Back
[01:31:40] <_abc_> I am puzzled over the difference between v-usb based devices, like the basic v-usb cdc, which has INT1 and a PD pin connected to USB, and things like USBasp which have two PD pins and an INT pin. I can't find a reference on this, does anyone know why this difference exists? The connected 2 pins are hard to route on 1 sided board for smd parts.
[02:27:32] <_abc_> Meep.
[02:28:18] <Casper> meep meep
[02:28:43] <_abc_> Hmm nice part, Si8441, cheapish too. About time the $$$ monopoly analog held on digital fast insulators was broken. *Thanks* Silabs!
[02:29:01] <_abc_> An equivalent ad part is about 10 times as much as the Si one
[03:25:50] <_abc_> Why is freenode so quiet lately? Summer vacation time?
[03:29:45] <_abc_> Oh I think I get it, usbasp uses v-usb modified to use usb data on PB while interrupt is still on a PD port.
[03:29:55] <_abc_> I wonder why. The pcb makes NO sense like that.
[04:13:54] <veek> oh man, 19+chapters on usb
[04:25:34] <_abc_> Hm?
[04:28:38] <veek> _abc_, USB complete 3rd ed
[04:29:21] <veek> would be nice if he just had notes on usb
[04:30:38] <_abc_> Yes well usb is a designed-by-committee thing.
[04:30:44] <_abc_> Was, is, and always will be.
[04:31:34] <_abc_> And usb3.0 is completely unnecessary, simply using existing gigabit ethernet + POE would have been enough, perhaps with a new connector, miniaturized RJ45. All the crap in USB3.0 is just to make someone money. Nobody needs it technically.
[04:32:03] <_abc_> The USB stack is almost as heavy as a full tcp/ip stack now.
[04:33:58] <_abc_> You can basically pass 10GBps and POE to the tune of 60 Watts through suitable cat6+ cable now, no need for another bullcrap "standard". It's just about money.
[04:47:10] <flutterbat> _abc_: i need a symmetrical connector. thats it idgaf about the speed.
[04:48:33] <_abc_> There are ways to make RJ like connectors symmetrical. That is not a good reason.
[04:48:33] <flutterbat> well maybe usb 3.1 will come someday...
[04:49:29] <flutterbat> _abc_: problem with ethernet is taht its completely different than usb. the compatibility would be a nightmare for a few years
[04:52:56] <_abc_> You mean, like, cough, compatibility between serial ports and usb when it was new?
[04:52:59] <_abc_> Ow, cmon
[04:53:11] <_abc_> tcp/ip stacks are more common than usb now, and less idiotic.
[04:53:25] <_abc_> not to mention tcp/ip is wan routable natively.
[04:53:37] <_abc_> USB is about money, money, money, and more money.
[04:54:34] <_abc_> Just for fun compare something like ESP8266 with any usb capable mcu, see what happens.
[04:58:34] <veek> Oo jan axelson is a gurl
[05:02:32] <flutterbat> _abc_: theres VERY expensive equipment out there that runs on usb
[05:02:46] <_abc_> Unfortunately.
[05:02:53] <flutterbat> and i dont mean the newest razer gamer headset
[05:04:19] <_abc_> I know, but usb sucks from inception. Ethernet has had IEEE compliant RFI and insulation specs since inception. USB still does not have it. Firewire also not. No WAY one will use usb on very expensive equipment without galvanic digital insulation. Care to price a 300MBps galvanic digital insulation solution? Just for fun.
[05:05:11] <_abc_> USB is was and always will be a way to sell bullcrap hardware at extortionate prices to punters. And you can rely on the "standard" being bumped up in hw requirements every time the makers stop making money due to market saturation or competition.
[05:05:34] <_abc_> Did you notice how RJ45 based ethernet has REMAINED hw cable and connector compatible since about 1975 or so?
[05:05:59] <_abc_> In despite of speeds going from 10 to 10000 MBps meanwhile and new features like POE being added?
[05:06:07] <_abc_> Happy coincidence? I think not.
[05:06:29] <_abc_> 803.3x are industrial standards made by IEEE & industry and they WORK
[05:06:53] <_abc_> USB is made by manufacturers who are squirming weasels doing everyting 24/365 to squeeze more money out of someone.
