**** BEGIN LOGGING AT Mon Apr 13 03:00:00 2015 Apr 13 03:34:10 veremit: i finally have it figured out. thanks for pointing me in the right direction. Apr 13 09:46:30 Hi,I'm using the Debian image for the BBB from here: Debian (BeagleBone Black - 2GB eMMC) This image has the Cloud9 IDE built-in. It works quite nicely for my purposes, but I can't figure out how to add a password. Anyone on the network can go to 11.22.33.44:3000 (not the actual IP address) and the IDE will automatically log them in as "John Doe" (No password requested). Is there a way to reques Apr 13 09:47:01 *password Apr 13 09:51:14 looks like a cloud9 question? Apr 13 09:56:36 : yes mate, sorry if its not related to this chat area. As it was inbuilt IDE in BBB, thought to ask some suggessions as I tried different things. Apr 13 10:10:08 Amil okay the question comes here from time to time and I did not see any answer Apr 13 10:13:05 : Yeah, its ok mate. appreciate your response. Will have a look. Apr 13 10:23:12 ok, its dnsmasq Apr 13 10:59:33 Hello Apr 13 11:01:34 I has beaglebone white Rev A6, I want to install Debian in that . I downloaded the image from Apr 13 11:01:35 http://elinux.org/BeagleBoardDebian Apr 13 11:02:08 and flashed the sd card using dd command. But i the device faied to boot from that SD CARD.. Apr 13 11:02:23 Please help me.. I am new to embedded system Apr 13 11:14:45 Is that Beagle bone require exact 4GB sd card .. Apr 13 11:14:52 BeagleBone white Apr 13 11:16:58 Please let me know wheather BeagleBone (White Rev A6) requires exactly4GB sd card side Apr 13 11:22:00 sridhar_ buy a serial cable, connect it and you will be enlighted Apr 13 11:22:30 hm btw. the white has serial console support out of the box Apr 13 12:06:21 Hi, does anyone know if you can do VLANs on the BBB Ethernet port? Apr 13 12:07:17 Peanut the same way as on any other linux too? Apr 13 12:07:40 woglinde: I'd think so, but I wonder if it requires any hardware/driver support. Apr 13 12:09:44 I'll ahve to try it out tonight once I get home, or take it with me to the office again tomorrow. Apr 13 12:12:51 Peanut linux can do that all in software Apr 13 12:13:48 Ok, thanks... I'll see if I can get my boss to sign for a pair of BBB's, enclosures and TCXO's. Apr 13 12:14:55 should be under 200 dollars Apr 13 14:55:46 I see that we got a new 'official' Debian image on 2015-03-01. Is there a place where I can find some overview of the changes in this version? Apr 13 14:57:00 Peanut ask rcn Apr 13 15:06:32 * tbr will need to pounce on rcn or someone else familiar with dtsi... somehow if I include this my kernel doesn't boot. http://pastebin.com/jU97tnNR (first two lines removed) Apr 13 15:11:02 Ah, found it: http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#2015-03-01_Release_Notes Apr 13 15:16:19 Peanut: that sounds like you're doing something interesting Apr 13 15:18:02 Hi ddrown, nice to see you again. We're retiring some old Sun machines, and they run NTP amongst other things. Most of that is now virtualized, but I don't want to put NTP on a VM (it doesn't work very well, and I'm a bit of a time-nut). Apr 13 15:18:21 cool, makes sense Apr 13 15:18:53 So my plan is to use two BBB with TCXO and your code as a pair of NTP servers (just getting it from upstream clocks, no GPS reception at all in our basement). Apr 13 15:20:10 I'm curious to see your results Apr 13 15:25:42 Peanut, sounds good .. hopefully you'll do a write-up :) Apr 13 15:28:11 Also, I've found a source for a VCTCXO to generate 24 MHz and lock it to the Rb now: http://www.abracon.com/Oscillators/ASTGX.pdf Apr 13 15:28:47 It's a bit more expensive than other VCTCXO (€ 65), but you can buy just a single one instead of a whole SMD reel. Apr 13 17:56:38 hello Apr 13 20:00:44 Hi guys! Apr 13 20:01:23 so, I was following this guide on creating a bootable microSD for BBB: Apr 13 20:01:36 http://archlinuxarm.org/platforms/armv7/ti/beaglebone-black