00:02.39 | DocScrutinizer05 | well, chances |
00:02.55 | DocScrutinizer05 | not decided upon yet |
00:03.15 | DocScrutinizer05 | depends on what WLAN/BT module we will get (that meets our needs) |
00:03.46 | ds2 | oh |
00:04.10 | ds2 | is there going to be an uncommited UART that is exposed? recall something about expansion in the FAQ |
00:05.49 | DocScrutinizer05 | UART pretty unclear, We a) have assigned all 3 UARTS iirc (IRDA for example), and when we go for 1GB RAM and NAND on bottom balls of SoC, then we lose one UART, one McBSP, and some GPIO. Not clear yet if we can do that at all |
00:06.21 | DocScrutinizer05 | if we got a UART to spare,we will route it to "expansion bay" |
00:06.31 | DocScrutinizer05 | just like I2C and power |
00:06.34 | DocScrutinizer05 | and USB |
00:06.40 | ds2 | so only I2C is sure to be there? |
00:07.16 | ds2 | not well integrated but if there is a bus, a BLE chip can be added |
00:09.59 | ds2 | who makes a MCP NAND/RAM combo with 1G of RAM? |
00:14.41 | *** join/#neo900 trench (trench@unaffiliated/trench) |
00:17.07 | DocScrutinizer05 | nobody, that's the problem |
00:17.16 | DocScrutinizer05 | unobtainium |
00:18.57 | ds2 | is there enough balls to get a 1G RAM only POP and route NAND through the regular balls? |
00:19.40 | DocScrutinizer05 | yes |
00:20.10 | DocScrutinizer05 | there's even a 1GB RAM PoP chip |
00:20.22 | ds2 | that's what I mean |
00:20.26 | DocScrutinizer05 | but we need to find a source |
00:20.29 | ds2 | I know 1G POPs exists |
00:20.48 | ds2 | is anyone besides micron willing to talk to you? |
00:21.11 | DocScrutinizer05 | http://projects.goldelico.com/p/neo900/issues/516/ |
00:21.23 | DocScrutinizer05 | micron not willing to talk to us, afaik |
00:21.42 | DocScrutinizer05 | at least not unless we buy a 10k or so |
00:22.01 | DocScrutinizer05 | digikey has MOQ:1k |
00:22.32 | DocScrutinizer05 | we need: 200 :-P |
00:22.42 | DocScrutinizer05 | well, 250 now |
00:23.05 | dos1 | DocScrutinizer05: yep; sorry, was afk |
00:23.13 | DocScrutinizer05 | OOOH! |
00:23.21 | DocScrutinizer05 | it's alive! \o/ |
00:23.36 | DocScrutinizer05 | ;-P |
00:23.55 | dos1 | funny chars, eh? ęóąśłżźćń - now you know why I want left part of space to be the compose key? :D |
00:24.09 | DocScrutinizer05 | yeah |
00:25.11 | ds2 | not giving much hope in the 1000 unit phase succeeding? |
00:26.08 | DocScrutinizer05 | well, 1k seems pretty possible. 200? :-S |
00:27.15 | dos1 | ds2: hope is one thing, the reality is another - we have to be prepared to both options |
00:27.21 | ds2 | asking people to break up trays is hard |
00:27.41 | ds2 | better to eat the extra part (so costs goes up 5x) |
00:27.52 | ds2 | i build small volume stuff and run into this :( |
00:29.43 | dos1 | with >500 it may be not that painful |
00:29.53 | dos1 | with just >200 however... |
00:30.14 | ds2 | what's the price at 1K pieces? |
00:30.37 | DocScrutinizer05 | 22$ |
00:30.40 | DocScrutinizer05 | EUR? |
00:31.09 | DocScrutinizer05 | steep, eh? particularly at *5 |
00:31.18 | ds2 | is this the same part used in the BeagleBoard xM? |
00:31.26 | DocScrutinizer05 | dunno |
00:32.04 | *** join/#neo900 pavi (~pavi@dslb-084-061-106-094.pools.arcor-ip.net) |
00:32.06 | ds2 | nope. BBX is 512M only |
00:32.16 | DocScrutinizer05 | http://www.digikey.com/product-search/en?x=21&y=16&lang=en&site=us&KeyWords=+MT46H256M32L4JV |
00:32.18 | ds2 | how much are the DM3730's? |
00:32.37 | DocScrutinizer05 | arind the same or 50% more iirc |
00:32.41 | DocScrutinizer05 | around* |
00:32.55 | ds2 | Hmmmm |
00:33.06 | DocScrutinizer05 | RAM: USD21.68775 |
00:33.18 | ds2 | wonder if anyone wants a BBX w/1G memory |
00:33.35 | ds2 | who's doing the assemble for you guys? |
00:33.46 | DocScrutinizer05 | sssshh! |
00:33.57 | DocScrutinizer05 | I dunno if I may disclose this |
00:34.21 | DocScrutinizer05 | but the same guys Nikolaus suggested to openpandora |
00:34.27 | ds2 | ahhhh |
00:34.29 | ds2 | I know them |
00:35.00 | ds2 | there are other guys around :D |
00:35.14 | xes | who? |
00:35.19 | xes | :) |
00:35.21 | DocScrutinizer05 | yes, but we found our way arounf with that particular company |
00:35.40 | DocScrutinizer05 | and the conditions are almost too good to be true |
00:35.54 | ds2 | small guys like folks I work with |
00:35.56 | DocScrutinizer05 | fab risk goes with fab, deferred payment |
00:36.19 | ds2 | they are based in the SF bay area |
00:36.26 | ds2 | the fab is $$$$$$$$ :( |
00:36.27 | DocScrutinizer05 | so we basically buy 100% yield |
00:36.46 | DocScrutinizer05 | aiui |
00:36.47 | ds2 | at the prices they want for it, it should be 120% yield |
00:36.53 | DocScrutinizer05 | HAHA |
00:37.24 | ds2 | I came within mm's of pulling the trigger on a custom DM3730 board build |
00:38.05 | DocScrutinizer05 | sorry? |
