#milkymist IRC log for Saturday, 2011-04-16

rohhey wolfgang02:18
wolfspraulroh: hi02:19
wolfspraulI wanted to keep in touch over the cases...02:19
wolfspraulwe talked about L/ON/R already, that's clear02:20
wolfspraulmy next worry was the button tolerances on the inside of the case, maybe you saw it02:20
wolfspraulI don't have enough cases to look at though, you know better how you feel about it02:20
wolfspraulah OK, then I repeat02:20
wolfspraulso here is my concern:02:20
wolfspraulthe button is made out of 3 pieces now, the big one on the inside preventing the button to fall out, and pushing it against the electrical switch02:21
rohi am quite busy these days with the move.. my status is: i ordered the metal, and contacted the laser contractor02:21
wolfspraulbut I've had quite some problems with cases where after screwing them together, the pressure on the switch was so high that the buttons were stuck02:21
wolfspraulsure great, I totally trust the things is moving on your side02:21
wolfspraulI'm not worried about that02:21
wolfspraulI just want to give some more feedback02:21
wolfspraulthis one is about 'stuck buttons'02:21
rohhavent heard back from the laser guy, but he said on the phone already he will be on the road the next too weeks (about a week ago).02:21
wolfspraulthe thing is that we are talking about 0.1/0.2 mm issues, so I'm not sure whether we should change anything, or maybe that would make things even worse02:22
rohi think the 'stuck button' thing will go away or be 'different' with the better machine (laser optics is more precise)02:22
wolfspraulok. do you understand my concern first of all?02:23
wolfspraulI am also wondering about tolerances on the PCB side02:23
wolfspraulmaybe if we make 80 sets, we will have 5-10 or so where some buttons will just be stuck02:23
wolfspraulwe could try to find a thinner inner plastic piece02:23
wolfspraulor we could move the electrical switch on the pcb into the board a few mil02:24
wolfspraulbut then what we have now is not bad, so we can easily make something worse than it is now...02:24
rohso if my test runs will show some 'tolerance issues' i will widen the button holes some fractions of a mm02:24
wolfspraulI'm talking about the thickness of the larger plastic piece02:25
wolfspraulthe one that sits between the acrylic (on the inside) and the electrical switch02:25
rohthat one is 1.0mm.02:25
rohi use 'precision acrylic' there to get real 1.0mm, since regular one has so much tolerance02:25
wolfspraulshould we move the switch into the pcb a few mil?02:26
rohit also depends on the fitting of the sides.02:26
rohnot neccessary, i can move out the side panel a bit if neccessary02:26
wolfspraulok, that's a good point02:26
wolfspraulI'd rather not make mechanical changes on the pcb, otherwise it will get messy over time02:27
rohso i guess its easier to keep the button where it is, which would also make the cases 'more universal'02:27
rohi will redo all the mechanical tests anyhow when i do the testrun on the other machine. no worries02:27
wolfspraulso then just remember this issue, and let's keep an eye onit02:27
wolfspraulit's a pretty nasty problem once it hits the user02:28
wolfspraulthey will reassemble the case, tighten the screws: suddenly the box won't boot anymore02:28
wolfsprauland then it's hard to fix too02:28
wolfspraulI am also wondering whether the electrical switches loosen over time02:28
wolfsprauli.e. the pressure to make them connect will go down?02:28
wolfspraulthere is probably a tolerance in that pressure too, so some of them may click/connect with less pressure than others02:29
rohhm. dunno.02:29
wolfsprauloh well :-)02:29
rohi think the mechanical 'distance from not pushed to pushed' should be spec-ed anywhere02:29
wolfspraulideally in a 'unpressed state' (with button from outside), the force applied to the electrical switch should be exactly 002:29
rohand i would wonder if it would change much02:30
wolfspraulbut the distance between 1.0mm plastic and electrical switch should also be 002:30
wolfspraulof course that's very hard to achieve02:30
rohi could just 'make it a bit wobblier02:30
wolfspraulif there is a small gap, yes I guess it would be wobblier02:30
wolfspraulI think in many cases we do in fact exert some pressure on the electrical switch, even if the user does not press the buttons02:30
rohi did quite some tries to get it 'right' maybe tolerances in reality are just a bit bigger than the 2 pcbs i have02:31
wolfspraulI may also be looking at old cases02:31
wolfspraulmine are from December02:31
rohi tried this stuff with the rc1 mechanical prototype and the rc2 i have02:31
rohwill just give it some more slack and see what happens02:32
wolfspraulalso we have to make sure that there is no glue residue coming out from between the 2 smaller button pieces and the larger one02:32
rohi will try another glueing method02:32
wolfspraulbecause if there is, it is exactly at the point that it will move the 1.0mm one further into the case02:32
rohthere is some chemical which softens the surface of acryllic and it is intended to be used for glueing02:33
rohwill do some tests with it and see if i can pre-glue the stuff properly.. maybe by using some fixture (thinking of an metal with a hole, which would be unaffected by the chemical)02:34
rohTetrahydrofuran or Dichlormethan02:37
rohi think i have seen some of the latter when packing the 'special glue box' earlier02:37
wolfspraulwow yes, lots of options02:42
rohwill test and report what happens02:51
rohas soon as i find some time between packing moving and unpacking stuff and machines ;)02:57
rohstill tearing down walls and mounting electric circuits.02:57
lekernelwolfspraul: PCB tolerances are usually small... they need to when you have this kind of track density and component pitch13:45
wolfspraulyes true, I wasn't very precise - I meant the positioning of the electrical switch on the pcb, and tolerances within the switch itself13:55
wolfspraulbut roh already gave me some confidence, so we won't change anything on that side, if needed we can move the entire side out a little13:55
--- Sun Apr 17 201100:00

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