer>
not perchlorate but persulphate
<whitequark>
through it ruins clothes even faster than FeCl(III)
<DocScrutinizer>
I guess I mixed it up somehow
<wpwrak>
now, HCl + H2O2 = clear solution, stays transparent, works at room temperature, you can choose how fast you want it to be, you can let it evaporate and then only have to dispose the solids
<wolfspraul>
that's your current process?
<wpwrak>
HCl + H2O2: ingredients extremely cheap and easy to find (hardware store and pharmacy)
<wpwrak>
yes, HCl+H2O2 is my current process
<wolfspraul>
you buy hcl and h202 separately?
<DocScrutinizer>
keep away any organic solvent from this though!!!!
<DocScrutinizer>
esp acetone
<wolfspraul>
hcl = hydrogen chloride
<wpwrak>
drawbacks: need to control the chemistry a bit. releases Cl gas.
<wpwrak>
hcl = muriatic acid ("salzsaeure")
<wolfspraul>
ah ok, hydrochloric acid, i.e. solution of hydrogen chloride in water
<DocScrutinizer>
and keep clear of the gas!
<wpwrak>
h2o2 = peroxide. at high concentrations, peroxide is used as rocket fuel.
<wpwrak>
at low concentrations, as wound desinfectant
<wpwrak>
fortunately, you'll only need the 5% variant
<DocScrutinizer>
easily found at barbers
<wpwrak>
you may want to keep a bottle of ~30% around to recycle your acid. but that's already advanced use.
<wolfspraul>
wait wait
<wolfspraul>
I'm still with the hcl acid
<wolfspraul>
what concentration of that?
<wpwrak>
any pharmacy has 5% peroxide. liters of it. probably hundreds ;-)
<wpwrak>
the "standard" concentration. i think it's 35%. could be 30%, though.
<wpwrak>
the one they have at hardware stores. it's used for plumbing work. not sure what exactly the plumbers do with it, though.
<DocScrutinizer>
cleaning copper ;-D
<wolfspraul>
peroxide is a liquid too?
<DocScrutinizer>
yup
<wolfspraul>
and I need a 5% concentration
<DocScrutinizer>
wasserstoffperoxid
<wolfspraul>
a bottle of 30% concentration for "recycling your acid"?
<wolfspraul>
don't understand
<DocScrutinizer>
better forget that
<wolfspraul>
then you mix the 35% hcl and 5% h2o2 ?
<wolfspraul>
in which way?
<DocScrutinizer>
30% H2O2 is actually dangerous
<wpwrak>
i think it was 1/3 hcl, 2/3 h2o2
<wolfspraul>
my #1 concern with all this would be that what I think I buy is not what I actually buy
<wpwrak>
but you may want to experiment a bit. you can also add a bit of either while etching.
<wolfspraul>
the sellers here operate with 0 knowledge, if anyone can imagine this
<DocScrutinizer>
wolfspraul: indeed
<wolfspraul>
I think you cannot, unless you lived in china for a while and get the feeling on how this works
<wolfspraul>
so I actually would only trust the concentration if I measured it myself, in some way
<wolfspraul>
and if the measurement equipment was imported :-)
<wolfspraul>
seriously
<wolfspraul>
but anyway
<wolfspraul>
let's keep on our over-knowledge path...
<DocScrutinizer>
anyway you'll easily check for general properties of your chemical. HCl is easily detecable by smell, H2O2 will act like sodawater on any wodden surface
<DocScrutinizer>
wooden*
<wpwrak>
the 30% H2O2 would be for increasing the H2O2 concentration if the solution gets stale. or if you've let it sit for a while. and yes, that's something you don't touch without gloves :)
<wolfspraul>
1/3 hcl, 2/3 h2o2 ?
<wolfspraul>
how much do you make typically?
<wpwrak>
wolfspraul: i doubt the pharmacy will have peroxide much stronger than 5%. maybe 10%, which would be used for hair bleaching. anything stronger and the customer feedback will be quick and unfriendly :)
<wolfspraul>
those are both liquids - any danger of them reacting with air, something else, evaporating?
<wolfspraul>
I go to the 'home improvement' market around the corner
<DocScrutinizer>
a basin large enough to hold your PCB, and 5..10mm high
<wolfspraul>
I wouldn't be surprised if they are selling used equipment from all sorts of nuclear facilities around beijing, without knowing what it is
<wpwrak>
HCl can evaporate out of the bottle (unless the bottle is good enough). so better keep the bottle outdoors
<wolfspraul>
the problem is not what I can get, but whether I know for sure what I get
<wolfspraul>
but that's my problem
<DocScrutinizer>
wolfspraul: yes, toldya DO NOT GET THEM in contact with acetone
<wolfspraul>
I think this is actually a small miracle. the semiconductor industry uses large amounts of really dangerous chemicals
<wolfspraul>
like arsine
<wolfspraul>
gallons and gallons of it
<wpwrak>
the 5% peroxide should be as harmless as it gets. i mean, i wouldn't drink it, but i don't think it makes a mess with anything else.
<wolfspraul>
but I think until today, there never has been 1 fatality
<wolfspraul>
but then, china is still behind in that industry, and if there were a fatality it would surely be covered up
<wolfspraul>
(in china)
<wpwrak>
they probably just seal off the valley and wait until the screaming stops. no bad news ;-)
<wolfspraul>
do you know how many and what large amounts of *really* dangerous chemicals are used in semiconductor making? I had no idea...
<wpwrak>
unpleasant to inhale, so you're not likely to do that long enough for it to be really dangerous
<wolfspraul>
but I described the real issue to you :-)
<wolfspraul>
not only the cigarette-smoke like air I breath every day
<wolfspraul>
and the tap water I use for cooking that contains an unknown number of unknowingly dangerous things
ezdagor has left #qi-hardware [#qi-hardware]
<wolfspraul>
and the fish with antibiotics
<wolfspraul>
and the pigs that are slaughtered when having high fever and also pumped full with medication
<wpwrak>
what's more of a concern is Cl gas escaping after you're done. so again, store this outdoors and away from metals.
