modem has quit [Remote host closed the connection]
modem has joined #neo900
che1 has joined #neo900
mvaenskae has joined #neo900
che1 has quit [Ping timeout: 250 seconds]
paulk-collins has joined #neo900
vakkov has joined #neo900
mvaenskae has quit [Ping timeout: 244 seconds]
mvaenskae has joined #neo900
che1 has joined #neo900
jl has joined #neo900
jl is now known as Guest72593
che1 has quit [Ping timeout: 250 seconds]
<Guest72593>
so, i saw this on slashdot yesterday, and thought 'AWESOME'....
<Guest72593>
then the fine print 1200 EURO --> or $1600+ CDN which is about the cost of 3 iphones... which is really unfortunate, you'll be lucky to get 50 sales, nevermind 500
<Guest72593>
the price is just waaay too high, I would EASILY pay the price of an iphone for such a great open source phone
modem has quit [Quit: Quitte]
<Guest72593>
I'm curious how many people will really get one, I hope it's more than I expect
<Guest72593>
i guess i'm the only person here
paulk-collins has quit [Ping timeout: 272 seconds]
che1 has joined #neo900
<bencoh>
Guest72593: you're not alone ! :)
<bencoh>
in this channel at least :)
che1 has quit [Ping timeout: 250 seconds]
<DocScrutinizer05>
Guest72593: I got news for you: we're at 100+ sales already
rjeffries has joined #neo900
<DocScrutinizer05>
Guest72593: and the price is what it costs to build such a device. with our 3 digits build volume we could use the cheapest crap we could find and leave out *all* the awesome geek stuff and we'd still be more expensive than an iPhone that even has own chip fab building customer/application specific chips for Apple
<ZetaR>
Guest72593: I intend to get one. Doc said that they were getting 2/hour, if you check the chat logs, with ~400 doners interested in the downpayment too. Is that about right Doc? That was before the ./ DOS.
<DocScrutinizer05>
for R&D and for injection-mold, chip, whatelse there's a HUGE difference in cost per device whether you build 500 or 5 million units
<DocScrutinizer05>
we're down to one per hour. But hey that's statistics. I hope for another peak at end of month when people got wages
<DocScrutinizer05>
the shop is running since 5 days now
<DocScrutinizer05>
and we're at 100+ units
<ZetaR>
DocScrutinizer05: Do you have a best guess as to how many orders you will have to fill once you are moving to manufacturing?
<DocScrutinizer05>
500
<DocScrutinizer05>
though I *hope* for 1000+
<ZetaR>
I saw that as the "goal", but I was wondering how much margin you had on that.
<DocScrutinizer05>
zilch
<ZetaR>
Ah.
<DocScrutinizer05>
well, it depends a bit on how good we get on getting down the final sales price
mvaenskae has quit [Ping timeout: 276 seconds]
<DocScrutinizer05>
I'm fully aware the device may not get any cent more expensive than what it's estimated now
<ZetaR>
My order would probably be placed around the end of the month, since that is when my birthday is and so it is a convenient time to pester my parents to buy it.
mvaenskae has joined #neo900
<DocScrutinizer05>
but when we can get the price down, I might consider a few percent of the bargain becoming our margin
<DocScrutinizer05>
ZetaR: fine :-)
<ZetaR>
So you mean producing more phones than orders if you can push the price down?
<bencoh>
even with 1000 devices price wont go down by more than 180e anyway
<DocScrutinizer05>
that's a *very* complicated topic. Includes much of statistics, mere guesswork, and gambling. Basically we can't afford to build too many devices
jonsger has joined #neo900
<DocScrutinizer05>
bencoh: definitely not
<DocScrutinizer05>
unless we get significantly better prices on sourcing from that
<bencoh>
(190* but you get the idea and what calculation I had in mind)
<bencoh>
the thing that puzzles me is the cost of electronic components, to be honest
<ZetaR>
DocScrutinizer05: From what I understand reading the logs and such, only a certain % of the components are hard to source and this is what would limit manufacturing for "last minute" orders. IIRC you were going to buy extra for some of these?
