Back
[00:32:28] <TechIsCool> So I have a weird problem. I have some example code that works perfect in its default state but as soon as I comment out a printf command it breaks everything
[00:33:17] <TechIsCool> Code is listed here
https://github.com/gcopeland/RF24/blob/master/examples/pingpair/pingpair.pde where I comment out it line 211
[01:28:08] <jlstrauss> Hi folks, anyone familiar with implementing qtouch? I have a basic prototype working but I'm getting latency issues. A touch registers very quickly but it takes 50-70 ms after I remove my finger before I can detect another touch
[04:10:26] <langoliers> h
[04:12:19] <OndraSter__> u
[06:48:09] <twnqx> langoliers: i have a question related to power supplies that's bugging me for a long time, maybe you know the answer.
[06:48:45] <twnqx> you probably now over/undervoltage clamp diodes... what does happen if e.g. 5V come from on of them to a 3.3V rail
[09:15:19] <rue_house> easy, the zener shunts enough current to fuse, shorting and taking down everything till its replaced
[09:31:01] <rue_house> jlaumann, you can re-tap that fast? (60hz+?)
[09:34:20] <Tom_itx> hello rue_house
[09:34:46] <rue_house> hello
[09:36:44] <Tom_itx> rue_house, any new recent projects?
[09:36:51] <Tom_itx> haven't heard much outta you lately
[09:37:26] <RikusW> Aany thoughs on the Intel G2020 cpu ?
[09:38:01] <RikusW> 330 Euro laptop
http://www.proline.co.za/our-products/Laptop/smart/W258CZ-L/technical-specs/
[09:38:07] <Tom_itx> i'm too far outta the intel loop
[09:38:19] <RikusW> me too
[09:38:33] <Tom_itx> mine are amd anyway
[09:38:38] <RikusW> seems decent
[09:39:04] <RikusW> how low power are the Atom ones ?
[09:39:14] <RikusW> since you had/have one
[09:39:18] <Tom_itx> i haven't measured one but i really like mine
[09:39:35] <Tom_itx> it's not in a lappy
[09:39:56] <RikusW> they seems cheap too
[09:40:03] <RikusW> $100 for mb+cpu
[09:41:12] <Tom_itx> i paid $70 ish for my MB & CPU
[09:42:15] <RikusW> does it compare to P4 in performance ?
[09:43:05] <Tom_itx> i don't have a p4
[09:43:42] <Tom_itx> it's better than my AMD Athlon 64 but that's not saying much
[09:44:02] <Tom_itx> that's how old my junk is
[09:46:01] <RikusW> heh
[09:46:05] <Tom_itx> http://www.newegg.com/Product/Product.aspx?Item=N82E16813153242
[09:46:13] <RikusW> I got a XP 3200+ for now
[09:46:14] <Tom_itx> not mine but similar
[09:47:01] <Tom_itx> a recomended one for linuxcnc was why i had it marked
[09:47:22] <RikusW> http://www.comx-computers.co.za/FXD270S-Foxconn-D270S-Motherboard-with-Integrated-Atom-DC-D2700-2-13Ghz-CPU-Information-Price-Buy-Cheap-p-65244.php
[09:48:49] <Tom_itx> i should start looking for a new fast one for my desktop
[09:49:01] <Tom_itx> but i wouldn't know where to begin looking nowdays
[09:49:22] <Tom_itx> morning abcminiuser
[09:49:34] <abcminiuser> Heyo
[09:49:42] <Tom_itx> any recomendations?
[09:50:14] <abcminiuser> Hva?
[09:50:41] <Tom_itx> oh somebody was having touchy issues earlier
[09:51:53] <RikusW> Tom_itx: I'm in about the same situation, totally lost track...
[09:52:22] <Tom_itx> yeah well we know Dean stays on the bleeding edge so that's why i was asking him
[09:53:16] <RikusW> hi abcminiuser
[09:53:31] <abcminiuser> Heyo :P
[09:53:36] <RikusW> any thoughts on intel G2020 ?
[09:54:18] <RikusW> much faster than my current Athlon XP 3200+
[09:55:08] <RikusW> or would you recommend replacing it with AT32UC ? :-P
[10:05:27] <twnqx> rue_house: you don't even need a zener
[10:05:40] <twnqx> you put a standard silicon or shottky from your input to VCC
[10:07:10] <twnqx> so your input will not go higher than VCC + diode forward loss
[10:07:51] <twnqx> that's how clamping is normally implemented in ICs...
