[2016-11-14 Mon 17:43:55] oh, and they removed the 8 bytes of user ram/otp memory in ~33. i guess that's where the config bits for the additional luts are. [2016-11-14 Mon 17:44:46] i get about 7 nA noise Ipp with ~1 m leads crawling around on my bench [2016-11-14 Mon 17:58:02] ~33 looks pretty good. lemme check if that pin N -> ioN renaming did anything weird ... [2016-11-14 Mon 18:05:37] azonenberg thinks it might have been a die shrink (makes sense for nasty new STQFN-22 (2.0 x 2.2 mm) ), though your find re I2C lock bit also suggests they used that opportunity to do a bugfix [2016-11-14 Mon 18:06:38] maybe you want to make friends with them and earn reputation, by already sending them a bug report? [2016-11-14 Mon 18:06:55] so either they do a respin early, or they just fix their docs [2016-11-14 Mon 18:07:07] either way they'll love you for it [2016-11-14 Mon 18:07:37] and isn't that what we all want, being loved ;-) [2016-11-14 Mon 18:08:17] ooh, what about the lock bit in 33? [2016-11-14 Mon 18:08:40] ah yes, i *knew* there was a channel i had forgotten to put on auto-join. grmbl ... [2016-11-14 Mon 18:08:47] hehe [2016-11-14 Mon 18:08:57] ##openfpga [2016-11-14 Mon 18:09:19] yup [2016-11-14 Mon 18:09:43] azonenberg is on his way home, so afk right now [2016-11-14 Mon 18:09:49] ~33 has that lock bit, too [2016-11-14 Mon 18:22:42] pin mapping looks sane, too. i.e., all the acmp and such are where we have them in ~31. [2016-11-14 Mon 18:23:25] what the holy fsck!?!!! I place a new dir into /home/joerg/stuff/joerg/ (aka /srv/www/htdocs/stuff/...) and it *doesn't show up* in http://neo900.org/stuff/joerg/ [2016-11-14 Mon 18:24:17] wrong host ? [2016-11-14 Mon 18:28:16] hardly? [2016-11-14 Mon 18:28:30] oooh damn YES [2016-11-14 Mon 18:31:18] https://neo900.org/stuff/joerg/tmp/help/ [2016-11-14 Mon 18:31:27] legacy [2016-11-14 Mon 18:31:51] the best I could find for that 33 [2016-11-14 Mon 18:32:56] https://neo900.org/stuff/joerg/tmp/help/GP5/SLG46533V/ [2016-11-14 Mon 18:35:56] do we want to change from ~31 to ~33 now ? [2016-11-14 Mon 18:46:39] yes [2016-11-14 Mon 18:47:08] this been an oversight by me that we used 31, to start with [2016-11-14 Mon 18:47:38] https://www.google.de/search?q=site:http://www.silego.com/uploads/Products+SLG46533&prmd=ivns&filter=0 finds SLG46533r103_09272016.pdf but that file doesn't exist anymore [2016-11-14 Mon 18:47:51] I wonder if it would be in cache [2016-11-14 Mon 18:47:57] of google [2016-11-14 Mon 18:48:11] i checked for availability at mouser, and they don't have either. so, no regression :) and digi-key don't have silego yet either. but as long as the silego shop works ... [2016-11-14 Mon 18:49:35] HAH!! http://www.silego.com.cn/uploads/Products/product_482/SLG46533r102_08302016.pdf [2016-11-14 Mon 18:49:46] ;-)) [2016-11-14 Mon 18:50:35] nfc, still pretty new [2016-11-14 Mon 18:51:14] and silego with their damn REFERRER check and "dynamic" content creation... :-/ [2016-11-14 Mon 18:52:06] http://www.silego.com/uploads/Products/product_482/ doesn't show that file for me [2016-11-14 Mon 18:53:05] THAT's what I call nasty website management [2016-11-14 Mon 18:54:13] yeah, there seems to be a rule in "modern" website design that it's only good if it's at least passive-aggressive, if not overtly aggressive [2016-11-14 Mon 18:54:25] (31 vs 33) it didn't matter to me as long as sufficient LUTs available [2016-11-14 Mon 18:54:26] err, hostile [2016-11-14 Mon 18:57:37] http://www.silego.com/uploads/Products/product_482/SLG46533r100_07142016.pdf [2016-11-14 Mon 18:57:40] 404 [2016-11-14 Mon 18:57:47] however it existed [2016-11-14 Mon 18:57:53] R100 [2016-11-14 Mon 18:58:15] so the 33 is not all that new [2016-11-14 Mon 18:58:58] it however didn't get obsoleted like the 31, so I wonder wtf... [2016-11-14 Mon 19:01:27] it seems that the only real change between old and new is that i2c bit, and the ~33 has it [2016-11-14 Mon 19:04:44] I wonder how they possibly could have *knowingly* messed that up for the 31 only [2016-11-14 Mon 19:05:13] so they have to phase out and come up with 37 a <6 months later [2016-11-14 Mon 19:05:54] maybe the ~31 went to the fab a few days before they found and fixed it. and when they made that URGENT call, it was already too late :) [2016-11-14 Mon 19:06:10] funny enough it seems I read the 33 DS regarding those lock fuse bits [2016-11-14 Mon 19:06:51] so you think they delivered vulnerable crap to keep a deadline? [2016-11-14 Mon 19:07:02] they're almost identical. there's just one more bit in the ~33/37/etc., that's all [2016-11-14 Mon 19:07:47] hard to say. could also be that they fixed it without realizing the implications. [2016-11-14 Mon 19:07:55] I think what's new for sure is that nastypackage [2016-11-14 Mon 19:08:39] oh yes [2016-11-14 Mon 19:08:42] so they might indeed have spun a crappy interim batch since they knew they will take a few more months for die shrink [2016-11-14 Mon 19:09:25] and since the 31 was buggy bit the 33 wasn't, only the 31 got phased out now [2016-11-14 Mon 19:10:02] both with shrinked new die, but only 31 needed the bugfix, thus new number only for that one [2016-11-14 Mon 19:10:25] industry, you got to love it [2016-11-14 Mon 19:11:07] if you allow, I'kk pastebin this convo for azonenberg [2016-11-14 Mon 19:12:57] starting at [2016-11-14 Mon 17:43:55] oh, and they removed the 8 bytes [2016-11-14 Mon 19:14:02] if you don't like that, simply answer "no" [2016-11-14 Mon 19:14:05] :-) [2016-11-14 Mon 19:14:22] sure, go ahead [2016-11-14 Mon 19:14:26] ta!