#qi-hardware IRC log for Saturday, 2014-02-01

apeletewpwrak DocScrutinizer05: will take a picture of the "brackets for power supply" when I get back from fosdem, maybe you'll have an idea by seeing it00:11
dos1Nuernberg gingerbread and cookies are yummy :)00:22
DocScrutinizer05but stricly a xmas thing00:25
wpwrakdos1: the rest of the year, they're highly toxic01:55
DocScrutinizer05rather not available. Those which *are* nevertheless actually might not be that tasty, dunno if they even are toxic. Maybe, at least for germs etc04:21
DocScrutinizer05ok, I know one place where you get fresh juicy stuff during the whole year 04:22
DocScrutinizer05maybe even a second, though that might have changed years ago and I wouldn't have noticed04:23
DocScrutinizer05I also dunno how much of the cinnamon is Cassia which has cumarin04:27
DocScrutinizer05cumarin itself is obviously highly controversial and interesting a substance04:31
DocScrutinizer05from medicine to toxin it has everything, broad range from carcinogen to healing cancer to killing rats to derivative warfarin (4-hydro-cumarin) blood thinner, to same substance killing kettle04:34
DocScrutinizer05kettle? dang! 04:36
DocScrutinizer05probably the best widely available mass product: http://www.haeberlein-metzger.de/ (of course you can find better stuff when you know where to ask for it)04:41
DocScrutinizer05Aachener Printen- und Schokoladenfabrik04:45
DocScrutinizer05Henry Lambertz GmbH & Co. KG04:45
DocScrutinizer05Borchersstr. 1804:45
DocScrutinizer05D-52072 Aachen04:45
DocScrutinizer05meh, as usual04:51
DocScrutinizer05anyway, get "Elisen-Lebkucjen". Those are no fake04:51
nicksydneymanaged to solder ATTiny13 now want to test continuity test....this is the pinout http://circuits.datasheetdir.com/18/ATTINY13-pinout.jpg ..... should it beep when i place the multimeter red cable to Pin 8 (VCC) and black cable of the multimeter to Pin 4 (GND) ?04:58
DocScrutinizer05actually you should be very careful with probing. You can even destroy components when the meter is cheap and poor quality or the component very delicate05:19
DocScrutinizer05particularly with continuity test which applies unknown voltage to the Device Under Test05:20
DocScrutinizer05some of the better multimeters have a continuity test and a diode test. The continuity test is usually safe in this case, and the diode test is less safe but still mostly safe05:21
DocScrutinizer05cheap multimeters however may not have 0.3V for continuity and 0.9V high impedance for diode test, but I seen even 9V(!) on continuity05:22
DocScrutinizer05from V+ to GND of your ATtiny anyway there shouln't flow more than maybe 1uA when not active (running program). This is definitely not enough for any continuity tester to start beeping05:26
DocScrutinizer05continuity tests should get done along traces and across solder points, but usually *not* across components05:28
DocScrutinizer05and you might want to think twice when deciding where to place red and black probe. E.G. when you test if pin4 is correctly connected to the PCB, you rather place red to the PCB and black to pin4, so the GND pin of chip is actually minus wrt any other point on PCB when it's NOT connected correctly05:30
DocScrutinizer05on a general note: I guess there are fewer people living on this planet that know how to probe correctly that those who really know c++05:32
DocScrutinizer05than thaose*05:32
DocScrutinizer05probing is a really high art05:32
DocScrutinizer05>>Only those who have measured a lot of rubbish will learn how to measure what's important with few measurements. <<05:37
DocScrutinizer05make sure your multimeter has no "high" voltage on the probe tips when in continuity test! easy way when you have no second meter: taste both tips same time. You can taste voltage that is dangerous for chips05:45
DocScrutinizer05(no, I'm not kidding you :-D )05:45
DocScrutinizer059V battery is already slightly achy, but still perfectly harmless. and you can easily detect even 1.2V05:46
DocScrutinizer05with your tongue05:46
DocScrutinizer05if you never before done it, practice with a 1.5V battery to get an idea how it "tastes"05:47
DocScrutinizer05an age old very accurate and sensible way to test for low voltages. Make damn sure you don't use this method on 220V! ;-P05:48
DocScrutinizer05actually I'd recommend to stay away from 12V already05:49
DocScrutinizer05ooh, and another iron rule (maybe too obvious but I don't want to have missed mentioning it): NEVER probe power sources of any kind for Ampere!05:52