[10:11:11] <tzanger> yep, it's good and cheap but does not protect your VCC line from potentially rising to the invalid/fault voltage
[10:11:55] <tzanger> if you've got a good enough power supply it'll usually bully its way and maintain your VCC. it all depends on what kind of fault you're trying to mitigate and the constraints of the system
[10:12:05] <twnqx> well
[10:12:07] <twnqx> right now
[10:12:19] <twnqx> i am feeding 5V from my AVR into my 3.3V powered FT230X
[10:12:28] <twnqx> whose inputs are "5V tolerant"
[10:12:39] <RikusW> twnqx: GTL2002
[10:12:42] <RikusW> or 2003
[10:12:43] <twnqx> which more often than not means "diode clamped, as long as the current isn't too high"
[10:12:46] <tzanger> I bet a simple 100 ohm resistor would be more than enough to protect the FT230
[10:13:01] <twnqx> i have no resistor since it's 5V tolerant:P
[10:13:10] <twnqx> but i wonder what happens with the excess
[10:13:27] <RikusW> on mosfet input, nothing
[10:13:31] <twnqx> nothing, because it's just some µA to drive the CMOS couple
[10:13:48] <twnqx> or more, because there are clamp diodes
[10:14:02] <tzanger> the AVR can source/sink 20mA, and the 100 ohm resistor will dissipate up to 4mW (I^2R). I don't know how strong the input protection diodes are on most ICs but they're usually not that great
[10:14:12] <twnqx> and when i read that my CAN transceiver (automotive grade) is +-200V resistant
[10:14:40] <twnqx> on a signal that has 2.5V +-0.7 or so (differential)
[10:15:04] <tzanger> what happens iwth the excess is you get a 2.7V drop across the (internal) protection diode and whatever current the pin is providing makes up a certain power level that the diode must dissipate
[10:15:15] <twnqx> yes
[10:15:35] <twnqx> but will my vcc voltage rise, potentially damaging the whole circuit?
[10:16:10] <twnqx> or is it just "the mosfets won't melt down immediately when epxosed to that voltage"
[10:16:22] <tzanger> now again, say 2.7V * 20mA is ~50mW. a 100 ohm resistor is cheap insurance and helps take the load off the internal diode
[10:16:51] <tzanger> twnqx: doubtful. you've only got 5V@20mA trying to push a stiff 3.3V supply up
[10:16:54] <twnqx> i'd limit it to 1mA
[10:17:03] <twnqx> my whole circuit uses 150mA...
[10:17:04] <tzanger> it's not going ot move, but the protection diode is what'll be dissipating that power
[10:17:19] <tzanger> so put a 100 ohm resistor in series with teh I/O between the AVR and FT230
[10:17:33] <twnqx> nah, in this scenario vcc can't go up
[10:17:44] <twnqx> and no, too late to fit a resistor :P
[10:17:57] <twnqx> i'm more worried about the genaral case, not this particular one
[10:18:01] <twnqx> general*
[10:18:08] <tzanger> you'll likely be fine. MAYBE shorten the life of the FT230, but probably not to any length you'd notice
[10:18:37] <tzanger> general case when designing these kinds of things is to put room for series resistors, you can always stuff them with 0 or adjust the gerbers to short them after if you really want
[10:18:43] <twnqx> but what if i really want to go and be, say, 1.5kV resilient
[10:19:10] <tzanger> 1.5kV has some bigger strike/creep clearances you have to be aware of, and you also need to look at what you're tryign to do
[10:19:34] <tzanger> 1.5kV doesn't mean anything. Are you using a human body contact type of model or worried about a 1.5kv line touching the board or ESD or?
[10:19:54] <twnqx> just ESD protection at those levels
[10:21:17] <tzanger> ESD protection is usually achieved with high speed external protection. ESD has no real current to worry about but you're trying to avoid the high voltage discharge from puncturing the oxide layer
[10:21:28] <twnqx> but automotive stuff could have some unusual voltage swings
[10:21:43] <tzanger> automotive has load dumps and other fun things to test for
[10:22:02] <tzanger> i.e. your supply going from 14ish V to upwards of a hundred or more
[10:22:12] <twnqx> yeah...
[10:22:22] <twnqx> it can handle 30V easily
[10:22:25] <tzanger> http://www.ti.com/lit/an/snva681/snva681.pdf
[10:22:38] <tzanger> that says peak load dump can be 60V
[10:23:08] <twnqx> ... 400ms at 120V
[10:23:15] <twnqx> yeah, that would probably fry this ciruit
[10:23:15] <tzanger> but right inside that document it says 120V and 400ms
[10:23:16] <tzanger> yeah
[10:23:26] <tzanger> so you use a shunt system then
[10:23:46] <tzanger> zener and fuse, or just a beefy zener and don't skimp on the trace width
[10:24:03] <twnqx> :/
[10:24:18] <tzanger> 400ms isn't enough to heat the traces, but make sure the zener can withstand the current for 400ms
[10:24:37] <tzanger> it's a small amount of time and if you are looking at transient protection you can get an AWFUL lot of current passed for a short time without damaging things
[10:25:00] <tzanger> when I worked in power electronics you'd see ratings on SCRs that were rated for say 100A continuous but had ratings of 10kA (!!) for one-half cycle
[10:25:14] <twnqx> yeah, noticed that quite often
[10:25:28] <twnqx> when selecting diodes for my SEPIC converter
[10:25:47] <twnqx> i don't care about 30V, circuit would handle that easily on a permanent base
[10:25:48] <tzanger> it's all about heat, if the pulse is short the wafer gets hot but doesn't get damaged since the energy is gone before it gets out of the SOA
[10:25:59] <twnqx> i thought i was kinda safe with that :(
[10:26:51] <twnqx> but at 120V... i'd have exploding diodes and capacitors
[10:27:11] <twnqx> maybe even the mosfets would die
[10:28:21] <twnqx> hm ok, 30V would be dnageoursly close to some limits
[10:28:55] <twnqx> so how would i dtermine the current that goes trhough it?