Apr 13 20:02:42 and, when I try to boot from the microSD, I get this error messages from u-boot: Apr 13 20:02:43 ** File not found /boot/zImage ** Apr 13 20:02:43 ** File not found /boot/zImage ** Apr 13 20:02:43 No zImage found Apr 13 20:02:43 ** File not found /boot/zImage ** Apr 13 20:02:45 No kernel found Apr 13 20:03:04 I checked the boot folder and zImage IS there :/ Apr 13 20:05:28 @all any thoughts? ^ Apr 13 20:08:27 yes, whoever wrote that guide likes to live dangerous Apr 13 20:08:56 1) using boot and root as directories for mountpoints is a bad idea Apr 13 20:09:12 2) using relative paths for mountpoints is another very bad idea Apr 13 20:09:35 i'm pretty sure you messed up somewhere and have the files on your local disk instead of your sd card Apr 13 20:11:39 KotH: thanks for the input Apr 13 20:11:51 but I had mounted them on /mnt Apr 13 20:11:53 :) Apr 13 20:12:20 gjsousa uboot has commands to show the content of filesystems Apr 13 20:12:28 use them Apr 13 20:13:04 and good nite Apr 13 20:13:31 schlaf gut, woglinde Apr 13 20:14:49 woglinde: tks, good night there... still afternoon here in Brazil :P Apr 13 20:42:55 hi vmayoral, have time time for a beaglepilot question? Apr 13 21:34:29 hi, so i accidentally put 5v into the i2c line on my BBB, (I think) and now i've got nothing but a single blue LED flashes once when i power on Apr 13 21:35:04 I understand that I "might not have blown the SOC" but i'm not sure how to go about repairing it Apr 13 21:35:18 or even what this behavior indicates Apr 13 21:37:15 Why does vi symlink to vim on the BBB but running vim yields `bash: vim: command not found` Apr 13 21:38:38 on what, debian? Apr 13 21:39:01 yes, just flashed the 4/09 jessie image Apr 13 21:39:15 does running vi work? Apr 13 21:39:29 i shouldn't say symlink since I didn't look. what i mean is running vi ends up running vim Apr 13 21:39:36 yes Apr 13 21:39:40 oh Apr 13 21:39:51 hmm I've seen that before Apr 13 21:40:50 endrift: in the past what was the issue Apr 13 21:41:08 well clearly /usr/bin/vim wasn't present Apr 13 21:41:13 I don't recall what the solution was though Apr 13 21:41:33 it happened on the 2/xx/15 console image as well Apr 13 21:41:41 when you run `file` on vi, what does it say? Apr 13 21:41:59 @arm:/opt/scripts/tools$ file vi Apr 13 21:42:00 vi: cannot open `vi' (No such file or directory) Apr 13 21:42:05 full path Apr 13 21:42:07 file `which vi` Apr 13 21:42:21 @arm:/opt/scripts/tools$ file `which vi` Apr 13 21:42:21 /usr/bin/vi: symbolic link to /etc/alternatives/vi Apr 13 21:42:27 interesting Apr 13 21:42:33 so it's an apt alternatives thing Apr 13 21:43:09 I don't know how those work; I haven't used debian enough Apr 13 21:43:09 and this is why don't use debian much :D Apr 13 21:43:26 My beaglebone runs arch haha Apr 13 21:46:58 also so this while i was running the update_kernel script Apr 13 21:47:00 saw* Apr 13 21:47:01 depmod: ERROR: Module 'hci_vhci' has devname (vhci) but lacks major and minor in formation. Ignoring. Apr 13 21:47:32 does that have any repercussions i should be aware of Apr 13 21:49:45 Hi, I have a BBB that has no power plug or USB, can power be provided over J1? Apr 13 21:55:13 to that end.. it would appear that running /opt/scripts/tools/update_kernel.sh on the latest debian image will brick network access Apr 13 21:57:16 so i was under the impression that the debian images were the newest/best/supported/whatever... where do you guys get your arch builds, or whatever Apr 13 21:59:53 trashguy: i don't think VCC is connected on the board anyway Apr 13 22:01:29 beest: since it has no power or usb port, on the pin out would i connect 5v to the 5v raw pin? Apr 13 22:19:00 hi, so i accidentally put 5v into the i2c line on my BBB, (I think) and now i've got nothing but a single blue LED flashes once when i power on Apr 13 22:19:03 I understand that I "might not have blown the SOC" but i'm not sure how to