Back
[00:00:05] <theBear> i think tom used a cheap pt100 himself
[00:00:15] <Valen> theBear: yeah, but i have like 6 projects happening now
[00:00:17] <braincracker> hot-air reflow is diy-able with a simple RTD or thermocouple
[00:00:29] <Valen> and one of them is actually ready for production
[00:00:31] <braincracker> IR reflow requires sensing magic
[00:00:49] <Valen> presently we are sitting a chunk of Al on the gas stove
[00:00:57] <theBear> Valen, ahh, that tightens the reins a bit
[00:01:11] <Valen> but yeah, not so good for double sided pcb either
[00:01:28] <Valen> the ebay reflow ovens look to only do one side and they sound like they are crap
[00:01:30] <theBear> Valen, maybe if yer nice to tom or leeloo or someone (and got a avr or a msp430 (from memory)) you can just use theirs more or less program, plug and play
[00:01:42] <braincracker> Valen <= well the third type is an industrial hotplate, it is fine for reflowing power leds on MCPCBs
[00:01:46] <braincracker> i have one too
[00:02:02] <theBear> well, thing with ovens is they tend to circulate at least a bit, with normal profiles even a single heater is likely to have fairly even top and bottom temps
[00:02:10] <theBear> normal reflow profiles aren't THAT fast
[00:02:15] <N1njaneer> If you are building a reflow oven from scratch, I would reccomend floating the thermocouple in space in the middle of the oven near the board.
[00:02:29] <Valen> we looked at that, even have a pair of flat plate elements we can use
[00:02:35] <Valen> but making the oven is easy for me
[00:02:41] <Valen> IE another person is doing it ;->
[00:02:50] <theBear> mmm, it came up the other night, i'm seriously considering various approaches involving free salvaged electric stove (curly kind that are happy to go low or glow red) elements
[00:02:54] <Valen> I just need something that will run a temp profile
[00:02:59] <N1njaneer> The batch oven here is generally about 3 minutes, belt oven is about the same on a 5-zone with standard FR4 boards.
[00:03:08] <theBear> Valen, you got a link to toms page ? pretty sure he documented 90% of it, maybe even got code on there
[00:03:13] <braincracker> Valen <= looking at somebody building something is indeed easy ;)
[00:03:50] <Valen> I was looking at something like this
http://www.ebay.com.au/itm/T962-INFRARED-SMD-BGA-IC-HEATER-REFLOW-OVEN-T-962-m9/200644679969
[00:03:53] <theBear> i like the idea of the elements in the open (apart from mounting/heat issues) 'cos there's much less thermal inertia both up and down
[00:04:04] <theBear> Valen, oooh, very fancy !
[00:04:16] <Valen> but yeah single sided only
[00:04:38] <braincracker> that is cheap
[00:04:44] <braincracker> grab one if you like it
[00:04:52] <Valen> I hear many bad things about them
[00:05:03] <Valen> like they vent 300C air onto your desk
[00:05:05] <braincracker> you can always hack it ;)
[00:05:12] <braincracker> that is no problem
[00:05:14] <braincracker> <;
[00:05:19] <Valen> it fails critical elements of my requirements
[00:05:22] <theBear> also the idea was half-birthed cos i'm currently thinking about a under-pre/general heater with one of these elements, and ir (and likely visual coincidentally due to supply/part/cost choices) focused/aimable on top for top of the line 'ghetto' bga reflows, like laptop gpus and xboxs and most importantly the wifi chip in my phone :)
[00:05:23] <Valen> IE double sided
[00:05:26] <N1njaneer> You pretty much just need stand-offs if you are hacking out double-sided boards on an oven that doesn't have a pin-belt. Beware, though, that ANY amount of thermal mass added to the oven will completely change the thermal profile.
[00:06:18] <theBear> i haven't found anything THAT close to the idea in diy-webland, but i can't see any major problems, and if people succeed with heatguns or hotair irons, shoddy shields and foil and kitchen ovens, i feel good about this approach
[00:06:18] <Valen> N1njaneer: those are IR, the back side of the board probably won't get done much and I want a reliable end result
[00:06:49] <braincracker> IR will cause uneven heating, and it will not be possible to measure temperatures
[00:06:52] <N1njaneer> Even a decent commercial batch oven like the one we have here, if we add small pieces of aluminum to hold a board up, unless those blocks of aluminum have been allowed to come up to temperature completely, they will absolutely suck the heat out and ruin the cycles. :)
[00:07:10] <Valen> http://www.ebay.com.au/itm/IR-Ceramic-Quartz-ACHI-SCOTLE-HONTON-LY-BGA-Rework-Repair-Reflow-Reball-Heaters-/160944472743?pt=LH_DefaultDomain_0&var&hash=item257909c2a7&_uhb=1 for the bga peeps
[00:07:25] <theBear> from memory azone or someone we know well does a 2stage approach, first side with lighter components (by design) then cool and flip, extra flux that side, and heat mostly from top again for big/heavy component side, so long as you don't bump it the bottom don't fall off
[00:07:28] <N1njaneer> You typically do double-sided boards in two passes, so you always have each side of the board pass through the "good" top :)
[00:07:57] <braincracker> smd glue bottom side.
