Humpelstilzchen has quit [Ping timeout: 256 seconds]
louisdk has quit [Ping timeout: 255 seconds]
sparetire_ has quit [Quit: sparetire_]
<ZetaR>
DocScrutinizer: You were a big help in my understanding of his position. Actually, just knowing that he would be doing layout allowed me to better understand why changes could raise the blood-pressure. =) I hope that I can foster a good relationship, since I am most interested in the low-level stuff that has relevance to layout.
<ZetaR>
DocScrutinizer05: Another option with the caps is that a switch from 100nF to 10nF could allow you to change X7R caps to NP0 caps without increasing the size.
<ZetaR>
That would be more expensive though, and it would negate the improvements in layout and filtering from decreasing the cap size.
<ds2>
more expensive...heheh
<ZetaR>
ds2: It is a serious consideration though, since making the phone better could increase its utility and lifespan... the question would be: would the cost per benefit be worth it? I don't think NP0 caps would add much total expense to each phone.
<ZetaR>
Each one is fractions of a penny.
<ZetaR>
Though there would be many many caps on the phone.
<ds2>
0.01uF NP0's?
<ds2>
they are like 2-3x if not more then X5R/X7R
<ds2>
most NP0's are in the pF range
<ZetaR>
True. You are right. That is why I didn't suggest that in my proposal that Doc was referring to.
<ZetaR>
I checked though, there are 10nF NP0s in 0603 package sizes.
<ds2>
they exists, just not cheap
<ZetaR>
Yes, but comparatively. I guess I could estimate how much cost it would actually add to each phone...
<ZetaR>
I personally don't think its worth it, but I thought I would bring it up.
<ds2>
for bypassing NP0 is overkill
<ds2>
i think you can do better with a higher voltage rating X7R
<ZetaR>
The main issue in my mind is that X7Rs are peizoelectric devices.
<ds2>
the MB is generally not exposed to physical strain (relatively speaking)
<ZetaR>
Which means that you can A) measure sound by measuring voltage (e.g. by the modem) and B) measure IC activity by measuring soind.
<ZetaR>
sound*
<DocScrutinizer05>
((estimate how much cost it would actually add )) the question is how much benefit it would add, as well
<ZetaR>
There have been papers published on breaking crypto using a microphone.
<ZetaR>
IIRC the attack measured the acoustic emissions by the power caps of the processor.
<DocScrutinizer05>
we have zhat on our radar
<ZetaR>
IDK how much of a threat it is outside of the lab though.
<DocScrutinizer05>
there *might* be worse threads: think BT to codec analog line buffer cap, which actually could pick up ambient noise
<ZetaR>
Actually, you could do something like: determine what ICs are "sensitive" and use NP0s to bypass them, and use X7Rs on everything else.
<DocScrutinizer05>
yes, that's the plan
<ZetaR>
Yeah, I meantioned that above.
<DocScrutinizer05>
_my_ plam at least
<ZetaR>
Ah, okay.
<ZetaR>
My calc shows that you could probably decrease the cap value on a lot of these ICs, meaning that NP0s would be more within reach.
<ds2>
wonder if there are ways to use pairs of caps to cancel out the noise
<DocScrutinizer05>
interesting idea
<DocScrutinizer05>
most audio lines ared ifferential anyway
<DocScrutinizer05>
differeantial*
<ZetaR>
ds2: Hmm, interesting but I don't think you could do that reliably without lab testing.
<DocScrutinizer05>
meh, my typing sucks
<ds2>
wonder if doing a study like that is enough for a PhD
<ZetaR>
Eh, we understand you Doc.
<DocScrutinizer05>
to much prolly ;-)
<ZetaR>
That is what is important.
<ZetaR>
((cancel noise)) also, if you screwed up cancelling it might actually make it worse.
<DocScrutinizer05>
thatks for pointing me at "high capacity" NP0 anyway
<DocScrutinizer05>
do you have a link?
<ZetaR>
Uhm, no, but I can find them again.
<ZetaR>
What is the quantity that you think you will need?