go about repairing it Apr 13 22:19:05 or even what this behavior indicates Apr 13 22:19:11 any thoughts? Apr 13 22:24:24 well, i2c can run at 5v, although i'm sure running it to VCC probably wasn't healthy Apr 13 22:24:28 did you try flashing it ? Apr 13 22:30:01 I am looking for some help regarding the default memory allocation done by loading the tilcdc driver. I am bumping into the limit when trying to increase my vyres for double buffering Apr 13 22:51:36 we are having trouble finding the point when the driver is loaded what values it is using to determine the amount of memory to allocate. It is allocating enough memory for the exact panel resolution and not enough to allow configuration of virtual resolutions greater than the panel to account for any double buffering Apr 13 22:52:41 This is being done using mainline 3.19.2 and the tilcdc driver. We have done through and attempting to configure the kernel with the proper setting outlined by ti Apr 14 00:15:47 treeherder / l4cr0ss: the BBB has absolutely _no_ 5V-tolerant I/Os Apr 14 00:16:31 treeherder: "blue led flashes briefly on power on" = BBB fried Apr 14 00:19:23 or more specifically it means the overcurrent protection of the power management IC tripped immediately after providing power to the processor Apr 14 00:20:41 well, the 5V input is 5V tolerant and it is an input... Apr 14 00:20:58 ds2: :P Apr 14 00:21:17 that's not really what people generally mean by I/Os Apr 14 00:21:55 got to keep an open mind... Apr 14 00:21:56 actually the am335x itself even has two 5V-tolerant pins: the USB Vbus inputs Apr 14 00:22:08 RESET is also an IO pin Apr 14 00:22:17 but I'll rephrase Apr 14 00:24:01 apart from pins 7 and 8 of P9, you may not apply a voltage greater than 3.3V on any pin of the expansion connector Apr 14 00:24:40 (+10% or something, see datasheet) Apr 14 00:25:06 trashguy: you can power the BBB via pins 7 and 8 of P9 Apr 14 00:25:22 trashguy: (not pins 5/6) Apr 14 00:26:57 the beagle paperweight is born once again :) Apr 14 00:26:58 ds2: in retrospect I retract my rephrased version, the original was fine :P Apr 14 00:27:10 zmatt +1 Apr 14 00:30:57 treeherder: to put things in perspective, the power supply lines of the processor's I/O banks are called "VDDHV" - High Voltage... because they may carry 3.3V Apr 14 00:31:12 uh... Apr 14 00:31:17 treeherder: internally stuff runs on 1.1 - 1.2V Apr 14 00:31:36 I'd rephase it as - do not exceed levels as described in the SRM. Apr 14 00:31:49 there are a few lines that cannot tolerate anything more then 1.8V Apr 14 00:32:29 even applications of 3.3V at the wrong time is fatal Apr 14 00:32:33 ds2: in fact, the GND lines don't really tolerate anything more than 0V. I gave an upper bound on tolerance, not a lower bound :P Apr 14 00:32:50 zmatt: -1 now :p Apr 14 00:32:51 lol Apr 14 00:33:17 zmatt .. you can have your GND wires at 3.3V as long as your VCC is above it .. Apr 14 00:33:30 I dispute that statement. GND is your reference Apr 14 00:33:35 that. Apr 14 00:36:31 ds2: the only point I was trying to make originally is that the BBB is really, really not 5V tolerant :P Apr 14 00:36:52 if you want 5v io .. get an arduino :p Apr 14 00:37:00 and use spi with level shifters :p Apr 14 00:37:03 and I am pointing out that with some creativity, you can get 5V tolerante input Apr 14 00:37:05 s Apr 14 00:37:20 or i2c .. whatever :P uart even! Apr 14 00:37:24 hence the reset example - reset works fine as a wake up input Apr 14 00:37:55 similar tricks can let you use the power input as a "wake up" signal Apr 14 00:38:23 ds2: I don't understand what you're trying to say Apr 14 00:38:46 ds2: how does that also perform output, and hence qualify as an input/output functionality? Apr 14 00:39:10 zmatt: for power management purposes, reset/power in can be (with some creativity) used as an input for wake up purposes Apr 14 00:39:16 not claiming output. just claiming input