DocScrutinizer05 changed the topic of #neo900 to: http://neo900.org | conversations are logged to http://infobot.rikers.org/%23neo900/ and http://irclog.whitequark.org/neo900 | 2013-11-04 - the day our fundraiser reached its goal | 2014-05-01 360 devices 75k€| 0712 183 ~30k | 0810 300 ~49k | 0914 346 ~56k
<wpwrak> unspecified seems a bit useless. when people want to power something from HB, they'll want to know what they can expect to work
<DocScrutinizer05> battery is specified as "charge with no more than 0.7C" but then system pulls side path current from VBAT_RAW
<DocScrutinizer05> we cannot tell them what will work since we don't know what battery they use and how much the system subtracts from the max current the battery can provide
<DocScrutinizer05> we can tell "this is identical to +-contact of battery, except we added a 1.5A polyfuse"
<DocScrutinizer05> when the mains outlets of your flat are rated "230V AC, 10A" that doesn't mean you can connect a 10A sink to all of them same time and expect that nothing bad will happen ;-D
<wpwrak> 1 C is also a bit weird and seems way too low. what ESR would we use ? and if we have a charger, then that's very different from a fixed charge
<DocScrutinizer05> *use* ESR?
<wpwrak> well, to calculate the current
<DocScrutinizer05> sorry you lost me
<wpwrak> (or i hope it works like that. i rarely use coulomb ... lemme check ...)
<DocScrutinizer05> VBAT_RAW is the battery plus contact protected by a 1A5 polyfuse. period.
<DocScrutinizer05> you may add that the connectors are rated 2A and we made sure that the copper traces are as well
<DocScrutinizer05> C!= coulomb
<DocScrutinizer05> C == 100% capacity of the battery cell
<wpwrak> oh
<wpwrak> so that's the energy
<DocScrutinizer05> industry speac is like "charge with 0.7C" which means "charge with a current that charges battery to 70% in one hour"
<wpwrak> Joule
<DocScrutinizer05> speak*
<DocScrutinizer05> when a battery has 1320mAh the 1C is 1320mA
<wpwrak> so we have 1.3 A for the standard battery, minus what the system needs for itself
<DocScrutinizer05> BL-5J is specified as "charge with 0.7C aka charge with 1320mA * 0.7"
<DocScrutinizer05> BL-5J can supply surges up to several C, but continuous current is limited to sth like 2C, maybe less
<wpwrak> perhaps we can spec recommended / likely values with system running and system off ?
<DocScrutinizer05> we simply don't know
<DocScrutinizer05> no, we cannot
<wpwrak> we should know with Isys = 0. i mean then it's just the battery
<DocScrutinizer05> "system running" is completely fuzzy
<DocScrutinizer05> and we don't have datasheets for the battery
<wpwrak> yes, but how much do we expect to be available on HB ? a ballpark number
<DocScrutinizer05> how much can you draw via a 1500mA polyfuse?
<DocScrutinizer05> c'mon, that been a rhetorical question
<wpwrak> what continuous current would be be comfortable to let HB have ? 100 mA ? more ? less ?
<DocScrutinizer05> HB?
<DocScrutinizer05> you mean VBAT?
<wpwrak> i'll put 1.5 A for peak current to HB, 0.9 A continuous and peak current from HB
<wpwrak> VBAT -> HB
<DocScrutinizer05> sorry, after first comma you lost me
<DocScrutinizer05> Neo900 can support 1500mA coninuous current
<wpwrak> if HB decides to feed power into the neo900 on VBAT_RAW, then it shouldn't try to send more than 900 mA (0.7 C), right ?
<DocScrutinizer05> you won't squeeze specs out of me that I don't have
<wpwrak> i'm anticipating the suffering of people trying to design something that connects to HB
<DocScrutinizer05> sorry, I don't know. That depends on battery, not on Neo900
<wpwrak> they'll have the big question whether they can power their circuit from HB or whether they need an external power supply
<DocScrutinizer05> and we don't have any datasheets for BL-5J
<wpwrak> assuming the standard battery. if we don't know the specs, how do WE charge it ? or does that come from one of the nokia blobs ?
