#garfield Logs

Dec 18 2020

06:55 AM polprog: linear CCDs are here
07:02 AM polprog: :)
11:42 AM polprog: picked up the kit from the uni
11:42 AM polprog: some nice stuff. Analog Discovery 2
11:42 AM polprog: jelly bean parts
11:42 AM polprog: pretty different opamps than what I have
11:43 AM polprog: OP07, TL084, TL082, ua741
11:43 AM polprog: and a 358 but ive got a couple of them
11:44 AM Tom_L: i'd rather have skittles instead of jelly beans
11:50 AM polprog: hmm
11:50 AM polprog: https://pdf1.alldatasheet.com/datasheet-pdf/view/22767/STMICROELECTRONICS/LM334.html
11:50 AM polprog: bottom of page 7, interesting RC filter feeding the opamp
11:51 AM polprog: unless thats a regular RC offset for parasitic components or something
11:56 AM aandrew: hah. when the extruder is actually clamping the filament correctly things work a lot better.
11:56 AM aandrew: the gear that grabs the filament is fine, but the clamping mechanism tightened up over time causing the idler to not keep consistent pressure holding the filament against the gear
11:56 AM aandrew: I've been spending a lot of time screwing with retraction settings trying to fix poor extrusion that seemed to occur mostly around retractions
11:56 AM aandrew: now when it's printing a flat "sheet" it's VERY nice (I was compensating with 105% flow and a slightly overstruded steps/mm on the extruder).
11:56 AM aandrew: now I'm dicking around with flow (down to 95% now and retraction tuning but they're looking considerably better
03:14 PM aandrew: ooh found this for rue: https://flux.ai/jharwinbarrozo/breadboard-wrist-watch
05:38 PM rue_mohr: ok
05:38 PM rue_mohr: well I think I passed it
05:38 PM rue_mohr: guess I'll see
05:56 PM Tom_L: yay
05:57 PM rue_mohr: polprog, what is the i2c?
05:57 PM rue_mohr: er spi
05:57 PM Tom_L: so does this mean you're all done?
06:04 PM rue_mohr: I need to wait a few days for the marks to come bak
06:42 PM polprog: rue_mohr: a LoRa modem and radio IC
06:42 PM polprog: i got it working tho
06:42 PM polprog: ^_^
06:43 PM polprog: now i gotta get it working with the peripheral SPi which is what ive been trying to do for a month
06:43 PM rue_mohr: ah
06:45 PM polprog: i measured the pinout of the laser supply
06:45 PM polprog: i think i have a matching pinout list
06:48 PM rue_mohr: so
06:48 PM rue_mohr: I want to try to make conductive paint
06:48 PM polprog: oh
06:48 PM rue_mohr: SCrarry enough
06:48 PM polprog: the paint?
06:49 PM rue_mohr: the internet recipe that sounds the best is an acrylic paint base with the guts of a carbon battery poured into it
06:49 PM polprog: lol
06:49 PM rue_mohr: what is the electrolye of a zink carbon battery?
06:49 PM Tom_L: what about graphite
06:50 PM rue_mohr: manganese dioxide
06:50 PM Tom_L: aluminum powder
06:50 PM rue_mohr: apparently tyhe electrolite works the best
06:50 PM rue_mohr: I was watching some youtube vid, not aht I trust it
06:51 PM rue_mohr: I'll give it a try
08:13 PM aandrew: well, 12 test prints later and I think I'm close to a set of values I'm happy with. I'm printing at 100mm/sec with good quality and reasonable dimensionality. I think I can tune it even better but I'm in the diminising returns now
08:26 PM rue_mohr: k
08:36 PM aandrew: it's funny; when you told me you had your flow turned down a bit and my flow was up a bit I decided to recalibrate e-steps and discovered that the retention clamp for the extruder had worked its way tighter
08:36 PM aandrew: bit of dry lube and loosening it off should help it stay I hope
09:14 PM rue_shop2: what did I come here to do again?
09:14 PM Tom_L: chat?
09:15 PM Tom_L: ran a steel part yesterday
09:15 PM rue_shop2: oooo
09:15 PM rue_shop2: did it bog at all?
09:15 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Czech_part.jpg
09:15 PM Tom_L: not at all
09:15 PM rue_shop2: nice
09:16 PM rue_shop2: did you do it light duty or did you just give 'er?
09:16 PM Tom_L: full width .250 em ~ .063 deep
09:16 PM rue_shop2: but hand countersunk?
09:16 PM Tom_L: no
09:16 PM rue_shop2: hmm
09:16 PM Tom_L: all on the mill
09:17 PM rue_shop2: so, the countersinks look a bit off?
09:17 PM rue_shop2: what are they parts for?
09:17 PM Tom_L: maybe a bit of chatter on one
09:18 PM Tom_L: front hinge on a piano
09:18 PM rue_shop2: ooh
09:18 PM rue_shop2: which ones did you make? all?
09:18 PM Tom_L: you can't tell?
09:18 PM Tom_L: :)
09:18 PM Tom_L: the bottom one
09:18 PM rue_shop2: I see what looks like a stress mark on the top one
09:18 PM rue_shop2: yea
09:18 PM Tom_L: the others are cast
09:18 PM rue_shop2: thats what I was starting to suspect
09:19 PM rue_shop2: but the bottom one has a cast look to it too
09:19 PM Tom_L: it's just that top crust layer on steel
09:19 PM Tom_L: what's it called?