<DocScrutinizer05>
a lot of the cost goes to "administrative" overhead too. E.G. it took weeks and weeks to bring up that dang shop
<DocScrutinizer05>
as cheesy as it looks
<bencoh>
DocScrutinizer05: doesnt that go to "r&d" ?
<DocScrutinizer05>
yes
<DocScrutinizer05>
sort of
<bencoh>
I'm talking about raw sourcing price
<DocScrutinizer05>
ooh
<DocScrutinizer05>
you lost me
<DocScrutinizer05>
ZetaR: exactly
<bencoh>
DocScrutinizer05: err, my bad. the 190* calculation was about R&D cost. me being puzzled was about the raw sourcing price.
<ZetaR>
The Neo900 administrative overhead is probably tiny compared to a big company because you don't have to coordinate tons of people while supporting a huge parasitic management structure.
<DocScrutinizer05>
so far we bought 1000 for the cheaper parts (e.g. display connectors) and now 500 of the RAM chips
<DocScrutinizer05>
ZetaR: a huge company doesn't mind 20 times the cost for administration when they sell 10000 times the device count
<DocScrutinizer05>
still the admin cost per device is 1/500 of ours
<DocScrutinizer05>
same for all fixed cost
<ZetaR>
DocScrutinizer05: If they were more efficient with a better customer relationship, the price would probably be even lower.
<ZetaR>
Though, better employee relationship would be good too, so they aren't effectively raping impoverished contries...
<bencoh>
ZetaR: sure, but they dont "care", especially considering the "risk" in doing so
<ZetaR>
(depends on the company of course)
<DocScrutinizer05>
bencoh: (cost of components) another tricky thing. We used the mouser/digikey/farnel public price lists. When actually sourcing, we of course will try the industrial grade sourcing which _sometimes_ can result in _dramatically_ better prices
<ZetaR>
bencoh: What do you mean by "risk"? I guess one of the things the management is good at is using external forces to maintain power in the market, e.g. lobbying governments.
<DocScrutinizer05>
on the bright side, you can watch the farnel or digikey component prices every day, so do we now basically, and we are pretty sure they can't get any higher than they are now
<bencoh>
ZetaR: trying to be more "efficient" (reducing structural/management cost ?)
mvaenskae has quit [Ping timeout: 258 seconds]
<bencoh>
DocScrutinizer05: can we really expect an "industrial grade" sourcing price for 500~1000 devices ?
<ZetaR>
DocScrutinizer05: I have a question about tantalum caps. I have read that most of this is mined in Congo with poor supply chain transparency (i.e. it is a conflict mineral). Is there a good way to prevent getting conflict tantalum with sourcing? I suspect you are using tantalum since the size is critical for a miniaturized application.
<DocScrutinizer05>
bencoh: I had no better words, but just as an example, the PoP 1GB RAM chips were cheaper than the 512MB when bought via Farnell
<bencoh>
oh, hmkay
<DocScrutinizer05>
plus all R&D on top which made them more expensive in total
<bencoh>
how much cheaper ?
<DocScrutinizer05>
iirc a few cent
<bencoh>
percents* ?