[05:12:35] <_abc_> I forgot to mention that the USB3.0+ main pushers are the same round corner patent pushers who pushed IEEE1394 for roughly the same reasons in the 1990s-2000s
[05:12:52] <flutterbat> usb 3.0 makes imho no sense
[05:13:18] <_abc_> That ALSO omitted galvanic insulation for "cost reasons" in despite of being used on >$2000 devices like video cameras, which I worked on at the time, and made a decent amount of money replacing exploded IEEE1394 interface chips.
[05:13:38] <_abc_> And that was ALSO obsoleted by... USB. No kidding. Can you see a pattern here?
[05:13:40] <flutterbat> well i can see that some people would benefit from it.
[05:14:49] <_abc_> Round corners (aka cut!), same company, "must have" shiny standard omitting industry wisdom (Ethernet magnetics are a must, as seen on its predecessor, aerospace ARINC networking equipment), funny connectors, molded cables with special requirements, unobtainium standards documents.
[05:15:09] <_abc_> Nothing to see here, it's just someone squirming to make a decent profit margin by all means. Move on.
[05:15:15] <veek> _abc_, yeah but this stuff happens with mobos as well and dentists/docs.. it's not like we get a refund every time they do a bad root canal
[05:16:03] <_abc_> And I meant
https://en.wikipedia.org/wiki/MIL-STD-1553 not ARINC. Sorry.
[05:16:09] <flutterbat> the most ironic part is that the interfaces promoted by apple tend to fail, or are total crap to begin with i.e. thunderbolt
[05:16:17] <_abc_> (that's 1973 technology for you)
[05:17:03] <_abc_> Ethernet can be seen as a watered down 1553 bus in many ways. Aka "civilian" version. Which is unsurprizing since DARPA probably had experience with 1553 when Ethernet came about in the Internet context.
[05:17:20] <_abc_> flutterbat: They are just squirming to make money like everyone else.
[05:17:33] <flutterbat> obviously
[05:18:05] <_abc_> Heh some new designs based on 1553 are moving to 1394? I hope they use magnetics!
[05:18:36] <flutterbat> we would only need a proper on the go ethernet rj45 replacement/ addition. a circular one would be nice
[05:18:43] <_abc_> I never understood why top of the line Panasonic, Canon and JVC prosumer camcorders used by professionals and costing upwards of $2000 easily used 1394 with no magnetics in the 1990s. It was a disaster.
[05:19:00] <_abc_> flutterbat: flat works well.
[05:19:09] <flutterbat> well it doenst for usb
[05:19:14] <_abc_> It does.
[05:19:16] <flutterbat> nope
[05:19:27] <_abc_> Look up reversible usb connector on ebay/alibababa
[05:19:28] <flutterbat> everyone i know is complaining about the orientation
[05:19:38] <_abc_> It has a magnetic thingy and you can insert it both ways
[05:19:41] <flutterbat> y that would be fine as well
[05:19:54] <_abc_> It exists. It's about $5 to $10 for a complete cable.
[05:19:56] <flutterbat> but round ala 3.5mm headphone jack looks better
[05:20:12] <_abc_> 3.5mm jacks are maxed out at 5 contacts.
[05:20:14] <flutterbat> and is even smaller
[05:20:25] <_abc_> You can actually wire usb on a 4 contact 3.5mm connector.
[05:20:31] <flutterbat> thats why i didnt suggest a 3.5 mm headphone jack :p
[05:20:39] <tavish_> hi, is anyone using mark3os over here?
[05:21:06] <_abc_> And I know people who did this for embedded things which need to be waterproof. The bung that covers the 3.5mm hole is much more secure against tearing out or water ingress than a large USB rubber bung.
[05:21:47] <flutterbat> jup. the size dimension would be similar to micro usb. but the roughedness would be closer to fullsize usb
[05:21:52] <_abc_> What does work really well is 3.5mm TOSLINK. You get fiber and you can "misuse" the space around the plug to add DC contacts.
[05:21:56] <_abc_> That would be NICE.
[05:22:19] <flutterbat> fiber isnt flexible enough
[05:22:24] <_abc_> Oh it is
[05:22:37] <flutterbat> not for "stuff it in your bag"
[05:22:47] <flutterbat> and throw books ontop
[05:22:50] <_abc_> It sucks for headphones but combined 3.5mm sockets with both metal contacts and a FO sender/receiver in the deep end exist.
