| xiangfu | (find a more precise fab) yes. totally agree. for fast, I just find random one in taobao. then they send back the PCBs in 3 days. | 09:03 |
|---|---|---|
| xiangfu | (a U-shaped ground trace) yes. I also notice that after I sent out those gerber files. | 09:03 |
| xiangfu | (one silo cap ) I ignore all those caps in first try. :-) | 09:04 |
| xiangfu | the first try has totally messed. but I learn quiet some on KiCAD, gerber files, making PCB etc. | 09:05 |
| xiangfu | the bad news is I never success with oven. | 09:06 |
| xiangfu | I made a working board today with hot air.(2 times hot air) | 09:09 |
| xiangfu | after burned 7 PCBs and 2 chips. :-) | 09:09 |
| kyak | xiangfu: hi! i guess qt4 still didn't build on buildhost? | 09:23 |
| xiangfu | kyak, checking... | 09:23 |
| xiangfu | kyak, great. the qt4 stuff compile fine now. | 09:24 |
| xiangfu | checkout here: http://fidelio.qi-hardware.com/~xiangfu/openwrt-xburst.full_system/bin/xburst/packages/ | 09:24 |
| xiangfu | kyak, sorry for late reply. :) we have 17 qt4 package now. | 09:25 |
| xiangfu | kyak, I think I will update the images and packages, just starting a continue build | 09:27 |
| kyak | xiangfu: yeah, it makes sense to re-release the image, how do you think? | 09:45 |
| xiangfu | yes. agree. | 09:47 |
| xiangfu | building now. | 09:47 |
| wpwrak | xiangfu: (working board) wow, that's great. congratulations ! | 13:36 |
| xiangfu | wpwrak, I can load configure bits file to that board. get some result from the firmware. :) | 13:59 |
| qi-bot | [commit] Xiangfu: qi_lb60: add ks7010 support (release_2012-10-11) http://qi-hw.com/p/openwrt-xburst/3f62494 | 14:02 |
| qi-bot | [commit] Xiangfu: qi_lb60: add ks7010 support (master) http://qi-hw.com/p/openwrt-xburst/c3be281 | 14:02 |
| wpwrak | xiangfu: not a bad result at all. i thought it would take you a lot longer to get something that works. did you only use hot air or also heat from the bottom ? | 14:31 |
| xiangfu | wpwrak, I heat the pcb and chip in oven for ~3 minutes. then use hot air on top. | 14:34 |
| xiangfu | I set the hot air to ~350C. | 14:35 |
| xiangfu | ~3 minutes hot air on top of chip. | 14:35 |
| xiangfu | wpwrak, oven: no matter what high temperature I set, the TIN only get soft. it doesn't reflow at all, | 14:38 |
| xiangfu | if I stick >220C for more then 2 mins. the PCB start burn. :) | 14:39 |
| xiangfu | TIN: I mean the tin ball of the chip. | 14:39 |
| larsc | xiangfu: with flames and everything? | 14:44 |
| xiangfu | larsc, the PCB starts to change color(black). and smoke come out. | 14:45 |
| xiangfu | larsc, http://downloads.openmobilefree.net/tmp/IMG_1393.JPG <-- checkout. | 14:46 |
| viric | what is a 'tin ball'? | 14:50 |
| larsc | xiangfu: nice toast ;) | 14:58 |
| larsc | viric: the small dots on the back of the chip | 15:01 |
| viric | mh back is where text is, no? | 15:01 |
| viric | or that is front? | 15:01 |
| larsc | I'd call it front | 15:02 |
| viric | ah ok :) | 15:04 |
| viric | you are not imaging chips as bugs then | 15:04 |
| viric | you mean the balls on the belly of the chip ;) | 15:04 |
| whitequark | xiangfu: uh oh | 15:05 |
| whitequark | are you trying to use pb-free solder? :) | 15:05 |
| xiangfu | whitequark, you mean the solder on the PCB? | 15:08 |
| xiangfu | viric, yes. what is the correct name of those 'tin ball'? :) | 15:12 |
| wpwrak | just "balls" | 15:17 |
| wpwrak | and chips don't normally have "front" and "back" but "top" and "bottom" | 15:18 |