<DocScrutinizer05>
piezo of X7R sucks big time, often
<DocScrutinizer05>
weird noise escaping the PCB and you got no idea where from
<DocScrutinizer05>
GTA02 did nasty click sounds during power up iirc
<DocScrutinizer05>
took a while until I understood it was the buffer caps getting charged
<DocScrutinizer05>
generally inside->out attacks via audio or whatever emissions from VCC are hard
<DocScrutinizer05>
you can't debug a CPU by minitoring the powersupply current
<ds2>
depends on how important is the info
<DocScrutinizer05>
unless you know *exactly* which code is running
<ds2>
start a conference call with the wrong folks and decypher a mail at the same time....
<DocScrutinizer05>
ds2: not really
<DocScrutinizer05>
well, I'm no math wiz, but I guess it's impossible to decode the mail content from the CPU's current consumption
<ds2>
not the mail but there have been attacks to determine keys
<DocScrutinizer05>
e.g. for a CPU power it's irrelevant if the data bus has a 01010101 or a 00001111 pattern, both consume same amount of current
<DocScrutinizer05>
yes, you can decode/uncover keys by long time monitoring the EM emission and stuff of a SIM
<DocScrutinizer05>
since you know the algo
<ZetaR>
Also, a proper implementation of a crypto algorithm will attempt to twart various side-channel attacks.
<DocScrutinizer05>
and you can check back which of your 5000 possible solutions actually works
<DocScrutinizer05>
that too
<DocScrutinizer05>
nowadays SIMs are hardened against such attacks, usually
<ZetaR>
e.g. it is considered highly improper to have your implementation take a different amount of time depending on the result of a calculation.
<ZetaR>
That is, you will get slapped by the crypto community.
<DocScrutinizer05>
and NB there's no way to monitor voltages or whatever sufficiently exact in Neo900
<DocScrutinizer05>
no implemented way that could get exploited
<DocScrutinizer05>
ZetaR: yes, deferred evaluation of comparison results is a usual method
<DocScrutinizer05>
e.g. instead a while ( i < X) i++; you do a for (j=0; j<max; j++){if (j==X) then i=0}
<DocScrutinizer05>
actually for (j=0; j<max; j++){if (j==X) then i=j; else i=0}
<DocScrutinizer05>
ugh, it's too late for this silly stuff
<DocScrutinizer05>
actually for (j=0; j<max; j++){if (j==X) then i=j; else dummy=0}
<DocScrutinizer05>
actually for (j=0; j<max; j++){if (j==X) then i=j; else dummy=j}
<DocScrutinizer05>
and ideally you define both i and dummy either as registrer vars or on two absolute memory addresses that have same number of 1s in their address
cybiko123 has joined #neo900
cybiko123 is now known as Guest30318
Guest30318 has quit [Client Quit]
<DocScrutinizer05>
when you wanna laugh, here's my latest story about my recent first encounters with windows (at least this century): the USB3.0 of that gigabyte board doesn't work as expected, so I checked and found the "eXtensible USB hub" or whatever is showing up twice in device manager, once with Intel driver, once with VIA driver. I disabled the intel and TADAA no more mouse no kbd ;-P
<DocScrutinizer05>
not even on the USV2.0 jacks
<DocScrutinizer05>
USB*
<ZetaR>
DocScrutinizer05: For NP0s in 0603 size w/ 10k quantity 1nF-10nF, here is the price curve (USD) for the least expensive products on Digikey: 1nF = 0.00615, 1.2nF = 0.016, 1.5nF = 0.0158, 1.8nF = 0.0174, 2.2nF = 0.0168, 2.7nF = 0.0206, 3.3nF = 0.022, 3.9nF = 0.0274, 4.7nF = 0.0399, 5.6nF = 0.0399, 6.8nF = 0.0399, 10nF = 0.056
<DocScrutinizer05>
I fuess that's USD?
<ZetaR>
Yes.
<ZetaR>
All 50-100V.
<ZetaR>
Checking other values.
<DocScrutinizer05>
we can look up the prices when we have a component part number
<ZetaR>
Most 5%, some 10% tolerance.
