| 00:02.27 | *** join/#edev prpplague (n=dave@72.22.147.70) |
| 00:03.44 | prpplague | ho ho ho |
| 00:18.03 | prpplague | MonMotha: ping |
| 00:27.35 | *** join/#edev ohsix (i=ohsix@bc111222.bendcable.com) |
| 00:57.59 | chouimat | lalalalalalalalalalalalalala |
| 00:59.14 | MonMotha | prpplague: pong |
| 01:08.44 | *** join/#edev davidc___ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 01:10.57 | *** join/#edev youam (n=youam@ciara.youam.de) |
| 01:41.59 | chouimat | night |
| 02:56.49 | *** join/#edev gandhijee (n=gandhije@c-24-147-213-182.hsd1.ma.comcast.net) |
| 03:57.47 | *** join/#edev EvilDevil_ (n=miau@p54A6D6D2.dip.t-dialin.net) |
| 04:48.46 | *** join/#edev chouimat (n=dieu@kde/developer/chouinard) |
| 05:58.47 | davidc__ | anyone know a bunch about jtag in here? |
| 06:16.41 | *** join/#edev redblue (i=star@ppp040.108-253-207.mtl.mt.videotron.ca) |
| 07:05.39 | *** join/#edev leoncamel (n=leoncame@219.238.107.107) |
| 07:23.25 | T0mW | davidc__: some |
| 07:23.46 | T0mW | davidc__: what are you looking to know? |
| 07:24.53 | davidc__ | TomW: eh, just fighting with this board trying how to get one device to bypass |
| 07:25.06 | davidc__ | I've hacked a jtag port onto an ipod and I'm trying to debug the arm cores :/ |
| 07:25.19 | davidc__ | [by hacking a jtag port on, I mean finding the right points and soldering ;)] |
| 07:25.26 | T0mW | oh |
| 07:25.34 | davidc__ | however, the things got 2 cores inside, and I need to bypass one of em |
| 07:26.15 | T0mW | Well, depends on the jtag unit you are using. I debug two ARM7TDMI-S cores that are daisy chained together on the jtag. |
| 07:26.46 | davidc__ | eh, this is just a dumb paralell port cable |
| 07:26.54 | davidc__ | I think my issue is getting the SW to do what I want it to |
| 07:26.56 | T0mW | For that, I have to specify the length of each chain. Then I select which device I'm working with. |
| 07:27.09 | T0mW | ah, I have a BDI2000 |
| 07:30.25 | T0mW | davidc__: when daisy chaining multiple cores, they chain the TDI & TDO signals. All other signals are fed from the same source (parallel fanout). |
| 07:30.42 | davidc__ | T0mW: yup, I've got that |
| 07:30.47 | T0mW | ok |
| 07:30.48 | davidc__ | T0mW: and I can see both cores in the chain |
| 07:30.55 | T0mW | good |
| 07:30.58 | davidc__ | aka, their 32 bit IDCODE regs |
| 07:31.08 | T0mW | davidc__: this is good |
| 07:31.23 | T0mW | you know the lenght of each chain then? |
| 07:31.34 | davidc__ | sorry, I'm lost - isn't it just one chain? |
| 07:31.38 | davidc__ | that runs serially through both? |
| 07:31.49 | davidc__ | [I'm learning jtag along with this hack, so excuse my ignorance :P] |
| 07:32.09 | T0mW | technically it is two chains cascaded. each device has a specific sized scan chain. |
| 07:32.21 | davidc__ | ah, k |
| 07:32.27 | davidc__ | sized as in the size of the IR for each? |
| 07:32.29 | davidc__ | its 4 IIRC |
| 07:33.04 | T0mW | overall, it would appear as if it were one chain, but the jtag driver needs to break it into two logical segments to control one or the other device. to do that, it needs to know where one chain ends and the other begins. |
| 07:33.24 | T0mW | not 4 |
| 07:35.50 | T0mW | davidc__: sorry, you are correct |
| 07:35.52 | davidc__ | you sure? all the arm docs say the IR length is 4 for an arm7/9 core? |
| 07:35.55 | davidc__ | ah, k, nm |
| 07:36.08 | davidc__ | and I see my total IR length as 8, which makes sense |
| 07:36.59 | T0mW | davidc__: http://rafb.net/paste/results/gS0Xek22.html |
| 07:37.42 | T0mW | bah, some of my file is truncated |
| 07:38.15 | davidc__ | eh, no worries |
| 07:38.44 | T0mW | http://rafb.net/paste/results/nEIMsO72.html |
| 07:38.58 | davidc__ | I think I've got enough sorted now to try firing up a debugger |