00:38.25 | ds2 | i had everything lined up for a custom DM3730 board build but the customer killed it at the last minute |
00:38.36 | DocScrutinizer05 | ouch |
00:38.37 | ds2 | fab was a major issue |
00:39.14 | DocScrutinizer05 | that's why we only build to order. And fab knows that, so we can give them a go with no upfront payment |
00:39.34 | ds2 | no that isn't the problem |
00:39.46 | ds2 | unless you order about 2000 or pieces, the fab is $$$$$$$$$$ |
00:40.46 | ds2 | have you considered an intermediate step where you rework N900 boards? |
00:40.50 | DocScrutinizer05 | yes, setting up and calibrating of the line is expensive, no matter if you run 200 or 200k devices after it works |
00:41.05 | DocScrutinizer05 | meh, not possible |
00:41.14 | ds2 | depop the OMAP3530, install a DM3730 + 1G RAM and redo the sysboot lines to boot from eMMC |
00:41.21 | ds2 | why's that? |
00:41.25 | DocScrutinizer05 | modem |
00:41.31 | ds2 | oh that #@%@$#@#@$ |
00:41.36 | DocScrutinizer05 | just to start with |
00:41.56 | ds2 | I have repop'ed beagles and they work fine |
00:42.05 | DocScrutinizer05 | sure |
00:42.09 | DocScrutinizer05 | can be done |
00:42.20 | DocScrutinizer05 | if you know your job |
00:42.46 | DocScrutinizer05 | yield like 40% on <10, maybe 70 on >100 |
00:43.30 | DocScrutinizer05 | I doubt you can get much higher than 70% yield |
00:43.55 | ds2 | better then that if the guys know what they are doing |
00:44.03 | ds2 | i had a BB done, no issues |
00:44.04 | DocScrutinizer05 | but anyway, no option for neo900 |
00:44.30 | ds2 | the main issue is the DM3730 wants an extra cap otherwise, it is a drop in |
00:44.41 | DocScrutinizer05 | maybe, yes |
00:44.54 | DocScrutinizer05 | no issue, we have our own PCB |
00:45.10 | DocScrutinizer05 | with our own nice LTE modem |
00:45.26 | DocScrutinizer05 | or combined UMTS&CDMA2000 |
00:45.33 | DocScrutinizer05 | (!!!) |
00:45.50 | ds2 | the power usage can be scary ;) |
00:45.58 | DocScrutinizer05 | with embedded GPS+GLONASS |
00:46.22 | ds2 | as long as the GPS is as sensitive (if not more so) then the n900 one |
00:46.33 | ds2 | most android GPSs suck |
00:46.36 | DocScrutinizer05 | whatever, any hints about sourcing HIGHLY appreciated |
00:46.54 | ds2 | sourcing GPS or what part? |
00:47.00 | DocScrutinizer05 | RAM |
00:47.02 | DocScrutinizer05 | NAND |
00:47.11 | DocScrutinizer05 | domesheets |
00:47.19 | DocScrutinizer05 | SMIA95 connectors |
00:47.37 | ds2 | Micron is your best bet |
00:47.46 | ds2 | the other folks won't even talk to you |
00:47.47 | DocScrutinizer05 | frikken little c-springs |
00:47.59 | ds2 | no idea what SMIA95's are |
00:48.19 | DocScrutinizer05 | molex 64pin connector OBSOLETE unobtainium |
00:48.29 | DocScrutinizer05 | latter one mission critical |
00:48.37 | ds2 | for the display? |
00:48.47 | DocScrutinizer05 | http://projects.goldelico.com/p/neo900/issues/515/#ic1521 |
00:48.51 | ds2 | wonder if hirose or samtec has a part that will cross |
00:48.52 | DocScrutinizer05 | MOLEX yes |
00:49.11 | DocScrutinizer05 | they all look a tiny bit different each time |
00:50.08 | DocScrutinizer05 | damn, not even a ticket for the molex |
00:50.17 | ds2 | ouch on the camera |
00:50.24 | DocScrutinizer05 | yeah |
00:50.32 | DocScrutinizer05 | "defined by Nokia and ST" |
00:50.37 | DocScrutinizer05 | KTNX! |
00:51.11 | ds2 | are you talking to micron directly on the memory? |
00:51.15 | ds2 | or just via digikey? |
00:51.24 | DocScrutinizer05 | Nikolaus doing that, I dunno |
00:51.56 | DocScrutinizer05 | should I point him to something? |
00:52.12 | DocScrutinizer05 | some not so known channel into micron, or sth? |
00:52.20 | ds2 | talk to avnet or arrow |
00:52.34 | DocScrutinizer05 | I think they said "sorry" |
00:52.40 | DocScrutinizer05 | avnet_ at least |
00:52.55 | ds2 | wtf? other folks have done similar volumes through them |
00:53.04 | DocScrutinizer05 | avnet is always-innovating no? |
00:53.11 | ds2 | NO! |
00:53.23 | ds2 | avnet is a major US electronics distributior |
00:53.26 | ds2 | like arrow |
00:53.30 | DocScrutinizer05 | aaah avnet, arrow. Sorry |
00:53.50 | DocScrutinizer05 | yeah, I had a pointer to avnet from Nikolaus |
00:54.03 | ds2 | is this done as a company or just a loose groip of people? |
00:54.03 | DocScrutinizer05 | he said "sometimes they can help" |
00:54.20 | DocScrutinizer05 | Golden Delicious GmbH |
00:54.37 | DocScrutinizer05 | to avnet |
00:54.43 | DocScrutinizer05 | to $sourcing |
00:55.02 | ShadowJK | Does the memory fit on any other SoC? |
00:55.16 | ds2 | no |
00:55.19 | DocScrutinizer05 | hardly, LPDDR(1) |
00:55.27 | ds2 | LPDDR1 is universal |
00:55.34 | ds2 | but the POP form factor is TI specific |
00:55.42 | DocScrutinizer05 | yes, but you don't want it on OMAP4 |
00:56.06 | ds2 | universal as in many SoC's will use it |