[10:28:56] <tzanger> like I said, reverse-polarity 25V SMB zener across the rails
[10:29:12] <tzanger> how thick are the power supply wires?
[10:31:58] <tzanger> the car will have a fuse for the power line you're on. the worst that will happen with the zener is that it'll pop. if it comes apart your board is susceptible to the next dump. if it shorts you'll take out the fuse, maybe the traces if the traces are too thin for the fuse rating
[10:32:46] <tzanger> the best part about designing this stuff is figuring out how to make sure the protection is still in place. How will you know if the zener opened up or has degraded?
[10:33:55] <twnqx> at 12V i draw 40mA...
[10:34:05] <twnqx> or <50 at least
[10:34:37] <tzanger> not what you draw
[10:34:46] <tzanger> what the circuit is capable of sourcing in a fault condition
[10:35:10] <tzanger> what's the fuse rating on the line? what's the length/gauge of the wire going to your circuit? how thick are the traces from the power connection to the zener?
[10:35:19] <twnqx> power traces ar 0.5mm
[10:35:41] <tzanger> just guessing but ou probably have a 10A fuse and 16 or 18AWG wire
[10:35:50] <twnqx> external wiring 0.35mm
[10:36:06] <tzanger> so make sure your traces can take 10A for a few second or so with a 40C rise over ambient
[10:36:22] <tzanger> the fuse should be open by then and everything can cool off
[10:36:23] <twnqx> and i have no fuse on my gear
[10:36:28] <twnqx> :X
[10:36:41] <twnqx> i think the power supply is 15A
[10:36:45] <twnqx> (the fuse in the car)
[10:37:29] <twnqx> but wouldn't it make sense to add a resistor to limit the current as well?
[10:38:34] <tzanger> a quick look online says you can get a 60C rise over ambient with ~8cm of 6mm trace width on 1oz copper at 20A
[10:38:48] <tzanger> the fuse should blow in milliseconds at that current meaning that you won't see that rise
[10:39:04] <tzanger> right, I'm saying the car has all lines fused
[10:39:11] <twnqx> well, you wouldn't want a fuse to blow for something you experience in normal conditions
[10:39:12] <tzanger> the fuse rating depends on which line you've tapped
[10:39:20] <tzanger> a 120V load dump is not a normal condition
[10:39:40] <tzanger> remember we're not worried about the fuse blowing for every load dump
[10:39:57] <tzanger> we're worried abot the zener shorting out due to a load dump that it could not handle
[10:40:27] <tzanger> the "worst case" load dump is 120V for 400ms. let's see what zeners are out there
[10:42:19] <twnqx> i think i'll build that tl431 based protection circuit as external frontend
[10:42:20] <tzanger> SMBJ24A is rated for 600W and comes in SMB package. it'll take a 100A surge for 8.3ms (60hz half-wave)
[10:43:21] <tzanger> it'll take over 80A as a non-repetitive pulse at 800ms
[10:43:28] <tzanger> that's a great diode for this
[10:44:38] <tzanger> it'll clamp to 26V which is a little close to your 30V max (I'd poke around and see what else there is that's a little lower) and in the case where it shorts, it'll blow the 15A fuse. what happens when it fails open though?
[10:44:43] <tzanger> you're unprotected
[10:45:17] <tzanger> the diode however is pretty rugged; it'd take a hell of a current pulse to have it fail open, and most semiconductors fail shorted, which blows your fuse and protects you
[10:45:19] <twnqx> hence the active circuit that shuts down supply
[10:45:30] <twnqx> according to the pdf you linked :)
[10:45:31] <tzanger> and when you replace the fuse it'll promptly blow the next fuse, forcing you to take a closer look. :-)
[10:46:14] <twnqx> too bad this PCB is already full
[10:46:42] <tzanger> there's always room for one more component
[10:46:54] <twnqx> on the bottom, yes :P
[10:47:11] <tzanger> queue the scene from monty python where the waiter pursuades teh obese man to eat just one more cracker
[10:48:50] <twnqx> i could throw out the unused bulk alectrolytic capacitor :P
[10:49:18] * twnqx always prepared for them but never fits them
[11:27:28] <TechIsCool> So I have a weird problem. I have some example code that works perfect in its default state but as soon as I comment out a printf command it breaks everything Code is listed here
https://github.com/gcopeland/RF24/blob/master/examples/pingpair/pingpair.pde where I comment out it line 211
[11:30:21] <RikusW> TechIsCool: a delay should fix it again...