<DocScrutinizer05> honestly when somebody considers building a HB extension that eats more than 4.0V * 1.5A then I'd rather be worried where to place the fan to cool that thing
<wpwrak> hehe ;-)
<DocScrutinizer05> and battery charging isn't part of the HB specs
<wpwrak> but we don't discourage charging through HB ?
<wpwrak> what i'm looking for are guidelines for HB users, so that they can tell what they're allowed to do and what not
<DocScrutinizer05> I think you can't get any better specs for VBAT_RAW than "this is identical to +-contact of battery, except we added a 1.5A polyfuse"
<DocScrutinizer05> I'm not writing a 42 pages guideline that includes stuff like "how to charge a LiIon cell"
<DocScrutinizer05> ~batteryfaq
<DocScrutinizer05> when we talk HB *specs* than that's "wire with 1500mA polyfiuse to battery PLUS", no more no less
<DocScrutinizer05> we won't explain how to charge a LiIon and we won't add USB specs incl all protocol and timing
<wpwrak> so we basically tell people to stay away from VBAT_RAW because it's unpredictable. okay.
<DocScrutinizer05> MEH!
<wpwrak> how about VBAT_SWITCHED ?
<DocScrutinizer05> sorry, I', wasted
<DocScrutinizer05> YOU WON'T GET ME TO DEFINE WHAT BL-5J CAN DO! I DON'T KNOW!
<DocScrutinizer05> sorry when this upsets you. Don't spread FUD like >>we basically tell people to stay away from VBAT_RAW because it's unpredictable.<< That's BS
<wpwrak> i think you know more than you let on ;-) again, the goal is to explain to prospective HB users what they can expect will work
<wpwrak> maybe the actual limits are much larger. but that's their problem then
<DocScrutinizer05> I D*O*N*T K*N*O*W!!!!
<DocScrutinizer05> afk
<wpwrak> things like: "for up to 100 mA, you can safely use VBAT. if you need more, you're on your own."
<wpwrak> i mean, what good is a power rail if you have no assurances at all about it ? that just doesn't make sense.
<wpwrak> as see above, i'm asking about recommended (!) conditions. not maximum values and such.
<wpwrak> or, putting it the other way around: if we found out that loading VBAT (via HB) with 100 mA does not work reliably, would we consider this a problem that needs fixing ?
<DocScrutinizer05> yes, when I found out that a wire doesn't work I would consider this a problem for my construction of reality
<DocScrutinizer05> and I'm not going to spit out any random mA value now, to later defend why it's not 50mA higher
<DocScrutinizer05> again, I DONT KNOW!
<DocScrutinizer05> the wire has a 1500mA fuse. PERIOD
<DocScrutinizer05> and even that doesn't mean that Neo900 guarantees that the plastic won't melt when you draw a constant power of 6W on your HB extension
<DocScrutinizer05> we can neither comprehensively allow nor rule out all possible scenarios that are working/breaking
<wpwrak> ok, "Note that the operation of the power rails ([...]) is not guaranteed and neither are there recommended operating conditions."
<DocScrutinizer05> that's up to user to evaluate if his circuit will melt the plastic or not
<DocScrutinizer05> we guarantee a wire with a 1500mA polyfuse
<DocScrutinizer05> how much more precise could it get?
<wpwrak> do we even want to say that ? someone may thing that means 1.5 A is okay, but that's not the case
<DocScrutinizer05> "all the things you can do with a wire rathed 1.5A. And all the things you better shouldn't do" ?
<DocScrutinizer05> why is that not the case?
<DocScrutinizer05> please tell me what's the problem with costant 1500mA current draw!