[00:08:06] <braincracker> or it may not even be neccesary
[00:08:06] <N1njaneer> Don't need glue.
[00:08:20] <theBear> braincracker, hmmm, suppose ir kinda messes with ir thermometer aimed at target... or does it ? hmm... now you got me thinking about ir thermometer and SHORT breaks in the ir source (for aimed/pro-reball-station style)
[00:08:31] <N1njaneer> Unless you have some really heavy components are or wave-soldering the SMT parts on the bottom, glue is unnecessary
[00:09:21] <braincracker> that *works* if you happen to know the emissivity and can aim well
[00:09:36] <theBear> braincracker, solder has good surface tension, so long as you don't have big fat tips and connectors and stuff on the bottom (which is pretty standard design approach anyway,) it should and apparently does work fine... if you ever done hotair it's VERY obvious how much surface tension there is, as the paste melts the components jump into place like the pins/pads are magnets
[00:09:42] <braincracker> (and the sensor senses correct wavelengths)
[00:10:18] <braincracker> sure i know that, i design boards that self-align components
[00:10:30] <N1njaneer> Some of the trick to good board reflow is good board design, too, like balanced copper and thermal-reliefs on pads. If you have uneven copper balance, passives will often tombstone on you due to one pad going liquid significantly before another. :)
[00:11:26] <theBear> braincracker, as usual i'm mostly thinking about me <grin> but mine will have at least half visible light, so easy to aim/focus/mask... tempted to do some very much school-style (basic and abstract) test/proofs with a 100w halogen i got lying around, but thinking one of those 250w 24v elc5 phillips integral reflector lamps would be good, very affordable compared to ANY ir/heatlamp, and run at (guessing) aro
[00:11:26] <theBear> und half power there should be a LOT of ir and minimal visible
[00:11:43] <theBear> N1njaneer, you mean balanced thermal wise ?
[00:12:07] <N1njaneer> And of course a lot of it also has to do with getting good quality solder paste on to the pads. The smaller the pads get, the better deposition you need.
[00:12:13] <N1njaneer> tb: Yep!
[00:12:20] <braincracker> hahaha yes, i have tried that thing too, halogens have about 5% visible light, but the IR is wrong wavelength
[00:12:38] <theBear> cool, on that note (reminded me) i learned with manual hotair and commercial products, the mask can help a LOT guiding the solder
[00:13:13] <braincracker> IR lamps use opal quartz
[00:13:22] <theBear> braincracker, i dunno about right w/l, but i DO know they put out a LOT of heat, and those dichroic integrated reflectors focus it TIGHT (pretty much criss-cross designed to go straight into focusing/handling lenses) and run HOT
[00:13:28] <theBear> opal eh ? interesting
[00:13:34] <N1njaneer> So for instance, worst case would be putting a copper pour that completely envelopes a set of pads of, say, an 0603 because you are pulling a bunch of 10K's to ground, with 8-mil traces going out the other side. Almost guaranteed to get tombstones on to the side opposite the fill :)
[00:13:52] <braincracker> http://gongtao.en.alibaba.com/product/618443059-214100976/Quartz_Short_Wave_Infrared_IR_Lamps.html
[00:14:09] <theBear> i notice/know the bathroom style ones run barely past red temp (just enough to look like soft/yellow sunlight), so i figured a halogen like i mentioned would be good
[00:14:24] <braincracker> milky quartz tube
[00:14:36] <theBear> i mean, little lens or two in front of them and you can burn holes in paper/hands 'instantly'
[00:14:57] <theBear> i don't need efficient, just affordable
[00:15:33] <braincracker> the quartz tube blocks and absorbs 95% of near infrared and visible light sent up from heat alloy wire, which make the temperature in the tube rise, producing pure silicon-oxygen bond of molecular vibration, radiation far from the red line. 95% of visible light, near-infrared light can be transformed into far-infrared radiation,far infrared ray with the wavelength>2.5 micron is named as far infrared far infrared.