[05:23:05] <_abc_> So it would work for headphones, data, and dc/charging too
[05:23:30] <_abc_> flutterbat: plastic fiber has been around for a long long time and withstands a lot of abuse like that.
[05:24:23] <_abc_> Basically I am very very wary of any new "standards" launched by Apple.
[05:24:28] <_abc_> Hw and sw alike.
[05:25:42] <_abc_> Also usb and firewire are ridiculous in that they limit network length to 4.5m
[05:25:56] <_abc_> Which is probably okay for a wiring cabinet or 19" rack at most.
[05:26:27] <_abc_> Ethernet will go 100m for 100MBps with the right cable, 10GB fiber can be stretched to 2km between repeaters and those are under $100 each
[05:26:33] <_abc_> (well under)
[05:26:35] <twnqx> pcie isn't much better :)
[05:26:41] <_abc_> I never said it was.
[05:26:52] <twnqx> and ethernet will go far, far further
[05:26:57] <twnqx> 10GBit on 70km
[05:27:05] <twnqx> 100m copper is 350
[05:27:09] <twnqx> meters
[05:27:15] <twnqx> gbit copper the same
[05:27:18] <_abc_> I wonder how they are going to move from 1553 to 1994 in airplanes. I think they are going to make only very small ones (max 4.5m long) so the 1394 wire length limit will not impact them >;)
[05:27:45] <_abc_> twnqx: yes but under realistic conditions those limits are not achievable with a decent margin.
[05:27:54] <twnqx> 50km 10gbit is pretty normal
[05:28:32] <twnqx> 200km if you use excessive stuff like ciena WDM systems :P
[05:28:35] <_abc_> Maybe in protected cable runs. In normal industrial environments and with locally made cables (the norm in installation now), you get to one cable drum length per span, which is usually 300ft.
[05:28:58] <_abc_> And people do not like splicing at all, with good reason.
[05:29:02] <twnqx> sure
[05:29:16] * twnqx was working with a network carrier for some time
[05:29:28] <twnqx> splicing was day to day business (not for me though)
[05:29:50] <_abc_> Did you tell the customer to ignore the error counts on spliced links ;)
[05:30:16] <twnqx> never hat problems, it just dampens the signal a bit
[05:30:34] <twnqx> so you just use ER optics and then have to insert optical dampeners anyway :P
[05:30:59] <_abc_> Oh you mean fiber, I was talking about copper.
[05:31:03] <twnqx> ah
[05:31:22] <_abc_> Well some smaller localities do get by on a single GPON fiber running through the whole town for triple play.
[05:31:46] <twnqx> we have 100mbit/s DSL links on copper here
[05:32:05] <_abc_> Which is very nice. Also one needs only one UPS at the distribution rack, and anyone else who has a proper UPS for modem+phone+computer at home will be able to call 112/991 if the power goes out.
[05:32:54] <_abc_> twnqx: Our internet in this house is 1GB fiber to a Huawei box then 100MB ether to custmers. 2 customers on I think.
[05:32:56] <twnqx> my current company recently found a wiring error in the UPS
[05:33:11] <twnqx> the border routers stayed up during the power outage
[05:33:18] <twnqx> so did the carrier's routers
[05:33:23] <_abc_> And that is bad?
[05:33:30] <twnqx> sadly, the switch connecting the routers to the inside turned off quickly.
[05:33:34] <_abc_> ah
[05:34:16] <_abc_> Yes, the famous painted red mains multi plugs in the computer room, connected to the always on power, and marked NO VACUUM CLEANER / KETTLE CONNECTION! which get used for just that every time.
[05:34:22] <_abc_> Sounds familiry.
[05:34:24] <_abc_> *ar
[05:34:53] <_abc_> I no longer have to see such things, fortunately.
[06:17:40] <veek> _abc_, what book would you suggest that wasn't math/tronics related but more tech related (like avr/usb)
[06:18:27] <Lambda_Aurigae> veek, what are you trying to learn?
[06:19:33] <veek> Lambda_Aurigae, electronics but nothing specific - i'm into diy stuff
[06:19:56] <veek> Lambda_Aurigae, i'm currently reading avr/usb
[06:20:13] <Lambda_Aurigae> how much electronics do you know now?
