« first day (1705 days earlier)      last day (3248 days later) » 

Anonymous
2:04 AM
 
I2C headache, if I i double the manufacturers claim for "maximum time before data" their sample code works, if I leave it be their sample code times out 50% of the time
@PatoSáinz There was a parade today and nobody was wearing this
 
Anonymous
@crasic Lesbian, Gay, Bisexual, Transgender and Engineers
 
May as well be, half of everything was sponsored by tech companies
 
2:23 AM
@PatoSáinz It's a sign that means "it's complicated".
@crasic Which I2C device are you grappling with?
 
@NickAlexeev atmel sha
The code is - send command - wait minimum time - try reading - if nothing poll for ready until maximum time hit
but it times out - consistently, every other assy
 
@crasic Does it hurt to increase the timeout time like 10x ?
... and call our neighbor Atmel on Monday.
 
They haven't been helpful
although I have bus captures of succesful and unsuccseful communications
it needs a wakeup command and they try to resync so it basically chokes on its own
Maybe with data I can get an engineer to talk to mne
@NickAlexeev I have to understand why and engineer a way to be resilient , this is a supposed to be for security lockout, if it might not work as expected and deletes our ip customers won't be happen
Ideally I would write comm code from scratch myself, at some point when there was too much to be done this device was pulled in from sample code and it has been living like that and misbehaving since
but infuriatingly inconsistently
 
 
5 hours later…
7:16 AM
@crasic Can't say that I have had any of those problems using it with Mega, Xmega, SAM or LPC. Though with LPC there were many other problems, but that's LPC :-D
And SAMs usually have something inside, so it doesn't happen often I use it with that, as well as XMegas. Those were mainly tests with experimenter boards.
 
 
2 hours later…
9:07 AM
@jippie only with free access to the necessary power lines...
but at least it has these nice red carry handles...
 
@PlasmaHH matter of some industrial grade metal cables and a grappling gun, isn't it?
 
@Asmyldof would need some space to run the cables thorugh...
 
9:28 AM
Air?
 
@Asmyldof and held there by siemens airhooks?
 
... what part of grappling gun did you not understand?
:-P
 
the part where it enables me to run a few km of cable freely through air? ^^
 
Surely you are going to have some cables near?
 
@Asmyldof 380kv? that would be quite a few kilometers outside the city
 
10:27 AM
Does it matter whether you get 230V or 18V?
It'll still be welding currents, so....
 
I can't draw more than 65kVA from anything thats within reach... pretty lame it would be
 
65kVa over a 1mm gap.. let alone 0.1mm... that's serious plasma, plasma
Titanium arc welding, etc
 
11:06 AM
@PlasmaHH Or, you could just have a blimp hold up the wires above your transformer... presto, done.
 
@Asmyldof hm, low on budget here, can I use a hydrogen one instead of helium?
 
Sure! Can't see why not
I hear Germans did that before with great success, by the way
 
 
2 hours later…
12:59 PM
@PlasmaHH The other day I was at the factory in Nijmegen where they make those transformers. Impressive beast if you stand next to one. Even a single coil is huge, not to mention the core material.
 
@jippie Hm, how big are the windings? Skin effect should be about 9mm at 50Hz so this might be getting an issue there... and what is the core material? surely not laminated steel or ferrite...
 
laminated steel.
windings are square copper, or at least flat so it stacks easier.
 
thats going to be a hell lot of layers
 
windings are wrapped in paper, best isolation for the job.
 
Just hope its not cheap newspaper ;)
 
1:06 PM
@PlasmaHH Royal Smit Transformers are one of the world's leading manufacturers, I don't think they use newspapers to wrap the copper.
 
@jippie thats good, I was just imagining the oil boiling and the thing exploding... it loosk nice on a small transformer, wonder how it looks like on those ^^
 
@PlasmaHH once a coil is wound, it is dried. By removing the moisture from the paper, the coil shrings about 10cm in height.
I'm really an electronics guy, but it is amazing to have a tour in a factory like this one.
you kinda feel very small ;o)
 
same for A380 assembly...
 