[00:15:33] <theBear> plus these days, like sodlering stupid tiny things with 1.6mm solder and a regular conical fullsize iron, doing successful/scientific 'ghetto' stuff is one of my hobbies :)
[00:15:58] <braincracker> well i have a toaster oven that has IR lamps, it was cheap, not ghetto enough?
[00:16:17] <braincracker> but still, sensing the board temperature is dufficult
[00:16:25] <braincracker> IR will cause uneven heating
[00:16:53] <braincracker> especially if board has fat black components
[00:16:57] <theBear> first up i was thinking some tests with my existing halogen 100w and some aluminium/shiny one side HIGH temp (250w hid originally) reflectors on the bench with the ir thermo and maybe some random chunks of sodler will give me a good idea, if that even gets close to melting solder in a reasonable time/profile shape, i'm sure one of the 250w will do ok
[00:17:43] <braincracker> depends on mass/size.
[00:17:44] <theBear> not even SURE that little cheaper halogens like this use quartz glass, and even if it's killing a lot of the ir, i know a lot gets out, i replaced ir filters in things that use them
[00:17:57] <braincracker> 1cmx1cm boards will reflow with a few watts :)
[00:18:16] <theBear> i got an old 'free' netbook to tryout once i'm confident, but the main object is a single approx 16mm*16mm bga in my fone :)
[00:18:39] <theBear> and remember the preheater underneath, i feel i can probly push that to 150c or so safely if that's what it takes
[00:18:46] <braincracker> 100-150C preheat from under, and top reflow
[00:18:52] <theBear> heh, great minds :)
[00:18:53] <braincracker> cover with tin-foil
[00:19:02] <braincracker> cut out bga shape
[00:19:12] <theBear> and toaster oven is almost ghetto enough, but i don't have one and curb-pickup isn't again for 5 months or so
[00:19:32] <theBear> anyway, can't sit anymore right now, back getting grumpy, and need a coffee
[00:19:40] <braincracker> junk-yard?
[00:19:42] <braincracker> <;
[00:19:55] <theBear> no junk yards round here these days, and tips are strictly non-scavenging rules
[00:19:56] <braincracker> and microwave ovens have often have grills
[00:19:57] <theBear> stupid oh+s
[00:20:06] <braincracker> not that cool IR lamp, just a heater, but add a fan
[00:20:07] <theBear> i never had a microwave and don't want one
[00:21:03] <braincracker> i had one, but i don't want one either
[00:21:39] <braincracker> i'm only IR reflowing, or hot-air reflowing cookies
[00:21:54] <braincracker> any any food
[00:22:23] <braincracker> a microwave oven would be good for industrial use, maybe creating diamonds?
[00:22:44] <braincracker> but those hot-spots ...
[00:24:07] <braincracker> it is nasty when you take out the spinner tray, cut some grapes in half, and slice them in to be a short-circuited RF antenna, and give them some microwave power
[00:24:44] <braincracker> instant plasma bolts, it comes to life
[00:25:50] <N1njaneer> $15 Harbor Freight heat-gun also works remarkably well if you are in a pinch and can channel the Reflow Oven Gods while using it. I've actually used one to put down 5x5mm 32-pin QFNs perfectly fine in a major pinch. :)
[00:26:23] <braincracker> it is nice to have a temperature controller hooked up to it
[00:26:38] <theBear> http://download.p4c.philips.com/l4b/9/924010520594_na/924010520594_na_pss_aen.pdf is a fairly unhelpful datasheet, for reference
[00:26:50] <braincracker> it is possible to DIY a better stuff than commercially available
[00:26:50] <theBear> 100 cri (not sure what that scale is) is positive, and a few other comments
[00:26:51] <N1njaneer> Have a Pace TF2000 here for the precision rework stuff :)
[00:27:00] <theBear> and they're cheap, and i can still get cost price from old work
[00:28:51] <braincracker> cri is color rendering index
[00:28:59] <braincracker> it has nothing to do with soldering ;>>
[00:29:10] <braincracker> 100 means perfect color rendering
[00:29:37] <N1njaneer> It is used to determine how awesome you look WHILST soldering, which is very important :)
[00:29:52] <N1njaneer> If you exude confidence, the solder joints will pick up on this and flow more nicely.