<ZetaR>
This is just to give you an idea of what is feasible.
<ZetaR>
-55 - 125*C
<ZetaR>
Checking for automotive.
<DocScrutinizer05>
just give me the link
<ZetaR>
These are lists.
<ZetaR>
I am just using the Digikey search.
<ZetaR>
Filtered for NP0/0603/1-10nF etc.
<ZetaR>
Is there a specific one you want a link to?
<ZetaR>
The CGA3E2C0G1H***J080AA / CGA3E2C0G2A***J080AA looks very nice to me. It showed up quite a lot near the least expensive of NP0s for various values, and yet is AEC-Q200 and has 50V and 100V models.
<ZetaR>
Looks like class 2 caps have an aging mechanism. For X7R the capacitance loss is only 5% over 10 years though.
<ZetaR>
Something to do with crystalline changes in the ferroelectric material.
<ZetaR>
Wait wait, I am misinterpreting this graph.
<ZetaR>
Looks like it is 5% per "hour-decade", where "decade" refers to a logarithmic scale. 5% loss after 1 hour, 10% after 10 hours, 15% after 100 hours, etc.
<ZetaR>
After manufacture that is.
<ZetaR>
After manufacture the cap goes through 4 "decades" of aging, or 1000 hours (41 days), and then during reflow the cap essentially gets "de-aged" by the temperature, and is intended to then be stable.
<DocScrutinizer05>
as I understand it it ages again after each de-aging
<ZetaR>
Interesting link. Fortunately NP0 caps are nearly immune to all of those effects.
<ZetaR>
Haha, but 10.9 USD!
<ZetaR>
Wow, it is huge too.
<ZetaR>
For a 1uF cap.
<DocScrutinizer05>
well, NP0 has a dielectric constant of only a few 100s, wsy 4000 for X5R and 16000 for Y5V
<DocScrutinizer05>
s/wsy/vs/
<ZetaR>
So would the Neo900 use Y5V for the power supply?
<ZetaR>
The class 3 caps go up to 50,000 according to Wikipedia, but for some reason they can't be made multilayer. Maybe something to do with the sintering process.
<DocScrutinizer05>
I think we'll try using X5R/X7R whenever possible
<DocScrutinizer05>
Y5V lose too much capacity with voltage
<ZetaR>
Is it significant at 1.8-5V?
<ZetaR>
I mean, significant enough that an X7R could have a similar capacitance in the same size.
<ZetaR>
Or X5R.
<DocScrutinizer05>
why would we use Y5R anyway?
<DocScrutinizer05>
err Y5V
<DocScrutinizer05>
when we have one slot with e.g. 100nF X5R on our component shooter, we will use that whenever possible
<DocScrutinizer05>
for the quantities we produce, streamlining the production process and machine setup is way higher importance than saving a few micro-cents per capacitor
<ZetaR>
Okay. I thought that you might be able to use it to buffer input voltage better or something.
<ZetaR>
i.e. through the battery or USB.
<DocScrutinizer05>
what would be the advantages on buffering of Y5V over X5R?
<DocScrutinizer05>
ooh, you meant for the 100uF or whatever, yes for those we might use whatever meets our requerements best
<DocScrutinizer05>
so possibly even Y5R
<DocScrutinizer05>
or even tantal
<DocScrutinizer05>
if nothing else works
<ZetaR>
Okay. I try to avoid tantalums though, because of the conflict mineral issue.
<ZetaR>
I always find it not only sad but ironic, because most of these minerals could be mined safely and efficiently locally in the developed countries, but instead these poor countries are raped for the minerals because it is a bit cheaper.
<ZetaR>
I have heard talk in the US, worring about depletion of reserves of minerals, e.g. lithium, and my response is usually something like "but the US has significant lithium deposits". These things are more social/political/economic problems than technological ones.
<DocScrutinizer05>
not US, but south america
<DocScrutinizer05>
there's a whole huge slatlake full of lithium salts
<DocScrutinizer05>
at insane altitude
<DocScrutinizer05>
Peru?
<ZetaR>
IIRC the US has significant lithium deposits as well, they just aren't as cheap to mine. Lithium is everywhere, and you can even extract it from seawater practically.