00:56.12 | DocScrutinizer05 | we could get a LPDDR2 RAM chip |
00:56.15 | ds2 | the AM33x will, the Freescale stuff will |
00:56.30 | ds2 | LPDDR2 is a diff voltage, IIRC |
00:56.32 | DocScrutinizer05 | even on PoP168 |
00:56.36 | DocScrutinizer05 | yes |
00:56.42 | ShadowJK | So basically only people building stuff with TI omap3 would use it? |
00:56.43 | DocScrutinizer05 | and differend CAS etc |
00:56.50 | ds2 | and then you need to hack MLO |
00:57.07 | ds2 | TI OMAP3/AM35x, DM37x |
00:57.21 | DocScrutinizer05 | :nod: |
00:57.32 | DocScrutinizer05 | afk bbl |
00:58.51 | DocScrutinizer05 | re |
00:59.53 | DocScrutinizer05 | Nokia used a damn phantom chip 1GB-RAM&NAND PoP in N9 |
01:00.08 | DocScrutinizer05 | aunt google never heard of it |
01:00.15 | DocScrutinizer05 | custom made |
01:01.16 | DocScrutinizer05 | I searched my fingers bloody |
01:02.52 | DocScrutinizer05 | you only learn about such negilgible details when you actually start sourcing, while before you do you think "oh my, there's a device that HAS DM3730 with 1GB RAM" |
01:03.47 | DocScrutinizer05 | friggin molex connector is the killer |
01:04.30 | DocScrutinizer05 | took me like two days to find the type, then another 2 days to finally accept it's *very* hard to source, or maybe impossible |
01:05.08 | DocScrutinizer05 | fekken 50ct conector killing the project - FIIINE |
01:06.01 | DocScrutinizer05 | again, we might build 1000 of those critters |
01:06.13 | DocScrutinizer05 | but not 200 |
01:06.36 | DocScrutinizer05 | even 1000 for sure cost more than gold |
01:06.47 | DocScrutinizer05 | 200: nogo |
01:08.39 | DocScrutinizer05 | http://www.digikey.com/product-detail/en/5000246471/5000246471-ND/2818746 |
01:09.08 | DocScrutinizer05 | (please check back to confirm part number before sourcing 5000 for us!) |
01:09.19 | DocScrutinizer05 | I might have picked the wrong half |
01:11.42 | DocScrutinizer05 | alibaba bunglers announce MOQ:100 and then tell "Sorry, onsolete, VERY hard to find" |
01:12.01 | DocScrutinizer05 | even obsolete |
02:05.07 | ds2 | who made that 1G chip? |
02:05.14 | ds2 | micron? elpida? samsung? |
02:20.13 | DocScrutinizer05 | the unobtanium N9 chip? |
02:20.36 | DocScrutinizer05 | I seem to recall it been elpida |
02:23.29 | ds2 | oh them |
02:23.47 | ds2 | those folks won't talk to us little fish :( |
02:24.36 | DocScrutinizer05 | http://wstaw.org/m/2013/12/05/plasma-desktopVg3743.png |
02:25.50 | DocScrutinizer05 | 110J1H0K |
02:25.50 | dos1 | wow, they use their nokia font even for schematics? :D |
02:25.59 | ds2 | definitely not micron |
02:27.33 | DocScrutinizer05 | I already considered buying broken N9 to slaughter them ;-P |
02:28.03 | DocScrutinizer05 | *of course* this chip also not available as spare part for N9 repair |
02:28.34 | ds2 | could it be a preelpida bankruptcy part? |
02:28.46 | DocScrutinizer05 | nfc |
02:29.08 | DocScrutinizer05 | all I have is that picture I pasted above |
02:29.32 | DocScrutinizer05 | even elpida been concluded by somebody else from (e1) |
02:29.39 | DocScrutinizer05 | (el)? |
02:30.18 | DocScrutinizer05 | iirc |
02:30.34 | DocScrutinizer05 | might as well have been some other guessing of that guy |
02:31.10 | DocScrutinizer05 | I have a hard time remembering *all* the useless facts and links I gathered during last 2 or 3 months |
02:31.18 | ds2 | there is a code you can read from the memory itself |
02:31.21 | ds2 | that will id the mfg |
02:31.27 | DocScrutinizer05 | oooh |
02:31.51 | DocScrutinizer05 | should that code show up in dmesg? |
02:31.57 | ds2 | nope |
02:32.06 | DocScrutinizer05 | thought as much :-/ |
02:32.07 | ds2 | look at the datasheet |
02:32.17 | ds2 | the n9 is a DM3730? |
02:32.18 | DocScrutinizer05 | datasheet? of what? |
02:32.27 | DocScrutinizer05 | err omap yes |
02:32.28 | ds2 | of the LPDDR1 |
02:32.34 | ds2 | no, omap or dm3730? |
02:32.42 | ds2 | they have different ram controllers, IIRC |
02:32.51 | DocScrutinizer05 | omap3630 I think |
02:33.08 | ds2 | that's a 3730 controller |
02:35.10 | ds2 | it uses a variation of the EMIF IIRC |
02:35.16 | ds2 | the 3530 uses a SDRC |
02:36.12 | DocScrutinizer05 | it's definitely a 36/37 cpu |
02:36.43 | DocScrutinizer05 | since it has "better" cpu than N900 |
02:36.54 | DocScrutinizer05 | which is 3430 |
02:37.01 | ds2 | 3430 == 3530 |
02:37.07 | DocScrutinizer05 | :nod: |
02:37.19 | ds2 | both of them HS parts? |
02:37.40 | DocScrutinizer05 | dunno, don't think so but that's only guesswork |
02:37.53 | ds2 | the N900 is HS, IIRC |
02:38.35 | DocScrutinizer05 | but the 3530 TRM has info about HS iirc |
02:38.39 | ds2 | just for stats, how many N900s shipped? 1million? |
02:38.46 | DocScrutinizer05 | no idea |
02:38.48 | ds2 | the HS parts need a special key for MLO |
02:38.55 | DocScrutinizer05 | yep |