<wolfspraul>
and the sellers of chemicals who have no clue what they are doing
<wolfspraul>
and filling liquids from one bottle into another at will
<DocScrutinizer>
just don't dare to mix your etching mix with anything else, esp not the thinner you're using to remove the toner
<wolfspraul>
writing on the label whatever they want, or maybe not
<wolfspraul>
and reselling
<wolfspraul>
and and and
<wolfspraul>
:-)
<wolfspraul>
I can read in wikipedia all I want
<wolfspraul>
still need to be careful not to touch open wires on the street, or else...
<wpwrak>
if you can seal the container, you can reuse the acid. but you may need higher-grade peroxide to reactivate it after storage.
<wolfspraul>
ok
<wolfspraul>
cl gas, hmm
<wolfspraul>
you mean out of the hcl acid bottle?
<wpwrak>
if you can't seal the container, i'd just let it evaporate. you'll get some bluish-green crystals that fit nicely in the trash. even though they probably shouldn't go there :)
<wpwrak>
both from the HCl bottle and from the etching solution
<wpwrak>
both release Cl gas
<wolfspraul>
that's my least concern, I help the trash on the landfill balance out better :-)
<DocScrutinizer>
once you want to get rid, I'd mix the shit with chalk or sth, maybe even flour, then just dispose it
cladamw_ has joined #qi-hardware
<wolfspraul>
ok but the eching solution is only around for some X minutes, no?
<wolfspraul>
whereas the bottle of hcl acid may stand there for days, weeks, months
<wpwrak>
DocScrutinizer: if you mix it, you may make it more stable and then you have a liquid to get rid of. much messier.
<wolfspraul>
they have bottles with plasticky covers on top, that snap around the glass
<wolfspraul>
doesn't look very tight, but not bad either
<wpwrak>
depends. you can use the etching solution just once or you can use it for years
<DocScrutinizer>
wolfspraul: nope, you usually keep the etching liquid, it's ok for several boards
<wolfspraul>
ah I see
<wolfspraul>
but then you would want to fill it back into a bottle?
<wolfspraul>
or keep in the etching container?
<wpwrak>
the actual etchant is generated in the process. HCl+H2O2+the copper of the PCB react first to form cupric acid. and then it's the cupric acid that takes over the etching.
<DocScrutinizer>
if it'sa a nice tupperware box
rejon_ has quit [Ping timeout: 272 seconds]
<wpwrak>
some people cultivate their cupric acid for ages ;-)
<wpwrak>
i tend to just let all the stuff evaporate, then toss the crystals. it's cents anyway ;-)
<wolfspraul>
sounds cleaner to me too
<wpwrak>
evaporation will take a while, though. depends on ambient temperature and humidity.
<wolfspraul>
a while = ?
<DocScrutinizer>
well, otoh it's probably not worth the hassle to keep such small amount of etching fluid, esp when it's based on H2O2 which is instable
<wolfspraul>
you take the board out with some tweezer, then under tap water?
<DocScrutinizer>
yep
<DocScrutinizer>
wear goggles
<wolfspraul>
with the tap water you probably wash a small amount of the etching solution off as well
<wolfspraul>
no problem with that?
<wolfspraul>
it wont' destroy the pipes?
<wpwrak>
DocScrutinizer: yes, the H2O2 escapes/decays. that's why you need higher grade peroxide to restore the acid if you store it.
<wolfspraul>
or say - damage
<DocScrutinizer>
unless you recycle your greywater in a pond in the garden...
<DocScrutinizer>
wolfspraul: nope
<DocScrutinizer>
way too dilluted
<wpwrak>
wear gloves. latex or plastic
<wolfspraul>
but you cannot just dispose of the entire etching solution in this way?
<DocScrutinizer>
you can
<DocScrutinizer>
if you don't mind
<wolfspraul>
I would mind damaging the pipes
<DocScrutinizer>
dillute 10:1
<DocScrutinizer>
no problem, really
<wolfspraul>
but then I live in a household where pretty much anything can and will break at any time
<wpwrak>
you can use tweezers or just a wooden stick (the one you use for grilling small chunks of meat ... don't remember the english word. "spiess" in german) to handle the pcb
<wolfspraul>
so it's hard to see causes :-)
<wolfspraul>
wooden stick = chopsticks?
<DocScrutinizer>
dillute, flush with 10L water
<wpwrak>
too big :)
<DocScrutinizer>
but that's absolutely not PC
<wolfspraul>
ah the problem with tweezers would be if they are metallic?
<wpwrak>
i wouldn't want to take the risk of releasing all that acid. you never know if it may not end up somewhere in a higher concentration that you expect.
<wolfspraul>
but maybe for quickly taking a pcb out it doesn't matter, just wears down the tweezer over time, so what
<wpwrak>
but the drop or two that you rinse off the board surely won't do anything.
<wpwrak>
no metals ;-)
<wolfspraul>
but it won't affect it so quickly, no?
<wpwrak>
also don't wear metal gloves ;-)
<DocScrutinizer>
you see people clean their toilets with HCl all over the globe
<wolfspraul>
sure
<wpwrak>
well, depends ... sure, it'll take a while for it to eat your tweezers (unless you go the the etch-the-board-in-seconds approach)
<wolfspraul>
seems I can dilute 50ml of that solution in 10L water, and down the drain
<DocScrutinizer>
even with concentrated HCl you usually can't etch iron
<DocScrutinizer>
or steel
<wolfspraul>
ok
<wolfspraul>
alright I think it sounds pretty clear now
<wolfspraul>
lemme think
<wpwrak>
what you have isn't HCl anymore. it's cupric acid.
<wpwrak>
and i don't know what that does to iron. or whatever else is there in your pipes.
<wolfspraul>
I wish I have the actual 35% hcl at the beginning
<wolfspraul>
the less uncertainty the better, so I will probably dispose of that etching solution
<wolfspraul>
or just into the garden
<wpwrak>
so, the shopping list: 35% HCl from hardware store, 5% peroxide from pharmacy (it's used as a desinfectant, just like alcohol)
<wolfspraul>
why does it matter where the 5% peroxide comes from?