| 07:38.59 | T0mW | that worked |
| 07:39.51 | T0mW | davidc__: on the BDI2000, the first CPU is at port 2001, the second processor I have at port 2002, then I just run two instances of Insight and debug both. |
| 07:40.17 | davidc__ | cool - I'm gonna try using an OpenOCD thing - it supposedly works with my cable [which is just a nasty bit-bang] |
| 07:41.03 | T0mW | davidc__: that may work ok |
| 07:41.19 | davidc__ | it theoretically should - I'm still at the proof of concept stage anyhow |
| 07:41.43 | davidc__ | nobody's hacked a port onto one of these suckers yet [cept apple] - so even seeing it with a debugger will be a step forward |
| 07:41.45 | T0mW | davidc__: I went with a BDI2000 as #1) I had the money at the time, #2) I didn't want to fuss with bit-bangers |
| 07:41.53 | davidc__ | how much did it cost ya? |
| 07:41.59 | T0mW | $2800 |
| 07:42.19 | davidc__ | ouch |
| 07:42.27 | T0mW | naw, it is an investment |
| 07:42.48 | T0mW | davidc__: I can always sell it on ebay, heh |
| 07:43.52 | davidc__ | heh... I think I'll stick with my slow bit banger for now |
| 07:43.53 | T0mW | davidc__: but, this way, I don't have to worry if there is software for linux to run a JTAG debug |
| 07:44.11 | davidc__ | yea, course - commercial stuff means support |
| 07:44.31 | T0mW | for a price, I took a pass on the support: $400 yearly |
| 07:44.52 | davidc__ | ouch |
| 07:45.11 | davidc__ | eh, maybe I'll get work to buy me one.. I can play with it there, see if its worth gettin my own |
| 07:45.55 | T0mW | davidc__: they tried to sell me a nifty, cool, Windows XP based IDE development environ: compiler, debugger software. All for the low low price of $4k. |
| 07:46.07 | davidc__ | hehe |
| 07:46.19 | davidc__ | eh, I've given up on proprietary IDE's - the ones I've seen so far are all crap |
| 07:46.20 | T0mW | lol, I already knew about Insight |
| 07:46.26 | davidc__ | gvim + make + insight do it for me now |
| 07:46.32 | davidc__ | eclipse if I really wanna be pampered |
| 07:46.33 | T0mW | same here |
| 07:46.55 | T0mW | xvim + bintools + gcc + make + perl + insight |
| 07:47.43 | davidc__ | eh, I'd use xvim if I could get away with a linux dev machine |
| 07:48.00 | T0mW | it's going to be 10 years on linux for me the end of this year. I've not regretted the decision to scrap windows. |
| 07:48.00 | davidc__ | the damn vendors gcc toolchain won't build under linux ... blah |
| 07:48.08 | davidc__ | oh, I use linux all the time - at home + etc |
| 07:48.22 | davidc__ | also - the Keil 8051 stuff doesn't do linux either |
| 07:48.25 | T0mW | I roll my own cross dev |
| 07:48.49 | T0mW | right now, the projects are the LPC2000 series from Philips |
| 07:48.54 | davidc__ | eh, I'd build my own toolchain - and they provided the gcc sources to their compiler - its not in the FSF tree yet - but it doesn't build on anything non win :/ |
| 07:49.04 | T0mW | lol |
| 07:49.18 | T0mW | ooo! ooo! cywin! <sigh> |
| 07:49.23 | davidc__ | ya, exactly |
| 07:50.02 | T0mW | the more I moved from MSDOS / Windows, the more productive I got. First productive thing was to get rid of the mouse |
| 07:50.23 | davidc__ | heh - yeah.. I don't think I could do without a mouse - but yes - I definitely noticed that as well |
| 07:50.24 | T0mW | ... and learn to touch-type. :-/ |
| 07:50.32 | davidc__ | used linux for a previous project |
| 07:50.56 | davidc__ | eh, I've got my own crazy typing system - its really insane apparently but its pretty good for programming stuff |
| 07:51.21 | davidc__ | waaay more productive then windows being strange - and I was running a bunch of the SW in wine :P |