02:38.58 | ds2 | more then 10K units? |
02:39.02 | DocScrutinizer05 | sure |
02:39.13 | ds2 | so 1K neo's is just 1% or less |
02:39.16 | DocScrutinizer05 | we still have ~30k active devices :-) |
02:39.32 | DocScrutinizer05 | last time we checked |
02:39.37 | ds2 | wonder if it make sense to shoot for 2000 boards |
02:39.48 | ds2 | large volume == better chance of getting parts |
02:39.50 | DocScrutinizer05 | how? |
02:40.01 | ds2 | parts go down by volume |
02:40.10 | DocScrutinizer05 | no $$ in kitty |
02:40.32 | DocScrutinizer05 | a feeble 50k, almost |
02:40.33 | ds2 | what is each unit costing right now? |
02:40.46 | ds2 | 50K seems a lot for 200 units |
02:40.58 | DocScrutinizer05 | we reckon a 700+ right now, for the bare board |
02:41.42 | ds2 | is there a spreadsheet of part costs you can share? |
02:41.44 | ds2 | seems high |
02:42.01 | DocScrutinizer05 | may go down when we complete phase5 (1k units) |
02:42.09 | ds2 | the main costs should be CPU/memory/modem, right? |
02:42.21 | DocScrutinizer05 | ds2: ^we don't even have such spreadsheet yet |
02:42.30 | ds2 | at 2K unit, your fab costs should be sub $5 if it isn't already |
02:43.14 | DocScrutinizer05 | yeah, might be correct |
02:43.48 | ds2 | memory + processor is about price of a beagle board (~$150) |
02:43.55 | DocScrutinizer05 | I have no data about that stuff, Nikolaus handles that |
02:43.59 | ds2 | so where is the other 500 or so going to? |
02:44.11 | DocScrutinizer05 | 100+ is modem |
02:44.16 | ds2 | whoa |
02:44.23 | ds2 | you guys buying them retail? |
02:44.26 | DocScrutinizer05 | PCB |
02:44.41 | ds2 | $100+ is like in singles |
02:45.04 | DocScrutinizer05 | there's not much difference between 1 and 100 modems |
02:45.24 | ds2 | 1 vs 2K should have a noticeable difference |
02:45.35 | DocScrutinizer05 | yes, that's why phase5 |
02:45.54 | ds2 | $100 still seems steep |
02:46.10 | DocScrutinizer05 | please look at http://neo900.org/#donate |
02:46.37 | dos1 | I think we should somehow better estimate what the price drop will be with 1k devices than "significantly" |
02:47.00 | DocScrutinizer05 | I think Nikolaus will not do that |
02:47.30 | DocScrutinizer05 | he refuses to come up with any hard numbers yet, and I know why |
02:47.43 | ds2 | is this a for-profit venture? |
02:47.49 | DocScrutinizer05 | not really |
02:48.04 | dos1 | sure, I don't expect hard numbers |
02:48.20 | DocScrutinizer05 | lemme put it this way: I hope for 5000 units to earn some money for the time I invested |
02:48.54 | DocScrutinizer05 | otherwise I could make more money as a coder monkey |
02:49.02 | ds2 | think the 200 unit goal is a lost cost |
02:49.08 | ds2 | lost cause |
02:49.31 | DocScrutinizer05 | hmm, how thatß |
02:49.35 | DocScrutinizer05 | ? |
02:49.40 | ds2 | bom cost too high |
02:49.53 | DocScrutinizer05 | that's why it will cost 700 |
02:50.05 | ds2 | you might be able to build 1K unit with a 60K or so kittle |
02:50.10 | ds2 | kitty |
02:50.19 | DocScrutinizer05 | o.O |
02:50.37 | DocScrutinizer05 | don't see that |
02:50.38 | ds2 | i believe a sub 500 BOM cost is possible |
02:50.52 | dos1 | there are about 300 gta04s produced |
02:51.02 | ds2 | no offense to digikey but avoid them |
02:51.10 | ds2 | digikey pricing will lead you astray |
02:51.49 | *** join/#neo900 Neros (~quassel@24-55-190-109.dsl.ovh.fr) |
02:52.59 | DocScrutinizer05 | 1k batch for 60k (even EUR) is quite a dream |
02:53.46 | ds2 | oops math error |
02:53.57 | dos1 | 60k / 1k = 60? |
02:54.35 | dos1 | we should easily get 1k orders for 60 eur :) |
02:55.09 | DocScrutinizer05 | even when we sell the board for 250EUR then, it's still a snap |
02:55.46 | DocScrutinizer05 | and that would be WAAAY more per board for *me* than I ever hoped for |
02:56.50 | dos1 | 190k profit? looks great for me :) |
02:57.29 | dos1 | ds2: please elaborate :D |
03:02.29 | dos1 | ds2: but now seriously, I think it wouldn't hurt to have some mail conversation between you three (you, Doc and Nikolaus) about all this sourcing and pricing stuff maybe? |
03:03.57 | dos1 | for my unexperienced eye you seem quite experienced, and I guess some outside PoV may be helpful :) |
03:04.40 | DocScrutinizer05 | for now we don't need mail, this stuff can get discussed in the open - we're not disclosing any secrets when ds2 tells us where to source cheap components (or where to source at all those components) |
03:05.03 | dos1 | DocScrutinizer05: the alternative is to get Nikolaus there :) |
03:05.15 | DocScrutinizer05 | ooh, right |
03:05.22 | DocScrutinizer05 | that's actually a problem |
03:06.03 | dos1 | it can be public mailing list maybe even |
03:06.50 | DocScrutinizer05 | this kind of dialog is only possible in chat. email is like writing letters |
03:07.00 | dos1 | huh, right |