<bencoh>
-s
<DocScrutinizer05>
and I have two sources for them, one being twice the price than the other, but with shipping included and proper waranty, while the other is not evaluated at all yet
<DocScrutinizer05>
anyway 100% to 200% you see what we're talking about
<bencoh>
kinda
<DocScrutinizer05>
in THEORY the components price could go down to 25% of what we calculated
<ZetaR>
"It takes longer to do things quickly; its more expensive to do it cheaply; and its more democratic to do things in secret." - Yes, Minister (paraphrase)
<bencoh>
meaning ... ~85e
<DocScrutinizer05>
but you better don't hope for that to _ever_ happen
<bencoh>
DocScrutinizer05: krkr :))
<bencoh>
okay
<DocScrutinizer05>
if we would build them critters in Shenzen, we could go shopping chips next door and get *really* good prices
<DocScrutinizer05>
and basically that's exactly what large companies do
<ZetaR>
DocScrutinizer05: So in your aspirations you are an optimist, but since you are actually a realist you decide instead to be a pessimist? ;)
* DocScrutinizer05
needs more coffee
<bencoh>
:D
<bencoh>
DocScrutinizer05: let's relocate then :>
<DocScrutinizer05>
Online Visitors in the last 30 minutes: 20
<DocScrutinizer05>
Active Shopping Carts in the last 30 minutes: 1
<DocScrutinizer05>
NeoNDOWN PAYMENT NeoN bare board-25
<DocScrutinizer05>
+1EURDown Payment top up-2981
<DocScrutinizer05>
ZetaR: (tantalum) it's almost impossible, but we don't use any tantalum capacitors in our device iirc
<DocScrutinizer05>
ooops
<DocScrutinizer05>
no ooops
<DocScrutinizer05>
my eyes skipped a line and thus mixed ZetaR and bencoh
<DocScrutinizer05>
ZetaR: usually capacitors on that level are X5R or X7R ceramic capacitors. Tantalum capacitors are used for huge values (like 1000uF) usually in PSU and regulators of PC mobos and similar stuff
<DocScrutinizer05>
our capacitors are basically physically too small volume to use tantalum
<DocScrutinizer05>
even 100uF you can get basically same size/volume/footprint X5R as tantalum. and X5R ceramic never ages, while tantalum is often dead after a few years
<DocScrutinizer05>
ceramic capacitors have other issues, like they lose capacity with voltage applied. And they are noisy and susceptible to noise - yes, they make mechanical noise and could act as "microphone", but we take care of that aspect to protect our users' privacy, by not using them in a way where you could exploit that side effect
<DocScrutinizer05>
loss of capacity is "dramatic". A X5R capacitor can have as little as 50% of nominal capacitance when charged to nominal voltage
<bencoh>
hmm
<bencoh>
meaning it draws current ?
<DocScrutinizer05>
no
<bencoh>
so it's just a capacitance reduction ?
<DocScrutinizer05>
yes
<DocScrutinizer05>
nonlinearity
<bencoh>
interesting
<DocScrutinizer05>
which makes them ceramics quite bad a choice for audio
<wpwrak>
solution: use caps with Vmax >> Vin_your_circuit
<DocScrutinizer05>
yes
<DocScrutinizer05>
hi wpwrak
<DocScrutinizer05>
either 2 times voltage or two times needed capacity
<wpwrak>
drawback: volume ~ V * C. further drawback: cost ~ volume^X where X > 1
<DocScrutinizer05>
the former is the better solution since it adds security margin on top
<DocScrutinizer05>
:nod:
<wpwrak>
i guess for the audio scenario, larger capacitance wuoldn't help to avoid distortions. but for bypassing, yes
<wpwrak>
(security) yup, that too
<bencoh>
larger, as in volume ?
<DocScrutinizer05>
for audio, if we absolutely need capacitors, we might actually go for electrolytic, maybe even tantalum
<DocScrutinizer05>
but I guess we don't need much capacitors anywhere for analog audio
<wpwrak>
is there any place ever where you don't need caps ? ;-))
<DocScrutinizer05>
IHF amp is class-D
<DocScrutinizer05>
for the amp input decoupling we go massive over-voltage
<DocScrutinizer05>
it's only a 1uF or somesuch
<wpwrak>
bencoh: larger capacitance (electrical storage). but yes, it also means larger volume. volume grows with capacitance and voltage. it's somehow proportional to the energy the cap can store
<wpwrak>
1 uF, pleasantly easy :)
<DocScrutinizer05>
yup
<DocScrutinizer05>
that's what they tought back when ;-P
<DocScrutinizer05>
thought*
<DocScrutinizer05>
"100uF?? nah! 1uF is way more convenient"
<jurov>
lol supercapacitor that can keep it 10 seconds running would be convenient for battery changes :)
<DocScrutinizer05>
jurov: that was actually palnned initially (though for < 1s). But it's not entirely clear if we can implement it, prolly not due to size restrictions
<bencoh>
fun thread
<bencoh>
(is in technically interesting, and funny)
<bencoh>
(as*)
<DocScrutinizer05>
I wasn't amused to learn my dream device of that time had fsckd up audio and thus no MP3 player for me
<DocScrutinizer05>
later on I came up with a bass-rework
<DocScrutinizer05>
since it was really impossible to push this bug fix beyond Tony, our chief EE
<DocScrutinizer05>
with board revision A8 I finally managed to get a 47uF iirc
paulk-collins has joined #neo900
<DocScrutinizer05>
and I learned the reason for their refusal been that the serial modem calibration for fab which is done via headset jack "broke2 with larger capacitors - since they forgot to consider switching the audio amp into high-Z state during modem calibration
<ZetaR>
DocScrutinizer05: Thanks for the info. I was thinking you probably had to use an electrolytic for the DC-DC conversion. Looking at my half-dozen used cell phones: I see 2-4 tantalum caps of a couple of them, but they are actually the older models. The newer boards look like they use only ceramics.
