<DocScrutinizer05>
ah, I'm discussing programming of silego chip with Werner, and this was the wrong channel to post this link, also werber is *offline* .-/
<DocScrutinizer05>
maybe the first is not real program but actually something special the chip uses to do the clock-calibration thing
<DocScrutinizer05>
though by mere optics both data bursts look pretty similar - as far as it gets with optically comparing a 1024bit waveform to another
<Oksana>
Maybe, change NVM data, and see how waveforms change?
<DocScrutinizer05>
alas those codes on pin17 are nowhere to be found in datasheet
<DocScrutinizer05>
the 1024 bit NVM data are more or less described and explained in datasheet though
<DocScrutinizer05>
we asked for the programming specs but they were not enthusiastic about disclosing them, like "we offer the eval board and the greenpak designer software. And the programming is a complex process you don't want to deal with yourselves anyway"
<DocScrutinizer05>
seems they also calibrate the internal oscillator to have proper timing on all their delays etc. That's what I guess the clock signal on pin14 is for
<DocScrutinizer05>
then they send NVM data again, just with adjusted clock generator tuning data
<DocScrutinizer05>
makes sense?
<Oksana>
A tad, yes
<DocScrutinizer05>
anyway nice handy little chips
<DocScrutinizer05>
emphasis on *little* :-)
<edwin>
can that nvm data be copied to other chips directly, or you need to calibrate/program each one individually?
<DocScrutinizer05>
err, I can program single chips in eval board, or send the 'program' to Silego and buy pre-programmed chips
<edwin>
ah, was wondering how that programming scales to hundreds of devices :)
<Oksana>
Latter is faster, especially since you will buy large quantities of them later anyway?
<DocScrutinizer05>
yes, we will buy a 500 of them, or more
<edwin>
can the pre-programmed chips still be updated in case something goes wrong and you need to update it?
<DocScrutinizer05>
no, OTP
<Oksana>
De-soldered and replaced?
<DocScrutinizer05>
but you can upload a temporary patch to the RAM of a already programmed chip ;-)
<DocScrutinizer05>
*IF* you know how to do that in-circuit. Which is exactly why I'm doing this little investigation
<Oksana>
Temporary patch is good for testing, I guess? Of course, temporary solutions become permanent when permanent solution is too troublesome to do.
<DocScrutinizer05>
edwin: generally in EE when "something goes wrong" you're doomed. A little often neglected detail that distinguishes hw development from sw development
<Oksana>
Like, if power is exceeded, traces burn, and new board is needed?
<Defiant>
Like when having a EMC problem?
<Defiant>
EMI
<DocScrutinizer05>
like when you did an oopsie, it's not 2let's write and push a patch", it's "let's think if we can rework a NNNN devices manually or we need (and can afford) to dump a huge pile of junk hardware and build everything anew"
<enyc>
ooo fun =)
<enyc>
DocScrutinizer05: are you discovering mor and morer liktle gotchas nad optimizations needed? ;-)
* enyc
thinking... this is one o those projects that will bbe released when itrl' "right" sort of thing =)
<DocScrutinizer05>
we just try to minimize number of board spins by thorough QA before we build a prototype
<edwin>
my DECT phone has "buzz", turns out the batteries leaked all over the place inside it. Never had batteries in a smartphone leak so hopefully thats not a concern for neo900 :)
<enyc>
DocScrutinizer05: nodsnods =) very important not badging things, make sure everything done reasonably and works together!! =)
<enyc>
DocScrutinizer05: hrrm... well its' not clear to some outsiders / potential investors what the roadmap to completion is =)
arossdotme-planb has quit [Ping timeout: 256 seconds]
<enyc>
DocScrutinizer05: what choices till to make, what qa works, etc... but thens i guess the nature of developement is you don't know everything needed !!
<DocScrutinizer05>
(DECT) I suppose those were NiMH or similar batteries, not a LiIon cell?
<DocScrutinizer05>
Neo900 will use Nokia BL-5J battery which is unheard of to ever have leaked
<enyc>
DocScrutinizer05: i wonder what the potential options are to have a ''big'' lithium battery on the back, not just 'wider' but osmething using much more space (excepting not covering the camera ;p)
<DocScrutinizer05>
instead of the LiMg button backup battery cell used in N900 which is known to always break after 12 months and often even leak, we're using a polyacene capacitor "supercap" cell which afaik cannot leak and is supposed to survive a few years longer than Nokia's solution
<enyc>
=)
<enyc>
do supercaps 'tend to fail at all???
<DocScrutinizer05>
(huge battery) google 'mugen'
<enyc>
i'm used to non-electrolytic caps lasting 'forever'
<DocScrutinizer05>
(fail) never heard of it
<edwin>
DocScrutinizer05: yeah not LiIon
<enyc>
soo.. neo900 (or n900??) will take 'anya' 3.7v li-ion battery ? is the interconnect to 3rd battery pin contact standardized?
<DocScrutinizer05>
no, it's a 80kR to GND which is needed to detect there's a battery inserted, basically
<DocScrutinizer05>
otherwise yes, every LiIon 3V7 will do
<enyc>
i see =)
<enyc>
i'm sure i asked about effective battery usage anyway!