Apr 14 00:40:13 there are plenty of ways of interfacing 5V-stuff indirectly, but 5V will not be supplied nor tolerated by the BBB on any expansion pin other than the four 5V supply pins Apr 14 00:40:35 ds2: what does that have to do with 5V tolerance? Apr 14 00:40:47 zmatt: it takes 5V in Apr 14 00:41:19 general purpose IO on the other hand.... Apr 14 00:42:36 hell no, the reset pin connects straight to the am335x and is absolutely not 5V tolerant Apr 14 00:42:48 where'd you get that idea? Apr 14 00:42:56 you are confusing what I am saying Apr 14 00:43:08 nevermind Apr 14 00:43:42 come to think of it... it can be used as more then just a PM wake up Apr 14 00:43:54 back to my other problem Apr 14 00:44:38 pins 7/8 can indeed be used as 5V input to trigger an irq via the pmic, assuming you normally power your bbb via usb or battery Apr 14 00:44:53 bit of a scenic route to get a 5V tolerant input though :P Apr 14 00:45:18 I would think a level shifter might be safer ... Apr 14 00:45:25 also, there's the little issue that the BBB will probably try to draw more power from it than you'll like Apr 14 00:45:35 scenic yes... but when you have limited sources... Apr 14 00:46:07 IIRC, the PMIC can be configured to work around stuff like that Apr 14 00:46:34 if you have the space to do a 10high stack of capes, then ignore what I am saying Apr 14 00:47:16 GenTooMan: a bus-switch is also nice, works bidirectional (provided the 5V-side is okay with receiving 3.3V signals) Apr 14 00:48:08 I don't recommend clever use of IO unless you absolutely positively can't do it any other way. Apr 14 00:48:14 logic thresholds for 5V usualy manage 3v3 fine.. the reverse, however, doesn't hold true ! Apr 14 00:48:44 ds2: it's a creative workaround that might make sense in some cases, but I don't think it would make much sense in general :P Apr 14 00:48:45 any protection diodes you Might have .. tend to get quite Hot. Apr 14 00:48:53 veremit: not protection diodes Apr 14 00:49:09 veremit: there been situations where a large expense was needed to correct that :/ Apr 14 00:49:20 veremit: busswitch clamps without drawing current Apr 14 00:49:25 zmatt .. yes I'm aware there are none on the am335x not surprisingly Apr 14 00:49:49 veremit: well there are in the am335x, but just for esd Apr 14 00:49:58 that's why you use interface IC's and input protection devices. Apr 14 00:50:00 when in doubt, figure the worse case values based on the datasheet. anything else is foolish. Apr 14 00:50:11 GenTooMan +1 Apr 14 00:50:39 GenTooMan: yup, raw cpu pins of a modern SoC like the am335x are.... not very user friendly Apr 14 00:51:25 depends on the 'user' imho ;p Apr 14 00:51:48 5v pins aren't friendly if your IO takes place at 12 or 24v :P or is AC not DC .. Apr 14 00:52:24 Yeah well been their blown that up :D Apr 14 00:52:42 so I think the general lesson is .. Do Ya Homework :) Apr 14 00:52:45 Nothing like a bit of dull red glow in the morning :D Apr 14 00:52:50 hehe GenTooMan :) Apr 14 00:53:34 Anyhow DS2 think about how complicated your solution is first. What can go wrong with it etc. versus a simpler hardware solution. Apr 14 00:53:45 veremit: true, but still an arduino can obviously take much more of a beating while surviving Apr 14 00:54:29 zmatt ... precisely, so its a better board to start out playing with electronics .. and gpio especially Apr 14 00:54:31 Well most of those people are use to things "gracefully" taking a beating. heh.. Apr 14 00:54:55 GenTooMan: which *is* a form of a user-friendliness Apr 14 00:55:05 being able to afford a bit of carelessness Apr 14 00:55:06 indeed Apr 14 00:55:08 GenTooMan: for what I am thinking of, that is the simpliest way... almost anything else would be considerably more complex Apr 14 00:55:30 ds2 .. think we're all on a different planet :P~ Apr 14 00:55:41 I'm from saturn you? :D Apr 14 00:55:52 *cough* uranus .. *snicker* Apr 14 00:56:04 GenTooMan: if you