[06:20:51] <veek> just wondering what to add.. platt sensors maybe (oh i have a degree :p but 0 practical experience)
[06:20:54] <Lambda_Aurigae> Getting Started In Electronics is the best beginner book I know which I learned with when I was like 10.
[06:21:10] <veek> mostly electrical stuff/python/c i know
[06:21:15] <Lambda_Aurigae> platt sensor?
[06:21:44] <veek> yeah encyclopedia of electronic components vol 3 is out so :p
[06:22:06] <veek> that's pretty good i think (his first 2 vols were awesome)
[06:22:26] <Lambda_Aurigae> engineer mini notebooks were my bibles for many years too.
[06:22:38] <Lambda_Aurigae> I was reading those 30+ years ago.
[06:22:50] <veek> ah yeah forrest mims
[06:23:01] <veek> i haven't read his stuff
[06:23:04] <Lambda_Aurigae> I still refer to them.
[06:23:18] <Lambda_Aurigae> and use GSIE and the mini notebooks when teaching electronics to kids around here.
[06:24:03] <Lambda_Aurigae> for learning AVR,,,read the datasheet.
[06:24:13] <Lambda_Aurigae> for learning USB, read the usb spec.
[06:24:14] <veek> nope i'm good with avr
[06:24:23] <veek> got naimi.. covers assembly/c
[06:24:45] <veek> but there's the PIC and Tis chips which i'm unfamiliar with
[06:24:57] <veek> so a book on pic/ti arch would be nice
[06:25:00] <Lambda_Aurigae> Teach Yourself C by K&R is my favorite C programming learning book.
[06:25:23] <Lambda_Aurigae> read the datasheets...those tell you more about the architecture than any other book.
[06:25:28] <veek> i like king better for c
[06:25:43] <veek> K&R is dated but better problems perhaps
[06:25:54] <Lambda_Aurigae> K&R wrote C!
[06:25:59] <Lambda_Aurigae> or, one of them did anyhow.
[06:26:08] <veek> so.. he's still got to teach it :p
[06:26:10] <Lambda_Aurigae> who better to learn from than the maker of the product.
[06:26:27] <veek> K.N.King! awesome
[06:26:28] <_abc_> Okay the mystery of why usbasp uses PC1,PC2 for USB plus INT0 is solved. Those pins are at pin 14 and 15 of the DIP28 package, which is what the original usbasp was designed for.
[06:26:53] <_abc_> So Fischl modified the v-usb code to work with those pins.
[06:27:03] <_abc_> This is not a good location for the SMD version of usb-asp
[06:27:10] <Lambda_Aurigae> _abc_, you can pretty much use any two pins on a single port plus the int0 pin to do v-usb.
[06:27:22] <_abc_> Lambda_Aurigae: Yes but I have to compile my own fw and maintain it.
[06:27:52] <_abc_> I managed to route a single sided 20x40mm usbasp board, the only 2 jumpers are for that, routing PC1,2
[06:27:57] <_abc_> Without them, no jumpers.
[06:28:31] <_abc_> The jumper ends consume a lot of space ;) Need decent sized pads for single sided.
[06:29:33] <_abc_> There is really one wire jumper anyway, the self programming RST one, but that does not count.
[06:57:34] <Emil_> Hi
[06:57:43] <Emil_> So, atmel provides a soft reset example
[06:58:10] <Emil_> Oh, they have fixed it now
[06:58:32] <Emil_> They used to have MCUCSR=0 in here
http://www.atmel.com/webdoc/AVRLibcReferenceManual/FAQ_1faq_softreset.html
[06:58:36] <Emil_> If I remember correctly
[06:58:58] <Emil_> but such a register is only once referenced in m32u4 and even then it has no register description
[06:59:05] <Emil_> But alas, they simply mean MCUSR
[07:02:29] <Emil_> Oh
[07:02:38] <Emil_> I didn't know you could program a device without active power supplied :D
[07:02:51] <Emil_> I think I powered it through the protection diodes
[07:49:41] <LeoNerd> Emil_: That's not really supposed to work; it's just a side-effect of those diodes being able to charge a decoupling cap long enough during SCK and MOSI pulses
[07:51:49] <Emil_> LeoNerd: yeah
[07:51:52] <Emil_> but cool nonetheless
[08:42:01] <l9> do i need too get an bench power supply?