1:28 PM
@PlasmaHH @jippie Technically, they could be using newspaper, with governament and market pressure on using receycled. Sure, it's not visible as newspaper any more, but, it's quite likely it is for 50%+
Do they use a single winding, or parallel sets for skinning and fringing? And in either case, about what size?
Facts man, facts!
Numbers.
With decimal points
:-D
 
@Asmyldof the problem with newspaper is the toner. It is mostly pure carbon and dissolves after a short time in the oil
 
@PlasmaHH No, no no, you were talking about the news_paper_ not the news_toner_
:-P
 
2:26 PM
@PlasmaHH liked your comment-answer to the HV cap question :-)
 
@Asmyldof I really never understood how people could ask questions at that level... It is like those "I want to learn electronics, what projects could I do" things
oh that next one is even bette
r
 
@Asmyldof They use six windings :-p
@PlasmaHH yeah they have this strange "worry" about arcing between windings, amongst other things because of contamination. Don't know why.
 
@PlasmaHH Luckily someone added the option to vote "I'm voting this off topic, because it's blatantly off-topic" to the last one
Which I just clicked
 
it needs some rep to do
 
I believe one of the reasons for their worry is that after the transformer has exploded, similar sized transport is required to transport the bits and pieces back to the factory for post mortem analysis.
 
2:36 PM
@jippie But whether that contamination comes from trees or newspaper, they'll have to stringently wash in all kinds of hydro-fluorides anyway, so I bet that they use waste paper as well to seem "clean on paper" (pun somewhat intended)
Lol
@PlasmaHH I know. I'm working on getting there ;-)
 
@Asmyldof not sure if you are LOL'ing about the transport thing, but that is actually a serious issue. You can't just send the thing back to the factory.
 
after it has arced...
 
@jippie I know, but it is funny to state that as primary fact of worry rather than the socio-economic issues concerned with something you built or tested&verified exploding with an energy level associated with 380kV @ whatever those things are allowed to peak at in current.
Wasn't it in Russia one of the universities had a small transformer explosion costing them half a wing mid 2000's?
 
@Asmyldof hundreds of MVA's, can't recall.
 
Believe it was only treatably woundeds, but still
That was one rated quite the same as the connection the TU/e has (or had, now EE had to move?)
 
2:43 PM
@Asmyldof another worry would probably be that it is bad for your company name in the transformer business ;o)
 
Ah, you can always Peugot it and move into consumer business. start making laminated bicycle frames
 
wait I saw this little video from KEMA the other day, it was 11kV I believe
 
We were always joking about "accidentally dropping a wrench over the railing"
Or once in a while "accidentally dropping one specific professor over the railing"
:-D
 
that one is tiny
and unavailable apparently :-/
 
?
I wonder what material went up in smoke there... likely not healthy
 
2:49 PM
just a Rittal'like cabinet I believe
DNV KEMA done some cool experiments in their labs :-p
 
yeah, if someone gave me a few thousand ultracaps I could do some cool experiments too...
 
@PlasmaHH they don't use caps.
 
sure, but for me it would be cheaper to use caps than to have some bigger power lines installed ^^
 
@jippie Guess who was there when they did that cabinet explosion
I could quickly see mr Ypma at the end of the look around as well
I think
 
3:13 PM
@PlasmaHH they don't use power lines either.
 
overunity circuits?
or a bunch of cr2032?
 
@Asmyldof John McEwan?
@PlasmaHH damn it was supposed to be a secret.
 
@jippie Well, it was either a very similar explosion to the one I witnessed together with some fat-cats and colleague recent ex-students and students, or it is the exact same, the back of the head and profile of one of the guys there reminds me of someone who was there. I just didn't see myself on the look-around, so it may be a coincidence. I know Edwin was filming with his digi-photo-cam and did a look around.
Nope, just saw all the way at the end who was laughing in suit, that's one of my predecessors at the student board, very likely the same explosion.
 
4:09 PM
I feel like baking
What should I bake and why?
Winning suggestion MAY be eligible for a DHL World Express package with part/one-piece. No warrantees to freshness and or mould-free-ness.
 
Anonymous
4:20 PM
@Asmyldof
 
@PatoSáinz Nope. No-bake is not bake, disqualified. Also 4th of July &/| US themed suggestions will by definition not win
:-P
 
Anonymous
hahahaha it's tasty tho, I made it once
 
Anything with strawberries is tasty, because strawberries are tasty, which is also why adding anything to them is a pointless exercise, which is why that's another reason the suggestion wouldn't have won :-P
 
Anonymous
 
4:26 PM
Denied
 
Anonymous
why
 
Because the suggestion to have me use a website to create a suggestion is a non-suggestion
And, also, mainly, because I don't do recipes. I do ingredients and targets and what comes in between is art
 
Anonymous
last try
 
Anonymous
is that good enough for you?
 