need to wake up from a completely unpowered SoC, any inputs you use HAS to be qualified with reset. the power input avoids that entirely for one thing Apr 14 00:56:05 You know I've heard their are rings around ... Apr 14 00:56:15 hahahaha Apr 14 00:56:30 ... Jupiter, yes .. :p Apr 14 00:57:31 ds2 serious question HOW does the BBB do that now? doesn't it control that with the TPSXXXXXX ic? Apr 14 00:57:51 GenTooMan: yes but the TPSXXXXX can notify the AM335x Apr 14 00:58:25 GenTooMan: Consider a battery power device where you hibernate (i.e. memory dump to SD) Apr 14 00:58:48 K I have made those ... Apr 14 00:58:58 the TPSXXXXX is accessible over I2C Apr 14 00:59:11 but this does assume you have a battery Apr 14 00:59:53 another thorny thing is sensing battery state... using the onboard ADCs is not that trivial for that Apr 14 01:01:46 veremit: this is an example what I meant with a bus switch -> http://www.ti.com/lit/ds/symlink/sn74cb3t3245.pdf Apr 14 01:02:30 will have a look in a sec .. gotta restarr my router .. its playin up .. but ty Apr 14 01:02:47 I used a bunch of 74-series level shifters between some SoM and TFT displays before :p Apr 14 01:02:53 brb tho' Apr 14 01:03:23 GenTooMan: the power up of the am335x is actually fairly unusual Apr 14 01:03:58 GenTooMan: (mainly as a result of the design of rtc-only mode) Apr 14 01:04:26 which I still think is a bit weird... the older solution of including an RTC in the PMIC made more sense to me Apr 14 01:04:54 the SOT-23 single gate versions of those are nice Apr 14 01:05:23 veremit: level shifters are generally not bidirectional though, the thing I linked to is Apr 14 01:06:32 ds2: "if you need to wake up from a completely unpowered SoC" -> you mean using e.g. the power-button signal on P9 pin 9 ? :P Apr 14 01:07:12 zmatt: that works too but that has a built in pull up Apr 14 01:07:13 I thought they used the same thing on this apparently not. Apr 14 01:07:48 ds2: use an OC driver? Apr 14 01:08:00 ds2: it had better, since the normal power supplies are unavailable at such a time :P Apr 14 01:08:28 GenTooMan: the problem there is it presents a voltage to your sensors which may not necessarily be what it tolerates. simple switches are fine Apr 14 01:08:57 if it wasn't for that, a 430 would make a nice wake up controller Apr 14 01:09:08 How would an open collector present "unexpected voltages"? Unless it's leaking or shorted that might be different. Apr 14 01:09:26 open drain I should have said really :D Apr 14 01:09:49 GenTooMan: there is often a nasty ESD diode Apr 14 01:10:44 what's "a 430" ? Apr 14 01:10:51 MSP430 Apr 14 01:10:53 microcontroller Apr 14 01:10:55 ah Apr 14 01:10:57 ehm Apr 14 01:11:21 the 430's have a low voltage tolerance Apr 14 01:12:43 Really? Apr 14 01:12:45 power button pin is low voltage too Apr 14 01:12:51 * GenTooMan uses them all the time. Apr 14 01:13:03 (trying to find a concrete number...) Apr 14 01:13:42 is there an upper limit? didn't see one when I looked Apr 14 01:14:07 ds2: well you're not supposed to source any voltage into it anyhow Apr 14 01:14:15 If you pull down the PWR_BUT input you can turn the unit on. Furthermore voltage is irrelevant. Apr 14 01:14:34 ds2: and its pull-up isn't going to hurt your msp430 either Apr 14 01:14:47 zmatt: the MSP430 has an upper limit on it like the 335x Apr 14 01:14:55 3.x IIRC Apr 14 01:15:46 Which msp430? and what is the pullup resistance in the SoC? The msp430F series has a range of 25k to 75k for the pull up and pull down resistance. Apr 14 01:16:11 the new NFC ones Apr 14 01:16:36 ds2: even if the pmic would try to pull it up to 5V, nothing will happen: the msp's protection diode will limit it, and nothing bad will happen because the max current supplied by the pmic is really really really tiny Apr 14 01:16:47 max 1 μA, typ 10 nA Apr 14 01:17:18 zmatt: I don't think TI documents the max current through the ESD diode Apr 14 01:18:09 