[08:46:58] <Snert> you need a good source of juice. That could be a bench supply or some sort of wall wart.
[08:48:14] <l9> i have today a 220 volt too 12 volt and converts that into 5 volt
[08:48:29] <l9> i have had problems running of that
[08:48:43] <l9> had no problems sorry
[08:50:24] <l9> reason i was asking now is cause i want too make something of battery cells, and knowing how much power it uses is critical
[08:52:19] <Snert> measure it.
[08:53:01] <l9> dosent the bench power supply measure it for me?
[08:53:26] <Snert> many do. Most do. But so does a DVM meter.
[08:55:02] <Snert> but is it strictly necessary to have a metered bench supply? no.
[08:55:22] <Snert> just measure the current drawn by the circuit.
[08:57:05] <l9> hell of alot cheaper than the bench power supply that is for sure.
[08:57:47] <Snert> oh for sure. I use my chinesium $8.99 DVM for everything.
[08:58:09] * l9 notes bench power supply on his dear santa list
[08:58:29] <Snert> if it's a critical value, I also measure it with my high quality meter. The 2 extremely rarely show any difference.
[08:59:22] <bss36504> I bought a $100 power supply from amazon, and the first time I exceeded 50% of its load rating for any real duration the decoupling caps exploded. 30V max power supply and they used 40V caps.
[09:07:22] <l9> wtf computer crashed...
[09:07:44] <l9> oh well good bye unity hello fluxbox...
[09:30:09] <carabia> this one guy called adolf had a good source of jews from poland and the likes
[09:30:25] <carabia> oh i totally fucked up that pun. I was supposed to write juice.
[09:31:29] <veek> carabia, good way to get banned :p
[09:32:13] <carabia> it's worth it in any case
[09:32:52] <veek> only if your not into avrs
[09:33:31] <carabia> it's not like this place is a mecca of geniuses, actually
[09:33:36] <veek> carabia, so what do you use mostly?
[09:34:09] <carabia> mostly, i use my brain, for the most part, mostly, --
[09:34:11] <veek> true dat, that's the nobel winners mostly
[09:34:55] <carabia> and to answer your question me nor any other part of me is not in an avr
[09:36:32] <carabia> now was it something you actually wanted to tell me, or just play captain obvious with your really not-witty-at-all -remark?
[09:38:29] <l9> bad day at the offic huh?
[09:41:59] <l9> riddle me this why the hell does my xterm -name "fluxie" go too hide him self.in(the_corner)
[09:44:57] <veek> l9 what desktop are you on.. i use olvwm.. does xterm run the usual way, minimize correctly (anyway #linux dunno how the guys here are about topic)
[11:16:44] <Jartza> evening
[11:16:55] <LeoNerd> Nonsense! It's barely afternoon
[11:30:34] <Emil_> LeoNerd: I like the way you think
[11:30:48] <Emil_> And besides, my sleeping rythm is all sorts of fucked up
[11:32:16] <Thrashbarg> 01:35 here
[11:32:28] <_abc_> http://www.dailymail.co.uk/news/article-3644931/Junction-dubbed-red-light-district-42-separate-traffic-lights-possibly-fewer-say-engineers-designed-it.html what the f*
[11:32:39] <_abc_> And it's not even in Romania where I expect such things.
[11:32:45] * _abc_ is in Romania
[11:33:33] <_abc_> I make an educated guess: the police is making a killing there since unless the lights are all synchronized the chances that all will be other than red at any moment are pretty much zero. Therefore, at any time, there will be someone who can be fined.
[11:33:34] <Thrashbarg> sounds like something they'd do here too
[11:33:46] <_abc_> Here where?
[11:33:51] <Thrashbarg> Adelaide, Australia
[11:34:18] <_abc_> I mean here where I am they do such things for nepotism reasons. Like the company installing the lights belongs to the local deputy mayor's cousin's driver's wife.
[11:34:42] <Thrashbarg> the police here are effectively tax collectors. They have quotas to achieve for fines, and the group who get the least fines are shamed with a 'wooden spoon' award.
[11:34:44] <veek> wow that's a lot
[11:42:20] <_abc_> Thrashbarg: ditto here except they don't turn in all the earnings.