[11:30:27] <RikusW> like 1ms or so
[11:31:36] <TechIsCool> I understand I could deplay but when all the other code is not dependent on that I would rather know what causes it to stop working
[11:32:49] <xorentor> TechIsCool: what does a debugger say?
[11:33:42] <TechIsCool> xorentor: I am not really sure how to debug it properly. I have a avr dragon with debugwire functioning what I can't understand is how check it
[11:34:15] <TechIsCool> my first idea is to just use mills and take it before the printf and then after to figure out how much time it takes
[11:40:17] <xorentor> printf takes about 50k clock cycles unoptimized on gcc, with O2 it's significantly less
[11:40:38] <TechIsCool> lol what a waste of time
[11:43:43] <TechIsCool> RikusW: with a delay of 25ms it works fine without problems
[11:43:50] <xorentor> as RikusW already suggested, radio.startListening(); is called too early after radio.write( &got_time, sizeof(unsigned long) ); (if printf is missing) .. so you should put some delay there, however, you need to find out exactly what the delay should be, so you don't slow it down
[11:44:14] <xorentor> that's where a debugger comes handy
[11:44:20] <RikusW> or fix the actual problem
[11:44:34] <RikusW> I got a dragon to
[11:44:43] <RikusW> but for realtime debugging its useless
[11:44:57] <RikusW> nearly anyways, like for soft uart...
[11:45:24] <TechIsCool> what do you use for realtime debugging?
[11:49:00] <RikusW> printf ?
[11:49:18] <RikusW> or custom data dumped via serial maybe
[11:49:26] <OndraSter__> yep
[11:49:29] <RikusW> breakpoints + realtime don't work too well
[11:49:34] <RikusW> or at all
[11:49:42] <OndraSter__> nope
[11:49:45] <TechIsCool> not at all
[11:49:45] <OndraSter__> try debugging USB stuff :)
[11:50:01] <RikusW> might be usefull for verifying that the code breaks at some point
[11:50:22] <RikusW> OndraSter__: dW got one HW breakpoint actually, use "run to cursor"
[11:50:40] <OndraSter__> :)
[11:51:18] <RikusW> USB is a nightmare
[11:51:36] <OndraSter__> hardware debugging does not help here :D
[11:51:54] <RikusW> cool toy
http://www.youtube.com/watch?v=I3jE-PXV-68
[11:52:13] * RikusW can use YT now, on 4Mbit line :)
[11:52:18] <OndraSter__> wow
[11:52:19] <OndraSter__> nice
[11:52:41] <RikusW> shared by 3 people
[11:52:49] <RikusW> 90 Euro/m
[11:53:28] <inkjetunito> RikusW: with/without a monthly transfer limit?
[11:53:39] <RikusW> uncapped
[11:54:01] <RikusW> in SA actually so ZAR900
[11:55:04] <TechIsCool> RikusW: How far are you away from the termination point?
[11:55:10] <RikusW> no idea
[11:55:14] <inkjetunito> yeah, that's about 2-3x the price in europe
[11:55:16] <RikusW> must be close
[11:55:35] <RikusW> how much is 10Mbit in EU ?
[11:55:53] <megal0maniac> $1
[11:56:08] <inkjetunito> RikusW: 40 euros i suppose
[11:56:14] <RikusW> nice
[11:56:16] <TechIsCool> for me about 40$
[11:57:02] <TechIsCool> http://pastebin.com/g2MrVq5r
[11:58:02] <OndraSter__> I can in theory get 100Mbit for $45
[11:58:02] <OndraSter__> or 200Mbit for $100 :)
[11:58:19] <TechIsCool> I pay for 36/3 for about 80$
[11:58:23] <megal0maniac> In your little village
[11:58:37] <OndraSter__> but I live too far from the Prague's centre :(
[11:59:31] <OndraSter__> in here we pay $17.5 for whatever works -- it used to be 24/24Mbit (half duplex), now it is 8Mbit-ish
[12:00:16] <TechIsCool> Thats not bad at all
[12:03:39] <TechIsCool> Ok next dumb question when I open the serial port it cause the code to crash well stop responding to the wireless module. How do I determine why
[12:07:08] <TechIsCool> nevermind I found it
[12:07:44] <OndraSter__> well
[12:07:44] <OndraSter__> you can judge it by my e-penis behind my nickname
[12:07:44] <OndraSter__> ;)
[12:07:44] <OndraSter__> how stable it is
[12:07:58] <TechIsCool> lol
[12:09:49] <megal0maniac> Is /that/ what they call it?
[12:09:52] <megal0maniac> :)
[12:11:02] <OndraSter__> does it do DTR/RTS/...?
[12:11:55] * megal0maniac slaps megal0maniac_afk around a bit with a large trout
[12:12:32] <megal0maniac> Dafuq?