<ZetaR>
Lol. My oldest one has several SOICs on it.
<ZetaR>
LTC1265 = DC-DC converter, 7133 = buffer/driver, rf2132 = amp, RF9908 = RF transceiver, 24C128 = I2C EEPROM. All are SOIC. The rest are TSSOP, TSOP, and QFP, with only one BGA (CPU). Date on it is 1999.
<DocScrutinizer05>
:nod:
<ZetaR>
Hm. I'm looking at the N900 schematic. It surprises me that they are only using linear regulators rather than buck DC-DC converters. They might be larger, but I think you would waste much more power. Wouldn't sacrificing battery volume (for example) to make room for a buck converter increase your effective capacity? Also I did not see any polarized caps, it looks like they use arrays of 100nf or 1uf ceramic.
<DocScrutinizer05>
ZetaR: check out datasheet of twl4030
<DocScrutinizer05>
the linear regulators are only used where very little continuous power is needed
<DocScrutinizer05>
all the main converters are inside twl4030 and are buck-boost
<DocScrutinizer05>
mostly buck
<ZetaR>
I see. Took me a moment to find out which thing you were referring to. So its this big block box called "Gaia" on the schematic. That makes sense considering the presence of L4201/2/3.
<ZetaR>
Confusing because below "Gaia" it says N4200, and not TWL5030, but Maemo wiki say it's TWL5030.
panais has joined #neo900
<panais>
hello people
<panais>
I'd like to contribute to the project
<panais>
for starters I intent to devote around 2-3 hours a week
<panais>
I have an engineering and programming background
<panais>
what can I help with? :)
<ZetaR>
I would have offered as well, but I think that I am enough of a noob that the hand-holding would just suck up the time of the real devs.
<ZetaR>
My level of engineering/programming is fiddling around with Arduinos.
<panais>
I think that's not a problem
<panais>
every little helps :)
<ZetaR>
Well, maybe if there is some task that doesn't require much help from the devs for newer people to do it.
<panais>
yes, and if you don't want to do programming I'm sure there are other things that people can help with
<ZetaR>
I have some programming experience but I have actually been reading a lot about transmission lines and signal interference lately.
<panais>
very nice! :)
<ZetaR>
I was going to try my hand at designing a digital board for ADC, etc. but I found out that I didn't really understand how wires work. =P
<ZetaR>
Though, most of it is probably compulsiveness.
<ZetaR>
(it often gets in the way of me doing real work)
<ZetaR>
What is your experience panais?
<panais>
hehe, I get you!
<panais>
I'm a full time programmer, although my educational background and personal interest I would say is engineering in general
<panais>
I'm also crazy about DIY stuff, whatever that might be
<ZetaR>
That's cool. What is your focus in programming?
<panais>
I work both on standalone and web server applications
<panais>
I use mostly Java and C/C++
<panais>
but I like low level stuff too
<panais>
how about you, what's your background?