| whitequark | xiangfu: both on pcb and on the chip | 15:18 |
| wpwrak | xiangfu: using the oven to pre-heat is an interesting variation | 15:18 |
| viric | oh | 15:20 |
| viric | I thought that 'tin' was a metal closer to iron | 15:20 |
| xiangfu | whitequark, the chip is pb-free. the PCB I don't know. some random fab. | 15:20 |
| viric | as in 'tin can' | 15:20 |
| whitequark | pb-free solder seriously sucks if you don't have an oven with programmable profile | 15:20 |
| viric | But then I looked at the dictionary for the english word of the solder soft metal, and I discovered it's tin :) | 15:20 |
| whitequark | Í'd say that you can preheat the PCB at 170, or maybe 180 deg almost continuously | 15:21 |
| whitequark | if you go higher, it'll fry sooner or later | 15:21 |
| viric | whitequark: there comes into play the openPID? | 15:21 |
| whitequark | so, I preheat it to 180 and then apply the heat gun on top | 15:21 |
| whitequark | set to 320 | 15:21 |
| whitequark | for 20 seconds and no more | 15:21 |
| whitequark | works very good on LQFP, but might not work so good on BGA | 15:22 |
| wpwrak | viric: maybe you can recycle tin cans in a pinch. heat them on the stove until they melt, add lead to lower the melting point, stir well, voila, kitchen-made solder :) | 15:22 |
| viric | you don't cook in that oven later, right? | 15:22 |
| whitequark | viric: haha, indeed | 15:22 |
| whitequark | heavy metals are quite poisonous | 15:22 |
| viric | good :) | 15:22 |
| whitequark | well I don't have one yet | 15:22 |
| whitequark | didn't have much time lately. | 15:23 |
| xiangfu | whitequark, thanks for share. | 15:23 |
| whitequark | actually, I have yet to turn on my M1 for the first time :/ | 15:23 |
| xiangfu | whitequark, then I also have to use heat gun from now on. | 15:23 |
| whitequark | xiangfu: probably. it takes some practice, but after twenty fried pcbs/chips or so I was able to do it correctly each time | 15:24 |
| wpwrak | whitequark: why even bother with the oven for LQFP ? they're trivial with a soldering iron | 15:24 |
| whitequark | wpwrak: naw, I used the heat gun from the bottom for preheating and top for actual soldering | 15:25 |
| whitequark | for both desoldering and soldering | 15:25 |
| xiangfu | whitequark, (it takes some practice). I already fired 7 pcbs and 2 chips. I still need 10 burn chips class. :-) | 15:25 |
| whitequark | wpwrak: fine-pitch lqfps aren't easy | 15:25 |
| wpwrak | ok, desoldering is a different story | 15:25 |
| whitequark | .5 mm ones for example | 15:25 |
| whitequark | I find it way easier to do this: | 15:26 |
| whitequark | 1. cover each pad with an appropriate amount of tin with the iron and lots of RMA flux | 15:26 |
| whitequark | relying on surface tension forces here | 15:26 |
| whitequark | 2. more flux, place the chip correctly, apply heat gun and push it _very_ gently | 15:27 |
| whitequark | 3. if the thingy doesn't work, probably due to not heating it enough, use _dry_ iron on every pin again. | 15:28 |
| wpwrak | i wouldn't put a lot of tin on the pads. else you have to maintain heat for a long time to make sure all pins actually sink into the solder | 15:28 |
| whitequark | wpwrak: you cannot actually place "a lot" of tin on pads | 15:28 |
| wpwrak | well, in the Z direction :) | 15:28 |
| whitequark | there's an amount it will absorb, and it will try very hard to get this exact amount | 15:28 |
| whitequark | you obviously need to have soldering mask, and the pads should be quite small | 15:29 |