<wpwrak>
next, latex/plastic gloves. also get some goggles or glasses, in case something spills
<DocScrutinizer>
if you can get 10%, I'd prefer that
<wolfspraul>
the normal gloves they have for household work shoudl be fine?
<wpwrak>
5-10% should both be okay.
<DocScrutinizer>
mind you, 65..70% of the HCl are water as well
<wpwrak>
(gloves) if they're thin enough that you can actually manipulate things
<wolfspraul>
and I keep the hcl and h2o2 away from acetone
<DocScrutinizer>
YES!
<wolfspraul>
ah they also have these factory gloves, they are very fine
<wpwrak>
and any other solvents, paint thinner and such
<wpwrak>
DocScrutinizer: have you had some of that stuff go off by accident ? ;-)
<wolfspraul>
I have a bottle of acetone here, which I would think is also slowly evaporating (don't trust the seal on the bottle entirely)
<DocScrutinizer>
not THAT stuff
<wolfspraul>
if it is acetone, well, it smells like :-)
<wpwrak>
fine gloves are good. as long as they're watertight.
<wolfspraul>
any danger of smelling the hcl or h2o2?
<DocScrutinizer>
ammonium-joidine though...
<wolfspraul>
trust me, even by know I already know 100 times more about hcl and h2o2 than any of the people I will talk to to buy from :-)
<wolfspraul>
by now
<wolfspraul>
what comes after etching
<wolfspraul>
rinse with water
<wolfspraul>
then?
<DocScrutinizer>
wolfspraul: H2O2 is "harmless" regarding gas, HCl you'll manage to stay away
<wolfspraul>
toner is still on, I guess
<wolfspraul>
DocScrutinizer: but smelling on a hcl bottle is a no-no?
<DocScrutinizer>
you don't do that, believe me
<wpwrak>
you'll understand very quickly why it's a bad idea ;-)
<wolfspraul>
alright
<wolfspraul>
how does the story continue?
<wolfspraul>
rinse with water
<wolfspraul>
then?
<wpwrak>
if you want to play chemical emergency, get some caustic soda, dissolve it in water (exothermic, so beware), open the window, then mix that solution with a bit of hcl. run ;-)
<wolfspraul>
nah
<wolfspraul>
I try to stay focused and have some knowledge safety around me
<wolfspraul>
my main difficult is starting with known chemicals
<wolfspraul>
difficulty
<wpwrak>
i experimented with the when trying to neutralize my FeCl. what a mess ...
<wolfspraul>
because they can easily mix and match whatever in their mess
<wolfspraul>
for example when I say I want "35% hcl", I am sure they will start some action with multiple bottles
<wolfspraul>
"let's take this bottle"
<wolfspraul>
"fill it over into this one, it's the right size for you"
<wpwrak>
35% HCl is pretty much the standard concentration.
<wolfspraul>
"then fill the liquid from this one into that one"
<DocScrutinizer>
aah yes, when contact to skin: rinse with plenty of water, best with some soap (no joke!)
<wolfspraul>
"and now hastily write over the label in some unreadable mumble jumble"
<wolfspraul>
plug on top, cash, done and out
<wolfspraul>
I'm sure I could return unused chemicals to them, and they would just fill it right back into the bigger bottles
<wolfspraul>
and so on
<wolfspraul>
(I mean I sell the unused part back to them :-))
<wolfspraul>
learning by doing
<DocScrutinizer>
don't order 35% HCl, just ask what they got!
<wolfspraul>
everything
<wolfspraul>
:-)
<DocScrutinizer>
ask for HCl, they will offer 30..35%
<wolfspraul>
hopefully, anyway that is my problem
<wolfspraul>
not the amount of knowledge I have
<DocScrutinizer>
btw all this is extremely error tolerant
<wolfspraul>
I could get something different, a little different, or a lot different
<wolfspraul>
yes I figured, of course
<wolfspraul>
it's a simple process, sounds like
<DocScrutinizer>
if you get 10% HCl, you just will need 5 times longer to etch
<wolfspraul>
sure
<DocScrutinizer>
and I guess there's rarely ever HCl with significantly higher than 40% concentation
<wolfspraul>
I'm not that stupid, I would be able to tell if some procedure depends on very fine tolerances, exact relations, temperature, pressure, etc. etc. and I probably wouldn't want to do it then.
<DocScrutinizer>
if the bottle starts smoking when you open it, be careful! ;-D
<wolfspraul>
yeah
<wolfspraul>
wpwrak: what comes after etching?
<wolfspraul>
the toner is still on after the water rinse?
<DocScrutinizer>
yep
<wolfspraul>
you eatch both sides together, I would think
<DocScrutinizer>
also yep
<wolfspraul>
you need to turn the pcb around in the etching solution, or not needed?
<DocScrutinizer>
not exactly needed, though you want to watch both sides
<DocScrutinizer>
so you'll turn it anyway
<wolfspraul>
why did wpwrak say a few times to get the h2o2 at the pharmacy?
<wpwrak>
and you normally have to turn it from time to time
<DocScrutinizer>
you however should find a small glass object to put on the floor of the container, so the PCB downside is in liquid, not on floor
<wpwrak>
because the acid will flow differently over the board. you have to move the board around in the acid.
<wolfspraul>
you mean a stand?
<DocScrutinizer>
sth like that, yeah
<wolfspraul>
yeah sure, I remember from seeing and I think that should be easy after some attempts
<wolfspraul>
and it sounds like you can use the same etching solution for a few hours?
<wpwrak>
(pharmacy) because they're everywhere and because they won't accidently sell you 80% peroxide ;-)
<wolfspraul>
or will it all evaporate by then?
<DocScrutinizer>
yep
<DocScrutinizer>
no
<wpwrak>
a few hours shuold be good. next day maybe not
<DocScrutinizer>
H2O2 might gone inactive though
<DocScrutinizer>
you add some more
<wpwrak>
you'll notice when the solution gets weak
<wolfspraul>
can I just fill in more h2o2 then?
