#garfield Logs
Jan 18 2022
#garfield Calendar
01:38 AM rue_bed: its... not 1
01:46 AM rue_bed: hmm
09:17 AM Tom_L: and of course the bot died again
09:18 AM skunkworks: granted I have to offset the ttl square wave to go +/- - but that is easy.
09:20 AM rue_mohr: but
09:20 AM Tom_L: don't be late...
09:20 AM rue_mohr: if its not analog, why use op-amps?
09:20 AM skunkworks: I have tried this circuit was some generic op-amps (with a pretty high output impedence) and it seems to work..
09:20 AM rue_mohr: its 0-6V?
09:21 AM skunkworks: +/- 6b
09:21 AM skunkworks: 6v
09:21 AM rue_mohr: k
09:21 AM skunkworks: basically using the op-amp as a comparator
09:21 AM skunkworks: (me being lazy)
09:21 AM skunkworks: and I only need 3 of these circuits ;)
09:22 AM rue_mohr: how much current does this draw....
09:22 AM rue_mohr: 22R
09:22 AM skunkworks: I have these coming
09:22 AM skunkworks: https://www.digikey.com/htmldatasheets/production/7437/0/0/1/l272.html?utm_adgroup=STMicroelectronics&utm_source=google&utm_medium=cpc&utm_campaign=Dynamic%20Search_EN_Focus%20Suppliers&utm_term=&utm_content=STMicroelectronics&gclid=CjwKCAiA55mPBhBOEiwANmzoQk_TAk-4vg0BlOPb4TDT4m1jfBbIbmO8-q9laj3I9ugXUTNOJfEm8hoCb8QQAvD_BwE
09:22 AM skunkworks: the coils are about 100ohms
09:22 AM skunkworks: not high current
09:23 AM rue_mohr: ah
09:23 AM rue_mohr: between 500 and 22mA
09:23 AM skunkworks: the opamps i am currently playing with have about 200ohm output impedence.
09:25 AM skunkworks: it is the position feedback for this machine. Alwasy wanted to get the scales working..
09:25 AM skunkworks: https://electronicsam.com/images/KandT/oldkandt.JPG
09:25 AM rue_mohr: heh
09:25 AM rue_mohr: thats a lot of space
09:26 AM rue_mohr: you know any pnp or npn transitors should work fine there eh?
09:27 AM rue_mohr: it all seems to be digital
09:27 AM skunkworks: https://photos.app.goo.gl/FFNTuU4SZwfRTwjf8
09:27 AM rue_mohr: pn2222 and 2N3906
09:27 AM Tom_L: painted all purty!
09:27 AM skunkworks: sure.. I just didn
09:27 AM skunkworks: sure - I just didn't want to deal with any more circuit than I had to.. Getting lazy in my old age..
09:28 AM skunkworks: If the op-amps don't work - I will certainly try the transistor H-bridge.
09:28 AM rue_mohr: well
09:28 AM skunkworks: I just didn't know if they were doing something fancy with the diode - as you initally thought - to float
09:28 AM rue_mohr: whats the accupin head look like?
09:30 AM skunkworks: https://electronicsam.com/images/KandT/conversion/accpinset1.jpg
09:30 AM skunkworks: not the greatest picture
09:30 AM skunkworks: the top is the slot the pins go in
09:30 AM skunkworks: plus it is only 250hz
09:31 AM rue_mohr: ... are those programming pins?
09:31 AM skunkworks: no - those are what the head reads for position.. each pin is .1" what gets returned from the circut is a signal offset based on it's position within the .1" pin
09:32 AM skunkworks: kinda a resolver in reverse.
09:32 AM skunkworks: the head moves along the pins.
09:33 AM skunkworks: https://electronicsam.com/images/KandT/conversion/accupinlineup.jpg
09:34 AM rue_mohr: and it stops when it gets to the pin thats within a fixed distance?
09:35 AM skunkworks: kinda - it was just a counter. It used a 200khz signal to 'measure' the distance from the injected signal and the return signal. So it would divide each pin into 1000 counts.
09:35 AM skunkworks: (or .0001")
09:36 AM skunkworks: it would keep track of each time it passes a pin and count up. so first pin would be 0-.099 - next pin would be .1 to .199 and so on
09:36 AM skunkworks: inches
09:36 AM rue_mohr: 20 and 78 in increments of two...
09:36 AM skunkworks: *250khz
09:37 AM skunkworks: yah - I think that was a fine adjustment of the quadurature signal?