<wpwrak> we don't say that there's a problem, nor do we deny it ;-)
<DocScrutinizer05> you sound like the dude with cat in microwave oven
<DocScrutinizer05> this are the specs of hackerbus. XXXXXXX jkdghdag WRC dkjgdhq. STAY AWAY FROM HACKERBUS WHEN YOU DUNNO WHAT YOU'RE DOING. It's not meant for kids to have their forst experience with electronics"
Pali has quit [Remote host closed the connection]
<wpwrak> that's actually sound advice :) let me phrase it a bit more gently, though
<wpwrak> "Warning: Hackerbus gives access to signals that can upset the operation of the Neo900 and incorrect use may cause permanent damage. Use with caution !"
<DocScrutinizer05> please keep in mind that complete HB is not a warranted product property of Neo900. It comes "as is" and "use on own peril"
<wpwrak> lemme say what can get damaged ...
<DocScrutinizer05> you want me to rethink the whole HB idea. I bet that's what you're after
<DocScrutinizer05> I won't
<wpwrak> "[damage] to Neo900 battery, and the circuit connected to Hackerbus."
<DocScrutinizer05> HUH? now how did you determine THAT?
<DocScrutinizer05> are you sure it's wise to list possible damage? what if the damage user suffers is different than what you listed? Will you compensate user for that?
<DocScrutinizer05> sorry, I'm thoroughly fed up with this discussion
<DocScrutinizer05> n8
<wpwrak> (rethink) i'm just disappointed about how vague we are there. it sounds as if we didn't understand ourselves how neo900 works. i.e., we build something that happens to function (hopefully), but it's not properly engineered. i don't like to give this sort of impressions.
<DocScrutinizer05> that's YOUR impression
<wpwrak> adding "including but not limited to"
<DocScrutinizer05> I think giving random made up operation linits and a random set of possible damage is much worse than telling that operation of HB is only for the savvy and subject to evaluation against the schematics
<wpwrak> if it also causes the world to end, that's too bad ;-)
<DocScrutinizer05> since doing such random rambling and FUD is *really* telling "we don't understand shit"
<DocScrutinizer05> and now I'm definitely gone
* DocScrutinizer05 pnders a sticky capton tape over HB, printed "Do not remove! DANGER! Courtesy Almesberger"
<wpwrak> (limits) why not ? we can say "you can draw current X from this point, because that's what we planned for." maybe you can draw a lot more, but that's not our problem. like a reservation in a restaurant. you'll have at least your table because the restaurant will make sure of that. but maybe there's a dozen other free tables. would you complain ?
<wpwrak> hehe, leaving my mark on it. i like that ;-)
<DocScrutinizer05> >>because that's what we planned for<< Does a Polyfuse 1500mA sound lime "we didn't plan for it" in your ears??
<DocScrutinizer05> like*
<DocScrutinizer05> honestly, we should stop this discussion
<wpwrak> will any of the other rails have fuses ?
<DocScrutinizer05> I think I made utterly clear yesterday that the traces will have to be wide enough for 2A and we get a polyfuse of 1500mA. What else planning do you want me to tell you?
<DocScrutinizer05> please read backscroll in other channel. I hate to repeat myself
<wpwrak> i'm actually more concerned about 1v8 (if we'll have it) and 3v3. these are what people will actually want to use.
<wpwrak> please just tell me the final results of your reasoning. it's painful and a waste of time to read through all the back and forth, then pick what look the most likely, and then you'll disagree with it anyway
<DocScrutinizer05> [2014-11-26 Wed 05:20:12] <DocScrutinizer05> oooh, there are nice FET switches with integrated OCP [2014-11-26 Wed 05:22:17] <DocScrutinizer05> whe could use those for V18 and V33 [2014-11-26 Wed 05:23:21] <DocScrutinizer05> should also conveniently protect the core system from bad things that happen on hackerbus, like short on VIO_18
<DocScrutinizer05> when it's too painful for you to read my rasoning then maybe I just send my specs per mail and we don't do any discussion
<wpwrak> so how many A are "nice" ?