<wpwrak>
DocScrutinizer: drawback of a stand: it increases the amount of acid you need
<DocScrutinizer>
sure, it should be 1..3mm max
<wpwrak>
to some degree, yes. if you add too much 5% h2o2, the water content of your acid gets too high and it doesn't work very well anymore
<DocScrutinizer>
a smal glas splinter
<wolfspraul>
understood
<wpwrak>
i'd just move the pcb over the floor of the container. the flow is usually stronger on the bottom than on the top anyway
<wolfspraul>
but if you have a higher percentage h2o2 it might work
<DocScrutinizer>
that#s why I suggested to use 10% if available
<wolfspraul>
ok
<wolfspraul>
after rinsing - what's next
<wpwrak>
yes. but careful: if you add, say, 30% peroxide to the mix, and you drop it right on top of the pcb, you may get some very fast etching on that spot before the h2o2 gets properly mixed
<DocScrutinizer>
you use.... Acetone ;-P to remove the toner
<wpwrak>
yeha, 10% is still safe and controlable
<wpwrak>
acetone or paint thinner/remover
<wolfspraul>
you mean the etching will undercut the toner?
<DocScrutinizer>
that's where my pun about going to jail for buying chemicalc kicks in
<DocScrutinizer>
yes, this can happen
<wolfspraul>
ok rinse with lots of water, then directly proceed to acetone?
xiangfu has quit [Ping timeout: 240 seconds]
<wolfspraul>
use the steel wool again?
<DocScrutinizer>
main reason why you want equal etching over the whole area
<wolfspraul>
how does it continue?
<wpwrak>
acetone is used for drug production. so it may be difficult to obtain at some places. e.g., in argentina, i have to bring an id document to buy it. some shops may even require a written declaration of the use.
<wolfspraul>
god
<wolfspraul>
should I praise the *freedom* of china?
<wolfspraul>
I could buy a hectolitre of acetone right in my neighborhood
<wolfspraul>
that won't get me in jail
<wolfspraul>
but 1 tweet about democracy would
<wolfspraul>
amazing how diverse our world is, no? :-)
<wpwrak>
;-)
<wolfspraul>
oh, and if I would make and sell drugs - death penalty
<wolfspraul>
of course
<wolfspraul>
1 week from 'catching' to execution
<DocScrutinizer>
mixing accetone with etching fluid will get you either in prison as well, or on cemetary
<wolfspraul>
with a mobile bus, organ harvesting included, of course
<DocScrutinizer>
;-D
<wolfspraul>
yes I understand
<wolfspraul>
but I just rinse with water, that sounds like enough to keep the 2 chemicals apart
<wolfspraul>
then acetone, right?
<wolfspraul>
then what?
<wolfspraul>
more steel wool?
<wolfspraul>
finished?
<wpwrak>
here it would be a slap on the wrist. unless you drug business was not successful enough to produce sufficient bribe money. or, of course, if it was too successful, the competition would remove you.
<wpwrak>
no steel wool
<wpwrak>
toner removal is all solvent + paper towel
<DocScrutinizer>
yes, enough to keep them apart
<wolfspraul>
paper towel, ok
<wpwrak>
you need lots of paper towels. also for the alcohol you used first to remove the steel wool
<wolfspraul>
oh that was before etching?
<wolfspraul>
you said steel wool, then clean with alcohol?
<DocScrutinizer>
ye
<wolfspraul>
clean with alcohol better than with acetone?
<DocScrutinizer>
yep
<DocScrutinizer>
err, for removing steel wool, I dunno if alcohol is maybe better
<wpwrak>
alcohol is for the steel wool
<wpwrak>
acetone is for the toner
<wolfspraul>
just a high-percentage one I would think - 70%? 95% ?
<DocScrutinizer>
after all alcohol usually has some water, acetone hasn't
<wpwrak>
95+% alcohol. you can get it basically pure. ideally in spray cans.
<wolfspraul>
ok
<wolfspraul>
alcohol before etching, then acetone after etching, then again alcohol with paper towels after toner is off
<wolfspraul>
cool
<DocScrutinizer>
sounds good
<wolfspraul>
do you seal the open copper traces etc in any way?
<wpwrak>
yeha, i usually swipe off the acetone with alcohol. not sure if it's needed.
<DocScrutinizer>
you'll see when copper is clean
<wolfspraul>
to leave only pads open?
<wolfspraul>
have you done any silkscreening?
<wpwrak>
well, or tather the paint thinner. which is assuredly a vile brew. and quite oily.
<wolfspraul>
tather?
<wpwrak>
rather
<DocScrutinizer>
you could varnish whole board with so called loetlack, dunno the english name. basically colophonium in alcohol
<wpwrak>
(i use paint thinner instead of acetone. no bureaucracy :)
<DocScrutinizer>
you don't probably want to do this for SMT
<wolfspraul>
wait, what is that for?
<wolfspraul>
now we have a clean board. toner removed with acetone/paint thinner, board and copper cleaned with alcohol and paper towels.
<wpwrak>
i seal the copper traces by "painting" solder on it. see the backlog from a few days ago ;-)
<wolfspraul>
now what?
<wolfspraul>
to avoid copper oxidization?
<wpwrak>
yes
<wolfspraul>
that's still conductive though
<wpwrak>
and it also improves solderability
<wolfspraul>
searching backlog, one sec
<wpwrak>
solder is conductive, yes ;-)
<DocScrutinizer>
you could also put the whole PCB in a galvanic bath not, and coat traces with silver or even gold
<DocScrutinizer>
s/ not / now /
<wolfspraul>
ah yes
<wolfspraul>
"I tin the board first" "tin = apply flux, then paint tiny amounts of solder"
<wolfspraul>
you recommend water soluble, and then after tinning you rinse with water?
<wpwrak>
a flux dispenser that looks like a felt pen
<wolfspraul>
you mean just a pen, like in which url?
xiangfu has joined #qi-hardware
<wpwrak>
after tinning, i toss it into the ultrasound bath, where cavitation and water will remove the flux
<wolfspraul>
ah, yes
<wolfspraul>
so you cover 100% of the copper in this way, with solder
<wpwrak>
pen like in the url. you can get flux in a number of containers. e.g., 5 l canister would be considerably less convenient :-)
<wpwrak>
yes
<wolfspraul>
which url? there were several
<wpwrak>
both looked similar
<wolfspraul>
ok
<wolfspraul>
seems clear
<wolfspraul>
pen :-)
<wolfspraul>
so the solder will be your oxidization protection
<wpwrak>
aye. and it'll make sure you already have a bit of solder under the pads when you place, say, a QFN.