[11:42:32] <Thrashbarg> yeah
[11:42:41] <Thrashbarg> I figure that's what the wooden spoon prevents
[11:42:47] <Thrashbarg> also it's all tracked
[11:43:01] <_abc_> Not advanced enough yet here (lalala)
[11:43:06] <Thrashbarg> yea
[11:43:12] <_abc_> Also this Chinese hardware keeps breaking down ;)
[11:43:17] <Thrashbarg> heh
[11:43:36] <_abc_> bbl
[11:53:23] <carabia> _abc_: I suspect traffic lights are few and far between down in romania, the gypsies' carriages aren't known to cause too many fatal accidents?
[11:53:46] <carabia> Damn, I was thinking out loud again.
[11:53:59] <carabia> thinking out... in writing
[11:54:52] <veek> _abc_, troll just ignore (or tailback)
[11:57:19] <carabia> no, on the contrary. In the early 20th century romania had quite the brains. Coanda, for example. Good stuff.
[12:00:21] <carabia> then again, it's been kind of degenerating ever after :)
[12:17:49] <_ami_> LCD of mobiles can be hacked via AVRs?
[12:22:31] <_ami_> probably i start with char LCDs first.
[12:22:59] <LeoNerd> I've been playing with the SSD1306 modules lately. Those little 128x64 OLED screens
[12:23:04] <_ami_> i got few old phones (samsung, nokia, motorola), may be one day i hack them
[12:23:09] <LeoNerd> So far not from an AVR, just a bus pirate. but they seem reasonable
[12:23:54] <_ami_> LeoNerd: 16x2 is a good starting point?
[12:24:08] <LeoNerd> The HD.. er.. Iforget the number
[12:24:32] <_ami_> LeoNerd:
http://www.microcontroller-project.com/16x2-lcd-working.html
[12:24:33] <LeoNerd> They can be, if you just want plaintext. They have the advantage that things like "hello world" are easier, because the module has a font, so you don't have to get into setting pixels and character generation and allsorts
[12:25:40] <_ami_> LeoNerd: i would like to start with basics one first.
[12:26:00] <LeoNerd> Then you have to weigh up the fun of there just being lots more wires to connect up of course :)
[12:26:04] <_ami_> although i think setting pixels is more intuitive to me
[12:26:08] <LeoNerd> Unless you get a fancier one with an SPI or I²C backpack
[12:28:00] <_ami_> btw LeoNerd, thanks for spanking me yesterday. i learned SPI though :)
[12:28:30] <LeoNerd> :)
[12:28:54] <_ami_> problem was in connection though. OE was not grounded :/
[12:29:16] <_ami_> so both bitbang and spi works fine! :D
[12:29:30] <LeoNerd> Ahyes, that can be annoying
[12:29:49] <LeoNerd> I wasted 2 hours the other day on an RS485 transceiver..I thought I had TE grounded, but the pulldown resistor turned out not to be quite soldered on one side
[12:29:56] <bss36504> Submitted the reflow controller PCB yesterday, parts are on the way. I have a toaster oven, some insulating pads and tape. Here's hoping it works! I did do something I rarely do which is to not worry about space and instead use plenty of solder jumpers and leave room for test points. If all goes well this board should work well for other projects t
[12:29:56] <bss36504> oo.
[12:30:01] <LeoNerd> So occasionally it would float and enable the TX module for a moment... knocked out the serial line :(
[12:30:09] <LeoNerd> It only did this when Iwaved my hand near the board
[12:30:48] <bss36504> https://goo.gl/gP2RoU
[12:31:57] <_ami_> LeoNerd: these mistakes happen a lot with me since its a hobby for me. After office hours, its too tiring to work again. although i enjoy every bit of it.
[12:31:59] <bss36504> Oh and I promise that isn't spam. Just realized google might warn you.
[12:32:32] <LeoNerd> _ami_: Oh, get used to them. This sort of mistake happens just as often to 20year+ industry pros as well. :) They just tend to be faster to recognise what the problem is and fix it
[12:32:46] <_ami_> :)
[12:33:11] <LeoNerd> You start to build little checklists: chip not working? have I checked all the right voltages? are all the control lines as they should be? does it have appropriate decoupling cap - sufficiently close by?