<wpwrak> sure :)
<DocScrutinizer05> honestly, 2h of discussing the possible dangers of a wire rated 1.5A. Incredible
<DocScrutinizer05> who's asking about the pain *I* suffer from this?
<wpwrak> naw, discussing what are usable interface specs
<DocScrutinizer05> no, evidently not
<wpwrak> nota bene, interface, not the physical limits of whatever is behind the interface
<DocScrutinizer05> you're insisting in a *guide*
DocScrutinizer05 has left #neo900 ["uptime counter overflow"]
<wpwrak> yes, that sounds nice. "how to use the hackerbus".
louisdk has quit [Read error: Connection reset by peer]
SylvieLorxu has quit [Quit: ZNC - http://znc.in]
nox- has quit [Quit: Leaving]
che1 has quit [Ping timeout: 250 seconds]
Kabouik has joined #neo900
wicket64 has quit [Quit: WeeChat 1.0.1]
Kabouik has quit [Ping timeout: 272 seconds]
<wpwrak> hmm, the joy of CSS (at the bottom): https://github.com/wes/slimCssBarChart/blob/master/css/styles.less
<wpwrak> maybe i should find something nicer for sourcing bar graphs ... pity that they'll be a list of arbitrary length, or i could just use gnuplot
jonwil has joined #neo900
NIN101 has quit [Quit: No Ping reply in 180 seconds.]
NIN101 has joined #neo900
b1101 has quit [Quit: b1101]
fling has quit [Ping timeout: 272 seconds]
fling has joined #neo900
b1101 has joined #neo900
b1101 has quit [Changing host]
b1101 has joined #neo900
mvaenskae has joined #neo900
freemangordon_ has joined #neo900
arcean has joined #neo900
Pali has joined #neo900
che1 has joined #neo900
mvaenskae has quit [Ping timeout: 255 seconds]
sparetire_ has quit [Quit: sparetire_]
<freemangordon_> Pali: looks like /proc/cpuinfo issues are going to be solved :)
<freemangordon_> Pali: BTW, do we really need the serial?
<freemangordon_> I know not having it is a regression compared to legacy boot, but is it used anywhere?
DocScrutinizer05 has joined #neo900
<Pali> freemangordon_: serial number is always zero (it is not set in legacy boot too)
<Pali> but revision number is set by NOLO
<freemangordon_> hmm, I am sure I've seen SN, maybe only KP/omap1 shows it
<DocScrutinizer05> Serial : 0000000000000000
<DocScrutinizer05> you probably seen it in OMAP3 TRM ;-)
<DocScrutinizer05> dunno if userland could read it out from CPU registers
<freemangordon_> DocScrutinizer05: /sys/power/idcode
jonwil has quit [Remote host closed the connection]
jonwil has joined #neo900
<DocScrutinizer05> aaahyes, I recall I've seen this before :-)
<DocScrutinizer05> we even tried to decode it
SylvieLorxu has joined #neo900
<DocScrutinizer05> IDCODE: 4b7ae02f
<DocScrutinizer05> Production ID: 00000000 00000000 000000cc cafeb7ae
<DocScrutinizer05> Die ID: 0c004012 040364fb 00000000 44f60004
<DocScrutinizer05> seems omly the die ID is unique
<freemangordon_> yep, the others are OMAP revision
<DocScrutinizer05> 128bit, heck!