<DocScrutinizer05>
panais: for coding you should talk to freemangordon, Pali, and jonwil when around
<panais>
cool. thanks for that :)
<panais>
but I can also help with other aspects of the project if needed
<ZetaR>
Well, I minored in programming but I haven't used it much, though I have some experience with POSIX. I have studied electronics theory as a personal interest, but my real passion is probably mathematics and philosophy.
<panais>
I understand that there might be a shortage of people
<Humpelst1lzchen>
shortage? really?
<DocScrutinizer05>
panais: on sw side we would need somebody finally setting up a complete build environment for maemo based on only FOSS stuff, and actually compiling maemo from scratch
<panais>
@ZetaR: very interesting, I like seeing people with a diversity of interests :)
<DocScrutinizer05>
I could provide git repo hosting if needed (I don't think there's shortage on public git hosts though)
<panais>
@Humpelst1lzchen: maybe not then :P. I'm really new into this and I haven't gone through the relevant threads
<DocScrutinizer05>
on hw side we're not really short of manpower, since most of our design is basically decided upon and just needs implementation by GDC into (augmented further) schematics and layout. And for that task Nikolaus is a passionate solitaire
<panais>
@DocScrutinizer05: cool, I'd like to help with that (the build environment). I have to go now I'm afraid. Who should I contact for this? the people you mentioned?
<freemangordon>
panais: which build environment?
<panais>
on the hw side, I'd really like to help but I guess my experience is not up there yet
<DocScrutinizer05>
on administration we are short on server admin, which is basically dos1's domain. But that's difficult to share
<DocScrutinizer05>
what's urgently needed are newletter writers etc
<panais>
that is also an option :)
<DocScrutinizer05>
(person to contact): freemangordon ^^^^
<panais>
I will be in touch next week
<panais>
just for newsletter writing or for everything mentioned?
<DocScrutinizer05>
for building a maemo from scratch
* freemangordon
won't be available for the next ~48 hours
<DocScrutinizer05>
we need a job or whatever that people could start and that downloads all sources needed (from git) and then builds a complete image, from kernel to hildon desktop
<DocScrutinizer05>
basically it already works OOTB n stock maemo, when you attach the modem GPS as "external GPS receiver"
<DocScrutinizer05>
assuming stock maemo would run on Neo900, you select either the right /dev/ttyUSB or /dev/ttyS0 or whatever, depending on whether you use USB or UART for the data from modem
<DocScrutinizer05>
prolly the stock maemo "external GPS" is made to only use BT GPS mouse, so we had to kicj out the BT layer I guess, or tune/patch from /dev/ttyBT (???) to /dev/ttyUSB
<DocScrutinizer05>
:-) easy and a nice target, no?
<freemangordon>
yep
ZetaR has quit [Ping timeout: 246 seconds]
xes has joined #neo900
xes_ has quit [Ping timeout: 272 seconds]
louisdk has joined #neo900
ZetaR has joined #neo900
sparetire_ has joined #neo900
<DocScrutinizer05>
freemangordon: *IF* liblocation and friends use detail info of WWAN like CellID or whatever, we can get some of that from Cinterion modem too. Ask me about details when needed
wazrus has joined #neo900
wazrus has quit [Quit: Lost terminal]
louisdk has quit [Ping timeout: 264 seconds]
rjeffries has quit [Ping timeout: 276 seconds]
rjeffries has joined #neo900
xes has quit [Quit: leaving]
<lexik>
Have you considered a community-run shop/marketplace with members-submitted parts/modifications? In case somebody will came up with custom super-cool 3D printed case or something, it would be nice having it distributed 'the official way'
xes has joined #neo900
<lexik>
Maybe even thread on TMO linked from neo900.org will handle that just fine
<lexik>
(Also, and most importantly, thanks all the involved for taking care of this project that far)
<DocScrutinizer05>
lexik: of course we will support this
lobito has quit [Quit: Leaving.]
lobito has joined #neo900
jonsger has quit [Ping timeout: 256 seconds]
paulk-collins has quit [Quit: Quitte]
illwieckz has quit [Remote host closed the connection]