| wpwrak | i keep my tin flat (microns). then lots of water soluble flux (less messy than RMA) and always a "wet" iron | 15:29 |
| whitequark | wpwrak: with wet iron, I tend to get a lot of solder bridges | 15:29 |
| wpwrak | ah, don't have solder mask in my diy boards. that would make things too easy :) | 15:30 |
| whitequark | which are incredibly hard to remove without ruining the pcb | 15:30 |
| wpwrak | apply flux and "paint" over the pins, away from the chip. surface tension will take care of the bridges | 15:30 |
| Action: xiangfu will read logs tomorrow. have to sleep. | 15:30 | |
| whitequark | as per solder mask, I find it comparable by cost to just purchase the pcbs at fab | 15:30 |
| whitequark | when you include the cost of my time spent fixing everything which could be just avoided by having a mask, not drilling anything, not having to remake the PCBs, etc. | 15:31 |
| wpwrak | ah yes, drilling is a pain | 15:31 |
| Action: wpwrak <3 his cnc mill ;) | 15:31 | |
| whitequark | drilling is a huge PITA, positioning can be hard, chemical processes are not health-friendly and not easy either | 15:32 |
| whitequark | unless you're doing lots of simple boards and need really low cost and small turnaround time, I wouldn't really recommend the laser printer process | 15:32 |
| wpwrak | for me it's the turnaround | 15:33 |
| wpwrak | and yes, i try to keep them small | 15:33 |
| whitequark | you could pay $30 for a moderately complex board and have it in 3 days here in Russia | 15:33 |
| whitequark | with a russian fab | 15:33 |
| whitequark | moderately complex = 2/4 layer, mask, drilling | 15:33 |
| whitequark | under 10x10cm | 15:34 |
| whitequark | well maybe $50. I mean, per batch of 10. | 15:34 |
| wpwrak | so $50 for 10 boards or for each of that batch ? | 15:35 |
| whitequark | for 10 | 15:35 |
| wpwrak | ok, those are good prices | 15:35 |
| whitequark | hm, let me make sure I'm saying the correct price | 15:36 |
| whitequark | well no, I was not entirely correct. | 15:37 |
| whitequark | 2 layers, mask, silkscreen on both sides: $60 preparation, $12 per dm², 3 days. | 15:37 |
| wpwrak | so $180 for the batch. and you have to wait :) but yes, if you actually need ten boards, it's more convenient. i tend to make two only identical boards at a time. | 15:39 |
| wpwrak | then the next generation learns from my mistakes :) | 15:40 |
| whitequark | +$10 delivery, or you could pick it up yourself | 15:41 |
| whitequark | well yeah, there's a chinese fab which lets you have 10 boards less than 10x10cm for $25 total | 15:42 |
| whitequark | plus delivery | 15:42 |
| whitequark | if you use anything except airmail, it'll have comparable ETA and will be at least thrice cheaper | 15:43 |
| whitequark | they have some other nice offers like 4 layer 5x5 for $65, or 10x10 for $100 | 15:44 |
| whitequark | which is very competitive to other fabs I've seen | 15:44 |
| LunaVorax | Hello! | 21:09 |
| qi-bot | [commit] Paul Cercueil: Hide the 'params' option from the user on the contextual menu (packages) http://qi-hw.com/p/gmenu2x/0fc6ac6 | 21:38 |
| qi-bot | [commit] Paul Cercueil: Use the '%f' token to insert the selected file on the params. (packages) http://qi-hw.com/p/gmenu2x/e7e21b4 | 21:38 |
| qi-bot | [commit] Paul Cercueil: Build file extensions filter from the MIME types present on the OPK (packages) http://qi-hw.com/p/gmenu2x/5f1cff6 | 21:38 |
| pcercuei | Wayland 1.0 released | 23:48 |
| --- Tue Oct 23 2012 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!