03:07.33 | dos1 | anyway, it's pretty late, need to get some sleep |
03:07.38 | dos1 | cya! |
03:07.49 | DocScrutinizer05 | I need to get some milk, and a pretzel |
03:07.54 | DocScrutinizer05 | cya, dos1 |
04:58.35 | *** join/#neo900 Herbstb3rt (~Herbstber@drms-4d0bd5fe.pool.mediaWays.net) |
05:51.54 | ds2 | be glad to help if I can |
05:52.01 | ds2 | usually busy doing other things |
06:01.19 | *** join/#neo900 anYc (anyc@home.hadiko.uni-karlsruhe.de) |
06:22.11 | *** join/#neo900 Mistah_Darcy_ (~MistahDar@71-9-57-207.dhcp.gvrb.ca.charter.com) |
06:32.41 | *** join/#neo900 Mistah_Darcy (~MistahDar@71-9-57-207.dhcp.gvrb.ca.charter.com) |
06:39.55 | *** join/#neo900 lufu (~user@5.254.134.56) |
06:59.38 | *** join/#neo900 anYc (anyc@home.hadiko.uni-karlsruhe.de) |
07:05.37 | *** join/#neo900 Taduro_ (~quassel@iesvidalibarraquer.xtec.cat) |
08:26.52 | DocScrutinizer05 | duh, been busy for 2h reading washington post and comments |
08:54.38 | *** join/#neo900 infobot (~infobot@rikers.org) |
08:54.38 | *** topic/#neo900 is http://neo900.org | conversations are logged to http://infobot.rikers.org/%23neo900/ | 2013-11-04 - the day our fundraiser reached its goal, 25k EUR | 2013-12-02 - 200 devices reached! | Now heading for 1000 devices to build! | latest news: OHSW live https://www.youtube.com/channel/UCP_U91ipVWsF_2EiG3s6BKA |
08:54.38 | *** mode/#neo900 [+v infobot] by ChanServ |
09:02.11 | *** join/#neo900 infobot (~infobot@rikers.org) |
09:02.11 | *** topic/#neo900 is http://neo900.org | conversations are logged to http://infobot.rikers.org/%23neo900/ | 2013-11-04 - the day our fundraiser reached its goal, 25k EUR | 2013-12-02 - 200 devices reached! | Now heading for 1000 devices to build! | latest news: OHSW live https://www.youtube.com/channel/UCP_U91ipVWsF_2EiG3s6BKA |
09:02.11 | *** mode/#neo900 [+v infobot] by ChanServ |
09:02.47 | *** join/#neo900 sequantz (~sequantz@194.11.24.44) |
09:38.47 | *** join/#neo900 archtimmy (~archtimmy@93-103-81-41.static.t-2.net) |
10:41.41 | *** join/#neo900 luka (~luka@62.113.223.155) |
10:42.53 | *** join/#neo900 luka (~luka@unaffiliated/luka) |
10:44.21 | *** join/#neo900 dos1 (~dos@unaffiliated/dos1) |
12:13.21 | *** join/#neo900 dos1 (~dos@unaffiliated/dos1) |
12:34.19 | *** join/#neo900 archtimm1 (~archtimmy@93-103-81-41.static.t-2.net) |
12:37.48 | *** join/#neo900 21WABU2IG (~quassel@6.Red-79-145-84.dynamicIP.rima-tde.net) |
12:43.57 | *** join/#neo900 Taduro_ (~quassel@6.Red-79-145-84.dynamicIP.rima-tde.net) |
13:15.13 | *** join/#neo900 Mistah_Darcy_ (~MistahDar@71-9-57-207.dhcp.gvrb.ca.charter.com) |
13:38.36 | *** join/#neo900 edgar2 (~edgar2@80-186-133-74.elisa-mobile.fi) |
14:10.05 | *** join/#neo900 Neros (~quassel@24-55-190-109.dsl.ovh.fr) |
15:27.21 | skrzyp | DocScrutinizer05: http://www.spidersweb.pl/2013/12/neo900.html |
15:28.20 | skrzyp | DocScrutinizer05: It's a crappy pseudo-IT polish blog, but have lots of visitors |
15:29.28 | skrzyp | I don't send their any request, they probsly wrote this article upon comments on my Wykop link. |
16:29.54 | *** join/#neo900 NIN101 (~NIN@p5DD2834F.dip0.t-ipconnect.de) |
16:42.13 | *** join/#neo900 hbib (~wurmel@pD9E0D245.dip0.t-ipconnect.de) |
16:45.57 | *** part/#neo900 dos1 (~dos@unaffiliated/dos1) |
16:52.13 | *** join/#neo900 jake42 (~jake42@webcache.rommel.stw.uni-erlangen.de) |
17:07.12 | *** join/#neo900 Taduro__ (~quassel@6.Red-79-145-84.dynamicIP.rima-tde.net) |
17:15.42 | *** join/#neo900 Taduro__ (~quassel@6.Red-79-145-84.dynamicIP.rima-tde.net) |
17:27.53 | *** join/#neo900 Taduro__ (~quassel@6.Red-79-145-84.dynamicIP.rima-tde.net) |
17:38.50 | *** join/#neo900 dos1 (~dos@unaffiliated/dos1) |
18:05.14 | *** join/#neo900 freemangordon (~freemango@46.249.74.23) |
18:24.32 | *** join/#neo900 sequantz (~sequantz@194.11.24.44) |
19:23.33 | *** join/#neo900 kolp (~quassel@212.255.118.236) |
19:41.39 | *** join/#neo900 Neros (~quassel@24-55-190-109.dsl.ovh.fr) |
20:18.01 | *** join/#neo900 lufu (~user@5.254.129.23) |
20:46.13 | *** join/#neo900 hbib1 (~wurmel@pD9E0D428.dip0.t-ipconnect.de) |
21:44.59 | *** join/#neo900 xes (~xes@unaffiliated/xes) |
22:25.40 | *** join/#neo900 chomwitt (~chomwitt@athedsl-4385504.home.otenet.gr) |
22:38.46 | *** join/#neo900 xes_ (~xes@unaffiliated/xes) |
22:53.15 | DocScrutinizer05 | howdy dos1 |
22:54.21 | DocScrutinizer05 | skrzyp: thanks! |
22:55.42 | FIQ | checkboxes at the donate page means how much progress has been done not by backers, but by the project leaders I guess? |
22:57.58 | DocScrutinizer05 | FIQ: that's a terrible mess I'm not happy with. Originally it been meant to show which donation campaign phase we're in. But recently Nikolaus and dos1 came up with an idea to also cram in some indication on how far internal development progressed - something I think isn't simple to answer in a long explanation and absolutely not compatible with the donation phases |