[12:13:31] <TechIsCool> OndraSter__: I think so its a arduino
[12:13:39] <TechIsCool> duemilanove
[12:13:58] <OndraSter__> btw TechIsCool weren't you the one who was using AP5100?
[12:14:18] <TechIsCool> Yes
[12:14:23] <TechIsCool> but I have yet to buy those parts
[12:14:44] <TechIsCool> I am stuck working on getting the wirless part of the project working on a bench before I buy parts. Why do you ask
[12:23:28] <OndraSter__> :P
[12:23:28] <TechIsCool> OndraSter__: Yup thats me
[12:24:34] <TechIsCool> OndraSter__: You still like them or are you having a problem?
[12:26:24] <OndraSter__> still liking
[12:26:25] <OndraSter__> I just had 3 minutes lag on IRC :D
[12:26:38] <TechIsCool> fail
[13:47:07] <TechIsCool> So when my serial console is open it causes issues
[14:23:11] <[z_z]> doing some usb using the asf. for some reason the vender/product strings don't show up in lsusb... even though they have numbers, and USB_DEVICE_MANUFACTURE_NAME is defined... any ideas/common mistakes?
[14:24:31] <RikusW> having _fun_ with usb ? ;)
[14:24:40] <RikusW> many things can cause that
[14:25:12] <RikusW> even looking at your screen the wrong way :-D usb is an overly sensitive creature :-P
[14:25:34] <RikusW> [z_z]: enumeration can fail for any number of reasons
[14:25:59] <RikusW> if you make even the slightest mistake in any one of the enumeration steps all will fail
[14:26:12] <RikusW> just ask OndraSter__ if he agrees....
[14:26:54] <[z_z]> I don't think enumeration is failing completely. The device is usable, just those damn strings...
[14:28:12] <RikusW> abcminiuser: is the local usb expert maybe he knows
[14:29:26] <[z_z]> lol. Its like abcminiuser is the local expert for everything.
[14:29:40] <abcminiuser> Aww I'm watching Game of Thrones
[14:30:19] <abcminiuser> The strings in lsusb come from usb.ids, not the device
[14:30:24] <abcminiuser> Which is damned confusin
[14:30:26] <abcminiuser> *confusing
[14:30:51] <RikusW> the code returning the strings might be at fault
[14:38:28] <[z_z]> udc_req_std_dev_get_descriptor calls udc_req_std_dev_get_str_desc which calls udc_req_std_dev_get_str_desc... all that happens in udc.c which i have made no modifications to... grr.
[14:38:46] <abcminiuser> > <abcminiuser> The strings in lsusb come from usb.ids, not the device
[14:40:34] <RikusW> abcminiuser: afaik the descriptor contains numbers
[14:40:51] <RikusW> then the OS queries the device with those numbers to get the strings ?
[14:59:29] <OndraSter__> no
[14:59:39] <OndraSter__> no idea how in linux, but on windows it queries the device
[14:59:45] <OndraSter__> but if drivers are installed it uses the driver name
[15:00:37] <OndraSter__> and yes, USB is really fun
[15:42:47] * RikusW didn't know pci-e x1 vga cards even existed -> http://www.sunix.com.tw/services/list_magazine.html?messageid=397
[15:49:23] <tzanger> hm, I could use a couple of those
[15:49:33] <OndraSter__> well
[15:49:38] <OndraSter__> my IGP can take two screens
[15:49:42] <OndraSter__> and my dedicated can take 4
[15:49:58] <tzanger> apparently (I've not tested it but do have a pretty good understanding of PCIe) that you can tape over lanes and the protocol will gracefully degrade
[15:50:06] <OndraSter__> pretty much, yeah
[15:50:40] <tzanger> and you can use any lane in a multi-lane connector
[15:50:43] <tzanger> I have tested that
[16:19:17] <RikusW> anice summary
http://en.wikipedia.org/wiki/LGA_1155
[16:29:50] <tzanger> hm, got the dragon, seems overkill for a programmer
[16:31:26] <[z_z]> hmm. so yes the strings printed by lsusb come from a file that maps product and vender codes to human readable names... the strings that come from the lsusb -v iSerial iProduct etc. are queried...
[16:32:31] <[z_z]> doesn't explain why iSerial shows the correct string, but iProduct and IManufacturer does not...
[16:33:14] <twnqx> blame the chinese cloners :P
[16:55:35] <langoliers> h
[16:57:02] <langoliers> o cool i'm doing 3 jobs in parallel
[17:05:03] <OndraSter__> RikusW, so appearantly my computer will have to be pretty much completely renewed for haswell/broadwell
[17:05:09] <OndraSter__> except I AM FINE WITH MY SANDY!
[17:05:17] <RikusW> heh
[17:05:30] <OndraSter__> 1155 socket, 1333MHz RAMs...
[17:05:53] <RikusW> I'm looking into MSI-Z77A-G43
[17:06:03] <OndraSter__> I have got G45
[17:06:07] <RikusW> about 100Euro
[17:06:10] <OndraSter__> yep
[17:06:47] <RikusW> does those combined PS2 ports have both mouse and keyboard clock and data ?
