Back
[04:34:42] <LeoNerd> Anyone looked at/used the (new?) ATtiny1623? It's a 20pin SOIC. It says on the DS that it has both the USI and a separate slave-only I2C module
[04:35:01] <LeoNerd> The pinout list doesn't seem to say where the I2C slave pins are though; it only lists SCL/SDA as USI pins in 2-wire mode
[04:35:23] <LeoNerd> er... 1634 even
[04:42:26] <LeoNerd> It has some rather cute features, such as a software-controllable CKSEL register, that the fuses just initialise at boot
[04:42:35] <LeoNerd> So you can change clock source entirely if you want
[04:45:40] <DO9XE> LeoNerd, xMega got those, too :)
[12:20:47] <Berra> Is it possible to connect to an RPI serial console using an AVR programmer?
[12:21:33] <LeoNerd> I doubt it; AVR programmers are usually SPI-like
[12:21:48] <Berra> Alright, thanks for the help
[13:50:32] <hypermagic> Berra, you can use gpio to bitbang anything
[13:50:50] <hypermagic> spi, i2c, sd card
[13:51:59] <hypermagic> and i think you could make a serial portwith gpio too
[13:52:09] <Berra> Yeah but it's no simple solution
[13:52:24] <hypermagic> it is timple...
[13:52:28] <hypermagic> software needed
[13:52:32] <hypermagic> and connect it
[13:52:52] <Berra> And it would give me console access on the rpi?
[13:59:12] <hypermagic> hmm
[13:59:29] <hypermagic> yes, if you write a simple terminal then you can access with whatever
[13:59:38] <hypermagic> even an ir remote control
[14:00:06] <Berra> So using an AVR programmer in a roundabout way as a USB->Serial would be quite easy?
[14:00:08] <hypermagic> but the rpi has an ethernet.
[14:00:16] <hypermagic> i'd just plug in a crossover cable.
[14:00:28] <hypermagic> ssh onto it
[14:00:31] <hypermagic> :P
[14:00:33] <hypermagic> superlazy
[14:00:36] <hypermagic> and fast
[14:01:02] <Berra> Yes that is what I'll do once I can read the SD card and set up the networking to allow for that kind of connection
[14:04:00] <hypermagic> well you could use a sd card reader and mount the thing in kvm
[14:04:08] <hypermagic> and set it up simply
[14:04:19] <Berra> Yes. Indeed
[14:04:28] <hypermagic> or just boot it using an usb card reader and another machine
[14:04:46] <hypermagic> or why not on rpi btw?
[14:05:16] <Berra> No HDMI
[14:05:23] <hypermagic> :/
[14:05:34] <hypermagic> you can get a hdmi- dvi cable
[14:05:46] <Berra> I have HDMI mini but not HDMI
[14:06:05] <hypermagic> and the rpi has tvout
[14:06:06] <hypermagic> ;>
[14:06:12] <hypermagic> you can plug it into old tv lol
[14:06:24] <Berra> Not the RPI2 as far as I know
[14:08:47] <hypermagic> so you might want to get a dvi cable
[14:09:16] <Berra> Fixing the networking with a card reader seems most sane
[14:09:17] <hypermagic> btw you dont want low power lcd for it ?
[14:09:56] <Berra> I am basically only going to use the rpi as a way to speak to an array of arduinos
[14:10:14] <Berra> arduinos / atmegas over tcp
[14:10:33] <Berra> Since the rpi has networking it will be the gateway to talk to a bunch of avrs
[14:10:42] <hypermagic> hm...
[14:13:19] <hypermagic> http://tuxgraphics.org/electronics/200612/article06121.shtml
[14:13:30] <hypermagic> Using the AVR microcontroller based web server
[14:14:04] <Berra> Thanks, will have a read
[14:14:20] <Berra> Watching mailbag now though : )
[14:40:17] <Lambda_Aurigae> your programmer is likely not a simple usb-serial adapter and would require some hacking to turn it into something to do that with.
[14:41:38] <hypermagic> Lambda_Aurigae, well i think hte onboard 100Mbps ethernet is superior to the serial port
[15:28:01] <Lambda_Aurigae> probably
[15:28:14] <Lambda_Aurigae> however, it does go through the usb controller....but,,,that's still faster than serial.
[16:35:05] <guest57627> Anhone know of a good working software I2C library using gpio?
[16:41:27] <LeoNerd> Send or receive?
