#qi-hardware IRC log for Friday, 2014-03-21

DocScrutinizer05quiet earth?01:05
DocScrutinizer05that's kinda scary when even the gaby #openpandora turns silent01:06
DocScrutinizer05gabby even01:06
whitequarkDocScrutinizer05: I was sleeping :p04:58
whitequarkso, yes, those holes *are* weird04:58
whitequarkI fixed the PCB with the clamp attachment system of my own invention04:58
whitequarkI used to use masking tape but it was too elastic, as you've said04:58
wpwrakmaybe try a single slot cut and see how it turns out05:00
whitequarkI can say one thing, if I mill it CW instead of CCW, the direction in which those gaps point inverts05:00
wpwrakthen add two parallel cuts at, say, +/- tool_radius/2 05:00
wpwrakhmmm. that doesn't sound too good05:01
wpwraksomething is moving05:01
whitequarkI'm suspecting cutter radius compensation a bit05:01
whitequarkthough, not sure05:02
wpwrakmaybe. anyway, those test cuts shouldn't be affected by radius compensation05:03
whitequarkok. lemme check with a slot.05:09
whitequarkwpwrak: http://i.imgur.com/sGPEBfN.jpg05:16
whitequarkwithout any cutter compensation05:16
wpwrakhmm, on both ends05:17
wpwraksomething is moving05:17
wpwrakwhat would a perpendicular cut look like ?05:18
wpwrakand i wonder whether your clamps really hold it well enough. better: make a "bridge" where each clamp has two screws. you already made nice holes for just that ... :)05:19
whitequarkthat wouldn't really work well with arbitrarily sized material05:19
whitequark(well enough) I can't feel any movement when jerking workpiece with my hands as hard as I can05:19
wpwraklet's worry about that later :)05:20
whitequarkI'd rather move the whole machine than workpiece against table05:20
whitequarkunlike my previous setup with masking tape & etc05:20
wpwrakhow about movement along the Z axis ? can you move the end that's far from the screws ?05:21
whitequarklet me just fix it with 4 clamps05:22
whitequarkabsolutely no difference whatsoever05:24
wpwrakah well, at least know we know it's not that05:25
wpwrakperpendicular ?05:25
wpwrakand is this the entry or the exit hole ?05:26
whitequarkleft is entry05:27
whitequarkperpendicular: exactly same thing rotated by 90°05:27
wpwrakwhat happens if you make two cuts in the same run ?05:28
wpwrak(two parallel ones)05:28
whitequarkwell... you've seen that on board from yesterday, haven't you?05:28
whitequarkI'm not sure what you're looking for05:28
wpwraki wonder whether the head may have some play that gets "consumed" when you start cutting05:29
wpwrakbut yes, given yesterday's results, maybe not05:29
whitequarkso on entry and exit cuts, head for some reason moves to the right relative to cutter movement 05:30
wpwrakthat would suggest that the head or the spindle has some play no matter what05:30
whitequarkno, I think not!05:30
whitequarkI have an idea05:30
wpwrakspace aliens ?05:31
whitequarkrightmost cut is better, isn't it?05:33
whitequarkwhat do you think changed?05:33
wpwrakspeed ?05:34
whitequarkyeah, feedrate05:34
whitequarkit's not entry/exit holes which are screwed up05:34
whitequarkit's the cut itself05:34
wpwrakbafore/after ?05:34
whitequarkthis endmill is *long*, it's almost 5mm05:34
whitequarkso it deflects when I move it at 300mm/min05:34
wpwrakwhat diameter ?05:34
whitequarkand yes, this is definitely not a drill05:35
wpwrakgrr. wasn't "go slower" the first thing i suggested ? :)05:35
whitequarkwell... that's first thing one usually does :p05:35
whitequarkanyway, I should've used my favorite endmill, it's carbide instead of HSS and just 3mm long05:36
whitequarkthat one doesn't deflect half a millimeter05:36
wpwrakoh, that 5 mm critter is HSS ? maybe it looks like an endmill but is a broken drill ? :)05:37
whitequarkwell, at least now we know why were they so cheap05:38
wpwrakhope your time has very little cost :)05:38
whitequarkanyway, they work much better as drills than the drills I've got, and they're perfectly fine for cutting wood05:38
wpwrakyeah, wood is quite forgiving. and versatile, too. a bit of pinewood goes a long way ...05:39
DocScrutinizer05(([2014-03-21 06:34:22] <whitequark> it's the cut itself)) toldya06:03
whitequarkright, right. I'll just move back to lower feedrate and proper endmill06:03
whitequarktalking about that... need to make another attempt at plating06:04
whitequarkhttps://pbs.twimg.com/media/BjOAu72CUAAPVGi.jpg "after disconnecting external monitor"06:13
whitequarkhow does this thing even happen? LCD driver gone completely mad?06:14
whitequarkhm, what if I make a cross-section of the board? seems to be used in industry for examining things unaccessible to eye09:54
whitequarkhrm. no clue whether it is legitimately that bad, or milling it in half with an endmill made it worse.10:05
wpwrak(monitor gone wild) seems that is figured out where you keep he "acid" :)11:16
Action: whitequark giggles11:17
whitequarkwpwrak: latest try: http://imgur.com/9EMucAl,73saqVL,YVPIst4#011:17
whitequarkquite beautiful, isn't it?11:17
whitequarkesp 2nd hole in lowest row11:18
wpwrakthey look pretty smooth indeed11:20
whitequarkstill needs more work. but, the direction is right.11:21
whitequarkperhaps a bath with more carefully set pH and pCl¯ would work better.11:21
wpwrakdoes the copper rub off ?11:22
whitequarkrub off? you mean like by fingers?11:22
whitequarkno, this time it's as strong as factory-plated copper11:23
wpwrakgreat !11:23
whitequarkin fact I was curious and inserted an 1.0mm endmill used to drill those holes and turned it a few times inside11:23