DocScrutinizer05you wouldn't believe how many users try to not only probe for the 220V in the mains outlet but also for the 16A written on the fuse in their fuse box. They actually switch the meter to Ampere and plug it into the mains outlet05:54
DocScrutinizer05there's a reason why better multimeters have a common ground jack but distinct Ampere jack and Volt/Ohm/blabla jack.05:56
DocScrutinizer05and the cheaper meters actually make it very easy to accidentally switch the meter range to Ampere from volt when turning the selector knob to "off" position. I've seen several meters explode because of exactly that05:57
DocScrutinizer05as a general rule you probe for ampere only where you have CUT UP a wire or trace or the like, and place the probes to both open ends of that cut wire05:59
DocScrutinizer05test points that are not supposed to normally touch each other are usually not suited for Ampere measurements between them06:01
nicksydneyDocScrutinizer05: just came back and saw your message :)06:49
nicksydneyDocScrutinizer05: hope I didn't fry the MCU as I tried earlier with continuity test and it didn't beep as you said06:50
nicksydneyDocScrutinizer05: what's the best thing to do to check if the MCU is ok ?06:50
DocScrutinizer05operate it06:51
nicksydneyDocScrutinizer05: when you say "operate it" .. you mean program it ?06:51
nicksydneylike try to flash a program into it ?06:51
DocScrutinizer05for example, yes06:51
nicksydneyahh ok06:51
nicksydneyso it's not like testing a 555 when you can test with multimeter 06:52
DocScrutinizer05hardly, since this chip has no special pins06:52
DocScrutinizer05I don't know what it's supposed to behave like when you power it up with no user program flashed. Does it come with a factory program that wiggles all pins? does it come with empty program storage and does nothing at all? I dunno06:53
DocScrutinizer05and I guess there's no "simple" special signal like a clock or sth that you can probe, that is supposed to always show sth recognizable06:54
nicksydneyunderstood..ok be right back06:55
DocScrutinizer05I'd think for safety reasons all pins will be in high-Z input state and nothing will happen. the chip should "act dead"07:03
DocScrutinizer05maybe Atmel shipped it with a factory test program though, that listens for a certain challenge on one pin and then runs some test program taht would wiggle pins or echo state of one pin to another, or sth like that07:05
nicksydneynah don't think so...but let me read the datasheet07:07
nicksydneyDocScrutinizer05: this is the board http://ctrlv.in/28987007:08
DocScrutinizer05I guess it's same like yesterday?07:09
nicksydneyit's populated now with MCU :)07:10
DocScrutinizer05it looks like your solder oxidized due to too high soldering temperature07:10
nicksydneywish that MCU vendors should upload some test program so that we can know whether MCU is working fine or not07:11
nicksydneyyeah still need to stabilise my soldering skills...it's still very rough07:11
nicksydneynot as smooth as wpwrak's07:11
nicksydneyi've see that video...i think the problem is my soldering tip is not that good...as i notice sometimes the solder 'sticks' to it07:16
DocScrutinizer05your iron is too hot07:17
DocScrutinizer05"add some more flux here, as you can never have too much"07:21
nicksydneyi put flux and it become too much :)07:22
DocScrutinizer05do you also think that the 2nd from right pin at 25:00 isn't soldered?07:26
DocScrutinizer05how do you know it's too much flux?07:38
DocScrutinizer05did it act as cooling liquid for your iron, lowering the temperature so much that you can't melt the solder anymore?07:39
DocScrutinizer0539:45 !07:43
DocScrutinizer05http://www.youtube.com/watch?list=PL2862BF3631A5C1AA&feature=player_detailpage&v=J5Sb21qbpEQ !!07:49
DocScrutinizer05nicksydney: http://www.youtube.com/watch?list=PL2862BF3631A5C1AA&feature=player_detailpage&v=fYz5nIHH0iY#t=326 08:24
whitequarkDocScrutinizer05: so I bought a new NE555N and fixed the spindle control board10:06
whitequarkit works, except there's a weird quirk: it works exactly in reverse10:06
whitequarkwhen the speed control pot is on the min position, it spins at maximal speed10:07
whitequarkand vice versa10:07
DocScrutinizer05you swapped upper and lower end of poti10:07
whitequarkI'm pretty sure I connected it the right way, because the cable has remembered the way it was twisted10:07
DocScrutinizer05obviously not10:07
whitequarkmoreover, if I reverse the pot connector, it stops working properly, more specifically10:08