22:58.41 | FIQ | mhm |
22:59.32 | dos1 | well, not really "how far internal development progressed" |
22:59.50 | dos1 | I rather saw it as a very convienient way to link to some results |
22:59.58 | dos1 | like last update |
23:00.04 | DocScrutinizer05 | for some unclear reason the idea came up that with recent presentation of a mech prototype somehow "phase-1" *development* got accomplished - something I ABSOLUTELY disagree |
23:00.31 | dos1 | it's not "finished, done, never touch again" |
23:01.02 | DocScrutinizer05 | neo900.org/#donate is pretty unclear reagrding thaz |
23:01.08 | DocScrutinizer05 | that* |
23:01.25 | FIQ | if phase 1 internal work isn't done, which I assumed by that page |
23:01.29 | FIQ | then the info is misleading |
23:01.41 | DocScrutinizer05 | I'd prefer to roll back any "shiny" or whatever attribute suggesting any form of development progress indicator |
23:02.09 | FIQ | and should not be there |
23:02.10 | DocScrutinizer05 | FIQ: absolutely |
23:02.49 | DocScrutinizer05 | development doesn't map to a phases model |
23:03.44 | dos1 | "Thanks to reaching 5 000 EUR we expect that we can do the schematics, component identification and initial mechanical PCB design." |
23:03.46 | FIQ | if you want to inform progress, it's probably better to try to elaborate shortly on what was just finished, and what you're going to do now (something which you already do by "news") |
23:03.58 | dos1 | well, right, schematics are not really there yet |
23:04.14 | DocScrutinizer05 | yes, and two of those 3 points are far from being finalized |
23:04.39 | DocScrutinizer05 | FIQ: ack |
23:05.05 | dos1 | initial mechanical PCB design is there, was ordered, was presented on OHSW |
23:05.09 | dos1 | isn't that it? |
23:05.17 | DocScrutinizer05 | no |
23:05.31 | dos1 | then what will be the "initial design"? |
23:05.38 | DocScrutinizer05 | it's ONE step of initial mech verification |
23:06.15 | DocScrutinizer05 | there need to be more of those PCBs with identified flaws squashed |
23:06.48 | dos1 | sure, but those aren't "initial" anymore |
23:06.53 | DocScrutinizer05 | we have no complete schematics at all yet, not even rudimentary, and we have not yet completed component id, by far not |
23:07.30 | dos1 | the most important thing I wanted to put there is the link to http://neo900.org/news-0005-phase-i-results-ohsw-showcase |
23:07.42 | DocScrutinizer05 | heck we haven't even decided yet on the WLAN module |
23:08.12 | dos1 | bbl |
23:08.35 | DocScrutinizer05 | you can keep the link, nothing wrong with it. But pretty please remove any shiny highlights and checkmarks that suggest we might consider done with err whatever |
23:08.45 | FIQ | btw, >1.5k€ donors would get prototypes.. would that be as part of phase 2 or 3? |
23:09.12 | FIQ | just curious |
23:09.26 | DocScrutinizer05 | we have phases in donation campaign, we don't have phases in development |
23:09.39 | DocScrutinizer05 | and the 1.5k offer is valid all the time |
23:10.39 | DocScrutinizer05 | hw development is a complex process where you always do several "phases" concurrently and actually never completely finalize any of them |
23:11.46 | DocScrutinizer05 | there might come in changes to schematics or even ME pretty late in development, even weeks before MP starts |
23:12.18 | DocScrutinizer05 | so neither schem nor ME-eval is ever "closed" |
23:13.09 | DocScrutinizer05 | and the prototypes we will hand out to 1.5ers are meant to be semi-working complete devices, of course not a dummy like the thing demonstrated for ME-eval on OHSW |
23:13.15 | *** join/#neo900 Mistah_Darcy (~MistahDar@71-9-57-207.dhcp.gvrb.ca.charter.com) |
23:13.51 | merlin1991 | I've got 2 atmel avr cpus here and want to connect them over i2c I can't find any definite info though what kind of resistors I should use as pull ups |
23:14.45 | DocScrutinizer05 | hmm, I2C electrical specs should be reprinted in datasheet of every single I2C-enabled chip |
23:15.03 | DocScrutinizer05 | iirc 1k, but don't quote me on that |
23:16.48 | DocScrutinizer05 | merlin1991: N900 using 2k2 |
23:17.34 | DocScrutinizer05 | (e.g. R4807, R4808) |
23:18.59 | DocScrutinizer05 | and 1k (R4805, R4806) ;-D |
23:20.16 | DocScrutinizer05 | and 4k7 (R4242-45) |
23:20.21 | DocScrutinizer05 | XP |
23:20.33 | merlin1991 | okay you're starting to confuse me ;) |
23:21.27 | DocScrutinizer05 | it's probably not really THAT relevant |
23:21.54 | DocScrutinizer05 | 1k is more on the safe highspeed side, while higher values are more on the optimized side |
23:21.57 | DocScrutinizer05 | I guess |
23:22.54 | DocScrutinizer05 | prolly also depends on real and phantom (trace length etc) load on your I2C bus |