whitequarkthe copper is still there, now just more shiny11:23
Action: whitequark *really* likes the color of freshly plated copper11:24
whitequarkoh by the way, it also solders really well11:25
whitequarkeven without tinning!11:25
wpwrakhmm, with that huge unbroken copper surface ? that should be nearly impossible to solder11:42
wpwrakwell, unlesss you preheat the board11:42
whitequarkwpwrak: nah, not very hard13:08
nicksydneywpwrak: completed KL25Z bootloader upgrade ..now can flash from Linux14:33
nicksydneywpwrak: are you able to use the serial from the SDA usb ?14:34
nicksydneyok found the serial looks like mine is called /dev/ttyACM014:48
nicksydneygot the sample code from here https://github.com/payne92/bare-metal-arm and the makefile covers everything from downloading the gcc to compiling and flashing..very convenient14:52
wpwrakwhitequark: perhaps i shouldn't ask about your soldering iron then :) does it even need electricity or does it come with an ex military power source with a "lifetime guarantee" ?15:08
wpwraknicksydney: congratulations. ah, it has serial ? never tried ;-)15:09
wpwrakyes, bare-metal-arm is very very convenient :)15:09
nicksydneywpwrak: yes it has serial...the sample from that github shows the accelerometer and touch values15:09
whitequarkwpwrak: a 30W iron I believe15:09
nicksydneythe serial is in the SDA USB15:09
wpwrakah, right. now i remember. briefly used it as well.15:10
nicksydneyok time to hit the bed 15:13
whitequarkwpwrak: very fascinating17:06
whitequarksince I stopped using NaOH for preparing copper surface and instead use isopropanol+5second etch, persulphate started to work really well17:07
whitequarkseems I inadvertently passivated it17:07
whitequarkalso, when persulphate works well, it emits a *ton* of SO2. I wanna a fume hood17:08
whitequarkpretty good irritant17:08
whitequarkgrmbl. 1 via out of 7 actually survived etching17:43
whitequarkI need better electrodes and better alignment of negative...17:44
wpwraki wonder what the industrial sequence for etching, plating, and drilling is19:38
wpwrakhmm ... "Conventional through-hole processing usually relies on 3 or more baths to prepare the hole walls for copper plating."19:38
wpwrak(and then they suggest their magic pixie dust to do it in one step)19:39
eintopfahhh another party night21:35
eintopfwpwrak: I get a 40421:37
wpwrakwhat kind of drink is this ?21:39
eintopfbecks gold, a drink made for womens... or for a pussy like me21:42
eintopfi mean with 404 your link21:43
eintopfmy next semester begins at monday...21:45
eintopfso this os my last weekendcin freedom21:45
eintopfand there are so many todos in 6lowpan21:46
wpwrakthe thinktink link ? should work21:47
wpwrakdoesn't seem to be too bad: http://blog.zeit.de/berlinjournal/?p=18421:50
eintopfok. seems that my phone cant parse it21:50
DocScrutinizer05>>CH29-0000-050Conductive plating ink - 50 grams (w/dispenser cap) $109.25 <<  hmm, sounds fair enough22:33
DocScrutinizer05I mean, how much of that stuff will get used by a thru-hole/via wall coted by a 0.005mm thick coating of that stuff? 50g will last a loooong way, I guess22:37
DocScrutinizer05wpwrak: prolly that magic pixie dust is worth each penny22:41
wpwrakdunno. i found the 3 baths intriguing.22:42
DocScrutinizer05not because of the "three kinds of carbon", not even because of the silver it contains, but merely because of the adhesive22:42
DocScrutinizer05or we all go the classical way again: http://www.electricalrivets.com/F6880/copper_rivets.html22:44
DocScrutinizer05will get a tad nasty for in-pad vias22:44
DocScrutinizer05but then, when you need in-pad vias, you also need 6 layers at very least22:45
DocScrutinizer05so this seems to be a non-issue for all that matters22:46
DocScrutinizer05actually I wonder if simple conductive ink to fill the complete via hole wouldn't over-accomplish the task we're discussing here, which seems to simply be: connect top and bottom copper layer traces of a 2-sided PCB22:48
DocScrutinizer05sure, won't result in proper thermal vias, but hey22:49
DocScrutinizer05for proper thermal vias you're probably better off with copper rivets anyway, even when comparing to via plating22:50
DocScrutinizer05seems I need a break - I just thought about dentists and amalgam filler22:53
DocScrutinizer05FSCK RoHS! ;-P22:53
wpwraki'm just waiting for the first exploding wire experiments :)23:06
DocScrutinizer05oooh, those been done ~30 to 40 years ago23:09
DocScrutinizer05pretty funny. Particularly when coating glass. It's amazing how strong it sticks to the glass23:10
DocScrutinizer05actually for such experiments unfiltered 220V mains is the most adequate power source. You should maybe use a fast 10A protective breaker23:11
DocScrutinizer05not that it usually would kick in, just in case you mess up stuff23:12
DocScrutinizer05finding the right switch for such stuff is quite an interesting task. You want it to be very fast on closing, with low ESR and no bouncing23:13
DocScrutinizer05surge killers are probably best for that task. the small ones with axial wires and a glass thingie of ~10mm diameter and 5mm height in the middle23:16
DocScrutinizer05trigger that critter with a spark of some kV, e.g. from car ignition23:17
DocScrutinizer05of course you rather want to use one for 110V rather than 220V, so it doesn't cut out immediately after the spark from ignition ended23:18
DocScrutinizer05moin dos123:32
--- Sat Mar 22 201400:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!