DocScrutinizer05or your chip has inverted logic ;-)10:08
whitequarkit spins at min speed in min position, but when I start to adjust it to the max, it spins faster for some time10:08
whitequarkthen at a certain position it stops spinning at all and the motor makes funny noises10:08
whitequarkthe consumed current also drops to almost zero at that point10:08
whitequark(inverted logic) I've desoldered an ST NE555N from there. I soldered an ST NE555N there.10:09
DocScrutinizer05are you sure you didn't mix up the 555 component with some other technology?10:09
whitequarkchips look exactly same except for lot number10:09
DocScrutinizer05lemme think a while about it10:10
DocScrutinizer05do you have a scope?10:15
whitequarkno :(10:15
whitequarkit's practically impossible to import and I don't have spare $1200 to buy $200 scope locally10:15
whitequarkI may have some friends with a scope though10:16
DocScrutinizer05recheck all your solder points and the connectors10:16
DocScrutinizer05honestly when you swap left and right end of poti, it MUST work in inverse way. Logarithmic potis may behave unpleasant regarding huge spreading at one end and extreme sensitivity at other end, but the baisc function can't change by swapping left and right end *only*10:19
whitequarkactually I think the cable is not symmetrical10:20
DocScrutinizer05for me however that sounds like you swapped completely and when you try to swap back, you actually swap middle with one end10:20
whitequarksee, the poti has a 3-pin molex connector on its wires. I can't mix it up completely, there are only 2 ways to plug it in10:21
whitequarkand the cable is bent in a particular way which demonstrates how it's been installed before, apart from my recordings10:21
whitequark(solder joints) rechecked, no faults I can observe. retouched them all with a bit of flux, let me check the behavior.10:22
whitequarkactually, hm, can RMA resistance lead to this weird effect? I can't see how10:22
DocScrutinizer05I told you to swap upper and lower end of poti, not to turn around connector10:22
DocScrutinizer05what's RMA?10:22
whitequarkrosin mildly activated10:22
whitequark(swap upper and lower end) ok, I'll do that, but why? obviously that won't correspond to original schematics?10:23
DocScrutinizer05sorry, I can't diagnose stuff from remote without eyes and without schematics and without a layout of PCB10:24
whitequarkoh, schematics and layout is easy, sec10:24
whitequarkDocScrutinizer05: layout: http://i.imgur.com/6Wf0Az3.jpg10:26
DocScrutinizer05hmm, and where is your poti plug?10:28
DocScrutinizer05how many pins are unused between plug and LED?10:29
whitequarkwhat do you mean?10:30
DocScrutinizer05the must be a lot of unused posts10:30
DocScrutinizer05how many of them between poti plug and LED?10:31
DocScrutinizer05sorry, this is a systematic question10:31
whitequarknone? PWM and STOP are unused, SW1 has a jumper between pins 2&3, LED is a LED, Rp1 is completely used by 3-pin poti plug.10:32
DocScrutinizer05none is correct10:32
DocScrutinizer05the plug is sym10:32
DocScrutinizer05turning it 180° mustn't change a thing10:32
DocScrutinizer05except the orientation of poti action10:32
whitequarkok, well, it does. -_-10:33
DocScrutinizer05jumper is in "lower" position, next to poti plug?10:33
whitequarknext to poti plug, yes10:34
DocScrutinizer05I have NFC10:34
whitequarkif it helps, I took a few pictures of the board: http://imgur.com/a/wWSIO10:35
DocScrutinizer05NE555 open collector output defect10:35
whitequarksorry for quality, only have a webcam10:35
whitequarkDocScrutinizer05: http://www.st.com/web/en/resource/technical/document/datasheet/CD00000479.pdf10:36
whitequark"The circuit may be triggered and reset on falling10:36
whitequarkwaveforms, and the output structure can source10:36
whitequarkor sink up to 200 mA.10:36
whitequark"source or sink" doesn't sound like OC to me10:36
whitequarkor just look at its schematics on page 210:37
whitequarkit is clearly push-pull10:37
DocScrutinizer05that's absolutely mad shit10:38
DocScrutinizer05but honestly, when your circuit doesn't work anymore when you turn poti plug 180° then poti is defect10:39
whitequarkooo, that can actually be true somehow10:40
whitequarksince 555 was killed by 110V breakdown from the transformer10:40
DocScrutinizer05a poti is an absolutely symmetric compnent, particularly when linear10:40
DocScrutinizer05there might be a short between middle and one end10:41
DocScrutinizer05or a break10:42
DocScrutinizer05what size is poti?10:42
DocScrutinizer05you got a DMM?10:42