09:38 AM rue_mohr: 250khz sine wave with an offset proportional to position
09:38 AM skunkworks: the 2 h-bridges get driven by 2 sets of square waves - one set being 90deg offset
09:39 AM skunkworks: The operaction section explains it pretty well.
09:39 AM rue_mohr: yea 90 degrees for the 2 phase
09:40 AM rue_mohr: what if all this could be put on an arduino?
09:40 AM rue_mohr: :)
09:41 AM skunkworks: that is the plan
09:41 AM skunkworks: I am outputing the quadurature from the arduino currently - then use the 16bit timer to do the measuring
09:42 AM rue_mohr: ok
09:42 AM skunkworks: (I am really winging it here)
09:42 AM skunkworks: https://patents.google.com/patent/US3491277A/en
09:42 AM skunkworks: that has more info...
09:42 AM rue_mohr: I dont have time to dig in now
09:42 AM skunkworks: not a problem - thanks for the help
09:42 AM rue_mohr: and I'v got myself burried in a sensorless motor positioning project
09:43 AM skunkworks: Oh - fun!
09:43 AM rue_mohr: you have an interesting project tho
09:43 AM skunkworks: I will keep you guys updated
09:43 AM rue_mohr: I love the old hybrid analog/digital things
09:43 AM rue_mohr: Tom_L, I'm on a bit of a different schedule today
09:43 AM skunkworks: the blueprints are dated aroun 62'
09:44 AM rue_mohr: :) I cant make out any of the years on them
09:45 AM skunkworks: they suck.. But glad we have what we have.
09:46 AM skunkworks: this is the balancing part I needed - the rest I think will be op-amps.
09:46 AM skunkworks: https://photos.app.goo.gl/c5dJLJ19AjreH6Mt6
09:48 AM rue_mohr: yup
09:53 AM Tom_L: well then you needed a morning project anyway
09:54 AM rue_mohr: need to come up with a PD loop for the motor EMF control
09:55 AM rue_mohr: or an integrator for the position estimate
09:55 AM rue_mohr: 1/2hr drive for an appt at 9:30
09:55 AM rue_mohr: 7:40 now
09:56 AM Tom_L: i gotta go fix a sump
09:56 AM rue_mohr: I should make sur the truck starts
10:02 AM Tom_L: yours or the companys?
10:08 AM Tom_L: k, i'm off to waller in the sump
10:10 AM Tom_L: need to add a deflector to the pipe discharge
10:11 AM Tom_L: the new crapper is a 'fast' flush thing that dumps the water all at once and causes the waste to fly out the pipe further landing on the sump float then the pump short cycles
10:11 AM Tom_L: so i'm gonna add a 45 to the discharge
11:15 AM Tom_L: ok hopefully the Sh** will run uphill smoothly once again..
01:37 PM polprog: the thing im designing next semester
01:37 PM polprog: a network controlled power distribution unit
01:37 PM polprog: considering whether the mains cords should be on the front or back
01:38 PM polprog: or maybe just the input on the back and 4 outputs on the front
01:38 PM polprog: or everything in the back
01:38 PM polprog: I want to put an RCBO on each output too
01:39 PM Tom_L: rack mount rear mains
01:39 PM polprog: input mains or all mains
01:39 PM polprog: it will be rack mount
01:39 PM Tom_L: your design dude
01:39 PM Tom_L: :)
01:39 PM polprog: :D
01:40 PM * polprog designs a tetrahedron with 1 socket on each wall
01:41 PM Tom_L: see i knew you'd figure it out!
01:42 PM Tom_L: so did you pass or fail?
01:42 PM polprog: pass what?
01:42 PM polprog: exams didnt start yet
01:43 PM polprog: just some tests from minor subjects
01:43 PM Tom_L: oh
01:43 PM polprog: i had material science basics today and LabView test
01:43 PM Tom_L: but they become major if you don't pass em
01:43 PM polprog: hehe
01:43 PM polprog: labview test was the most stupid assignment ive seen to date
01:43 PM polprog: questions like "what is the default value of an integer variable"
01:44 PM polprog: (in lab view a variable is just a square.. with the value in it)
01:44 PM polprog: so unlike a normal programming language. you create it and the default value is inside the square, its zero
01:44 PM polprog: 7 pages of crap like that
01:44 PM Tom_L: fun
01:44 PM polprog: the only sensible question was karnaugh table, that i actually forgot how to do
01:44 PM polprog: since i lerant it last year i never had to use it
01:44 PM Tom_L: real "need to know" stuff
01:45 PM polprog: yeah, amazing 30 year old tool
01:45 PM polprog: ill see if i passed that, if not then its gonna be fun
01:45 PM rue_mohr: polprog, how did the test go?