ds2: no, it's kind of implicit: they spec a max voltage, but you can't reach that voltage if you can't supply enough current Apr 14 01:18:29 zmatt: you going by the abs max rating? Apr 14 01:19:15 ds2: I'm not really going by any rating... anything that has a protection diode is not going to be hurt by a line having an insanely weak pull-up, even if it's to a too-high voltage Apr 14 01:19:42 since the diode prevents actual overvoltage, and the weakness of the pull-up prevents significant current Apr 14 01:20:12 if a protection diode is absent, then things are different Apr 14 01:21:30 zmatt: the concern I'd have is - will that trip the scr and latch up and will they stand behind it with a value that doesn't trip it Apr 14 01:21:54 you can't latch up anything with near-zero current Apr 14 01:22:17 (TI usually does specify latch-up current btw, at least for their SoCs) Apr 14 01:23:13 sounds worthwhile to revisit Apr 14 01:23:37 If you are worried about it use something like SN74AUP1G07 to buffer the msp430 output? Apr 14 01:24:28 finally it timed out lol Apr 14 01:25:07 GenTooMan: actually a 2N2222 or a logic level FET would work just fine for that but that's another component that takes space Apr 14 01:25:40 ds2: for example, am335x latch-up performance at 105 °C is min 45 mA Apr 14 01:27:15 ds2: dunno what the msp430 can take, but I'd suspect it can probably take the max 1 uA from the pmic :P Apr 14 01:27:40 <--- conservative. will design by specs Apr 14 01:28:02 clamp it with a diode yourself Apr 14 01:28:35 RF430FRL series is so weird.. Apr 14 01:28:40 might as well use a logiclevel FET Apr 14 01:29:00 GenTooMan: weird yes. but there is so many potential uses for that Apr 14 01:30:02 omg this is the best ident Apr 14 01:30:07 ^5 snowy Apr 14 01:30:30 !song Apr 14 01:30:33 that what you are playing with? Apr 14 01:30:40 ~a friend with weed is better....~ Apr 14 01:30:49 ooops .. wrong channel rofl Apr 14 01:30:55 sorry Apr 14 01:31:02 ds2: I'd check the ds for latch-up performance first, or ask e2e... I really do not think the pmic's pull-up current poses any threat whatsoever here Apr 14 01:31:02 that went up in smoke. Apr 14 01:33:55 GenTooMan: btw, you wanted to know how exactly power up/down works on the BBB ? Apr 14 01:35:19 hmm yeah curious Apr 14 01:35:34 ds2 the RF430 IO is rated to 1.65V tops Apr 14 01:36:14 hmmm? Apr 14 01:36:22 let me look...that doesn't sound right Apr 14 01:37:00 The antenae is rated to 3.6 but the io is 1.95 (0.3 + Vddb which is 1.65) Apr 14 01:38:00 read section 5.1 Apr 14 01:38:09 Those are of course MAX. Apr 14 01:38:48 GenTooMan: initially when power is applied the TPS will initialize to "OFF" mode. however, it also sets an irq-bit such as "usb power detected" Apr 14 01:39:35 GenTooMan: this is by default a wakeup source, so it enters power-up sequence Apr 14 01:40:14 zmatt hmm so when you power down it shutdown things in an orderly manner and waits for the the wakeup from the button or power cycle? Apr 14 01:40:30 initially it only powers the VRTC of the am335x Apr 14 01:40:58 let's not skip ahead :P Apr 14 01:41:41 there are 2 chips Apr 14 01:41:46 one can do the 3ish volts Apr 14 01:42:08 tps talks to the cortex-mZ doesn't it? Apr 14 01:42:11 the RTC of the am335x is a completely isolated power domain Apr 14 01:42:33 when its power is stable the tps releases it from reset Apr 14 01:42:36 zmatt .. whats the current consumption of the rtc stage? Apr 14 01:43:26 it, by default, asserts an output that requests the TPS to initiate main power on Apr 14 01:43:43 which triggers the whole-soc power up sequence Apr 14 01:43:48 veremit: more than it should :P Apr 14 01:44:28 veremit: a dedicated RTC on a backup battery will consume a miniscule fraction of the power consumed by the am335x rtc Apr 14 01:45:05 zmatt .. yeah that seems to be a universal problem with SoM boards :? Apr 14 01:45:16 GenTooMan: note that at this point the