whitequarkDMM sure10:42
DocScrutinizer05probe end to end10:42
DocScrutinizer05or can you read what's written on poti?10:43
whitequarkpoti is this: http://www.openobject.org/physicalprogramming/images/b/b5/Potentiometer.jpg10:43
whitequarkno markings on it10:43
DocScrutinizer05wrong side10:43
whitequarkok, lemme look if I can unscrew10:43
DocScrutinizer05value engraved into shielding pot10:43
DocScrutinizer05usually other side of contacts10:44
whitequarkunscrewed, it's B10K10:44
DocScrutinizer0510k, makes sense10:44
DocScrutinizer05probe if it has 10k between ends10:44
DocScrutinizer05unplug for probing10:45
whitequarkyep, unplugged. end to end is 12.2k10:45
DocScrutinizer05WTF NOW10:45
DocScrutinizer05probe from middle to one end, turn poti slowly10:45
DocScrutinizer05watch resistivity10:46
DocScrutinizer05should be ~5k in middle pos10:46
whitequarkdoing it now10:46
whitequarkeither I'm doing it wrong or values make no sense, lemme recheck10:46
DocScrutinizer052.5k at 9o'clock pos and 7.5 at 15:00 pos10:46
whitequarklooking at DMM while adjusting it. according to my DMM it's sometimes open and sometimes in 0-50k range seemingly randomly10:48
DocScrutinizer05open the shield pot (has 4 latches)10:49
DocScrutinizer05you'll find scortch marks on the carbon trace10:49
DocScrutinizer05oops, yours has 5 latches10:49
whitequark4 latches actually, the image is a bit different10:50
whitequarkanyway, opened it10:51
whitequarkthere are indeed some scratch marks, lemme photo it10:51
whitequarkoh, burn10:53
whitequarkhm, no, nothing of kind10:53
whitequarksee that trace-like things on carbon? the entire surface is like that. but no, no burns10:53
DocScrutinizer05those are from moving contact, normal10:54
DocScrutinizer05check the 3 connector rivets, closely10:54
DocScrutinizer05often when they burn, they do there10:55
whitequarkrivets seem fine10:56
DocScrutinizer05you could connect a 5V series 1kR to both outer contacts, and a LED between outer and middle10:56
DocScrutinizer05better control than a DMM10:56
DocScrutinizer05acoustic would be even better still10:56
whitequarkok, will do LED10:57
DocScrutinizer05you can connect line out to both outer, and line in of next device (amp) to outer GND and middle10:57
DocScrutinizer05should work perfectly to adjust audio volume between 0% (almost), and 100% (also almost)10:58
DocScrutinizer05without any noise10:58
whitequarkok, there's a lot of noise11:05
whitequarkappears to be shitty soldered rivets11:07
DocScrutinizer05no, they burned11:07
whitequarkshitty riveted rivets or shitty soldered contacts11:07
whitequarkwhy is stuff inside intact then?11:08
DocScrutinizer05rather the carbon trace (there actually metal, I guess) burned under the rivet11:08
whitequarkmetal, yes11:08
DocScrutinizer05it's usual failure pattern for poti which seen too high current 11:09
DocScrutinizer05the rivets are the weakest part, the contacts of them to the carbon tace element11:10
DocScrutinizer05*sometimes* you can fix by using a pair of pliers to fasten the rivet and spread it a little bit11:11
whitequarkDocScrutinizer05: http://i.imgur.com/INOf6Ya.jpg11:11
DocScrutinizer05I can't see anything. But the problem is under the metal latches11:12
DocScrutinizer05around the rivet11:12
DocScrutinizer05the carbon trace element is basically a simple PCB11:13
DocScrutinizer05coated with a special "copper" layer11:13
whitequarkI see11:16
whitequarkok, so I wanted to make it PWM anyway11:17
whitequarklet me see if that even works on the stepper control board11:17
whitequarkI need vodka or something11:26
whitequarkfor cleaning the PCBs of course11:26
whitequarkthey passed the law that you can't buy pure ethanol in >50ml quantities...11:27
wpwrakwhitequark: (flux) yes, flux very be nicely conductive. easily a few ten kOhm, maybe even more11:31
wpwrakwhitequark: i found out about that then i put a nice external 100 kOhm pull-up on an nRESET input of atusb. when i tried to turn on the still flux-loaded board, the damn chip wouldn't come out of reset ... took me a while to realize that the chip was perfectly healthy ...11:33
whitequarkwpwrak: what flux wsa that?11:35
wpwraki think that was with a "water soluble" flux. but i guess they all can do that to a larger or smaller extent.11:38
whitequark"water soluble" ones are apparently much worse at this...11:38
whitequarkthat makes total sense, they must be polar if they need to be soluble in water11:39
whitequark=> mobile charge carriers11:39
whitequarkthis RMA residue doesn't have any resistance my DMM can measure11:40