[17:06:48] <Steffanx> "MSI Z77/H77 Motherboards have many unique features which can maximum the performance of Intal i3, Intel i5, Intel i7 CPUs".. intal i3 :D
[17:06:56] <OndraSter__> RikusW, no idea, I use USB
[17:07:04] <OndraSter__> lol Steffanx
[17:07:18] <RikusW> I'm considering Pentium G2020 for now
[17:07:22] <RikusW> 50E
[17:07:33] <RikusW> i3 3220 110E
[17:07:38] <Tom_itx> still debating?
[17:07:43] <OndraSter__> I have got i5-2500k
[17:07:51] <OndraSter__> in my server I have got G540 or G550
[17:08:03] <RikusW> my brother's got a i5-2500 too
[17:08:25] <RikusW> don't need quadcore right now
[17:08:25] <OndraSter__> G550 in the server
[17:08:27] <OndraSter__> :)
[17:08:31] <OndraSter__> 2500 or 2500k? :P
[17:08:36] <RikusW> not sure
[17:08:37] <OndraSter__> I paid extra for the unlocked multiplier
[17:08:41] <OndraSter__> (10€-ish more)
[17:08:47] <OndraSter__> if I need power, I just overclock it to 4.5GHz
[17:08:56] <RikusW> heat ?
[17:09:04] <OndraSter__> Thermalright ultra 120
[17:09:12] <OndraSter__> kit for 1155 socket cost me 10€ :)
[17:09:14] <OndraSter__> err
[17:09:16] <OndraSter__> 5€
[17:18:42] <disq> Hi! I need a 1mhz clock signal to fix the fuses on an atmega64 rc transmitter. Would it be possible to use an atmega8 (a non-working but still-flashable) ESC unit to provide the signal? (Which pin to use, how to toggle it, any tips or examples?)
[17:23:59] <disq> well, this was fast, I was already pointed to the answer on rcgroups, even with the clock code! i'll hang out a bit anyway :)
[17:25:26] <langoliers> hi
[17:25:30] <langoliers> what precision ?
[17:26:25] <langoliers> ok you probably want some 1-2ms crap, then rcgroups has numberless things of those
[17:26:59] <Tom_itx> disq, yes.
[17:27:40] <Tom_itx> langoliers, no he want's to recover his fuses
[17:27:50] <disq> yeah, precision isn't that important, just to allow me to set the fuses on the m64 back
[17:28:04] <Tom_itx> my programmer has that built in :)
[17:28:14] <langoliers> oh then grab an 555 and a resistor, and maybe a capacitor and let it live
[17:28:31] <Tom_itx> or a fast blink program on his spare mega8
[17:28:32] <langoliers> other solution is a cd4093, 40106 or similar
[17:29:02] <Tom_itx> PORTB ^= 0xFF;
[17:29:03] <Tom_itx> delay
[17:29:05] <Tom_itx> repeat
[17:29:12] <langoliers> or an lm393, lm358 in astable mode
[17:29:40] <langoliers> or an astable multivibrator using 2 transistors, a few resistors, 2 capacitors
[17:31:22] <langoliers> or a cd4060 with some quartz, or rc osc, or resonator, you get divided clocks too
[17:31:54] <langoliers> 2hz from 32.768kHz crystal
[17:32:59] <Tom_itx> the clock rate needs to be 4x the spi program frequency
[17:33:31] <langoliers> what is the spi freq btw?
[17:33:44] <Tom_itx> what you set it to
[17:33:49] <Tom_itx> default is 125khz
[17:33:55] <langoliers> parallel port probably does not do good above 50kHz ;>
[17:34:36] <langoliers> btw what is the fmax of the osc pins ?
[17:34:48] <langoliers> 100MHz ?
[17:36:14] <disq> bored, eh :)
[17:36:15] <langoliers> if div/8 is on then giving 160MHz to the xin is within specs
[17:37:44] <Tom_itx> i put out a 4Mhz clock for such purposes
[17:38:21] <Tom_itx> default freq for most chips is 1Mhz
[17:38:24] <OndraSter__> or a common mode choke!
[17:38:29] <OndraSter__> with a transistor
[17:38:35] <OndraSter__> oscillates too.. :P
[17:38:47] <OndraSter__> (battery joule thief)
[17:40:15] <langoliers> a pushbutton?
[17:41:48] <langoliers> OndraSter__<= ok, 1 transistor simplified colpitts then ?
[17:42:06] <langoliers> that can do FM 108 too
[17:52:36] <langoliers> http://www.rapidonline.com/Electronic-Components/Crystal-oscillator-modules-72920
[17:52:43] <langoliers> :)
[17:52:51] <langoliers> this chip has one purpose
[18:01:52] <disq> fuses restored :)
[18:04:13] <langoliers> what solution you used? a pushbutton, and kept pressing?