irq isn't cleared yet Apr 14 01:45:51 GenTooMan: this is easy to see: power on with sd button but no sd card to prevent booting, then long-press the power button to force power-off Apr 14 01:45:55 it'll instantly power on again Apr 14 01:46:10 because the irq that originally triggered power on is still set Apr 14 01:46:40 huh ... that seems like it could be a bug or a feature depending on how you can annoy people with it :D Apr 14 01:47:02 interrupts have to be cleared normally or acknowledged .. Apr 14 01:47:06 ds2: I'm confused what parts your use using? RF430FRL or some other NFC 430 IC? Apr 14 01:48:01 power off is accomplished by speaking to the tps to clear (or disable) the irqs, then request the tps to enter OFF mode or SLEEP mode Apr 14 01:48:24 the latter keeps the RTC on Apr 14 01:48:54 the RTC can optionally be configured to drive the power-on request in response to the configured RTC alarm, allowing timed power-on Apr 14 01:49:33 the wakeup cortex-m3 is a different story, it's not really involved in any of the above Apr 14 01:50:30 The msp430 you use the sleep mode which leaves RTC the 32khz oscillator running. heh yet another thing to make it complicated :D Apr 14 01:50:47 it is held in reset initially since it needs the cortex-a8 to load firmware into its memory Apr 14 01:51:08 micro boot strap? Apr 14 01:52:12 the m3 is needed to assist in partial power-down modes of the SoC Apr 14 01:52:31 a full power-down doesn't need to be particularly orderly since all state is going to be lost anyhow Apr 14 01:52:42 so the cortex-a8 can do that without help Apr 14 01:53:03 same goes for rtc-only mode Apr 14 01:54:26 however "deep sleep" is a different matter: there the wakeup domain of the SoC stays powered on, which means a proper power-down sequence needs to be coordinated with the PRCM module Apr 14 01:55:35 this has been a source of headache in the past, and deepsleep has for example been defeatured entirely in the dm81xx parts due to problems getting it to work right Apr 14 01:56:05 so for the am335x TI came up with a neat solution involving the wakeup-M3 Apr 14 01:59:54 normally once the M3 firmware is initialized, it enters WFI until further notice. when the cortex-a8 wants to enter sleep, it leaves the cortex-m3 some final instructions w.r.t. wakeup sources, sets the "module mode" of its own clock-control register to "disable", makes sure all memory transactions have completed, and executes WFI Apr 14 02:00:39 changing the module mode causes PRCM to assert the idle-request to the Cortex-A8 subsystem Apr 14 02:01:31 once WFI is entered and the processor is fully quiescient and internally clock-gated, it will acknowledge the idle-request to PRCM Apr 14 02:02:02 * veremit makes notes Apr 14 02:03:35 as a corner case: if a wakeup irq source occurred in the mean time, it'll fall right through the WFI and proceed with the resume sequence (starting with settings its module mode back to enabled) Apr 14 02:04:37 however, if all goes well, then once PRCM gets confirmation the cortex-a8 has entered idle state it'll send an irq to the wakeup-m3 Apr 14 02:05:56 zmatt .. so this allows you to 'startup' from the power switch like acpi ? Apr 14 02:06:26 ie . its powered.. but 'off' Apr 14 02:06:55 the m3 can now turn off power based on the final instructions left by the cortex-a8: since the m3 is in the always-on wakeup domain it can freely reign over all remaining power controls of the chip Apr 14 02:07:10 which parts are powered off depends a bit on the situation Apr 14 02:07:41 (such as which peripherals are supposed to be able to wake up the device from sleep) Apr 14 02:09:13 when a wakeup source is triggered, the m3 powers stuff up again, re-enables the cortex-a8 (which continues the resume sequence) and enters WFI to pick its nose until needed again by the cortex-a8 Apr 14 02:09:18 GenTooMan: RF430CL330 Apr 14 02:10:02 veremit: this scheme is for sleep, not "off" Apr 14 02:10:24 mkay Apr 14 02:10:36 veremit: ddr3 is put into self-refresh mode, the pmic remains in ON mode and maintains all supplies Apr 14 02:10:53 but internally in the am335x a very large part of the SoC can be completely powered off Apr 14 02:11:26 the deepest sleep modes shut off all power domains except the wakeup domain, disables all PLLs *and* the main oscillator Apr 14 02:13:19 veremit: powering on the power button is handled by the pmic itself: assumes the button is configures as wakeup source in the PMIC (it is by default), it is monitored by logic in the always-on domain of the PMIC Apr 14 02:13:30 pressing it triggers an irq Apr 14 02:13:47 I see Apr 14 02:13:55 and from there things are the same as initial power up Apr 14 02:15:43 the button and the two "external" power supply sources are the only wakeup events the PMIC has btw Apr 14 02:19:43 deepest sleep mode consumes 3 mW typical according to the datasheet btw (obviously excluding power drawn by external components like the ddr3 memory) Apr 14 02:22:53 DS2: use INTO for your purposes on the PWR_BUT that should do what you want without anything special read 5.7.1 of the RF430CL data sheet. Apr 14 02:24:15 wakeup sources available in sleep are gpio0, timer1, usb resume signalling, touchscreen pen-down, uart0, i2c0, and rtc alarm Apr 14 02:24:24 3mw is a bit much of power then again it's a big IC good night gentleman. Apr 14 02:24:47 3 mW is not much power for a PC in sleep ;P Apr 14 02:24:52 (I think?) Apr 14 02:25:22 Hmmm I thought you were talking a BBB not a PC :D Apr 14 02:25:30 well, small-ish PC then Apr 14 02:25:55 the am335x ain't a microcontroller Apr 14 02:27:01 nor is it an omap (which typically have insanely fine-grained power management but therefore also require a big investment of effort into power management to reap the benefits) Apr 14 02:33:28 the am335x is not a true low power processor Apr 14 02:37:12 it is intended to be usable for portable applications though, and they clearly put some thought into power management Apr 14 02:37:20 I guess "low power" is kinda relative Apr 14 02:39:15 compared to x86.... Apr 14 02:39:22 ;) Apr 14 02:40:48 I honestly don't really have any good reference points... like I have no idea how long the am335x would last in sleep mode when using a standard li-ion cell Apr 14 02:41:09 and I honestly don't really care enough to look it up either ;) Apr 14 02:43:09 it depends on how you implement it Apr 14 02:43:42 yup, you'd need to pay attention to all components Apr 14 02:44:58 "The battery charger will remain active for as long as either USB or AC supply is connected to the device. Please note that all register values are reset as the device enters in SLEEP state, including charger parameters." Apr 14 02:45:03 ... Apr 14 02:46:19 sometimes I really have no idea what designers are thinking Apr 14 02:46:28 (or smoking) Apr 14 02:46:41 they probally slapped together something based on blocks they have on hand Apr 14 02:48:09 tbh the whole am335x looks like it has been done by an intern or something, either that or they were really in a hurry Apr 14 02:49:10 lot of things are really messy compared to previous SoCs Apr 14 02:50:41 probably done to a budget set by a handset/STB manufacturer/etc Apr 14 02:50:49 so its what could be "lashed together" Apr 14 02:51:49 STB ? Apr 14 02:52:02 set-top-box Apr 14 02:52:33 euh, the am335x is not for either of those applications Apr 14 02:52:47 true Apr 14 02:52:54 but it must ahve come from somewhere .. lol Apr 14 02:52:56 have* Apr 14 02:53:45 industrial applications Apr 14 02:54:04 one of its "feature" applications is EtherCAT Apr 14 02:55:20 ick yeah our servo drives have that as one option Apr 14 02:55:27 we use straight tcp though :) Apr 14 02:58:27 basically the only thing that's really new in the am335x is the new PRUSS version that's EtherCAT-capable **** ENDING LOGGING AT Tue Apr 14 02:59:58 2015