wpwrak(polar) yes, but there may be other fun things in the formula11:41
wpwrakbtw, i found a ""No Clean  Water Washable  Tack Flux": Chipquik SMD4300TF10. at digi-key as SMD4300TF10-ND11:43
wpwrakworks quite nicely. but a bit too tacky - you feel dirty working with it rather quickly11:43
whitequarkok, I think I'm dumb12:12
whitequarkthis board has a component installed, it's a DC-DC converter12:12
whitequarkfrom 5VDC to 5VDC12:12
whitequarkoh, it's isolating, nevermind12:12
whitequarkcool, the board has properly populated and connected ports for limit switches, probe, pwm and charge pump12:22
whitequarkoh and spindle direction12:22
whitequarkawesome. I now have a working machine and it even can do more than before13:20
whitequarkat 24V my adjustable PSU has enough beef to drive 2 steppers and spindle at 100%13:20
whitequarkand I now even have control of spindle via EMC2!13:22
wpwrakin a way, they sent you exactly what you really wanted - a CNC mill you could modify at your heart's content without any remorse13:36
whitequarkwpwrak: by the way, do you have any idea how to measure spindle speed?14:02
whitequarkto calibrate it14:03
rohwhitequark: optocouple14:05
whitequarkroh: gotcha14:08
rohwpwrak: there are no unmodded mills... everybody mods14:42
rohnice you got speed control. we only have a pot on a box14:42
rohbut the motor also eats 800W dc at something between 100 and 250V14:43
whitequarkroh: (pot on a box) I had it too. but turns out that motor PCB has PWM in, and control PCB has PWM out14:55
whitequarkso it was a matter of two wires and correct emc2 config14:56
rohwhitequark: yeah. nice.15:37
qi-bot[commit] Werner Almesberger: ircstat/mlabs: not very smart script to download monthly log of #m-labs (master) http://qi-hw.com/p/wernermisc/62bd73e22:51
qi-bot[commit] Werner Almesberger: ircstat/README: change date format from MMYY to YYMM (master) http://qi-hw.com/p/wernermisc/ebb0d2722:51
qi-bot[commit] Werner Almesberger: ircstat/mlstat, stat: add the new year (2014) (master) http://qi-hw.com/p/wernermisc/037d65c22:51
qi-bot[commit] Werner Almesberger: ircstat/ML: update for 2014-01 (master) http://qi-hw.com/p/wernermisc/57b53c622:51
whitequarkwpwrak: seq -f '%02.f' 1 1222:54
DocScrutinizer05whitequark: (calibrate spindle speed) use a magnet in spindle - as "tool". and a coil you get near to the rotating magnet. Then listen to the audio output of that coil and compare to your piano ;-)22:54
whitequarkDocScrutinizer05: I actually rather like roh's suggestion22:55
whitequarkbecause it means emc2 will maintain spindle speed regardless of load22:55
DocScrutinizer05err that suggestion is basically identical, just it uses more complicated sensor22:55
DocScrutinizer05oooh, you want to permanently control spindle speed, not calibrate it?22:56
whitequarkI actually think measuring it permanently is better22:57
DocScrutinizer05then go for the principle "scope across the 2 pcs 1R on" 22:58
DocScrutinizer05though I think a feedback loop to regulate spindle speed is massive overkill22:58
whitequarkthat feedback look is basically free22:59
whitequarkbecause emc2 already implements it22:59
DocScrutinizer05nfc what's emc222:59
whitequarkaka linuxcnc22:59
DocScrutinizer05oh lol, no, that won't fly23:00
DocScrutinizer05implementing a PID-regulator in software is sort of.... a PITA, given your hw limitations23:00
DocScrutinizer05your loop turnaround time is way too high23:01
DocScrutinizer05IOW the phaseshift is way too huge23:01
DocScrutinizer05and it's definitely not 2for free"23:02
DocScrutinizer05you can implement a dynamic calibration23:02
whitequarkwait, why is turnaround time too high?23:02
DocScrutinizer05which is sort of a PID with ultralong integral and zero differential component23:03
wpwrak(seq) duh. new-fangled stuff ...23:03
DocScrutinizer05anyway I'd use voltage from electromotoric effect of spindle motor during PWM power off periods to conclude the current spindle speed23:05
DocScrutinizer05or, even better, the inevitable noise in motor current from the motor switching coils while spinning23:07
DocScrutinizer05actually some automotive components (cassette player, radio) use same effect from car generator to get engine RPM and thus background noise and adjust volume automatically23:09
DocScrutinizer05in the end, your spindle turns at the RPM you've set it to via your PWM23:10
DocScrutinizer05you don't really need any regulator loop23:10
whitequarkthat may well be true23:10
--- Sun Feb 2 201400:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!