[18:05:06] <disq> lol
[18:05:32] <langoliers> hahaha it won't work btw
[18:08:51] <disq> ...unless you have a robotic arm that can push 1 million times a second?
[18:09:11] <disq> or you just hack the pushbutton and solder a 555 there?
[18:10:29] <langoliers> in theory if you could keep the button continuously bouncing, it'd be all right, but the clock must be stable
[18:10:40] <langoliers> atmel don't like unstable clocks
[18:13:19] <disq> ok, i have another question for you :)
[18:14:01] <disq> what's up with usbasp misreading the "device signature" as 0x102 and complaining it's not 1e962?
[18:14:20] <langoliers> nah finished eating
[18:14:25] <disq> i shouldn't use bad usbasp clones from china? :)
[18:14:44] <langoliers> use a proper cable
[18:15:16] <langoliers> kill a junk ata cable, 40 or 80 pin, and connect every second to gnd, so you don't have cross-talk problems
[18:15:27] <disq> apparently it doesn't like my keyboard's usb hub
[18:15:59] <disq> connected it to the powered hub of the monitor, finally worked (everything has a built in usb hub nowadays)
[18:16:16] <langoliers> it was an usb transmission error? ;/
[18:16:56] <disq> yes, but a consistent one.
[18:17:22] <langoliers> well vusb doesn't have parity checking
[18:17:49] <disq> though a proper cable or a port on the device is in my todo list. currently I have 3-pin servo connectors sticking out that I plug into the 10-pin cable
[18:19:14] <disq> I was under the impression that the seller would also send a 10 to 6 adapter but apparently that was a trap and I didn't notice it
[18:20:18] <Casper> disq: avr programming is quite timing sensitive
[18:20:23] <langoliers> you can desolder goldplated jumpers from a motherboard
[18:20:30] <Casper> all those hubs add delays and jitters
[18:20:56] <langoliers> the female pair is not used on them though
[18:22:17] <langoliers> ah usb hubing adds lag that's sure
[18:26:40] <disq> I need a connector to affix to the plastic case. maybe I should use something like a 6-pin din, since it would be very easy to mount
[18:26:58] <[z_z]> hm... if usb spec says device strings are in unicode... whatwill happen if i specify an ascii string?
[18:28:09] <langoliers> disq<= will it be programmed that often ?
[18:28:23] <langoliers> RJ-45 ?
[18:29:12] <langoliers> http://en.wikipedia.org/wiki/File:Ethernet_RJ45_connector_p1160054.jpg
[18:29:20] <langoliers> fair quality connector with 8 pins
[18:29:33] <disq> no, but it's also used to read eeprom, to change settings on the computer instead of fiddling with nasty switches and bad ui :)
[18:30:22] <langoliers> oh yea
[18:30:25] <disq> wouldn't the female part of the rj45 be flimsy and hard to mount on a hard plastic case?
[18:30:33] <langoliers> you can add an usb cable
[18:30:53] <langoliers> usb is crap too if it gets to mounting.
[18:31:09] <langoliers> they used to break off laptop boards
[18:31:18] <langoliers> *from
[18:31:22] <disq> a usb-b connector (square-type) might work, but I would have to get another programmer
[18:32:22] <langoliers> programming needs 6 pins, usb has 4)Ö
[18:32:42] <disq> (i could embed a programmer inside)
[18:32:57] <langoliers> hm... you can also use a bootloader
[18:33:18] <langoliers> not saying usb is cool
[18:33:23] <langoliers> it is shit
[18:33:34] <langoliers> you are happy if it works
[18:34:23] <disq> a db-9 could also work, but it's so boring
[18:34:35] <langoliers> :)
[18:34:41] <langoliers> no it is ok
[18:35:03] <langoliers> if you want 15 pins then you can have the vga connector too
[18:36:21] <disq> i could settle for a 6 row female pin header thingy, provided i don't have to crimp it (and affix with hot glue)
[18:37:08] <langoliers> i usually try to use as few wires i can
[19:22:13] <tzanger> wow I can't get this damned dragon to work with hvsp
[19:23:02] <tzanger> my wiring is right, I've tied +5 to VTG on the ISP header (I'm using the HVPROG headers) and avrdude says broken pipe
[19:23:17] <tzanger> if I don't have the jumper to VTG it communicates but says no power detected
[19:26:01] <Tom_itx> i tried HVPP one time on a dragon
[19:27:56] <tzanger> yes but it seems that avrdude is not very well run in hvsp :-/
[19:28:03] <tzanger> I guess I'll try it in windows
[19:28:11] <tzanger> with avrstudio, bleh
[19:30:17] <Tom_itx> hey, if it works you can't really knock it
[19:30:55] <tzanger> yes i can :-)
[19:32:49] <tzanger> hm, avirstudio can't see it either
[19:33:03] <Tom_itx> does it do hvsp?
[19:33:14] <Tom_itx> maybe the dragon needs a firmware update
[19:33:25] <Horologium> you will probably need avr studio, not atmel studio...like avr studio 5 or something.