<ZetaR>
Practically just doesn't translate to competatively.
<ZetaR>
That is interesting, why is it at high altitude?
<ZetaR>
I wonder how the lithium got into the lake. Leeching from the soil, I guess?
<DocScrutinizer05>
or was it "Salar del Hombre Muerto at around 13,123 feet above sea level on the border of the northern Argentine " ?
<ZetaR>
Wow, mountains of lithium salts.
<ds2>
mountains of lithium metal would be more impressive ;)
<DocScrutinizer05>
I guess what I meant was "Uyuni salt lake" in Bolivia
<ZetaR>
ds2: It would be, especially since it corrode before your eyes, and the heat from a mass of that size would probably be enough to make it burst into flames.
<ZetaR>
I would hate to have that job though, lithium is pretty toxic.
<ds2>
could be worse... could be a mountain of cesium
<ZetaR>
Lol. Or it could be EVEN worse: a mountain of francium.
<ZetaR>
Technically an alkali metal, but it is so radioactive that a mountain of it would probably kill you near instantly and then would be gone by the time your body was cold.
<ZetaR>
That is, assuming your body wasn't just vaporized.
<ZetaR>
The ironic thing about lithium toxicity is that it is a pretty good mood stabilizer.
<ZetaR>
And the effective dose is close to the toxic dose.
<ZetaR>
DocScrutinizer05: What do you think about epoxy underfill? I have seen it used for caps and for BGAs.
<DocScrutinizer05>
you mean glue?
<DocScrutinizer05>
dunno, that's a question for the fab engineer
<DocScrutinizer05>
so: Global components
<DocScrutinizer05>
my optinion: dang annoyingly hard to unsolder
<DocScrutinizer05>
-ly
<ds2>
it supposely adds rigidity
sparetire_ has joined #neo900
<ZetaR>
I have been thinking a lot about longevity, and BGAs and caps do have problems with mechanical stress. IIRC the N900 had a problem with this and ICs under the keyboard, which AFAIK is fixed in the Neo900, but there could still be long-term problems from board flex.
Pali has joined #neo900
<ZetaR>
The caps are probably small enough that it isn't a significant issue though. The BGAs might be susceptible though.
panais has quit [Read error: Connection reset by peer]
`sno has joined #neo900
sparetire_ has quit [Quit: sparetire_]
`sno has quit [Ping timeout: 264 seconds]
`sno has joined #neo900
Pali has quit [Remote host closed the connection]
Pali has joined #neo900
`sno has quit [Ping timeout: 252 seconds]
mvaenskae has joined #neo900
paulk-collins has joined #neo900
che11 has joined #neo900
che11 has quit [Ping timeout: 265 seconds]
che1 has joined #neo900
louisdk has joined #neo900
che1 has quit [Ping timeout: 258 seconds]
louisdk has quit [Ping timeout: 245 seconds]
`sno has joined #neo900
nicksydney has quit [Read error: Connection reset by peer]
pigeons has quit [Ping timeout: 276 seconds]
cnj- has joined #neo900
paulk-collins has quit [Ping timeout: 272 seconds]
k4m has joined #neo900
<k4m>
hey yo, wassup!
<k4m>
I mean last news item was from 6 months ago. surely stuff has happened since then?
<`sno>
agreed, I was this project to succeed,with more pronouncements and details on progress it will get more exposure and hopefully more investment
<`sno>
was = want
<k4m>
disregard my question, I just noticed the microblog, `sno
<`sno>
the curtains are closing even on hackers being able to have a privatr conversation free of interference and snooping from malicious and tyrannical entities
<`sno>
the sooner this handset finds it way into my possession the better
<k4m>
it's nice. however I can't afford 990E+150E for a phone
<`sno>
if thats the price of freedom i'm ready and willing
<k4m>
imo the price needs to be cut in half (at least)
<`sno>
i dont even use a mobiel device anymore
<`sno>
don't trust anything
<comradekingu>
k4m: please subsidize everyone
<k4m>
comradekingu: I don't understand, subsidize who?