<DocScrutinizer05> maybe 64bit for Die serial, 64bit for wafer serial
<freemangordon_> could be
<freemangordon_> I guess it is described in the TRM
jonwil has quit [Quit: ChatZilla 0.9.91 [SeaMonkey 2.30/20141013232806]]
SylvieLorxu has quit [Quit: ZNC - http://znc.in]
mvaenskae has joined #neo900
<Pali> that Serial value in procinfo comes from serial ATAG filled by booloader (in our case NOLO)
<Pali> and NOLO does not set any serial value into atag
<Pali> freemangordon: look at linux-n900 git at v3.18-rc6-n900 branch
<Pali> there more your patches which fixing some bugs but are not in mainline kernel
<Pali> if they should go mainline let me know which and I will send them
che1 has quit [Ping timeout: 240 seconds]
<freemangordon_> Pali: iiuc you just pulled upstream on top of what was already there
<freemangordon_> but you'd better look at my repo, there are just a couple of patches on top of 3.16
<freemangordon_> (SGX, cmt-speech, gpio driver)
<freemangordon_> it is way cleaner IMO
<freemangordon_> unfortunately, nothing besides cmt-speech can be upstreamed AFAIK
<Pali> there is some mutex_unlock patch
<freemangordon_> it is upstreamed already iirc
<Pali> I used patches from v3.16-rc1-n900 on top of v3.18-rc6
<Pali> and that patch was not in 3.18-rc6
<freemangordon_> which one?
<Pali> "ARM: OMAPFB: panel-sony-acx565akm: fix missing mutex unlocks"
<Pali> bfbd31b828b4fa59c09f40dc2805416d0eeb9d70
<freemangordon_> hmm, it is
<freemangordon_> lemme try to find it
<Pali> another patch: "omap3-n900.dts: fix i2c busses numbering"
<freemangordon_> yep, we can try to upstream that
<Pali> another: "RX-51: Add missing vaux2 dependency to vdds_csib regulator supply"
<Pali> "RX-51: Add MUX configuration to the board file"
<Pali> "omap3isp: bring dpll4_m5 rate divisor back"
<freemangordon_> that's fixed upstream (ISP)
<freemangordon_> MUX is for legacy boot afaik
<freemangordon_> the same for vaux2
<freemangordon_> nothing re mutexes there
<freemangordon_> hmm, I don;t see that mutex fix either
<freemangordon_> Pali: oh, wait, it is not about tsc, but about acx565
jake42 has quit [Ping timeout: 244 seconds]
<freemangordon_> Pali: LKML thread https://lkml.org/lkml/2013/12/30/193
<Pali> ok, it is qued for 3.19
<freemangordon_> hmm, it shoud be upstreamed since february, why queued?
freemangordon_ has quit [Quit: Leaving.]
<Pali> 3.18 was not released yet
<Pali> so if it for 3.19 since february, then it is funny :D
che1 has joined #neo900
NIN101 has quit [Ping timeout: 255 seconds]
mvaenskae has quit [Ping timeout: 255 seconds]
mvaenskae has joined #neo900
SylvieLorxu has joined #neo900
NIN101 has joined #neo900
arcean has quit [Quit: Konversation terminated!]
arcean has joined #neo900
jake42 has joined #neo900
sparetire_ has joined #neo900
Kabouik has joined #neo900
xes has joined #neo900
_whitelogger has joined #neo900
che1 has quit [Ping timeout: 255 seconds]
mvaenskae has quit [Ping timeout: 258 seconds]
Kabouik has quit [Ping timeout: 255 seconds]
che1 has joined #neo900
Pali has quit [Remote host closed the connection]
Pali has joined #neo900
<wpwrak> hmm ... 394 pieces USB3322C-GL-TR at Arrow, and 394 at Mouser. do we believe in that much coincidence ?
jonwil has joined #neo900
Kabouik has joined #neo900
louisdk has joined #neo900
arcean has quit [Read error: Connection reset by peer]
wazrus has joined #neo900
wazrus has quit [Quit: Lost terminal]
Kabouik_ has joined #neo900
Kabouik has quit [Ping timeout: 258 seconds]
chainsawbike has quit [Ping timeout: 258 seconds]
norly has joined #neo900
chainsawbike has joined #neo900
chainsawbike has quit [Quit: yep... i broke it good that time...]
chainsawbike has joined #neo900
nox- has joined #neo900