[16:41:51] <LeoNerd> I have half of one, that talks a modified version of I2C
[16:42:22] <guest57627> master I2C
[17:00:52] <N1njaneer> Whoops, forgot to log back in here. Hi all! :)
[17:05:16] <LeoNerd> Afternoon/evening/morning...
[17:05:40] <N1njaneer> Late afternoon here. About to go take a break for a goathike, then back to jam on things tonight :)
[17:08:43] <LeoNerd> N1njaneer: Ah, that reminds me. I was ranting about the tiny1634 earlier
[17:09:07] <LeoNerd> In summary: It has a USI and a real I2C slave-only module, but the datasheet doesn't *actually* say where the pins for the I2C are.
[21:08:01] * N1njaneer yawns, back from hike.
[21:08:30] * Casper throws a ghost pepper in N1njaneer's yawning mouth
[21:08:38] <N1njaneer> Yum! I love those things!
[21:09:08] <Casper> saw a video of a kid that shallowed one and instantly regretted it
[21:09:17] <Casper> I felt wrong of laughting
[21:10:33] <N1njaneer> Finally got video of the new pick and place running production if anyone is interested! Only took a week for install, calibration, and training. :)
[21:10:54] <Casper> :) but
[21:11:07] <Casper> do you know stuff about compressor pressure switch reassembly? :(
[21:11:37] <N1njaneer> Not specifically, though I've probably had to fiddle with them at some point.
[21:11:41] <Casper> had to dissassemble mine, it exploded due to the springs... appear to be reassembled correctly, don't turn on
[21:12:03] <N1njaneer> Check contacts for cleanliness?
[21:12:13] <Casper> it don't click
[21:12:26] <N1njaneer> Not sure what else to check other than the obvious stuff. I assume it needs to be recalibrated after assembly, etc.
[21:12:35] <Casper> if I push the kind of test button it kick in, then should work until full, then stop and won't kick back on
[21:13:02] <Casper> it have 2 metal pieces
[21:13:09] <Casper> connected via a spring
[21:13:32] <Casper> when the bottom one get lifted, it cause the second one to kick/flip up
[21:13:44] <Casper> when it goes down it should make the top one kick down...
[21:13:49] <Casper> ... it do not
[21:14:33] <N1njaneer> Aww :/
[21:15:59] <Casper> yeah
[21:16:08] <Casper> must be a lost spring or spacer or something
[21:16:15] <Casper> can'T see what I could have misassembled...
[21:16:19] <N1njaneer> Hit it with a hammer till it works.
[21:16:24] <N1njaneer> Or yell at it.
[21:16:50] <Casper> the bottom part... fit in a notch... the top part also have a notch
[21:17:01] <Casper> they can't fit in any other way
[21:17:15] <Casper> and the spring only have one place it can hook to
[21:21:07] <Casper> also the bottom part need to move way too much...
[21:32:48] <N1njaneer> Video:
https://www.youtube.com/watch?v=ObPvU5fru7k
[21:36:09] <Casper> the video is somewhat corrupt it seems
[21:36:13] <Casper> show as 2:14
[21:36:17] <Casper> but last 4 seconds
[21:36:51] <N1njaneer> Runs fine for everyone else I'e send the link to :)
[21:37:15] <N1njaneer> Odd. Maybe bad YouTube server? Runs fine here
[21:38:46] <N1njaneer> Try reload?
[21:50:20] <hypermagic> hi
[21:51:19] <N1njaneer> Hi Hypermagic
[21:51:21] <N1njaneer> !
[21:51:34] <hypermagic> whats up N1njaneer ?
[21:51:47] <N1njaneer> N'much! Got our new pick and place running!
[21:51:54] <hypermagic> hmm
[21:51:57] <N1njaneer> Exciting times here!
[21:52:07] <N1njaneer> And now debugging some ARM stuff
[21:52:35] <hypermagic> so now you are gonna make boards after filling the tubes with smds and drinking your cffee ?
[21:53:01] <hypermagic> and clicking a button to start
[21:53:03] <N1njaneer> All the parts are on generally on reels, but we do have a vibe stage for legacy tubed parts :)
[21:53:27] <N1njaneer> Pretty much. It runs with little to no intervention, which is nice change. No more babysitting.
[21:53:30] <hypermagic> and will it replace boards too ?:)
[21:53:38] <hypermagic> and reflow them ?
[21:53:44] <N1njaneer> Yep!