01:46 PM polprog: the material science was OK,
01:46 PM polprog: the labview one.. i knew some of the answers
01:46 PM polprog: but its like, you know. you don't learn these things by heart. it was mostly gut or experience, which could be wrong
01:47 PM rue_mohr: k
01:47 PM polprog: material science subject was pretty chill, i had it as an extra, it's a subject for 1st year chemistry students
01:47 PM polprog: did you know that berylium brass does not spark? at all?
01:48 PM polprog: i can see youre enjoying the discussion about branchless fizzbuzz :v
01:48 PM rue_mohr: oh, his goal is branchless...
01:48 PM aandrew: one of my all-time favourite books is Ivan Amato's Stuff
01:48 PM polprog: ... yeah
01:48 PM rue_mohr: huh
01:48 PM polprog: her, i think, even
01:48 PM rue_mohr: oh, didn't dig
01:48 PM aandrew: I picked it up on a whim, and could not put it down. It's all about the history of materials science
01:49 PM polprog: interesting :o
01:49 PM aandrew: yeah it is. I can't find an electronic version though or I'd happily share it
01:49 PM polprog: i found my book on metallogeny yesterday, someone was talking about unique books
01:49 PM polprog: it's ore formation science
01:51 PM rue_mohr: why branchless?
01:51 PM rue_mohr: thats evil
01:52 PM polprog: no branches = no branch predictors to do shady stuff
01:52 PM rue_mohr: huh
01:53 PM polprog: though i would not call that truly branchless, because it is convoluted to calculate cycles
01:53 PM polprog: yes they are using a jumptable to call procedures
01:53 PM polprog: but ive seen algorhitms that are branchless *and* jumpless
01:53 PM polprog: which makes the execution time very, very predictable
01:53 PM polprog: thats not the case here
01:54 PM polprog: wait, no
01:54 PM polprog: its just the char* thats being looked up
01:54 PM polprog: thats even cooler then
02:00 PM rue_mohr: hmm
07:45 PM Tom_L: so how was the day?
07:45 PM Tom_L: seemed it started off a bit bumpy
07:51 PM rue_mohr: yea, I'm eating and heading to bed soon
07:52 PM rue_mohr: see if I can sleep till 1
07:53 PM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/temp.png
07:53 PM rue_mohr: check this out
07:53 PM rue_mohr: its the EMF of the motor after the pwm pulse
07:53 PM rue_mohr: I caught that one cause its got a commutation spike
07:53 PM rue_mohr: I think I have a plan on filtering it
07:54 PM rue_mohr: I can grab up to about 15 adc samples over where it settles
07:54 PM rue_mohr: if I grab 8, track the min and max, and throw out the whole reading if its more than about 20 difference
07:54 PM rue_mohr: they dont really happen that often
07:55 PM rue_mohr: what I also see, is that it looks like the windings are different
07:55 PM rue_mohr: there is a jump in the EMF level
07:55 PM rue_mohr: the two flats should be the same level
07:56 PM Tom_L: green is the spike?
07:57 PM Tom_L: what's white?
07:57 PM Tom_L: hmm
07:58 PM rue_mohr: at, white is pre-amplifier
07:58 PM rue_mohr: green is post amplifier
07:58 PM rue_mohr: ~35x
07:59 PM Tom_L: ok
07:59 PM rue_mohr: so, yea, that spike is it switching armature coils
08:00 PM Tom_L: ahh i see
08:01 PM rue_mohr: not sure how to time the ADC readings
08:01 PM Tom_L: i wondered what exactly about it caused the spike
08:01 PM Tom_L: anything that would cause the field to collapse i suppose
08:01 PM rue_mohr: its the brushes briefly lapping over both commutators
08:02 PM rue_mohr: its only a 3 pole motor, so its pretty prounounced
08:07 PM rue_mohr: I have all sorts of caps
08:07 PM rue_mohr: but ponder this
08:08 PM rue_mohr: this is a 3V motor, running at about 60-100rpm
08:08 PM rue_mohr: at 1.5V id does a few thousand rpm
08:08 PM rue_mohr: these voltages are really tiny
08:10 PM rue_mohr: the pwm is 5V, at about 9%
08:10 PM rue_mohr: so drive voltage is on average .45V
08:10 PM rue_mohr: the motors backemf is about 70mv
08:10 PM Tom_L: gotten a bit livelier last few days it seems
08:11 PM rue_mohr: see what this schedule shift does for me
08:11 PM Tom_L: heh
08:11 PM rue_mohr: I should be tired when I need to sleep now
08:11 PM Tom_L: i'll be up at 4:30 my time
08:11 PM Tom_L: 2:30 for you?