[12:33:32] <LeoNerd> If it's any sort of serial thing:is the clock speed slow enough for S&H times, ... and so on
[12:33:36] <_ami_> yeah
[12:34:30] <_ami_> also some times, giving a break is also important. If problem is not solved after debugging few hours, leave it there. Try next day or take a walk or do something else.
[12:34:31] <_ami_> :D
[12:34:36] <LeoNerd> Oh definitely
[12:34:50] <LeoNerd> I like having a list of projects I'm working on. If I get stuck on one I just switch to something else
[12:34:57] <LeoNerd> So often the solution comes at unexpected times
[12:35:04] <_ami_> indeed. so true!
[12:35:16] <LeoNerd> The number of times I've solved a problem in the shower...
[12:35:32] <LeoNerd> I think all engineering offices should have an onsite shower, for engineers to go and think about problems. Or a bathtub
[12:35:38] <_ami_> oh boy, you think abt it every time. :D
[12:35:49] <_ami_> LeoNerd: good idea! haha
[12:35:49] <bss36504> Oh that's just what we need, another reason to stay at work longer
[12:36:10] <_ami_> suggest this to HR Team. they might like this idea.
[12:36:11] <bss36504> I get your idea though :) I personally come up with my best thoughts in the car
[12:36:11] <_ami_> :D
[12:36:43] <_ami_> bss36504: Focus on driving mate! its bad habit, IMHO
[12:36:48] <_ami_> :P
[12:37:03] <LeoNerd> Ahyes, I'm public transport, so I don't have that problem to worry about
[12:37:17] <LeoNerd> I do sometimes find I get a bit distracted by my thoughts when driving
[12:37:50] <bss36504> _ami_ I am, I swear! Sometimes I just have moments of clarity since there is, overall, less stimulus that being at home/in the office.
[12:38:14] <bss36504> But shower is a close second for moments of clarity.
[12:39:21] <_ami_> bss36504: true, sitting at any place does not generate solutions
[12:41:33] <_ami_> time to crash though.
[12:41:36] <_ami_> nn guys!
[12:49:06] <rue_house> the amount of sleep gained is roughly proportional to the amount of time spent in bed
[12:49:23] <rue_house> so I call acception
[13:50:17] <GeneralStupid> I recently switched from Windows Atmel Studio to Linux. Is there a eclipse plugin or something? Please, i love vim but i need some arguments to my colleagues
[13:52:50] <carabia> LeoNerd: You can drive a hitachi hd44780 controller (various sizes, 16x2, 20x2, 20x4 etc.), with a minimum of 6/7 pins. 4 data bits, enable and register select. If you implement dummy delays in your code instead of polling the screen for busy states, you can tie r/w to ground to only perform write operations. Ergo, 6 pins, not really that much more than driving a screen with spi.
[13:52:55] <LeoNerd> 44780
[13:52:57] <LeoNerd> that's the one
[13:53:16] <LeoNerd> Yah.. but still more annoying than the 2 of I²C
[13:53:25] <LeoNerd> And it's Yet Another Custom Protocol, rather than /literally/ being SPI or I²C
[13:54:04] <carabia> And it's not really a big deal. You can literally write a driver for it from scratch in 5 minutes.
[13:54:15] <LeoNerd> Ohsure. /I/ can
[13:54:29] <carabia> Who cares about anyone else anyway.
[13:54:47] <LeoNerd> You probably dont' want to generate keys on such a machine, but if you could provide a pregenerated keypair, it should be OK
[13:54:49] <carabia> How fast can you drive screens with i2c?
[13:54:54] <LeoNerd> Having a good source of entropy would be fun
[13:55:01] <LeoNerd> ... wait wrong channel >.>
[13:55:45] <LeoNerd> Well, at absolute best you're probably looking at 400kbit/sec.
[13:55:59] <LeoNerd> So 50kbyte/sec presuming 8 data bits, 1 ack, and one bit of spare timing
[13:56:25] <LeoNerd> A 128x64 monochrome screen has 1024 bytes of data in it.
[13:56:40] <LeoNerd> So... a little under 50fps if we totally saturate the bus to its maximum possible rate
[13:57:07] <LeoNerd> Obviously faster if we're not drawing the -whole- screen every time, but instead just doing delta updates where required
[13:58:00] <carabia> Yeah, theoretically 50 fps. But you won't be landing anywhere near it due to having to wait for the lcd controller itself
[13:59:54] <carabia> But as said, the 44780 interface isn't really what you would call "custom", as other controllers adopt the exact interface, too.