| 07:51.54 | T0mW | I was two / three fingers for years. Then I forced myself to re-learn, I am glad I did that as I now longer have to stare at the keyboard while I type. heheh |
| 07:52.20 | davidc__ | eh, I use most of the fingers on both hands and dont have to look at the keyboard |
| 07:52.29 | davidc__ | but its definitely not something that was ever taught in a school |
| 07:52.35 | T0mW | ok, time for some sleep for me. cul |
| 07:52.39 | davidc__ | heh, later |
| 07:52.43 | *** part/#edev T0mW (n=Tom@24.238.68.20.res-cmts.sth.ptd.net) |
| 08:59.57 | *** join/#edev bd2 (n=ufo@81.222.177.146) |
| 11:53.29 | *** join/#edev [1]gandhijee (n=gandhije@c-24-147-213-182.hsd1.ma.comcast.net) |
| 12:08.21 | *** join/#edev chewy (n=dieu@r2351064.cidc.net) |
| 12:13.52 | *** join/#edev GPSFan (n=kenm@12.10.255.248) |
| 12:13.52 | *** mode/#edev [+o GPSFan] by ChanServ |
| 12:19.49 | *** join/#edev Pavlov_ (i=avictor@maxim.org.za) |
| 12:35.10 | *** join/#edev gandhijee (n=gandhije@mail.win-ent.com) |
| 12:43.14 | *** join/#edev prpplague (n=billybob@72.22.146.214) |
| 12:53.21 | *** join/#edev tuxmaniac (n=aanjhan@60.254.67.17) |
| 14:01.35 | *** join/#edev tuxmaniac (n=aanjhan@60.254.67.17) |
| 14:35.43 | *** mode/#edev [+o prpplague] by ChanServ |
| 15:43.54 | prpplague | http://money.cnn.com/2006/06/27/news/companies/intel.reut/index.htm |
| 15:44.04 | prpplague | marvell to buy intel communications division |
| 15:54.09 | *** part/#edev Pavlov_ (i=avictor@maxim.org.za) |
| 15:55.13 | *** join/#edev gandhijee (n=gandhije@mail.win-ent.com) |
| 16:04.32 | *** join/#edev dfarnsworth (n=dfarnswo@65.200.49.156) |
| 16:08.27 | jbevren | there goes any linux support for the intel wireless cards |
| 16:08.28 | jbevren | :P |
| 16:08.32 | prpplague | yea |
| 16:08.45 | prpplague | jbevren: thats what i'm thinking |
| 16:09.38 | dfarnsworth | the sale doesn't include intel's wireless networking chips does it? |
| 16:09.55 | jbevren | isnt that part of the communications div? |
| 16:09.58 | jbevren | I dunno |
| 16:10.21 | prpplague | dfarnsworth: well it doesn't specifically say, other than communications division |
| 16:10.36 | dfarnsworth | I don't know either, but I got the impression that they just sold the xcale chips used in pdas and cell phones. |
| 16:10.43 | prpplague | dfarnsworth: which intel's web page lists their flash, processor, and wireless stuff under |
| 16:11.05 | dfarnsworth | I certainly could be mistaken. :) |
| 16:11.24 | prpplague | dfarnsworth: as could i |
| 16:11.55 | dfarnsworth | Marvel's no fun to work with |
| 16:12.17 | prpplague | http://www.thhttp://www.theinquirer.net/?article=32670einquirer.net/?article=32670 |
| 16:12.33 | prpplague | looks like intel _is_ keeping the wireless stuff |
| 16:12.38 | prpplague | http://www.theinquirer.net/?article=32670einquirer.net/?article=32670 |
| 16:12.42 | jbevren | yay |
| 16:37.05 | jbevren | bp; |
| 17:03.48 | *** join/#edev bd2 (n=ufo@81.222.177.146) |
| 17:41.43 | *** join/#edev bd2 (n=ufo@81.222.177.146) |
| 17:42.16 | *** join/#edev T0mW (n=Tom@24.238.68.20.res-cmts.sth.ptd.net) |
| 17:50.07 | prpplague | T0mW: hey hey |
| 17:50.39 | prpplague | T0mW: got a question for ya |
| 17:50.53 | T0mW | go |
| 17:50.56 | prpplague | T0mW: http://download.intel.com/design/pca/applicationsprocessors/manuals/27869401.pdf |
| 17:50.59 | prpplague | T0mW: grab that |
| 17:51.07 | prpplague | T0mW: look at page 55 section 5-1 |
| 17:51.07 | prpplague | 4-1 |
| 17:51.10 | prpplague | sorry 4-1 |
| 17:52.21 | T0mW | ok |
| 17:52.26 | prpplague | hehe |
| 17:52.29 | prpplague | T0mW: got it? |
| 17:52.37 | T0mW | yeah |
| 17:53.05 | prpplague | T0mW: ok so do what do you think about how they are suggesting to do the +5v power detection? |