<comradekingu>
i can afford to pay 1k, but dont let that hinder you in helping out the others in reaching your goal
<comradekingu>
The ones you think should pay half, great idea
<k4m>
it is
<comradekingu>
With more takers, the price goes down, spread the word
<k4m>
I will probably donate anyway. that is, unless donations remain "suspended"
panais has joined #neo900
<panais>
hello people
<panais>
what up?
pigeons has joined #neo900
pigeons is now known as Guest53003
<k4m>
panais: not much bro. u?
<comradekingu>
my funding was accepted and processed
<panais>
same here I guess :)
<panais>
If anybody is interested or has any advice, I started my research on compiling maemo from scratch
<k4m>
why u compile maemo bro? why u no compile qtmoko?
<panais>
it was mentioned by joerg that this is something that will be needed
<k4m>
ok bro!
<panais>
I thought I should give it a go
<k4m>
do you have a BeagleBoard to test it?
<panais>
yes, I have a BBB
<k4m>
w00t
<panais>
beaglebone black#
<panais>
though I currently have know idea whether/how it will work on it
<ZetaR>
AFAIK, the plan is to make the Neo900 close enough to the N900 that most of the software work done in Maemo can be reused, while just pulling and replacing the proprietary drivers and such. I think this is mainly because the dev team doesn't have that much funding or manpower, so they need to economize their efforts as much as possible while still accomplishing their goals.
Guest53003 has quit [Ping timeout: 256 seconds]
<panais>
sounds right
<k4m>
it sounds right to me too!!
pigeons has joined #neo900
k4m has left #neo900 [#neo900]
pigeons is now known as Guest25661
Guest25661 has quit [Ping timeout: 244 seconds]
pigeons_ has joined #neo900
arcean has joined #neo900
pigeons_ has quit [Ping timeout: 265 seconds]
pigeons has joined #neo900
pigeons is now known as Guest78951
Guest78951 has quit [Ping timeout: 265 seconds]
paulk-collins has joined #neo900
jonsger has joined #neo900
pigeons_ has joined #neo900
mvaenskae has quit [Quit: Lost terminal]
panais has left #neo900 [#neo900]
<ilon>
Have anyone recently trying a fresh reinstall of CSSU on the n900, following the wiki guide?
<ilon>
it seems broken
<jurov>
i did
<jurov>
there are broken download links for nokia's original 1.3 images, but it wasn't hard to find the files elsewhere
<DocScrutinizer05>
#maemo-ssu
<jurov>
otherwise it went as described
<ilon>
DocScrutinizer05: sorry doc
<DocScrutinizer05>
np :-)
<ilon>
jurov: i just had troubles actually installing the cssu package
<ilon>
and some broken repositories
<jurov>
remove all broken repos before installing cssu. that was mentioned on the wiki iirc.
mvaenskae has joined #neo900
<ilon>
"Make sure you have no too weird catalog settings in Application manager, esp it's recommended you never disable the core Nokia SSU repositories."
<ilon>
only mention of them
<DocScrutinizer05>
seems that's not really correct anymore, since Nokia trpos are fown
<jurov>
^ i agree
<ilon>
Yeah
<DocScrutinizer05>
repos*
<DocScrutinizer05>
well, I am author of this sentence but feel free to update it after you discussed the issue with guys in #maemo-ssu
<ilon>
I just felt that maybe the guide should be updated slightly, to actually work for users doing it for the first time
<varu|zZz>
^ would be nice, i never got cssu working for similar reasons
<ilon>
:P
<ilon>
Well, I'm not really a fan of writing how-to's, but when i do, I make sure they work for an idiot like myself
<ilon>
:D
<DocScrutinizer05>
this guide is like 4 years old
<ilon>
yep, I know
<ilon>
And you can feel it when you try to follow it and find dead links :D
<DocScrutinizer05>
duh, dead links?