[19:34:06] <Tom_itx> funny, i've got 5 versions installed
[19:34:25] <Horologium> none here....no windows to install it on.
[19:34:59] <Tom_itx> 3 linux boxes 2 windows and 1 dos box here
[19:35:10] <Tom_itx> dos has nt4 on it
[19:35:17] <Tom_itx> dual boot
[19:35:35] <Horologium> just 3 linux boxes here...and the wifey's win8 notebook.
[19:35:44] <Horologium> but I don't put dev tools on her computer.
[19:35:47] <Tom_itx> if i counted all those....
[19:35:58] <Tom_itx> 3 more lappys
[19:36:13] <Valen> nt4 could run directX w00t
[19:59:50] <tzanger> I have AVR studio, did a firmware update, unfortunately no change
[21:10:39] <tzanger> balls. my avr is fine. buspirate reads it fine in isp mode
[21:10:47] <tzanger> dragon is very disappointing
[21:13:43] <Valen> don't say that
[21:13:50] <Valen> I am about to start using one lol
[21:14:20] <tzanger> I don't know if ISP works, but HVSP is a disaster so far
[21:14:42] <tzanger> I can talk to the dragon just fine, but dragon can't seem to talk to the attiny13
[21:16:16] <Valen> on the supported list?
[21:22:07] <tzanger> yes of course
[21:22:19] <tzanger> neither avrstudio nor avrdude can read the attiny13
[21:22:29] <tzanger> I've verified the wiring, I guess I have to break out the scope tomorrow
[21:22:54] <Valen> logic analiser might do a better job if you have one
[21:26:51] <Casper> tzanger: are you sure that you use the right pins? some avr use non-standard one
[21:27:03] <tzanger> yes
[21:27:03] <Casper> and some tiny use PDI instead of SPI
[21:27:11] <tzanger> this is HVSP not ISP
[21:27:15] <tzanger> ISP works fine with buspirate
[21:52:32] <TechIsCool> tzanger: What is wrong with your dragon?
[21:53:56] <TechIsCool> you are not using pdi by chance?
[22:03:53] <tzanger> no I am not using PDI
[22:03:57] <tzanger> I'm using HVSP
[22:07:40] <TechIsCool> tzanger: Alright pdi requires a 50pf cap between reset and ground I found out after head banging for about 3 days not documented anywhere
[22:08:02] <tzanger> really the atmel docs seem to suggest removing any capacitance
[22:08:19] <tzanger> and 50pF is barely anything, the dragon seems to be a really marginal design
[22:08:40] <TechIsCool> Don't ask my why another user on irc tipped me off to it and it works perfect. Can't read the device before the cap but after it reads just fine
[22:09:06] <tzanger> hm
[22:09:12] <TechIsCool> He found it because his scope probes where 25 pf a piece
[22:09:17] <tzanger> maybe I'll tack on a 50pF cap on the reset line here
[22:09:33] <TechIsCool> I know it might not be your issue but as a dragon owner its worth knowing
[22:09:44] <tzanger> yeah for sure
[22:10:18] <tzanger> I already found an obscure reference to tying the +5 from the dragon to the VTG on the ISP header when using HVSP or the dragon says the target is not powered
[22:10:54] <TechIsCool> Yah I ran across that aswell but I don't use hvsp
[22:19:49] <tzanger> the dragon seems needlessly complex for what it is. two mcus, external memory... odd design
[22:20:12] <TechIsCool> find that it was made that way to make it cheap I know it sounds dumb but
[22:20:29] <tzanger> not sure how that design achieves those goals
[22:21:07] <tzanger> oh, the dragon does full emulation of a bunch of devices too
[22:21:09] <tzanger> that explains it
[22:21:10] <TechIsCool> I use isp and it seems to work most of the time. Windows 8 was not supported until the beta
[22:25:04] <tzanger> yeah I can't use ISP in this design
[22:25:11] <tzanger> it's a tiny13a and I'm using all 6 I/O lines
[22:26:02] <TechIsCool> is the DWEN stuck?
[22:30:21] <tzanger> nothing is stuck, the tiny13a works fine with a buspirate (but I can't blow the reset disable fuse)
[22:30:52] <TechIsCool> since you have access with the bus pirate did you ever do debugwire?
[22:31:15] <tzanger> never tried debugwire
[22:31:38] <TechIsCool> you have a dump of the fuses set on it?
[22:34:08] <tzanger> yep
[22:34:13] <tzanger> nothing out of the ordinary
[22:34:20] <tzanger> as I said I can connect and program it fine with the bus pirate in ISP
[22:34:49] <TechIsCool> but you can't program from isp via the dragon?
[22:35:09] <Tom_itx> sure you can
[22:35:27] <Tom_itx> he probably want's to reset something require hv
[22:36:09] <tzanger> no
[22:36:28] <tzanger> I currently just want to see the dragon identify the tiny13a via hvsp
[22:36:46] <tzanger> once that is done I can safely blow the fuse and start development