<wolfspraul>
any further sealing of the traces?
<wolfspraul>
I mean where you don't want a solder pad...
<wpwrak>
so it's a good thing to do. even if using parts that are easier to solder
<wolfspraul>
and it protects the copper, which is necessary I would think?
<wpwrak>
you cover all the copper. also the traces.
<wolfspraul>
sure
<wolfspraul>
but they are still conductive everywhere now
<wolfspraul>
do you seal this off further?
<wpwrak>
of the copper oxidizes, the oxide is very difficult to solder
<wpwrak>
s/of/if
<qi-bot>
wpwrak meant: "if the copper oxidizes, the oxide is very difficult to solder"
<wpwrak>
normally no further sealing
<wpwrak>
if you want to isolate, you can apply acrylic, silicone, PUR, etc., coating at the end
<wolfspraul>
but you would need to keep the pads free?
<wolfspraul>
or you mean after soldering all parts on?
<wpwrak>
oh yes, isolation would be very last. after soldering, testing, rework, ....
<wpwrak>
basically when you're ready to forget the board ;-)
<wpwrak>
it's something you rarely need. if you want to protect the board from touch, you may want to make a real case that also prevent mechanical stress of the components.
<wpwrak>
and inside the case, you don't need isolation
<wpwrak>
well, for low voltages. things change at you approach mains voltage
<wolfspraul>
ok we see. one day I want to try with some hardening epoxy
<wolfspraul>
that should be easy and give mechanical stability as well
<DocScrutinizer>
hmm, considered comlicated
<wolfspraul>
like the blobs you have on cob mounted chips
<wpwrak>
it also makes repair impossible ...
<DocScrutinizer>
some guys claim the shrinking of epoxy will break your circuit
<wolfspraul>
I think there's a lot of epoxies that are relatively easy to mix and use, could be wrong though. won't try this now.
<wolfspraul>
sure could be, that needs trying
<wpwrak>
once you have that CNC mill, you can make your own cases :)
<wolfspraul>
but there are serious vendors for these things and I'm sure we wouldn't be the first ones
<DocScrutinizer>
there are commercialy available varnishes to seal PCB, in spray can
<wolfspraul>
any link?
<DocScrutinizer>
kontakt chemie
<whitequark>
(kontakt) +1 on that
<whitequark>
(HCl) AFAIK if you'd somehow get >35% solution of HCl the HCl itself would instantly begin to evaporate from it until it'd be 35%
<whitequark>
(epoxy) true, this depends on exact type of epoxy
<wpwrak>
notice how the boiling point goes down as the concentration increases
<whitequark>
wpwrak: just calculated. it's 45% at 0°C and 42% at 20°C
<DocScrutinizer>
toldya, be careful when the bottle starts smoking when opened ;-D
<wpwrak>
and around here, days with 40 C aren't all that common in summer ...
<wolfspraul>
ok I think we got it all
emeb has quit [Quit: Leaving.]
<wpwrak>
s/common/uncommon
<qi-bot>
wpwrak meant: "and around here, days with 40 C aren't all that uncommon in summer ..."
<wolfspraul>
now the only thing missing is to try and go through
<wolfspraul>
:-)
<DocScrutinizer>
you forgot your vias ;-)
<wpwrak>
will you go straight for CNC or try with a dremel first ?
<wpwrak>
via = tiny hole, a matching wire, a bit of solder / solder paste, and numerous curses :)
<DocScrutinizer>
yep
<DocScrutinizer>
or a rivet
<wpwrak>
messy and large
<DocScrutinizer>
hollow rivets are rather cool, if you can find matching ones
<DocScrutinizer>
bit you'll need such sophisticated stuff only for RF
<DocScrutinizer>
like FM-transmitters
<wolfspraul>
good question [dremel]
<wpwrak>
hmm. i take my chances with a small hole and wire. you can make the vias pretty flat
<wolfspraul>
I will try bottom up, most likely
<wolfspraul>
and buy cheap tools I find on the market
<wolfspraul>
that's one part of the value for me, with all (seeming) disrespect for the chinese cutting corners everywhere, I want to give myself some time to understand their hacks, no matter how brutal they are :-)
<wpwrak>
hehe :)
<wolfspraul>
if I want to buy something where everybody just shakes heads, I won't be the arrogant "know better" type
<wpwrak>
manual tools are also good for small corrections.
<wolfspraul>
I have learnt to let the beauty of brutal hacks sink into my mind, even if it takes a little :-)
<wolfspraul>
how small are the via holes?
<wolfspraul>
if I would use a dremel, you use the dremel for the vias, or a drilling machine?
<wolfspraul>
what's the difference actually?
<wpwrak>
i use 28 AWG wire for them
<wpwrak>
and the drill is ... looking ..
<DocScrutinizer>
(vias) silver wire is nice for that, it's "soft" so you can easily press it flat on both sides, so it both sticks to the hole and has nice contact and level to the pads
<wolfspraul>
pure silver?
<wpwrak>
0.0135"
<DocScrutinizer>
yep
<wolfspraul>
how does that connect to the copper?
<wpwrak>
you solder it down
<wolfspraul>
you mean just heat up the silver? or apply regular solder at the junction?
<wolfspraul>
ah you are tinning anyway
<wolfspraul>
that probably takes care of it
<wpwrak>
i do this: 1) apply solder paste on both sides. 2) stick the wire through the hole. it fits precisely, so it's held by friction. 3) solder one side, then the other.
<DocScrutinizer>
you choose a wire diameter that somewhat sticks to the hole by itself, the cut so both sides have 0.5mm overlength, then flatten it with a pair of pliers or a small hammer and bolt
<DocScrutinizer>
the solder, of course
<wpwrak>
4) cut off the wire ends.