| 17:53.22 | *** join/#edev davidc___ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 17:53.49 | prpplague | in the sea and on plates |
| 17:56.40 | T0mW | from what I read, they are saying that a level shifter must be used to knock the signal level down from +5 to +3.3v, from what I see, they are simply detecting the presence of the externally applied +5v (USB power) to determine if you are connected to an external device / host. |
| 17:57.04 | prpplague | T0mW: yep |
| 17:57.09 | prpplague | T0mW: right on target |
| 17:57.30 | prpplague | T0mW: question is, what method would you use to knock it down from +5v to +3.3v |
| 17:57.42 | T0mW | since the USB spec calls for +5v power to be supplied from the host, -and-, the PXA does not need / use that supply, you still need to detect the presence of the external host via the application of the +5v |
| 17:57.46 | prpplague | T0mW: with regards to bounce and noise from the host |
| 17:58.03 | prpplague | T0mW: yea |
| 18:01.12 | jbevren | just use an rc circuit to debounce the line |
| 18:01.13 | jbevren | hehe |
| 18:01.15 | jbevren | <- lazy |
| 18:01.48 | chouimat | hi everybody |
| 18:02.08 | T0mW | actually, that 470k resistor, IMO, should go to +3.3v not to GND as shown. The MAX6348 is open collector. |
| 18:02.52 | T0mW | um, that 470k is prolly a bleed-down load resistor... |
| 18:03.39 | prpplague | T0mW: so would you follow their suggestion? |
| 18:03.43 | T0mW | the "+5v to +3.3v" block doesn't show the electronics, so, there is prolly a pullup to the +3.3v inside that block |
| 18:04.03 | prpplague | T0mW: yea its very vague |
| 18:05.32 | T0mW | yeah, you are prolly going to use an interrupt edge to detect the presence of the connection to the host. That signal may bounce a few times as the cable is inserted. You can use a cheaper interface but debounce the pin in software. |
| 18:05.43 | prpplague | T0mW: yea |
| 18:05.46 | prpplague | T0mW: funny thing |
| 18:05.53 | T0mW | the idea is this: 1) detect the +5v presence. 2) avoid applying +5v to a non-tolerant pin (+3.3v max input). 3) debounce the sense circuit |
| 18:06.08 | prpplague | T0mW: sharp, samsung, intel all recommend using this method of detecting the usb slave connection |
| 18:06.18 | prpplague | T0mW: not one of them implement it in their dev boards |
| 18:06.22 | T0mW | for that matter, you probably could get away with using two resistors |
| 18:06.34 | prpplague | T0mW: yea |
| 18:06.47 | prpplague | T0mW: i was thinking two resistors and feeding it through a buffer |
| 18:06.48 | T0mW | yeah, well, "do as I say, not as I do" |
| 18:07.06 | prpplague | T0mW: we have some spare ahc04 lines |
| 18:07.35 | T0mW | or, use two resistors and a clamping diode. The cable may introduce some ringing when you attach it. |
| 18:07.55 | T0mW | prpplague: ever work with Transils? |
| 18:08.07 | T0mW | Transil is a very fast Zenor diode |
| 18:08.56 | T0mW | you would use a uni-directional transil to clamp anything below -0.3v (undershoot below GND) and clamps anything above +3.3v |
| 18:09.15 | T0mW | then debounce the GPIO input via software |
| 18:09.52 | T0mW | take a 1K and put it in series with the +5v & GPIO |
| 18:10.10 | T0mW | then take a 1k and put sink the GPIO to GND |
| 18:10.31 | T0mW | that will give a 1/2 divider |
| 18:10.44 | T0mW | Then, place a transil from the GPIO pin to GND |
| 18:10.53 | prpplague | T0mW: yea thats was another suggestion |
| 18:11.28 | T0mW | cheaper solution is to use two 1N4148 diodes instead of the transil |
| 18:11.32 | prpplague | T0mW: i was just curious what you thought |
| 18:12.07 | T0mW | tie the diodes end to end: cathode to anode so you get a 2x 1N4148, ok? |
| 18:12.18 | prpplague | got it |
| 18:12.32 | T0mW | -->|---->|--- |