<enyc>
a lot down to software it would seem =)
<enyc>
but new modem / gps should be rather less hungry AIUI
<enyc>
n900's 3g modem eats power!
<DocScrutinizer05>
in former times, the third contact was an NTC to test battery temperature. Seems that's not exactly the case anymore for BL-5J
<enyc>
kk
<enyc>
''laptops'' seem to have more interconnect wires etc
<DocScrutinizer05>
power consumption of Neo900 should be comparable to N900, maybe a tad less, maybe a tad more, depending on usage pattern. Except for LTE for which I don't have decent values yet
<enyc>
nods!
<enyc>
I spent ages to reccomend a USB LTE modem to a friend...
<enyc>
collected lists of supported in various gadgets...
<enyc>
then found what was easy to get
<DocScrutinizer05>
and yes, UMTS data is power greedy since it keeps TX on for several seconds after every byte you transfer
<enyc>
then ... discovered that some helpful german tested these =)
<enyc>
so settled on a ZTE MF823 =) as its' well supported in cheapo-routers, works in all OSes (usb-ethernet if), works with mikrotik routerboard, almost certainly will work with openwrt =)
<enyc>
but what a mess that was to sort out!
<enyc>
I can understand the trouble you have will all the littel parts...!
<enyc>
I do wonder how much n900 display backlight eats power comparatively-speaknig and if there are good ways to offer better 'dimming' control =)
<DocScrutinizer05>
http://wiki.maemo.org/N900_Hardware_Power_Consumption and there's a convenient gadget to control backlight brightness with a single click: "simple brightness applet" or somesuch. couldn't do without anymore, a must for N900
<DocScrutinizer05>
noteworthy: brightness control of N900 does not adjust according to ALS when set on level5
<DocScrutinizer05>
IOW "5" means "max brightness, unconditionally"
<enyc>
aha!
<DocScrutinizer05>
thus >>Backlight on max +150 mA, +40mA on 4 of 5 steps in simple brightness applet & low ambient (/sys/class/backlight/acx565akm/brightness = 63)<<
<enyc>
same backlight controller in neo900??
<enyc>
is that on the display sub-board or main board I wonder
<DocScrutinizer05>
err, the backlight controller is irrelevant
<enyc>
kk yes down to setting it in software =)
<DocScrutinizer05>
it's a PWM chip
<enyc>
but i ean, will the safteware driving it need to be rejiggled, new linux driver or whatever
<enyc>
indeed!
<DocScrutinizer05>
which is pretty weirdly controlled both by LCD controller and by CPU directly
<DocScrutinizer05>
backlight brightness setting is a command to LCD controller
<enyc>
good to keep the pwm frequency not too low... =) causes some peopl ehadaches/trouble
<enyc>
it seems some things have horribly low freqs getting closer to 100hz but better gadgets use 200 500 or more ....
<enyc>
affects saccades and reading and allsorts!
<DocScrutinizer05>
yes, I know
<enyc>
ooooo have you found any interesting research on this?
<DocScrutinizer05>
anyway, see above. LCD controller doing that
<enyc>
kk
<enyc>
I noticed, tridonic 'professional' LED dimmable drivers, take the approach of lovering the current down to a certain dimming level, then using 500hz pwm below that to get down to nothing...
<DocScrutinizer05>
I think N900 AKMblabla LCD controller does closer to 500Hz iirc
<enyc>
kk
<enyc>
keep up the good work =) want to see this project complete =)
* DocScrutinizer05
too
<DocScrutinizer05>
I need a looong holiday after that
<enyc>
=)
<enyc>
any idea of timescale hehe
<DocScrutinizer05>
no, I'm afraid I will not be able to take any holiday in 2016 either, not matter if we roll out in Q1 or not
<enyc>
howso?
<enyc>
may it help to take time out for a few weekssanyway then come back to?
<DocScrutinizer05>
prolyl after rollout I have to collect the pieces of my own life and try to recover from all this
<enyc>
yes, important to make time for that!
<DocScrutinizer05>
I'm working on a too high stress level since >2 years now
<enyc>
not making time to sort yourself out causes breakage anyway in my expreience
arossdotme has joined #neo900
<DocScrutinizer05>
alas already beyond that point
<enyc>
hrrm
<enyc>
so, whats' the best recoveryi plan now?
<enyc>
I can only suggest mbeing vey meticulous to list carefually everything needed to complete and avoid feauturitis!
<enyc>
and delberately asking for help fromthose also involved, they can hove soe things to work on while you have a defined break 2 weeks out or so.... ??
<DocScrutinizer05>
feature freeze been a 9 moths ago
<enyc>
you spoke of help combing back (dos1??)when not busy orn Pyra ??
paulk-collins_ has quit [Ping timeout: 250 seconds]
jonsger has quit [Quit: jonsger]
<Oksana>
(feature freeze been a 9 moths ago) So how is it going? Will the keyboard backlight be RGB? Will the touchscreen be capable of distinguishing finger from stylus (or not tested yet?) ? Will there be a sensor for stylus inside or taken out?
paulk-collins has joined #neo900
paulk-collins has quit [Remote host closed the connection]
mathox has joined #neo900
mathox has quit [Client Quit]
<DocScrutinizer05>
rgb: we will see. finger/stylus: already implemented in N900 microB, sensor: we will see