<DocScrutinizer>
with normal tin solder
<wolfspraul>
you use the dremel for drilling the via holes?
<wolfspraul>
(I mean without cnc)
<DocScrutinizer>
s/the/ten/g
<DocScrutinizer>
then* DANG!
<DocScrutinizer>
dremel with stand
xiangfu has joined #qi-hardware
<DocScrutinizer>
you don't wanna do this without stand
xwalk has joined #qi-hardware
<wpwrak>
if you have a steady hand and a flexible axis, you don't really need a stand. and the stand may be wobbly if made i china ;-)
<wolfspraul>
I see you internalize the culture here
<DocScrutinizer>
ouch
<wolfspraul>
the other day I met this foreigner who is working with electronics and complained how you can get anything done if even all test equipment fails
<DocScrutinizer>
drilling without stand is really really a PITA, esp when you got no pinpoint to center your drill to
<wolfspraul>
if every voltmeter or temperature sensor or *anything* is inaccurate
<wolfspraul>
so he starts to import those things now, to establish at least a testing baseline
<wolfspraul>
:-)
<DocScrutinizer>
:nod:
<wpwrak>
DocScrutinizer: if drilling manually, you drill after etching. that way, the center holes in the vias guide the drill
<DocScrutinizer>
apropos pinpoints, you *might* consider getting thise in your etching already
<DocScrutinizer>
exactly :-D
<wpwrak>
making pcbs, in less than 100 easy steps :)
<DocScrutinizer>
LOL
<DocScrutinizer>
I'm too old for that shit ;-D
<wolfspraul>
I think actually the steps are very simple, it just comes down to guiding someone correctly
<wolfspraul>
it doesn't matter that for every step there are 5 or 10 alternatives
<DocScrutinizer>
sure
<wpwrak>
yeah, that adds a bit of color ;-)
<wolfspraul>
in fact if you include outsourced fabbing you are actually buying into many different, hundreds, of automated processes at those factories
<DocScrutinizer>
one of the major problems with doublesided is to match both sides exactly
<wolfspraul>
and at least what I do is in no way meant to replace or 'compete' with a factory
<wolfspraul>
but I do think a very simple process of quickly making pcbs to hookup some chips can be valuable
<DocScrutinizer>
you might want to drill two adjusting holes even prior to toner transfer
<wolfspraul>
it's also so much better to outsource something after you understand it better
<DocScrutinizer>
indeed
<DocScrutinizer>
btw when you inspect the transferred toner and find some suspicious sparse covering, you can use an Edding fabric pen to improve the varnish layer
<wolfspraul>
the edding ink will withstand the etching as well?
<DocScrutinizer>
yep
<DocScrutinizer>
I used to draw whole boards like that
<wolfspraul>
how about waterproof pens like for writing on cds?
<DocScrutinizer>
that's basically this I think
<DocScrutinizer>
of course there are dedicated pens for that as well
<wolfspraul>
I still think I try with an atben clone, it seems the simplest thing I can use myself
<wpwrak>
i'd try some mcu. something really easy to play with.
<wpwrak>
atben is demanding to make. and you have no direct feedback of what's going on.
<wpwrak>
wolfspraul: do you have a scope ?
<wolfspraul>
not yet :-)
<DocScrutinizer>
oops sorry, tha last link didn't use the edding for etchstop
<wpwrak>
good answer ;-)
<wpwrak>
(double-sided) for boards that need high precision, like atben and atusb, i etch one side at a time and cover the other with adhesive tape. that way, i can position the toner transfer paper very precisely.
<wpwrak>
i.e., i put the paper on the table, turn the board to the right right, attach the adhesive tape to the back, then put the board such that the outlines on the paper meet the edges of the board, and finally push down the adhesive tape that extends outside the board down on the paper, to fix the board in place. add more tape if necessary, to cover the entire back. then, after etching, i remove the tape, scrub the board to remove all the
<wpwrak>
tape's glue, and repeat the procedure with the other side.
<wolfspraul>
a litle hard to follow but I think I get it
<wolfspraul>
the good news is that all these experiments are quick
<wolfspraul>
no matter how wrong one is, I just fail to see how one can be too unimaginative to not have it worked out after a few days at most
<wolfspraul>
but of course the various steps add up
<wolfspraul>
should be ok
<wpwrak>
yeah, relatively speaking. it sucks if you kill a board with a mistake at the very end, though.
<wolfspraul>
wpwrak: if pressed, what would be the turnaround you have now for going from proven and good design files to a pcb ready for soldering?
<wolfspraul>
it sounds like the whole process should not take more than a few hours, even 1 hour maybe if everything is setup and ready?
<wpwrak>
there are a few things that may need some experimenting. such as the transfer paper. some are very bad. some are almost right but still not good enough.
<wolfspraul>
yeah
<wpwrak>
also, you change the characteristics of the paper when you clean it.
<wpwrak>
time depends on the required precision and such. let's say 30 minutes for CNC, including setup, then 5 min for cleaning, toner transfer can vary - depends on when i'm happy with a print, maybe 10-30 minutes, then the actual transfer, 5 min. since i usually do two boards in parallel, that would be 10 min.
<wpwrak>
then 15-20 min for etching, with setup (fetching the acid from the terrace, mixing with peroxide, carrying the box over from the workshop)
<wpwrak>
after toner transfer also visual control and touch-ups. if the transfer is too ugly, i redo it. after etching also visual control. then toner removal, tinning, maybe some repair, then soldering the components.
<wpwrak>
toner removal is maybe 5 min. tinning too. soldering depends on the complexity of the board :)
<wpwrak>
qfns tend to add a bit more work. oh, did i mention that atben has a qfn ? :)
<wolfspraul>
which one, the crystal?
<wpwrak>
(atusb has two. that board is borderline DIY-able, because of all the ground areas)
<wpwrak>
the RF chip
<wolfspraul>
wpwrak: which dremel tips would I need?
<wolfspraul>
ah can't see so well. the chip on my atben (tuxbrain made) looks like it has leads, but maybe not
<wpwrak>
the crystal also needs some special attention. i use solder paste for it,to make sure these pads really connect
<wolfspraul>
solder paste?