| 18:12.34 | prpplague | T0mW: i'm sure my boss will do something, but i was trying to educate myself, hehe |
| 18:13.03 | T0mW | take the diode string an put it across the +3.3v and GND near the GPIO pin |
| 18:13.17 | T0mW | ANODE to +3.3v, CATH to GND |
| 18:13.30 | T0mW | nope |
| 18:13.50 | prpplague | T0mW: ?? |
| 18:13.54 | T0mW | GND-->|----GPIO--->|---+3.3v |
| 18:14.14 | prpplague | T0mW: ahh |
| 18:14.19 | T0mW | any voltage > 3.3v on the GPIO will be shunted into the power supply. |
| 18:14.39 | T0mW | any voltage less than 0v will also be sent into the power supply |
| 18:15.06 | T0mW | the power supply has capacitance, low impedance, and will "snuff" the excess energy |
| 18:15.11 | jbevren | actually, any voltage > 4v (silicon diode) will be shunted ;-) |
| 18:15.17 | prpplague | T0mW: by power supply you mean the 5v input from the usb? |
| 18:15.37 | T0mW | jbevren: good point, yes, silicon is typically 0.3v drop under no load |
| 18:15.58 | jbevren | silicon is .7, germanium .5 |
| 18:16.10 | T0mW | the more load you put on silicon diodes, the greater the voltage drop: e.g. 0.6v |
| 18:16.12 | jbevren | at any rate, most 3.3v gpio's can handle 4v |
| 18:16.27 | jbevren | tom: yeah, no-load. It doesnt take much load to get to .7 |
| 18:16.36 | T0mW | correct |
| 18:16.41 | jbevren | anyway |
| 18:16.43 | jbevren | point made :) |
| 18:17.14 | T0mW | the 1N4148 will handle a few milliwatts (joules) of power |
| 18:17.51 | T0mW | transils are designed to handle WATTS of power (watt seconds) |
| 18:18.34 | T0mW | 1n4148's are prolly 9cents each and a 3.3v transil is about 40 cents |
| 18:19.31 | T0mW | DON"T fsck with MOVs, they decay chemically and die, they also are not as fast in attacking spikes as the silicon clamps |
| 18:19.57 | jbevren | http://www.techeblog.com/index.php/tech-gadget/smart-laser-scanner |
| 18:20.10 | T0mW | but, MOVs handle WATTS of power and are cheaper than silicon: poor man's lightning protection |
| 18:21.20 | jbevren | prpplague: Laser trackign of objects in 3D space |
| 18:22.01 | T0mW | yeah, like I'd walk into a room filled with low power laser light... |
| 18:22.16 | jbevren | tom: the laser is spread through a 1cm plane |
| 18:22.24 | jbevren | er, 1cm thick plane |
| 18:23.04 | prpplague | that rocks |
| 18:25.04 | prpplague | T0mW: thanks for the info |
| 18:25.16 | T0mW | np |
| 18:26.27 | T0mW | prpplague: BTW, check the maximum ratings on the GPIO to ensure it will tolerate 3.9v and -0.6v before choosing the 1N4148. Otherwise use a transil |
| 18:27.00 | prpplague | T0mW: yea, these aren't that tolerant |
| 18:38.30 | *** join/#edev redb1ue (i=star@ppp203.108-253-207.mtl.mt.videotron.ca) |
| 18:40.39 | *** join/#edev ctaylor (n=ctaylor@static-70-106-125-178.pskn.east.verizon.net) |
| 18:42.22 | *** join/#edev redb1ue (i=star@ppp189.109-253-207.mtl.mt.videotron.ca) |
| 18:44.49 | prpplague | dfarnsworth: i can't remember, are you still with MV? |
| 19:00.40 | *** join/#edev bd2 (n=ufo@81.222.177.146) |
| 19:10.31 | chouimat | hmmm ok this is really bad |
| 19:22.04 | *** join/#edev bioeng (n=bioengin@68.216.132.14) |
| 19:22.05 | bioeng | Hi |
| 19:22.09 | bioeng | Is anyone around? |
| 19:22.30 | youam | nope |
| 19:22.51 | bioeng | But you are, clearly |
| 19:23.18 | youam | no, i'm just a bot |
| 19:23.32 | bioeng | No you aren't |
| 19:23.40 | youam | sure i am |
| 19:24.06 | bioeng | lol |
| 19:28.53 | *** part/#edev bioeng (n=bioengin@68.216.132.14) |
| 19:54.36 | *** join/#edev chewy (n=dieu@r2351064.cidc.net) |
| 20:38.34 | *** join/#edev bd2 (n=ufo@81.222.177.146) |
| 21:37.03 | dfarnsworth | prpplague, yes, I'm at MV |
| 21:38.36 | prpplague | dfarnsworth: ahh |
| 22:21.40 | *** join/#edev gandhijee (n=gandhije@c-24-147-213-182.hsd1.ma.comcast.net) |