08:11 PM rue_mohr: I'll be up 1am or earlier
08:11 PM Tom_L: earlier if today was an indication
08:12 PM rue_mohr: hope that was a flitch
08:12 PM rue_mohr: glitch
08:27 PM rue_mohr: hmm
08:27 PM rue_mohr: there dont seem to be flags I can watch to know if the pwm is on
08:29 PM rue_mohr: I may have to try to synchronize another counter to trigger the start of the adc readings
09:36 PM skunkworks: sorry not sorry... https://photos.app.goo.gl/oinAJPMnd5tSbMW88
09:38 PM skunkworks: rue_mohr: random - have you looked at this? I don't understand the code but what it does is amazing..
09:39 PM skunkworks: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ResolverToQuadratureConverter
09:42 PM Tom_L: he'll likely catch it in the morning. trying out a new sleep schedule
09:45 PM Tom_L: interesting idea though
09:46 PM skunkworks: hey - rue has some stuff on hackaday
09:46 PM Tom_L: several
09:47 PM Tom_L: i vaguely see what they're doing but i haven't coded avr in quite a while and never did much arduino speak
09:47 PM Tom_L: did all mine in c or asm
09:48 PM skunkworks: Yah - I don't understand all the bit shifting when setting registers.
09:49 PM Tom_L: i'm sure he can explain it
09:49 PM skunkworks: Seems all to flashy.. 'Get off my lawn'
09:49 PM Tom_L: getting the phases to align at the right time for the quad pulse
09:49 PM skunkworks: That was written by andy - I always wondered how his brain works
09:50 PM Tom_L: heh i still wonder
09:51 PM Tom_L: i think all the sbi cbi are asm commands rather than arduino
09:51 PM Tom_L: it has to be fast
09:51 PM skunkworks: You know I am 40 something and I just learned that the 'jack of all trades, master of none' has more to it..
09:52 PM skunkworks: Yah - I mean - I have been getting more familliar with it.. But it is painful for me coming from mostly VBA and a little C
09:52 PM Tom_L: i'm self taught and haven't done any in quite some time now
09:53 PM skunkworks: Me too - computers was my hobby...
09:53 PM Tom_L: started cutting back on things and chose mechanical over electronic
09:53 PM skunkworks: now I am a IT manager of a small company.. My degree is in electronics.
09:53 PM Tom_L: he'll be up around 1 est or a bit sooner
09:53 PM skunkworks: Jack of all.....
09:54 PM Tom_L: i hear ya
09:54 PM skunkworks: life is weird.
09:55 PM Tom_L: we did projects all the time in 'the good ole days'
09:55 PM Tom_L: now it's kinda thinning out
09:56 PM Tom_L: i'm closer to jt's age
09:57 PM skunkworks: I remember you showing up to the witchita fest but I don't think I actually introduced my self to you. I remember you not being there very long.
09:58 PM skunkworks: but then again - that was a long time ago now
10:00 PM Tom_L: no i didn't have alot of time and wasn't sure what i was getting into :)
10:00 PM Tom_L: stuart showed me around a bit which was great
10:01 PM skunkworks: That is what I remember. - stuart and kim stopped at dads shop when he was travelling through. Few years ago
10:01 PM Tom_L: i heard the cincinatti wasn't there now
10:01 PM skunkworks: What did you 'Think' you were getting into?
10:01 PM Tom_L: wasn't really sure to be honest
10:02 PM Tom_L: i generally work alone or online
10:02 PM Tom_L: bad hearing doesn't help
10:02 PM Tom_L: i spoke to andy a bit
10:03 PM Tom_L: i think i met jeff & seb
10:04 PM skunkworks: it was fun.. good memories.