<wpwrak>
no leads :) only tiny tin corners
<wpwrak>
the stuff used for reflow
<wpwrak>
you can also get it in syringes
<wpwrak>
1) solder paste on pad. 2) place crystal. 3) heat with hot air until the solder melts.
<wpwrak>
the crystal is hard to solder with an iron because you have very little lateral attack surface. the pads are flat and you have the metal case right above. so with an iron, you're likely to solder to the case, which in 50% of all cases not what you want. (the other 50% are ground pins)
cladamw has quit [Quit: Leaving]
Openfree` has quit [Ping timeout: 245 seconds]
rejon_ has joined #qi-hardware
<DocScrutinizer>
FYI: [copy from #openmoko-cdevel]
<DocScrutinizer>
[2012-05-15 04:54:52] <pabs3> hmm I got a weird email from sean
<wolfspraul>
or you heat up the entire board from the bottom?
<wolfspraul>
DocScrutinizer: Sean is quite chinese nowadays, no? :-)
<wolfspraul>
words are always on a different, removed, level from actions
<DocScrutinizer>
indeed
<wolfspraul>
it's like a bird singing, for example. just enjoy the flow, think about actions separately. needs some mental relaxation to get used to in meetings... :-)
Openfree` has joined #qi-hardware
<wolfspraul>
I am quite OK at it nowadays, I think. I can have a 'conversation' with someone, think about and decide about actions in a totally separate mental thread, and just purely enjoy the good food on the table as well.
<DocScrutinizer>
hehe
<wolfspraul>
it's like a symphony, words and actions and smiles and good life
<wolfspraul>
not bad actually
<wolfspraul>
no reason to get worked up - WHILE THE FOOD TASTES SO GOOD!
<DocScrutinizer>
hmm, food
<wolfspraul>
I unsubscribed from the om lists recently, I felt there was not much constructive thought left
<wolfspraul>
only the same old slogans, almost like from a tape recorder
<wolfspraul>
even the silence on the Qi channels is more refreshing :-)
cladamw_ has quit [Quit: Ex-Chat]
<wolfspraul>
wpwrak: what dremel bits do you think are commonly needed?
<wolfspraul>
just get a wide variety first? anything special to look out for? what are the ones you need the most?
<DocScrutinizer>
via drills
<DocScrutinizer>
via drills
<DocScrutinizer>
and
<DocScrutinizer>
via drills
<wpwrak>
first, you want drill bits. a variety of sizes, from maybe 3 mm down to as small as you can get
<wpwrak>
if you pick a specific wire diameter (for the vias), get lots of drill bits of that size. at least 10. you'll use them up quickly :)
<wolfspraul>
hardened? anything special to watch out for?
<DocScrutinizer>
then you might find a cutter disk useful
<wpwrak>
that was my next item ;-)
<wolfspraul>
if you use a dremel to cut the board outline - which bits?
<DocScrutinizer>
though I always found it's incredibly hard to get straight cuts with them
<wpwrak>
these discs some in a little cylindrical box. 20 or so.
<wolfspraul>
bbiab - lunch...
rejon_ has quit [Ping timeout: 272 seconds]
<wpwrak>
you also need the axis bit. with a screw at its end to fix the disc.
<wpwrak>
and then i'd add some bits for grinding and such.
<wpwrak>
dremels are usually available as kit with lots of drill bits, discs, etc. may be best to get one of these. then you have much of the basics covered. plus it creates a sense of empowerment, even if you'll only use 20% of all that stuff ;-)
<DocScrutinizer>
indeed
<wpwrak>
i would also get a flexible axis. that way, you don't have to move the heavy and bulky tool but only the much lighter axis
<wpwrak>
increases the chance of getting straight cuts really straight ;-)
<DocScrutinizer>
re steel wool: I now recall I always used "Wiener Kalk" or some other abrasive cleanser powder
<DocScrutinizer>
like AJAX, VIM (sic!) :-D
<DocScrutinizer>
ooh nope, it was ATA
<DocScrutinizer>
not AJAX
<wpwrak>
wolfspraul: regarding heating the crystal: just from the top. it should be better with also heating from the bottom, but i don't have a good solution for that
<DocScrutinizer>
warm e-oven plate
<wolfspraul>
got it all, thanks
infobot has quit [Ping timeout: 244 seconds]
<DocScrutinizer>
sigh, infobot is under permanent attack since some months now
<DocScrutinizer>
e-stove*
xiangfu has quit [Quit: Leaving]
phirsch has quit [Ping timeout: 244 seconds]
phirsch has joined #qi-hardware
phirsch has quit [Excess Flood]
phirsch has joined #qi-hardware
<whitequark>
DocScrutinizer: what's with infobot?
cladamw has joined #qi-hardware
<DocScrutinizer51>
offline every few days
<DocScrutinizer51>
some idiots running DOA attacks against it
xwalk has quit [Ping timeout: 244 seconds]
pabspabspabs has joined #qi-hardware
pabs3 has quit [Ping timeout: 272 seconds]
pabspabspabs is now known as pabs3
jurting has joined #qi-hardware
cladamw has quit [Ping timeout: 272 seconds]
cladamw has joined #qi-hardware
panda|x201 has quit [Ping timeout: 248 seconds]
panda|x201 has joined #qi-hardware
pabspabspabs has joined #qi-hardware
Textmode has quit [Ping timeout: 255 seconds]
pabs3 has quit [Disconnected by services]
pabspabspabs is now known as pabs3
cladamw has quit [Quit: Ex-Chat]
cladamw has joined #qi-hardware
cladamw has quit [Quit: Leaving]
cladamw has joined #qi-hardware
wolfspra1l has joined #qi-hardware
wolfspraul has quit [Ping timeout: 245 seconds]
cladamw has quit [Ping timeout: 252 seconds]
cladamw has joined #qi-hardware
xiangfu has joined #qi-hardware
pabs3 has quit [Quit: Don't rest until all the world is paved in moss and greenery.]