[00:30:03] <braincracker> this is some elite studio halogen rated at 50 hours operational life :)
[00:30:44] <braincracker> the $1 type heats same way.
[00:30:56] <braincracker> it only lives a few hundred times more
[00:34:42] <Valen> N1njaneer: our heat gun blows the bigger parts around :-<
[00:35:07] <N1njaneer> Come in slower and from further away, and put it on low :)
[00:37:09] <Valen> it was as low as it went and as far away as i could hold my hand in front of it
[00:38:42] <N1njaneer> Carefully rest the tip of some tweezers on them in the process, then, and hold the parts down :)
[00:41:41] <theBear> braincracker, i know what it means, but a halogen shouldn't have 100, and that's no doubt why they listen that qualifier (no doubt a scale/measurement method) after it, and i'm no glass scientist (budding amateur maybe) but uv-open quartz glass, 900c max bulb temp and 3400k also sound encouraging, and like i said, i KNOW you can 'instantly' burn paper with one of these focused, heck, we all know how hot the
[00:41:41] <theBear> light from a halogen lamp is, even with various kinds of glass inbetween the bulb and your hand... quietly confident
[00:42:53] <theBear> braincracker, and that 50hr on the philips (most popular/common brand)surprises me, and i know they usually last MUCH longer, most of them are 500hr rated which isn't AMAZING, but for 5 or 10 bucks and locally available, if my existing non-integrated reflector halogen tests are encouraging, it can't hurt to try
[00:43:40] <theBear> i got enough psu parts and a suitable transformer or two, big fat 15003/4 on a decent sink and a pot, easy to implement
[00:43:50] <N1njaneer> Why not just take the heating elements out of a toaster oven - ones made to put out long-wave IR instead of visible light :)
[00:44:12] <N1njaneer> You know, for heating. :)
[00:44:13] <theBear> oooh, AND i got some nice tweezers at the 2 dollar shop too, metal, various angles and one medical style springs-shut one
[00:45:39] <theBear> N1njaneer, cos i don't have access to a cheap/free toaster oven, and when yer cripple-retired you gotta have hobbies so you don't go even crazier than you may have already been <grin> and like i said earlier, one of mine recently is succeeding at 'silly ghetto' approaches to technical stuff, with low-mid backing of actual science :) so far the approach seems to make me 'do better' than all those instructab
[00:45:40] <theBear> le and forum monkies that i dislike so :)
[00:46:04] <N1njaneer> tb: Are you in the US?
[00:46:05] <theBear> and what i can do with 1.6mm solder and a fullsize iron is verging on impossible recently :)
[00:46:12] <theBear> me ? .au
[00:46:23] <N1njaneer> Got Craigslist over there? :)
[00:46:45] <N1njaneer> Sometimes people practically just give 'em away.
[00:47:26] <N1njaneer> Dunno, was a thought :)
[00:51:17] <theBear> nah, afaik there's nothing like craigslist, there'll be another 10-50 within 10 streets the next pickup, but this last one a month ago was greenwaste, so another 5 months give or take till the next regular one
[00:52:18] <theBear> and last round i was obsessed with getting lcds and fixing/frankensteining them together for sale so i ignored toaster ovens and other stuff i 'know i need eventually' cos space was tight... of course then i worked out that it's almost impossible to sell a working lcd monitor for more than about 20 bucks, so now i've just got a bunch of spare/extra monitors all over the house :)
[01:04:04] <braincracker> theBear <= i'm not telling a big secret i got my IR toaster oven for $25, it has 2 IR heater lamps, 700W total.
[01:04:40] <braincracker> and it has a few thousand operating hours already
[01:04:59] <theBear> i don't have $25 to spare tho
[01:05:32] <braincracker> but this is an oven, temp control needs to be hacked in it and done
[01:06:08] <braincracker> glass window with IR filter :)
[01:06:14] <braincracker> fancy
[01:08:19] <braincracker> by default is is junk, but hackable
[01:11:08] <braincracker> a crap hot-plate is about $15 too :)
[01:11:18] <braincracker> it only needs temp control
[01:11:49] <braincracker> and you have an industrial hot-plate
[11:19:03] <darsie> If I want only 8 bit ADC, can I make the ADC stop after 8 bits?
[11:21:38] <BJfreeman> no
[11:22:28] <darsie> thx
[11:24:31] <BJfreeman> you can get a Spi or I2c 8 bit ADC
[11:26:50] <OndraSter_> well
[11:27:23] <OndraSter_> you *can* read the result sooner
[11:27:41] <darsie> and then stop the adc.
