**** BEGIN LOGGING AT Tue Mar 07 03:00:02 2017
Mar 07 09:33:14 <rah> "Dr. Nikolaus Schaller has been pushing the limits of what you can do in a pure community project"
Mar 07 09:33:17 <rah> umm..
Mar 07 09:33:19 <rah> what the fuck?
Mar 07 09:34:31 <rah> plainly Mr Welte needs to be educated on the non-community nature of GTA04 hardware
Mar 07 10:23:41 <PaulFertser> pabs3: DocScrutinizer05: I'm told that Jason Kridner is sometimes available on #beagle and that he might know the right hw production engineers as they manage to produce BB xM with DM3730.
Mar 07 10:24:25 <DocScrutinizer05> wow, thanks. Who's that?
Mar 07 10:26:46 <PaulFertser> BeagleBoard.org Co-founder
Mar 07 10:27:26 <PaulFertser> There're other connections but they all had experience with older omap3, unfortunately.
Mar 07 10:28:39 <PaulFertser> Or you can write him an e-mail as listed on http://beagleboard.org/about . This page says he works for TI.
Mar 07 10:29:11 <PaulFertser> Not sure if that's a good sign or not. But the fact they are actually able to produce the boards gives some hope.
Mar 07 10:32:11 <PaulFertser> They're using the CBP version too.
Mar 07 10:32:28 <DocScrutinizer05> well, _been_ able, once. But yes. 
Mar 07 10:32:55 <PaulFertser> They have several revisions and have been able to use different RAM chips too.
Mar 07 10:33:18 <PaulFertser> Micron and Numonyx
Mar 07 10:33:35 <DocScrutinizer05> I don't buy the dependency on RAM PoP type
Mar 07 10:36:04 <DocScrutinizer05> https://irclog.whitequark.org/neo900/2017-03-01 
Mar 07 10:37:40 <PaulFertser> 12:33 < thinkfat> talking to my former colleague, he doesn't quite remember how we did the omap3 pop, but 
Mar 07 10:37:43 <PaulFertser>                   we didn't solder vapor phase but just reflow
Mar 07 10:37:45 <PaulFertser> 12:34 < thinkfat> he thinks we first soldered the ram onto the omap and then the whole stack onto the pcb
Mar 07 10:37:47 <DocScrutinizer05> https://irclog.whitequark.org/neo900/2017-03-01#18951868;
Mar 07 10:38:19 <DocScrutinizer05> HEHE, *exactly* what I spotted as the most likely solution
Mar 07 10:41:16 <PaulFertser> Yep
Mar 07 10:42:52 <PaulFertser> Isn't it too hard to reball omap and ram if testing goes wrong?
Mar 07 10:44:43 <DocScrutinizer05> well, it can get done. the question is if it's cheaper to use a new chip instead of reballing, particularly considering the time it takes
Mar 07 10:46:03 <DocScrutinizer05> http://www.ti.com/lit/an/swpa182c/swpa182c.pdf#19 "2-pass reflow: Pre-stack tray for FCPOP"
Mar 07 10:46:07 <PaulFertser> I see cell phone repair guys here do that in no time, heh, probably 30 minutes most. Not sure if they can do 0.4 though.
Mar 07 10:48:42 <DocScrutinizer05> >>* soldering them to a stack before production is not possible because we have no holder for it<< [Quote Nikolaus]
Mar 07 10:49:16 <DocScrutinizer05> I guess this "holder" is the magic bullet
Mar 07 10:49:50 <PaulFertser> I'm really glad to see you being optimisitic about this issue.
Mar 07 10:50:24 <DocScrutinizer05> though I fail to see why you couldn't do experiments with pre-stacking using whatever contraption to keep RAM PoP on SoC and the reflow that single package
Mar 07 10:50:29 <DocScrutinizer05> even kapton tape
Mar 07 10:50:59 <PaulFertser> Can kapton help with the warpage?
Mar 07 10:51:00 <DocScrutinizer05> and then*
Mar 07 10:51:22 <PaulFertser> I can understand how glue in between memory and soc can help :)
Mar 07 10:51:28 <DocScrutinizer05> no, but it could keep the SoC on top of the inverted RAM PoP
Mar 07 10:52:45 <DocScrutinizer05> actually I *think* when you invert the stack (PoP down) then you simply place the whole "sandwich" carefully "into the oven" to reflow and it doesn't need _any_ holder
Mar 07 10:54:08 <DocScrutinizer05> there should be more than enough factors (solder cohesion etc) that pull the SoC exactly into place during such top-down pre-stacking reflow
Mar 07 10:54:18 <DocScrutinizer05> no holder needed
Mar 07 10:54:52 <PaulFertser> Just as it's not needed for soldering a regular BGA chip to a PCB.
Mar 07 10:55:28 <DocScrutinizer05> on a more generic topic I also wonder if the drastic temperature profile in simple VP reflow wuld agravate the warp issues
Mar 07 10:55:42 <DocScrutinizer05> yep exactly :-)
Mar 07 10:56:48 <PaulFertser> I wonder what else Nikolaus might be doing when he talks about repairing the failed boards. Isn't it the most natural thing to do when manually assembling a board like that?
Mar 07 10:56:54 <DocScrutinizer05> a careful preheat phase prior to FP reflow is mandatory, I think
Mar 07 10:57:42 <DocScrutinizer05> VP*
Mar 07 11:00:20 <DocScrutinizer05> a pity Nikolaus isn't ever available via IRC. Email is such a PITA for discussing such stuff
Mar 07 11:04:23 <PaulFertser> China sells even 0.3mm bga stencils for reballing and afaict those cellphone repair guys can get fine results with a simple hot air station.
Mar 07 11:14:20 <DocScrutinizer05> for reballing of course. basically all you need is a stencil and right size solder balls
Mar 07 11:15:52 <DocScrutinizer05> http://www.ti.com/lit/an/swpa182c/swpa182c.pdf#44 >>Manual TMV Removal and Re-ball <<
Mar 07 11:21:09 <DocScrutinizer05> I'd answer to the dang ML (community@tinkerphones.org) but for unclear reasons I seem to not receive any mail from it despite been registered and registering again. And thus I also lack a thread ID to 'reply' to
Mar 07 11:39:47 <DocScrutinizer05> ~logs
Mar 07 11:39:47 <apt> All conversations are logged to http://apt.rikers.org/%23openmoko%2dcdevel/ Lines starting with spaces are not logged. Logs are updated daily  ---  livelogs are on http://logs.nslu2-linux.org/livelogs/openmoko-cdevel.txt
Mar 07 11:40:09 <DocScrutinizer05> fuck nslu, only livelogs of today
**** ENDING LOGGING AT Wed Mar 08 03:00:01 2017