4:28 PM
I like the Death-cookies part
 
Anybody know of a bus analyzer that works with precaptured data?
 
Not so much a fan of the juxtaposition of them being OBL Death Day Cookies and they then needing hate
C/C++ and a compiler?
So, effectively, no
Oooh, or Python
 
Yeah that was my plan B
 
Python+PyGraph+PyMath+PyAwesomesaucePlugin
 
I have the libraries written to load in the scope data and format it nicely, wanted to avoid banging out an I2C reader
PyI2CTranslator
 
4:30 PM
Should be easy
 
The stupid thing is that my scope does the bus analysis for me, but refuses to save the translation
 
Easy enough to do, if it were my problem. Which, luckily, it isn't, so I can safely continue thoughts about baking
Rigol bug?
 
2GB of trace data though
 
<-- has 0 Rigols
 
No DPO
er tek DPO
 
4:31 PM
Hm
Heyho
Eh
Is the data saved in a manner that's compatible with.
Eh
Crap, forgot the name
Free SW or maybe even Open Source hardware based PC USB Logic Analyser 24MHz
Saleae
Sal...
Something with Sal and too many sounds after it
 
Anonymous
saleae
 
I were rightly!
 
Anonymous
@Asmyldof do you know what is the hardest part of building that little circuit for 2x the frequency?
 
Anonymous
scavenging the damn resistors
 
?
I have a cheap fake Chinese one... Very happy with it for travel purposes
 
Anonymous
4:34 PM
I have the caps, the mosfets, the diodes
 
Carrying big 32 channel machine with laminated iron core transformer = meh. Carrying tiny plastic box and laptop I needed anyway = yay
 
Anonymous
but I have a bag full of old pcbs with through-hole resistors
 
Anonymous
and it's just too much to start filtering through to get the values I need
 
0.01% value increments
 
Anonymous
@Asmyldof nope, here in Chileistan, getting electronics is a hassle and shipping is a hassle
 
4:36 PM
or 0.1% if you don't want to spend more than a dollar per piece. 1% if not more than five cents. 5% if not more than a cent
Getting a gun though....
Gun --> cheap electronics
 
Anonymous
there are a couple of radioshack-esque stores BUT they all are progressively stopping selling components and now they sell arduinos and shields
 
Anonymous
@Asmyldof soldering gun I presume?
 
No. Bullet gun
pewpew-type
 
Anonymous
soldering bullet gun
 
Anonymous
at the end of the day, bullets are made from lead, and so is solder
 
4:40 PM
Yes, but bullets go faster than solder. Faster is better
Next time shoot your components, QED
 
Anonymous
speaking of damaging components: I'm tired of only having a 500W soldering gun
 
Anonymous
I need to get a soldering station with regulated irons and things
 
500W?
Mount a blower above it, you'll have a hot air gun
 
Anonymous
somewhere in between 250 and 500
 
Anonymous
I do have a hot air gun haha
 
Anonymous
4:42 PM
750W
 
Anonymous
or something
 
INstallation wire wrapped around a nine inch nail?
 
Anonymous
@Asmyldof not good enough thermal transfer
 
Anonymous
the best part is that the little transformer inside the gun heats itself too after a while
 
Then why 250 to 500?
That's normal
 
Anonymous
4:43 PM
@Asmyldof because I can't quite remember the exact value
 
Anonymous
yea but it's overkill and not precise at all
 
Anonymous
so I end up heating up some components and I fear I may damage them
 
Anonymous
and dry some electrolytic caps
 
Weller WS-50 is a good replaceable tip station
 
Which, as we all know, should be done with the use of a reflow oven stuck in pre-heat due to badly designed PID
 
4:45 PM
WES51D
I have a metcal and hakko and the tips are just amazing
 
Or any of a number of stations that do not cost less than $50
 
Desoldering tweezers are a great addition
 
I have about the entire weller 1970 - now range, skipping some redesigned models
because who needs glossy teal rectangular WTCP50 if you have three in-places-molten mint green MagnaStat-50 's
I like the new smart stations. Sounds like a lot of money for nothing until you start using them and then at some point look across to the WPS81 (I think?) and think towards it: You foolishly simple toy. Caveat: To a hobbyist they are most certainly going to be money wasted.
 
the wes51 is basically the magnastat with a temperature knob
 
Except that with a temperature knob it's not the magnastat
 
4:51 PM
It has the locking mode with magnet pen
 
Egh
That's...
Nope
Glad I skipped that one
You either have nubcakes you don't trust, and you don't give them a knob, or have one that can be locked, like in the 80/81 series, or you accept that you have proper engineers in your workplace and give them a freely movable knob
 