[21:53:50] <hypermagic> ha
[21:54:08] <N1njaneer> Well the oven does that, but it goes out the conveyor to it so it can continue on with the next
[21:54:17] <hypermagic> sounds cool
[21:54:30] <hypermagic> no interaction needed then, 10k boards a day
[21:54:49] <N1njaneer> See if this video works, wasn't loading for Casper:
https://www.youtube.com/watch?v=ObPvU5fru7k
[21:55:22] <N1njaneer> If the boards had about 16 parts each on them, we could in theory do about 10,000 per day on it I suppose
[21:55:50] <N1njaneer> 17 seconds to assemble the 40 parts on the boards in those video, 12 per panel.
[21:55:59] <hypermagic> works for me
[21:56:01] <hypermagic> hacked it
[21:56:13] <hypermagic> whats with the hexagon mega red led light ?
[21:56:16] <Casper> now it appear to work
[21:56:28] <N1njaneer> High-speed line scan camera looking at the components on the heads.
[21:56:48] <hypermagic> alignment ?
[21:56:53] <N1njaneer> Yep!
[21:57:27] <N1njaneer> The track to the corners at 1:03 is the cameras checking any geometry change in the frame due to temperature.
[21:57:35] <Casper> look like the firmware could be quite optimised
[21:58:31] <hypermagic> https://www.youtube.com/watch?v=CRSLbo_8nTQ
[21:58:35] <hypermagic> this is the low end
[21:59:40] <N1njaneer> Casper: The machine actually does extremely well on doing heavy optimization for the placement of the feeders, pathing, and which parts get picked up on which pass on which heads.
[22:00:47] <N1njaneer> And if you have a bunch of them in-line, it will balance parts out to different machines in order to properly balance the line. It's quite a nice setup!
[22:00:55] <Casper> hypermagic: that one need serious optimisation
[22:01:31] <Casper> N1njaneer: yeah look better after a while
[22:01:41] <Casper> it looked intially as if it couln't move diagonally
[22:02:01] <N1njaneer> That's because it has to run over the line-scan camera on the way to placement
[22:02:45] <Casper> yeah, but past it, it seemed to mostly move in straight lines, but then it started to go diagonally
[22:03:10] <N1njaneer> The YS12 is the same machine as this, but has 10 spindles and the camera rides on the head so it does a scan while tracking to the placements, but at the expense of smaller components and less height capability. Typically you'd put a YS12 in front of the YS12F to chip-shoot passives, then use this one to lay the more off parts.
[22:03:38] <N1njaneer> Casper: Yes, because it's placing 5 parts per pass, so move moves will be in straight lines, as it's in block-build mode and assembling each image on the panel in succession.
[22:03:52] <hypermagic> the optical alignment thing looks interesting what does it do ?
[22:04:08] <hypermagic> simply outside bounding the image ?
[22:04:35] <N1njaneer> It looks at the components that are on the spindles to ensure that 1> it got the component 2> correct component 3> placement where the component is to correct x, y, theta
[22:04:45] <hypermagic> haha
[22:05:01] <hypermagic> it actually looks at the component and recognizes it ?
[22:05:06] <N1njaneer> No, it does a full suite of image feature recognition algorithms based on the part and the geometry.
[22:05:07] <N1njaneer> Yes.
[22:05:17] <N1njaneer> Camera accurate is down to a couple microns.
[22:05:22] <N1njaneer> +accuracy
[22:05:25] <hypermagic> ok but...
[22:05:32] <hypermagic> it is just a dumb pc
[22:05:36] <Casper> so just check that it got the right package then
[22:06:13] <hypermagic> i think it will be hard to recognize the component
[22:06:18] <hypermagic> with only 1 camera
[22:06:37] <N1njaneer> Right package, but also a lot of other criteria - can do things like bent pin detection, lead placements, etc. Absolutely critical for things like 01005 and microBGAs, etc, when you have to land the part within a few dozen microns for placement.
[22:06:45] <hypermagic> or your hexagon might lit it in 3 -6 different ways and scan it in 3d?
[22:06:46] <N1njaneer> It's a line-scan camera. It works extremely well :)
[22:06:50] <hypermagic> tlight
[22:06:53] <N1njaneer> No reason to.
[22:07:26] <N1njaneer> But the lighting can be adjusted based on the component geometry to get the best lighting for the image recognition.
[22:07:37] <N1njaneer> The connectors in particular are where the side-light kicks up.
[22:07:57] <hypermagic> and if you get diffeerent vendor packages and they come in white you must recalibrate ;)
[22:08:36] <N1njaneer> Not really, depends on what you're looking at. On those connectors we actually look at the pins and the chamfers on the ends of the pins.