23:24.34 | DocScrutinizer05 | the smaller the pullup, the higher the power consumption but also the higher the upper freq limit |
23:25.05 | merlin1991 | the chip only supports 400kHz speed |
23:25.49 | DocScrutinizer05 | then you're probably safe with 4.7k, but when you don't care to much about the marginal increase in power consumption then go for 1k nevertheless |
23:26.03 | DocScrutinizer05 | too much* |
23:26.20 | DocScrutinizer05 | and RTFDS ;-) |
23:26.31 | merlin1991 | I'm sitting in front of it |
23:26.33 | DocScrutinizer05 | :-D |
23:26.45 | merlin1991 | and it has a highly confusing table which defines some min / max values for the pullup |
23:26.54 | merlin1991 | which depends on a billion factors |
23:27.02 | DocScrutinizer05 | yes |
23:27.21 | DocScrutinizer05 | that's what it is like |
23:27.41 | merlin1991 | and it doesn't even fit on 1 screen in readable size :D |
23:27.51 | DocScrutinizer05 | depends on real and phantom load, on upper freq, on driver impedance and whatnot else |
23:28.38 | DocScrutinizer05 | I'd simply pick the most safe value (for highest freq etc), which most likely is 1k |
23:29.55 | DocScrutinizer05 | 1k makes for an average consumption while active of 3.6mA (@1.8V) |
23:30.09 | DocScrutinizer05 | 1st approach |
23:30.36 | DocScrutinizer05 | with higher values you can cut down on this consumption |
23:31.06 | DocScrutinizer05 | at the expense of other factors becoming more limited, like upper freq etc |
23:31.54 | FIQ | would neo900 support any kind of OC? |
23:32.04 | DocScrutinizer05 | aczually I'm wrong, that#s peak consumption, not evarage |
23:32.20 | DocScrutinizer05 | FIQ: like beagleboard |
23:32.55 | DocScrutinizer05 | we're not actively counteracting OC |
23:32.55 | FIQ | not planning to decrease life of my potential neo900, but I've begun slight OC experimentation on my n900 given that it will most likely fail anytime soon, and unlikely to be caused by cpu failure |
23:33.18 | FIQ | DocScrutinizer05: ok |
23:34.24 | merlin1991 | hm the max for the resistor is defined as 300ns / Cb Cb is capacitance of the bus line in pF but what's ns in that case? :D |
23:35.51 | DocScrutinizer05 | FIQ: we however will think about ways that tell an EE who examines the device that it has been OCed |
23:36.24 | DocScrutinizer05 | clock width? |
23:36.24 | FIQ | EE? |
23:36.32 | DocScrutinizer05 | ~ee |
23:36.32 | infobot | [ee] for gEEks. Electronics Engineer |
23:36.37 | FIQ | ah |
23:36.50 | FIQ | you totally meant the former definition |
23:37.15 | DocScrutinizer05 | ns is nanoseconds? |
23:37.24 | DocScrutinizer05 | 0.3us |
23:37.33 | dos1 | DocScrutinizer05: I added different checkmark style per Nikolaus request. please bring it to him rather than to me; but for now I can of course remove it |
23:37.35 | DocScrutinizer05 | 3MHz |
23:38.00 | DocScrutinizer05 | dos1: yes, please do |
23:38.23 | DocScrutinizer05 | I'm rather sure I know there's been a misunderstanding between the both of you |
23:38.45 | DocScrutinizer05 | started by your referenced to "phase1 completed" |
23:39.14 | dos1 | hmm? I used it in news, because Nikolaus used it in presentation |
23:39.16 | mashina | Neo900 ist liebe, Neo900 ist leben |
23:39.20 | FIQ | hm |
23:39.31 | FIQ | site doesn't work in n900 :D |
23:39.41 | FIQ | a bit ironic |
23:39.43 | dos1 | FIQ: why? it works on mine |
23:39.45 | FIQ | well, progressbar |
23:39.53 | dos1 | oh, right |
23:40.15 | dos1 | too old microb engine for recent jquery versions :( |
23:40.20 | FIQ | not only is the red and "lighter yellow" not showing up (which I can live with) |
23:40.47 | FIQ | but the 200 devices text overlap the black 200 devices border :P |
23:41.00 | dos1 | checks |
23:41.02 | FIQ | jQuery... |
23:41.30 | dos1 | I guess I could make at least progress bar working |
23:41.47 | dos1 | the whole page uses bootstrap, which uses jquery |
23:42.00 | DocScrutinizer05 | I looked at it yesterday and it wasn't too bad |
23:42.31 | dos1 | (allowed to massively speed up doing it at the beginning, but tbh I'd choose something else if I'd have to start from scratch) |
23:42.36 | FIQ | also hovering the phases doesn't work |
23:42.45 | dos1 | yeah, that's bootstrap |
23:43.01 | FIQ | DocScrutinizer05: the site *mostly* works |
23:43.07 | dos1 | maaaybe it'll work with earlier jquery version? from 1.x branch? |
23:43.13 | dos1 | I'll try, but not today |
23:43.16 | FIQ | a few things, mostly regarding donate part of the site fails |
23:43.31 | DocScrutinizer05 | (phase-1) well, Nikolaus and I refer to those phases internally sometimes, since we both know those are "metaphors" for something we want to accomplish. This shouldn't get used publicly to describe progress in development |
23:44.11 | merlin1991 | hm according to this datasheet 1k would be cutting it close to the minimum, min is defined as Vdd -0,4V / 3mA which would give me 967 Ohm (Vdd is 3,3) |