<jurov>
lol
<ilon>
Sorry, might be the n900 reflash guide i had in mind with dead links
<DocScrutinizer05>
there shouldn't be any dead links except to Nokia repos abd those are not explicitly linked afaik
<DocScrutinizer05>
ooh that one had daed links since quite some time, yes
<ilon>
sorry, the cssu guide didnt have dead links
<DocScrutinizer05>
since it needs to point to Nokia proprietary stuff
<ilon>
DocScrutinizer05: do you happen to have a backup of libwildmidi0 0.2.2-2maemo1?
<ilon>
DocScrutinizer05: its missing from the repos it seems
<DocScrutinizer05>
never ever heard of it
<ilon>
its needed for cssu upgrae
<ilon>
upgrade*
<DocScrutinizer05>
wow
<DocScrutinizer05>
i got it in my backups
<ilon>
:P
<ilon>
The following packages have unmet dependencies: mp-fremantle-community-pr: Depends: gstreamer0.10-plugins-bad-extra (>= 0.10.14-git266-0maemo10+0cssu6) but it is not going to be installed
<ilon>
gstreamer depends on it
<ilon>
i dont know what the proper way would be to get it into community repo
<DocScrutinizer05>
I don't think gstreamer0.10-plugins-bad-extra are actually something that should get shipped per default with any official release
<DocScrutinizer05>
not sure, about all that gstreamer(good|bad|extrabad|ugly|nasty) stuuf
<DocScrutinizer05>
stuff*
<DocScrutinizer05>
lemme put it this way: when libwildmidi0 was a CSSU dependency in any way, then *yes* that for sure is a bug in my book
<DocScrutinizer05>
anyway I upgraded from errr CSSU-T7 to 10(???) a month or so ago, and it went flawlessly
<DocScrutinizer05>
after I disabled all the broken/missing Nokia repos
<jurov>
"but it works for me" .. you know that doesn't help nor explain anything?
<jurov>
i confirmed you can't install CSSU on nokia's last image due to missing libwildmidi0 .
<jurov>
(it was some weeks ago, forgot about that at first,t oo)
ZetaR has joined #neo900
<ZetaR>
jurov: gstreamer plugins bad is the set of plugins that for one reason or another do not meet their code standards, or are not yet reviewed.
<jurov>
anc CSSU depends on them?
<ZetaR>
I don't think that it means that they are "bad" in general, though.
<ZetaR>
CSSU itself, or the update?
nox- has quit [Quit: Leaving]
<ZetaR>
If it is the CSSU system itself, then it would be kinda weird, because IIRC gstreamer is a bunch of modules for handling analog or encoded analog data.
<jurov>
not sure ywhat do you mean?
<ZetaR>
I mean: are you having trouble setting up CSSU itself, or is it the update process that is having trouble?
<ZetaR>
Because if it is the update process it might just be a missing repository.
<jurov>
1. i flashed the phone with RX-51_2009SE_20.2010.36-2_PR_COMBINED_MR0_ARM.bin
<jurov>
2. user the icons on the wiki to add CUSS repos and removed the broken ones
<jurov>
it did not allow me to update, so i added everything from CSSU and maemo.org
<jurov>
still no cigar
<ZetaR>
Let me see if I still have my notes from when I did that process.
<jurov>
then i found out about Problems tab and libwildmidi0
<jurov>
when i installed libwildmidi0 , the CSSU upgrade succeeded
<jurov>
*installed libwildmidi0 by dpkg, to be precise
<ZetaR>
Hm, in my notes the thing about CSSU just says something like "go to wiki.maemo.org/CSSU, click stable one-step install, accept when the manager asks for confirmation, after a while it will ask again for confirmation, run the new community SSU app."
<jurov>
yes i ran it
<ZetaR>
So it gives you an error about gstreamer when you try to run it?
<jurov>
no, i did not say anything about gstreamer
<jurov>
it gave me error "missing dependency libwildmidi0"
<ZetaR>
Oh, pardon, I confused your message with someone elses.
<ZetaR>
What is the version of the EMMC and COMBINED that you flashed?
<ZetaR>
Mine was 10.2010.13-2 for the EMMC and 20.2010.36-2 for the COMBINED
<jurov>
i did not flash emmc
<ZetaR>
Okay, sorry, I just woke up and I am groggy.