pabs3 has joined #qi-hardware
jivs_ has joined #qi-hardware
xiangfu has quit [Ping timeout: 272 seconds]
Openfree` has quit [Remote host closed the connection]
cladamw has quit [Ping timeout: 252 seconds]
cladamw has joined #qi-hardware
jurting has quit [Ping timeout: 260 seconds]
LunaVorax has joined #qi-hardware
<LunaVorax>
Hello everyone!
xiangfu has joined #qi-hardware
cladamw has quit [Quit: Ex-Chat]
xiangfu has quit [Remote host closed the connection]
Ayla has joined #qi-hardware
LunaVorax has quit [Quit: Quitte]
DocScrutinizer has quit [Disconnected by services]
DocScrutinizer has joined #qi-hardware
<wolfspra1l>
hi
<roh>
DocScrutinizer: bullshit
cladamw has joined #qi-hardware
<DocScrutinizer>
thought so
<roh>
DocScrutinizer: sean is the one not answering important questions. one thing is for sure. when stuff moves to the us, i will not lift another finger anymore.
<roh>
DocScrutinizer: also: all the data on openmoko.org servers is currently under german or bavarian data protection law. means: if the stuff moves to the us, i will wipe all MY personal data (mailboxes, homedirs etc.)
<DocScrutinizer>
could we talk to Harald, as he seems more down to earth with his idea about how the world works
<roh>
DocScrutinizer: so. be aware.
<DocScrutinizer>
roh: sure thing
<DocScrutinizer>
please wipe my mail dir as well
<DocScrutinizer>
if such a move is pending
<roh>
noted.
<DocScrutinizer>
I'm already working on migrating my main mail addr to some dedicated domain, just pondering how to implement proper spam protection
<roh>
DocScrutinizer: to be fair, i am not in the mood anymore to discuss with such a crowd anymore. no problem if somebody talks to me, or sends me a mail, but the bs going on is nothing i will continue anymore. gismo is basically out already (annoyed and not in the mood for unpaid bullshit)
<DocScrutinizer>
I can feel with you
<DocScrutinizer>
this rah guy killed my last nerve used for maintaining anything OM related
<DocScrutinizer>
he had proper help by... Sean?
<roh>
well.. if you ask for my opinion who was most influencial in killin om over time? sean himself
<DocScrutinizer>
ack
<roh>
just answering mails when it is important in less than a week would have been enough.
<DocScrutinizer>
what are you currently running for antispam on mail.OM ?
<DocScrutinizer>
or rather, what would you suggest if I were to set up my own mailserver?
<roh>
oeh.. if there is still some part of that working its spamassassin
<roh>
i dont have one on my own servers. i just filter the worst of it out by proper mail setup (whats not a real mailbox cant send me mail without being whitelisted) and the rest is filtered by my thunderbird after imap
<roh>
basically its a real simple exim + dovecot and saslauthd
jurting has joined #qi-hardware
cladamw has quit [Quit: Ex-Chat]
<wolfspra1l>
roh: how is raumfahrtagentur going, and the new space?
<wolfspra1l>
I'm really looking forward to visiting one day, maybe only after I move back to Germany though :-)
<wolfspra1l>
one by one I am acquiring some more usable hardware skills :-)
<wolfspra1l>
you and gismo still working there, and having a good enough stream of paid projects?
<wolfspraul>
I though fairwaves was doing pretty good work, didn't know about matlab
<wolfspraul>
maybe that's a compromise they are willing to make for the time being, but I am pretty sure from the people they would love to go all-open on the tools as well
* pabs3
wonders if octave is compatible enough
<wolfspraul>
sure just start using that and if needed improve it :-)
<wolfspraul>
a lot probably also has to do with inertia, I know this very well from logic board EDA tools
<wolfspraul>
young people < 20yr grow up with tool X or Z and after so many years in the industry, telling them to use something else feels like an amputation of sorts
<wolfspraul>
then they come up with all sorts of protective 'reasons' why their tool is so much better and the only one that meets "professional" standards, etc. etc.
<wolfspraul>
in reality there may not be much to it, but the degree of efficiency one can achieve with any complex tool over a number of years still precludes a change, for them
<wolfspraul>
I tend to just ignore it nowadays and start with the free tools, ignorance is bliss :-)
<wpwrak>
and free tools mean that they're not likely to die under your hands. vi, LaTeX, and make forever ! :)
emeb has joined #qi-hardware
<wolfspraul>
ah hi :-) I came up with one more question about the pcbs... what is the minimum trace width you feel comfortable with now?
<wpwrak>
hmm, lemme check ..
<wpwrak>
8 mil. leaves a bit of margin for error.
<wpwrak>
so that's ~0.2 mm
<wpwrak>
i use a clearance of 8 mil as well but i try to have a slightly wider spacing. e.g., 16 mil clearance for traces where space isn't extremely critical
<wpwrak>
at chips, the spacing may get tighter. e.g., QFN have a pitch pf 0.5 mm / 19.7 mil, sometimes also 0.45 mm / 17.7 mil. but then i fan out to a more relaxed spacing.
GNUtoo-desktop has joined #qi-hardware
<kristianpaul>
wpwrak: yeah 1V2 and 3V3 mix is not very nice tought..
phirsch has quit [Ping timeout: 272 seconds]
<roh>
wolfspraul: we are still here :)
<roh>
wolfspraul: its hard sometimes, but somehow we managed it so far. mostly it stuff.
<roh>
(which pays the bills)
panda|x201 has quit [*.net *.split]
rektide has quit [*.net *.split]
rektide_ has joined #qi-hardware
Jay7 has quit [Quit: Conversation terminated]
Jay7 has joined #qi-hardware
panda|x201 has joined #qi-hardware
rektide_ is now known as rektide
phirsch has joined #qi-hardware
LunaVorax has joined #qi-hardware
phirsch has quit [Read error: Connection reset by peer]
Jay7 has quit [Read error: Operation timed out]
kristoffer has joined #qi-hardware
phirsch has joined #qi-hardware
phirsch has quit [Read error: error:1408F119:SSL routines:SSL3_GET_RECORD:decryption failed or bad record mac]