[11:27:48] <darsie> But that would require very careful timing.
[11:43:48] <BJfreeman> wording it different, the output is not avalible till the end of conversion
[11:45:02] <darsie> I'll just wait till it completes and then use 8 or 10 bits, dunno.
[13:30:13] <taru> I am trying to run this file on WinAVR but it doesn't run.
http://pastebin.ca/2440576
[13:30:38] <taru> It compiles fine but doesn't show anything on proteus
[13:50:26] <endikos> So. If CTC is better than norml mode with overflow - why is overflow there at all? In what scenario would an overflow timer be better?
[14:45:26] <braincracker> endikos <= simple
[14:45:37] <braincracker> divides by 256
[15:48:11] <N1njaneer> Relative CALL to subroutine
[15:48:36] <ambro718> yeah but what is the dot?
[15:49:37] <N1njaneer> http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&p=792091
[15:50:11] <N1njaneer> . means the current value of the program counter
[15:50:17] <N1njaneer> That article explains is pretty well
[15:50:31] <N1njaneer> +it
[15:50:33] <ambro718> I see, thanks
[17:17:40] <ambro718> omg that's how gcc shifts an int32_t by 15 to the right
http://ideone.com/Esn1cF
[17:29:12] <N1njaneer> Yes, no barrel-shifter on AVR :(
[17:46:51] <Kev-> I need a source for 6pin isp cables... What are they called exactly so I can look for it on ebay?
[17:48:39] <braincracker> try a jumper array
[17:48:48] <braincracker> andcut an ide or floppy cable
[20:43:49] <spec__> Hi all. I'm having a bad time getting some code to work on the ATtiny2313
[20:43:56] <spec__> Source is here:
http://pastebin.com/NkrtcZQk
[20:44:11] <spec__> It's a single channel DMX LED driver.
[20:46:00] <spec__> On start up, I *think* I'm setting timer 1 to fast PWM (mode 15) and setting the duty cycle to 50%. On power up my output flashes briefly then bubkis.
[20:47:20] <spec__> Any help would be appreciated.
[20:47:22] <N1njaneer> spec: For one, I don't see you checking for a break condition to restart the DMX frame reception, or the zero byte start-of-packet that indicates normal DMX channel data :)
[20:48:50] <N1njaneer> I would get your PWM routine working correctly with manually set data or a ramping variable before attempting to recieve DMX correctly. Isolate the problem and work each part in succession until you know it is good.
[20:49:20] <N1njaneer> The DMX reception code section will not work as written, for one.
[20:49:37] <spec__> OK.
[20:49:47] <spec__> The break detection was torn from this page: blog.wingedvictorydesign.com
[20:51:17] <N1njaneer> You are also not checking for the packet type byte, which will cause exceedinly erratic behavior if you try to run this on any equiptment that sends RDM packets
[20:51:40] <spec__> Ok, I set OCR1AH and OCR1AL durring initialization, and it flashes briefly, then goes dark. SHouldn't it run continuously until another value is written?
[20:53:13] <spec__> I'm not too concerned about the packet type at the moment. The device driving it only sends brightness values.
[20:53:39] <N1njaneer> That's also a terribly error-prone way to sense break conditions.
[20:54:00] <spec__> Why is that? I'm open to better methods.
[20:55:34] <N1njaneer> Either use the Framing Error bit if the UART of the device supports it (I don't believe the 2313 does) or else put the UART in to 9-bit reception mode and check if the 9th bit comes up zero, which would indicate a break. Normally 9th bit would be the stop-bit, which should be high in a normal byte reception.
[20:58:02] <N1njaneer> Your timer method consumes a humungeous amount of computing power, and is subject to jitter and other timing abnormalities, plus 88us is defined as minimum DMX break but can be an order of magnitude longer, which may cause issues with the implementation as written without some additional flagging :)
[20:58:33] <N1njaneer> breaks as OOB data are pretty standard, so UARTs generally provide easy ways to detect them.
[21:00:32] <spec__> Ok. I'm going to forget about receiving DMX at the moment. Any idea what's wrong with the PWM generated with timer 1?
[21:04:26] <N1njaneer> Would have to pick through the code to make sure the settings are correct, gotta scoot for a bit though. Timer stuff is easy to misconfigure, though. It's usually something simple. I assume you have set the appropriate pin to OUTPUT in the DDRx registers and have excercised the pin manually to make sure the off-chip hardware is working correctly?
[21:06:54] <N1njaneer> bbiab