[14:00:11] <LeoNerd> Well sure
[14:00:19] <LeoNerd> all 44780s are the same. As are all HT11s all the same
[14:00:26] <LeoNerd> And all Dallas 1Wire the same
[14:00:27] <LeoNerd> and ....
[14:01:20] <LeoNerd> I'll still take an SPI unit over anything else any day of the week... or I²C if I can't get the SPI
[14:01:50] <carabia> LeoNerd: can't really compare onewire. I believe it's only used by dallas semi, right?
[14:01:54] <LeoNerd> My ideal MCU-based project has the AVR sat in the middle of the SPI bus with every other GPIO line being an SS or IRQ line. :)
[14:02:18] <LeoNerd> Most times I shove a 74'595 in to drive LEDs. That sits on SPI
[14:02:34] <LeoNerd> ThoughI've become something of a fan of the PCF8574 I²C expander lately. It's quite nice
[14:04:54] <carabia> I would just probably throw the 8 pin devices out the window, and get a micro with more i/o, really
[14:05:03] <LeoNerd> Hrm?
[14:05:04] <carabia> as i recall you playing a lot with the attinys
[14:05:08] <LeoNerd> Even megas
[14:05:11] <LeoNerd> Even large megas
[14:06:06] <carabia> yes? if you use serial comms as much as you can, do elaborate...
[18:17:12] <Lambda_Aurigae> http://hackaday.com/2016/06/16/evaluating-the-unusual-and-innovative-perf-protoboard/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+hackaday%2FLgoM+%28Hack+a+Day%29
[18:46:35] <carabia> who's this guy who performed this "test"? And how in the hell did he manage to make it look so shit, with a component count less than 20????
[18:48:47] <Lambda_Aurigae> hehe.
[18:48:53] <Lambda_Aurigae> I like the perfboard though.
[18:49:07] <Lambda_Aurigae> the horiz and vert bus bars between the holes are nifty.
[18:50:36] <carabia> well the bus bar only goes along one axis, obviously. not both
[18:53:31] <carabia> oh wait what they do run both ways, my bad
[18:54:02] <Lambda_Aurigae> horiz on one side and vert on the other.
[18:54:04] <carabia> soo, if i got that right it runs along one axis top side and the other on the bottom?
[18:54:09] <Lambda_Aurigae> yes.
[18:54:26] <carabia> yeah alright. I was already thinking that would've been a better idea
[18:54:32] <Lambda_Aurigae> 3 solder joints to connect any two holes on the board.
[18:54:51] <Lambda_Aurigae> hmm..4
[18:54:59] <Lambda_Aurigae> 2 on one side and 2 on the other side.
[18:55:15] <Lambda_Aurigae> can eliminate a lot of jumper wires with it.
[18:55:22] <Lambda_Aurigae> would take some planning
[18:55:31] <Lambda_Aurigae> but shouldn't be too difficult to do some nice layouts.
[18:55:33] <carabia> yeah, 4
[18:55:49] <carabia> cause you need to run it through to the other side and basically be wired to diff. buses
[18:57:45] <carabia> well, it'll only be mf'd in small quantities, don't really understand where the market is
[18:57:58] <Lambda_Aurigae> might make some myself for the fun of it.
[18:58:13] <Lambda_Aurigae> only real problem would be the plated through holes.
[18:58:13] <Lambda_Aurigae> but
[18:58:24] <Lambda_Aurigae> I might have some made up by dirty cheap pcbs
[18:58:27] <carabia> perfboard gets the job done, and well, that thing costs $19 for 2x 68x100 (24x36 holes) + shipping
[18:59:07] <carabia> over twice the price i can get single sided perfboard
[18:59:21] <carabia> and for "neat", pcbs are so dirt cheap nowadays.
[18:59:31] <Lambda_Aurigae> yeah.
[19:00:13] <Lambda_Aurigae> can do 35x35 hole 10 boards for 25 dollars.
[19:00:42] <Lambda_Aurigae> 10x10 cm boards...protopack of 10 double sided boards is 25 dollars.
[19:00:51] <Lambda_Aurigae> takes 1 to 8 weeks to get it though.
[19:01:29] <Lambda_Aurigae> my last two orders only took a week and a half though.