10:04 PM Tom_L: for sure
10:06 PM skunkworks: we had started going when it was down in galeburg IL
10:07 PM skunkworks: https://electronicsam.com/images/KandT/cncworkshop/cncworkshopemc.jpg
10:08 PM Tom_L: i don't think i'd actually started using lcnc at that point or had just started but i've followed it a long time
10:08 PM Tom_L: i recognize a couple faces
10:09 PM Tom_L: i think :)
10:09 PM skunkworks: Lol - some exscape me
10:10 PM Tom_L: is the guy in white the pico guy?
10:10 PM skunkworks: yes - Jon E
10:11 PM Tom_L: dunno the big guy left rear gray t but he was in wichita
10:11 PM Tom_L: as was the right side black tee
10:11 PM Tom_L: pretty sure anyway
10:12 PM Tom_L: was that at your place?
10:14 PM skunkworks: Left to right steve stallings, Matt shaver, JMK, jepler, Uhhh - I know him, Joh E , Steven Uhh , dave engvall, Don't know, Cradek
10:14 PM skunkworks: No, it was in galesburg IL, Can't remember his name either
10:15 PM skunkworks: there he is - Roland https://electronicsam.com/images/KandT/cncworkshop/roland.JPG
10:15 PM skunkworks: he somehow was involved or started what would be centroid? I don't remember the whole story. And it might be wrong..
10:23 PM Tom_L: huh
10:24 PM Tom_L: yeah i recognized jepler and pretty sure cradek
10:24 PM Tom_L: black tee?
10:24 PM Tom_L: matt was in wichita
10:25 PM Tom_L: i left before peter got there i think
10:25 PM skunkworks: Peter - I got to meet peter!
10:25 PM Tom_L: smart guy
10:25 PM skunkworks: Very
10:25 PM skunkworks: You knew John K?
10:26 PM Tom_L: not sure
10:26 PM Tom_L: i don't think so
10:26 PM skunkworks: Another very guy. Worked for rockwell automation making high power drives..
10:26 PM skunkworks: Like 50hp stuff.
10:26 PM skunkworks: He is the one who wrote HA:
10:26 PM skunkworks: HAL
10:27 PM Tom_L: oh?
10:27 PM Tom_L: nice
10:27 PM skunkworks: *Another very smart guy'
10:27 PM Tom_L: too bad it doesn't rub off...
10:27 PM skunkworks: Right?!
10:27 PM skunkworks: Little bits do...
10:28 PM skunkworks: I have certainly learned alot.
10:28 PM Tom_L: i think i felt a little intimidated walking thru the door
10:29 PM skunkworks: Yes - a lot of brain trust.. But everyone was very nice.
10:29 PM Tom_L: agree
10:29 PM Tom_L: alot to absorb
10:30 PM Tom_L: i've learned quite a bit since then i think
10:31 PM skunkworks: Steve Stalling (pmdx) would always get upset when we said we were thinking of cnc'ing our monarch 10EE..
10:31 PM Tom_L: heh
10:32 PM Tom_L: so where are you in the pic?
10:32 PM Tom_L: hiding behind the camera?
10:32 PM skunkworks: Hmm - I think dad took that picture..
10:32 PM skunkworks: He always had a nice camera..
10:32 PM skunkworks: *has
10:33 PM skunkworks: https://electronicsam.com/images/KandT/cncworkshop/memazak.JPG
10:34 PM skunkworks: wow - I look young...
10:34 PM Tom_L: hah
10:34 PM Tom_L: i'm sure we all did
10:34 PM skunkworks: atleast 30lbs lighter
10:34 PM skunkworks: 40 maybe
10:34 PM Tom_L: married with kids...
10:35 PM skunkworks: I don't think so.. Married maybe.. No kids yet.
10:35 PM Tom_L: luckily? mine are close to leaving the nest
10:35 PM Tom_L: both in their 20's
10:36 PM skunkworks: How are they doing?
10:36 PM Tom_L: oldest lost his job 2x from covid and went back to learn hvac. just finished and has an interview thurs
10:37 PM Tom_L: youngest is still at his first 'real' job
10:37 PM Tom_L: they both were there for a while
10:37 PM skunkworks: Ugh - and good! now you need to talk him into installing geo-thermal..
10:38 PM Tom_L: not alot of that around here that i know of
10:39 PM skunkworks: I have to get some sleep - thanks for the invite here. Sorry for filling it up with fluff.
10:39 PM Tom_L: stick around..
10:39 PM Tom_L: ttyl
10:39 PM Tom_L: we don't care. it's pretty random here