Its movable but you can lock the temp with a pen
 
For children's workshops I prefer the Magnastat no-knob solution on account of kids break everything that is supposed to be lockable (proven fact!)
Then it's not magnastat
 
fair enough
 
And too over-engineered for my taste
 
4:57 PM
its their mid range current model
 
Nice $3.50 magnastat tips are great for kid's workshops
 
Anonymous
a propane torch works, too
 
Nobody needs to know what's going on. No parents freak out about "Oh god, it's 300 degrees!"
Fckn parents
 
Its a black box that glues wires
no heat here twiddles thumbs
to be fair though, my parents got me my first soldering irons as a kid
 
I have guided 6k+ children in soldering workshops to date.
1 blister
On me
I was allowed my dad's Weller when I was four.
 
Anonymous
4:59 PM
@crasic same
 
I've neer burned myself on an iron, but I have burned myself picking up desoldered parts
 
Anonymous
@Asmyldof is that volunteer work?
 
But look at society now: If you send your kid to school with a long blister on his/her arm you get hit by the inquisistion
@PatoSáinz yup
 
Anonymous
@crasic I've only burned myself with a clothes iron and with my metal forge
 
Well, yes and no
If the organisation pays all other activities $$$$, then I'm billing too, if it's partly volunteer but with sponsoring, sometimes I bill material cost, if it's mainly volunteer I bill naught or near-naught
 
Anonymous
5:01 PM
@Asmyldof and what kind of organisation is that organisation?
 
4 to 6 I definitely went to school with blisters!
 
Sounds lovely
 
Depends on what organisation wants me to do a workshop, dunnit?
 
Anonymous
@Asmyldof oh, i see
 
If it's a First LEGO League regional that needs some schedule filling, or a HOT-Final or regional I'm libel to "forget to send the bill"
If it's a BeNeLux final of same, they're getting the full one, because the organising group pays themselves full time to let volunteers do the work and then show up at the end and go "Gee, look at what we made here, we are so fckn awesome, we should get a raise!"
But when someone (like me) bails out a year they get call after call from the person following them up, because they have no clue what needs to be arranged
And if a Dutch organisation for the promotion of technology needs me, they get the full bill too, because they pay every single person involved, so why not me?
And if a local community group has an activity day and they know about me and invite me I never even talk about cost, up to a certain amount of participants
etc etc
 
Anonymous
5:06 PM
alright
 
Very dependant on who, what and where.
 
Anonymous
well, sushi time
 
Anonymous
bbl
 
Yes plx
 
Anonymous
you should go bake something
 
5:08 PM
After dinner is done
About which, bbl too :-)
 
 
2 hours later…
6:54 PM
All of this work to realize that we are not waiting for I2C as long as we thought we are
 
Hm, it takes a considerable force to move the mouse... time to clean up the lab desk to make some room for it...
 
7:50 PM
LOL
at both
@crasic Told you I never had problems with the Atmel SHA :-P
 
The benefit of this exercise is that I can now read I2C packets in my head off a data trace
 
Do you go bip for 1 and boop for 0?
I do
When I see pulses on a line I go Bip Bip Boop Boop Bip Boop Boop Bip, ah that's 0xC9
 
8:06 PM
lol I'll try that
Any idea how many nacks the kernel listens for before returning NACK for read?
 
Eh
 
because I'm 90% sure that it does 4 read attempts on the kernel side, before giving it back to us
 
Depends on the kernel module
 
our read loop polling does bursts of 4 read attempts on the scope
which makes me thing the kernel is filling in the 4
 
That is possible
 
8:08 PM
Ok, so not going crazy
have to find what exact i2c module we are using
 
IF you are using a pre-created module that hosts an I2C device it may have a retry scheme
 
its some real time compatible i2c (generic) kernel driver
 
And usually that's set to 3, 10 or 25 depending on the environment and exact bus
 
which I haven't looked at yet, and then higher level code from atmel does read attempts until the max execution time of a sha command
 
But people sometimes make mistakes and don't count your forced try, so 4, 11 and 26 would also make sense :-)
 
8:09 PM
which is anywhere from 4 to 40 ms
o.O
ok gotcha, thanks for the info
 
No worries
 
Atmels code assumes a certain execution time for the read command, I think our driver is polling longer = not waiting long enough = timin out
 
I think RPi default module did 3 or 5
But...
 
so screw this and I will use real system time to check the read polling
or rather polling shorter
 
Isn't there a call back?
 