23:44.22 | ds2 | DocScrutinizer05: yesterday, you mention sourcing dome sheets...what's that for? |
23:44.47 | DocScrutinizer05 | when Nikolaus used "phase-1" in his presentation, then that's not meant to state we accomplished anything that can get described by the dination campaign phases model |
23:45.00 | DocScrutinizer05 | ds2: we need kbd |
23:45.12 | ds2 | why can't we reuse what is on the existing keyboard? |
23:45.38 | DocScrutinizer05 | s/dination/donation|fundraiser/. |
23:45.59 | FIQ | dinations... dining |
23:46.02 | dos1 | DocScrutinizer05: see his reply to my mail with latest news piece and later one in "200" thread |
23:46.03 | DocScrutinizer05 | because we don't have existing kbds |
23:46.13 | dos1 | DocScrutinizer05: for me it looks like a clear request to do that |
23:46.32 | ds2 | I am confused... are we talking about the keyboard on the current N900 or is this an enhancement? |
23:46.32 | DocScrutinizer05 | then I missed to protest that |
23:46.41 | DocScrutinizer05 | we don't disclose development progress in that way |
23:46.49 | DocScrutinizer05 | that would be dishonest |
23:47.50 | DocScrutinizer05 | Nikolaus also only watches occasionally and loosely what's going on at the PR front |
23:48.28 | DocScrutinizer05 | ds2: we are building complete devices. Where from shall we get the kbd? |
23:48.40 | FIQ | wut |
23:48.51 | ds2 | DocScrutinizer05: I thought this was a refit board for the N900? |
23:48.59 | DocScrutinizer05 | o.O |
23:49.02 | FIQ | just noticed http://www.prisjakt.nu/pryl/tele/4599_neo900_aer_inofficiella_efterfoeljaren_till_nokia_n900 |
23:49.05 | FIQ | interesting... |
23:49.16 | DocScrutinizer05 | http://neo900.org/faq#complete |
23:49.19 | ds2 | you are reusing plastic, right? |
23:49.22 | FIQ | (swedish price comparation page) |
23:49.50 | ds2 | so you are replacing the keyboard as part of the 'refurb'? |
23:50.20 | DocScrutinizer05 | ds2: please elaborate how we shall deliver a complete working device to users who order one, when we have no domesheets |
23:50.43 | ds2 | maybe I should ask - where are the plastics coming from? |
23:51.00 | DocScrutinizer05 | it's not clear yet where from we will get all the "old" parts, we might buy cases for 18EUR |
23:51.07 | DocScrutinizer05 | and LCD for 45EUR |
23:51.16 | DocScrutinizer05 | and digitizers for 23EUR |
23:51.22 | ds2 | Ohhh... I thought you were buying up old N900's |
23:51.27 | DocScrutinizer05 | maybe |
23:51.30 | DocScrutinizer05 | maybe not |
23:51.36 | ds2 | which would imply you get 1 KB, 1 camera, 1 screen, etc per old unit |
23:51.45 | DocScrutinizer05 | that's a question decided upon during sourcing |
23:52.01 | ds2 | I vaguely recall coming across a custom dome sheet maker at a trade show |
23:52.11 | ds2 | have you considered a flex circuit with soldered on domes? |
23:52.24 | DocScrutinizer05 | why flex? |
23:52.29 | ds2 | thickness |
23:52.35 | DocScrutinizer05 | irrelevant |
23:52.53 | ds2 | isn't the current one on a flex? never took mine apart but the pictures seems to suggests it is |
23:53.04 | DocScrutinizer05 | we can solder the clik domes directly on UPPEr PCB |
23:53.11 | DocScrutinizer05 | no |
23:53.41 | ds2 | hmmm ok |
23:53.43 | DocScrutinizer05 | http://maemo.cloud-7.de/Gallery-N900-exploded/ |
23:54.01 | ds2 | still confused but i am doing 10 things at once |
23:54.20 | DocScrutinizer05 | http://maemo.cloud-7.de/Gallery-N900-exploded/platine_01.jpeg |
23:54.42 | dos1 | DocScrutinizer05: anyway, bring it to him - he asked for that and now it's not there anymore |
23:54.57 | dos1 | and if I really misunderstood him, then he'll just say so :) |
23:55.07 | DocScrutinizer05 | yes, I will discuss it with him, but for now we're fine with removing it |
23:55.16 | ds2 | Hmm... let me show you an item |
23:55.52 | dos1 | (bummer, that glow looked cool :( ) |
23:56.02 | DocScrutinizer05 | dos1: irrelevant who misunderstood whom. We're not doing fingerpointing, we try to do the right thing[TM] and that is to remove any such misleading indicator |
23:57.24 | DocScrutinizer05 | btw I'd prefer you sending Nikolaus to me when he has requests about our public appearance |
23:57.33 | dos1 | sure, I know, I meant "it won't hurt anyway" :D |
23:58.23 | ds2 | http://www.digikey.com/product-detail/en/EVQ-PQHB55/P8090STR-ND/ |
23:58.30 | ds2 | what about an array of those on the PCB? |
23:58.41 | ds2 | more expensive for production but... |
23:58.53 | dos1 | DocScrutinizer05: tbh I thought you're ok with that since you haven't protested |
23:58.54 | DocScrutinizer05 | prolly exathe the ones Nikolaus planned as one of three alternatives |
23:59.12 | DocScrutinizer05 | I simply missed that detail, sorry |
23:59.17 | dos1 | np |
23:59.19 | dos1 | :P |