#qi-hardware IRC log for Wednesday, 2014-06-11

whitequarkwpwrak: so I switched to ferric chloride02:21
whitequarkas expected, it is *much* easier to work with than persulphate02:21
wpwrakwell, it probably etches faster but it's a dirty mess02:26
wpwrakbut now that you've mastered your fear of chloride, you can make the step to HCl. of course, you'll spend the rest of the week kicking yourself that you haven't chosen that from the start :)02:28
whitequarkit's not a dirty mess02:30
whitequarkI mean, why is it?02:31
whitequarkI put the PCB into a polypropylene box (mildly hydrophobic), put stirring stick inside, close it, put it on top of my magnetic stirrer, turn it  on.02:31
whitequarktake it out 20 min later02:31
wpwrakyou don't find FeCl messy ? all the black sludge and such ? and that you have to heat the whole mess ?02:32
whitequarkI don't heat it02:32
whitequarkblack sludge? um, where?02:32
wpwrakare you sure you got FeCl ? :)02:32
wpwraka red-brown liquid02:32
wpwraknot very transparent02:32
whitequarkI made a 50% solution02:32
wpwrakadd copper and you get black sediment. after just a bit of etching, the black goo dominates and you can't see anything02:33
whitequarkoh, I solve that simply02:33
whitequarkthe box is transparent. I illuminate it from the bottom02:33
whitequarkwhen the copper etches away, traces are visible on the bright background02:34
wpwrakjust try HCl and you'll finally know true happiness ;-)02:35
whitequarkha! now *that* is a mess, corroding everything in sight02:37
whitequarkand I never feared chloride, I *started* with FeCl302:38
whitequarkand if I even wanted to switch, I'd spend the rest of the week trying to smuggle HCl somewhere02:38
wpwrak(corrode) you don't have to use it to mark your territory ;-)02:39
wpwrakand wasn't the reaction: 1 bottle vodka + some rubles = 1 bottle HCl ? :)02:39
whitequarkI'm not good at bribing people02:40
whitequark(o°¡°o5 ;; what the fuck is up with this, I do not understand02:45
whitequarkI can't *quite* combine the holes in the PCB and the printed negative02:46
whitequarkeither the CNC or the printer is broken02:46
whitequarkit's within tolerances, i.e. the board will be fine, but it tickles my OCD02:46
wpwrakit's not a bribe. it's sharing vodka with a friend. a friend who does you a friendly favour.02:48
wpwrakif you bring two bottles, you can drink together. on your friendship :)02:48
whitequarkhm, seems like I kinda overpowered my UV leds02:51
whitequarkearlier I needed 180s of exposure, with the current I have set now it's 20...02:51
Action: whitequark shrugs02:51
whitequarkit's even convenient.02:52
wpwrakhow close to burning up are they ? :)02:52
whitequarkwell, in theory they're rated at 3W. I put in 5.6W02:53
whitequarkand since I turn them on just for 20s it's not really bad02:53
whitequarkwow, the solder mask not only doesn't develop, it's basically solid02:56
whitequarkI belatedly admit this may have been a slight overkill02:56
whitequarkI should rather expose it for like two seconds at this power02:57
whitequarkI mean, it's solid enough on *non*exposed areas that it takes significant force to puncture it at the places where it simply covers a hole02:59
whitequarkyou could probably not take it off from exposed areas with a jackhammer02:59
whitequark... oh. no. I'm a moron03:00
whitequarkI didn't take off the PTFE protective film on the top03:00
whitequarkhrm. I think my tinning solution got bad03:33
whitequarkit was yellow. now it's white.03:33
whitequarkalso, it doesn't work.03:33
Action: whitequark shakes the solution wildly03:36
whitequarknow it seems to work...03:36
Action: whitequark wonders if this is even going to work: http://i.imgur.com/jz26wdq.png03:40
whitequarkseems too simple to fuck it up, but who knows03:41
whitequarkhmm, I think I know what the unidentified "tin salt" is04:02
whitequarkit contains sulphur, though I'm not sure in which oxidation state. there's also a reducing agent in the solution, so anything could go to H2S04:02
whitequarkaha, I'm correct: http://books.google.co.uk/books?id=m8sJBIMtETgC&pg=PA318&lpg=PA318&dq=tinning+solution+composition&source=bl&ots=2_-hGNZbMf&sig=id1ijZTC8WRr9lWkrAcfpDGbNDQ&hl=en&sa=X&ei=CtWXU8OYMsTjOvCkgfAO&ved=0CCYQ6AEwAQ#v=onepage&q=tinning%20solution%20composition&f=false04:04
whitequarkwow, that book looks like something I want to get04:06