[22:08:46] <hataketsu> Hello, I have a question with 12Mhz crystal: does it require a couple of capacitor connected to GND?
[22:09:09] <N1njaneer> Yes, the crystal requires proper-value resonance caps.
[22:09:26] <hataketsu> I have a usbasp programmer and its crystal does not
[22:10:04] <hataketsu> The crystal simply connects to XTAL1 and XTAL2
[22:10:30] <hataketsu> I really confuse.
[22:14:00] <N1njaneer> hataketsu: If it has a ceramic resonator, some are built with integrated resonance caps. Check the datasheet for the microcontroller - it explains all of this :)
[22:15:50] <hataketsu> Atmega8, Ok, I will read it
[22:19:01] <hataketsu> I just have a very bad night, I use arduino as ISP, connect usbasp to it, pin 10 of arduino to reset pin of atmega8. Some first time I can easily upload firmware to usbasp but now it does not work anymore. avrdude can read its signature :/
[22:22:58] <hataketsu> Its signature is read as 0xffffff or 0x000000 ,crap :|
[22:28:49] <N1njaneer> Make sure you are lowering the clock speed you are attemping to do ISP at.
[22:29:15] <N1njaneer> 125Khz is a good bet to start - ISP rate needs to be at 1/8th or less the clock rate the device is runnign from.
[22:29:58] <N1njaneer> If you switched the fuses to run on an external crystal, you need to have the external crystal working correctly (with resonance caps) else the ISP connection will fail to work. The ISP state machine requires a valid clock signal in order to function.
[22:32:45] <hataketsu> I try avrdude with -B 1,2,4,8,16... but have no luck here
[22:36:54] <hataketsu> As errata in datasheet said, my atmega8 seems to be erased signature, anyone met this situation before?
[22:48:26] <N1njaneer> Did you attempt to run the micro from an external clock crystal?
[22:48:45] <N1njaneer> i.e. Did you reprogram the fuses?
[22:49:02] <hypermagic> why you use tungsten probes for wafer instead of stainless steel for example?
[22:49:47] <N1njaneer> Likely because of strength and better conductivity.
[22:50:21] <N1njaneer> I think tungsten may also be a lot more ductile.
[22:50:31] <hypermagic> well it is brittle
[22:50:38] <hypermagic> stainless stell is flexible
[22:52:09] <hypermagic> would you use SiC or WC for probe tip? :)
[22:52:50] <hypermagic> those are comparable to tungsten in hardness
[22:55:46] <hataketsu> No, I didnt change anything, I just program it
[22:58:52] <Valen> they normally use ruby for probe tips?
[22:59:14] <N1njaneer> For CNC probe tips, yes. My Renishaw has one :)
[23:05:02] <hypermagic> EEVblog #395 - World's Most Expensive Hard Drive Teardown
https://www.youtube.com/watch?v=CBjoWMA5d84
[23:05:23] <hypermagic> Valen, i meant electrical probes :)
[23:05:59] <Valen> ahh
[23:31:30] <N1njaneer> Conductive ruby probes, clearly!
[23:31:55] <vsync_> wow, eevblog
[23:32:16] <vsync_> does he have new $100k scopes already?
[23:32:24] <vsync_> that he can show off?
[23:33:28] <vsync_> the guy's background seems vague at best
[23:33:50] <hypermagic> EEVblog #501 - Sinclair C5 Electric Car Teardown & Test Drive
https://www.youtube.com/watch?v=xS6q27VOTOk
[23:54:43] <hypermagic> N1njaneer, why red leds for camera?
[23:55:17] <N1njaneer> Cheap, and I believe the camera is filtered for that wavelength. Less interference from ambient room light.
[23:55:28] <hypermagic> thne IR ?
[23:55:39] <N1njaneer> A lot of high-speed cameras stuff uses red for illumination.
[23:55:41] <N1njaneer> Nope, just red.
[23:56:27] <N1njaneer> https://www.youtube.com/watch?v=ObPvU5fru7k
[23:57:07] <N1njaneer> That's a YS12 - the 10 head version of the same machine, but you can see a camera is on a carriage below the spindles on the head so it doesn't have to run the line-scan for most parts.
[23:57:09] <hypermagic> saw that
[23:57:54] <hypermagic> the camera watches from the middle of leds no?
[23:57:56] <hypermagic> from below
[23:58:01] <N1njaneer> So higher-speed placement and double the number of spindles, but part size restriction.
[23:58:10] <N1njaneer> Camera is a linear sensor that looks straight up.
[23:59:20] <hypermagic> scanner ;>