8:12 PM
Its all done in a busy loop
 
I know Atmel doesn't always use the.... best route from a RT perspective in their examples
 
its more that they assume and define a time for the return of a single read
and then use this to keep track of the number of ms since polling started
 
Hm
I like writing call-backs for things like that
 
yes, same, ideally I would rewrite this
but step 1 was figuring out why it isn't working
using the existing code
 
I find it hilarious that ASF (Atmel Software Framework) is modded for ATSAM procs, but uses 0 DMA controllers
 
8:14 PM
ironically, the definition is so perfectly fit, that we are on the razor edge of timing out
sometimes 16 polls, sometimes 17, we time out at 16
 
hm
 
actualyl waiting until the full "max execution time" in the sha datasheet makes everything work well.
and you can see on the scope where there is some preemption going on
 
which number do you use?
 
the clock halts high/low for a few ms
well, 40ms for a write command according to the datasheet
16 polls ~ 11ms
 
I mean, device number
 
8:17 PM
204a
 
huh, they didn't make any new numnbers?
When it just came out I think they had 204 without the A that I sampled?
What'd you mean with pre-empting? Your controller leading the request, data leading the clock or ?
No! wait! The first SHA I ever held started with a 1?
Ugh. My brain stores only stuff it uses...
 
Yeah SHA 204 is an upgrade for 1 something something
204a is a revision bump that is supposed to pin compatible
its a die shrink effectively
 
Hm
But, can't you keep polling it, or am I misremembering that?
Or did they change that?
 
You can
we were timing out earlier than we should
their logic is "wait minimum time, check for read, then poll until max time, resync, try again, if failed return err"
 
I seem to remember it even keeps resending the same stuff until you send another command, but that may have been limited to none sensitive commands
 
8:32 PM
we never give it a chance to respond and consistently underestimating elapsed time for timeout checks
 
:-D
 
also stupid things like running the program through ssh would work sometimes, but from a remote script it won't
why: because of preemption due to print statements delaying the logic
intermittent bug finding at its best
 
Right, well, that's
yu[
yup*
One of the more funny things in programming that spans just about all the alnguages
"My timing is f'ed up"
> "what did you do?"
"I added a debugging printf"
> "There's your problem"
(or printf --> whatever it is in the language)
 
IO queue slaps you
 
Awesome thing in Altium, for example
If you do a change report pre-check on my PC: 0.1 second
Same report, execute: 4s
Difference? On screen log that cannot be disabled
:-D
 
8:36 PM
hehe
 
Someone should make a mute tool
that just queue --> /dev/null for every stdout call
 
lol
yeah I was gonna say you can find the fd for stdout for that child process in /proc
and redirect it
 
Well, yes, but not in Windows
Which is a problem if you are both Altiumming and Atmelling and Windows devving for customers/employers
not just both
but allathemming
 
I've been 100% linux for personal systems for a few years. Starting a job with windows machines means I now need to know the intricate details of cygwin just to maintain my work flow
 
 
2 hours later…
10:23 PM
Help, please!!
0
Q: JFET - Pinch off voltage value

JuanI have a JFET channel N and I have N_A, N_D, V_TH and the width of the channel without any polarization applied. I need to know the cut-off tension. The cut-off tension is the pinch-off tension, isn't it? I know that I have to solve when I_DS=0. When I_DS=0, V_GS=V_P. And, as the channel is N, ...

 
Anonymous
10:35 PM
@crasic I'd feel the same... but make it worthwhile and learn PowerShell: it's not bad
 
10:55 PM
@crasic The joys of having an all SSD laptop :-) I only have a virtual machine for what I don't feel like hard-installing. Switching over to Debian or Ubuntu (yes, both! shutup!) from Windows or back takes less time than starting something like virtual box now
Only 4 months ago I did nearly everything in virtual box. Either in Windows because I also needed Adobe, or in Linux because I wanted to compile loads of nix-work on all threads
Also... I just installed MS Office between my last message and this one, because DOCX
I should become Samsung SSD spokesperson
 
11:55 PM
@Asmyldof I mean I run a Xen hypervisor occasionally, but generally just haven't had the need for windows. Since we use windows only utilities and I love me a POSIX shell for editing I make do without vm's
 

« first day (1705 days earlier)      last day (3248 days later) »