whitequarkit also explains through-hole plating!04:06
whitequarkwpwrak: http://lab.whitequark.org/notes/2014-06-11/producing-pcbs-using-photolitography/04:21
wpwrakhmm, shouldn't there be a diode across the relays ?04:24
wpwrak"simple and quick" ... a CNC machine :)04:25
whitequarkquick? sure04:26
whitequarksimple? depends on your definition of "simple"04:26
wpwrakwhat, no silk screen ? and where's the stencil ? :)04:38
whitequarkstencil? I have a process for stencils, but this is through-hole04:41
whitequarkand no, no silk-screen yet. I even have the UV paint ready, but not the mesh yet04:42
wpwrakand i hope you have either a few spare transistors or are adding that diode :)04:53
whitequarkoh, crap, diodes on relay coils04:59
whitequarkgrmbl, wonder whether I have some spare diodes05:01
whitequarkI have so many curious things stashed under the ceiling. for example, handcuffs05:04
wpwrakthe mistress will be pleased05:04
whitequarkthey appear to be the sex shop kind, yes05:05
whitequarkI wonder why the hell do I have them05:05
wpwrakmaybe it has something to do with that night you don't quite remember, but after you somehow got out off the cuffs, you stepped on an empty vodka bottle on the floor and almost broke your neck ?05:07
whitequarkgrmbl. no diodes. I swear I had some 1n4148's05:12
whitequarkwell but I do have a dozen of transistors, so... let's plug it in and see what happens :]05:12
whitequarkugh, doesn't work at all06:09
whitequarkI used a conveniently placed 24V tap on the board06:15
whitequarkhowever it turns out that the chinese used *one* halfperiod of input 18Vac to power the control board and *other* halfperiod to power the fan (which is what the tap was for)06:16
whitequarksince I no longer power it with ac... yeah06:16
whitequarkwpwrak: uhhh, wtf06:37
whitequarkit kind of works now, by which I mean... it switches on and stays on06:37
whitequarkuntil I power cycle it06:37
whitequarkthe BE voltage is .6V somehow (the base is floating in air) and the CE voltage is 8V06:51
whitequarkit's almost exactly BC547's BE on voltage. no wonder it remains in some indefinite state06:56
whitequarkhm, if I short base to ground, the CE voltage actually rises. wat?07:01
nicksydneyHackers: the Internet's immune system  -- https://www.youtube.com/watch?v=erCAp_Bd0AQ#t=100 .... oh yeah !12:57
wpwrakdamn. how i sometimes wish to have one of those ultra high-res 27" monitors ...16:10
DocScrutinizer05C-B diode semi-fried by relay recoil already20:36
DocScrutinizer05never switch inductive load without a clamp diode20:39
-:#qi-hardware- [freenode-info] please register your nickname...don't forget to auto-identify! http://freenode.net/faq.shtml#nicksetup20:40
wpwrakespecially not with a transistor :)20:40
DocScrutinizer05BC547 V(CEO)=45V abs max20:40
DocScrutinizer05adding a capacitor of maybe 1uF to B-E would be the only viable alternative20:41
DocScrutinizer05but really a poor alternative20:42
wpwraki think the "good" circuits have both20:42
DocScrutinizer05BC547 V(CBO)=50V - maybe use C-B diode for clamp, when you got so many of those BC547 ?20:43
DocScrutinizer05breaktrough/leakage on C-B diode makes transistor open up at 66mV and thus Vcollector goes down. Nice feedback20:46
DocScrutinizer05you managed to tune the BC547 from a V(CE) of 45V to 8V20:47
wpwrakthe circuit designer's companion suggests: diode or zener or RC20:47
DocScrutinizer05RC snubber, sure20:48
DocScrutinizer05or clamp diode20:48
DocScrutinizer05or Zener, HAHAHA20:48
DocScrutinizer05maybe when you got surplus zeners20:48
DocScrutinizer05B-E capacitor is a usual means to stretch rise/fall times20:50
wpwrakzener for preventing the switch to get kicked below the rail20:50
wpwrak.. which is often undesirable :)20:50
DocScrutinizer05downside of stretching slew rate is increased power dissipation in transistor20:52
DocScrutinizer05benefit is better EMI characteristics of your whole circuit20:52
DocScrutinizer05so yes, usually you want both: clamp diode and a maybe 10..100nF capacitor, depending on your needed slew rate20:53
DocScrutinizer05for a relay you can use 10uF even20:53
DocScrutinizer05unless that fries your transistor20:54
DocScrutinizer05with a clamp diode you don't need a Zener when you can make sure your VDD is << V(CE) + V(fwd) of clamp diode20:56
--- Thu Jun 12 201400:00

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