09:19 PM rue_shop2: slag :)
09:19 PM Tom_L: probably
09:19 PM rue_shop2: k
09:19 PM rue_shop2: cool stuff
09:20 PM rue_shop2: I have to break my mind out of school mode
09:20 PM Tom_L: i was in a hurry so i squared it up manually on the cnc
09:21 PM rue_shop2: well, "2 days in vancouver"
09:21 PM rue_shop2: I sure hope I passed that exam
09:22 PM Tom_L: how's van?
09:22 PM rue_shop2: ugh
09:22 PM rue_shop2: with masks
09:22 PM Tom_L: i'm gettin tired of that
09:22 PM rue_shop2: aka ugh+masks
09:22 PM rue_shop2: its normally just ugh
09:22 PM rue_shop2: every all high strung and snotty
09:22 PM rue_shop2: we went to a big mall
09:23 PM rue_shop2: it was just about empty
09:23 PM rue_shop2: and I'm not talking about people
09:23 PM rue_shop2: I'm talking about stores
09:23 PM Tom_L: huh
09:23 PM rue_shop2: it had about 1/3 active stores in it
09:23 PM rue_shop2: the rest were borded up
09:23 PM Tom_L: yeah we still have some bare shelves
09:23 PM rue_shop2: and empty, like not comming back
09:23 PM Tom_L: yep
09:24 PM Tom_L: quite a few things here have closed
09:24 PM rue_shop2: will *this* trigger the reccesion?
09:24 PM rue_shop2: this cant be artificially propped up forever
09:24 PM Tom_L: first my electronics guy closed a while back now my tool guy is getting really old and shakey
09:25 PM Tom_L: i'm worried he may not make it long
09:25 PM rue_shop2: take on the store :)
09:25 PM Tom_L: no
09:25 PM rue_shop2: hahah
09:25 PM Tom_L: he literally had to hold his hand still to use the keyboard with 1 finger
09:25 PM Tom_L: felt kinda bad for him
09:25 PM Tom_L: never seen him that bad off
09:26 PM rue_shop2: oh dear
09:26 PM rue_shop2: time, time, time,
09:27 PM Tom_L: it sucks when alot of your friends aren't around anymore
09:27 PM rue_shop2: I need to get the position sensors going, make pressure sensors with them, then get the double acting muscles going so I can get a robot body assembled
09:28 PM Tom_L: one in the robotics club we used to have went last year
09:28 PM rue_shop2: work out the rest in a hurry
09:28 PM rue_shop2: its important to have a steady intake of new faces
09:28 PM rue_shop2: its too bad its just not that easy
09:28 PM Tom_L: i don't really need friends, i need sources
09:29 PM rue_shop2: hah
09:29 PM rue_shop2: or both?
09:29 PM Tom_L: meh
09:29 PM Tom_L: they just want stuff
09:29 PM rue_shop2: I have to work hard and keep my energy up
09:30 PM rue_shop2: or I'm gonna get tided over
09:30 PM rue_shop2: it sucks looking back and realizing what you have been at the forefront of and not known
09:30 PM rue_shop2: when was mach1 released?
09:30 PM Tom_L: no idea
09:31 PM Tom_L: i remember lcnc back when it was still a govt project though
09:32 PM rue_shop2: I wrote my cnc software in 1996 and was going to write cad software for it
09:32 PM Tom_L: one of my kids is going back for HVAC since he's been layed off 2x due to the covid
09:33 PM rue_shop2: if I had, I could have been the guy authoring the software all the hobbists used
09:33 PM Tom_L: good to have a backup nowdays
09:33 PM rue_shop2: ouch
09:33 PM rue_shop2: from cnc cadcam?
09:33 PM Tom_L: yeah
09:33 PM rue_shop2: ouch
09:34 PM Tom_L: this town goes thru that anyway but it's alot worse
09:34 PM Tom_L: we still produce more aircraft here than anywhere in the world
09:34 PM rue_shop2: 4 in the last year? :)
09:35 PM Tom_L: and have the biggest pool of talent available anywhere
09:35 PM rue_shop2: There is a big shift comming up, I'm really curious how its going to go
09:35 PM Tom_L: they said something about supersonic planes in the news but i didn't catch the full story
09:36 PM rue_shop2: its awing how much tech north america relies on, that it cant produce
09:36 PM Tom_L: we can, we just can't afford it when we do so we outsource it
09:36 PM Tom_L: just like everybody else
09:37 PM rue_shop2: no, once its been outsourced this long, nobody has the knowledge, tools, or materials
09:37 PM rue_shop2: we have a paper mill here that makes pulp
09:37 PM rue_shop2: they shut down the paper line quite a few years ago
09:37 PM rue_shop2: nobody knows how it works, it'll never be started back up again
09:37 PM Tom_L: probably not TP though :)
09:37 PM rue_shop2: no
09:37 PM rue_shop2: it was newspaper
09:38 PM rue_shop2: rolls about 80 feet wide
09:47 PM Tom_L: hmm
09:47 PM Tom_L: i remember going thru the paper here once and seeing the big rolls
09:50 PM Tom_L: https://www.ksn.com/news/local/supersonic-planes-to-fly-over-kansas-on-invisible-racetrack-seven-miles-high/
09:59 PM Tom_L: 380 planes from boeing alone
10:00 PM Tom_L: 485 cessna
10:32 PM rue_shop2: :)
11:02 PM rue_shop1: ok
11:02 PM rue_shop1: so that position sensor,
11:02 PM rue_shop1: I cant do it with a 4-20mA interface
11:02 PM rue_shop1: cause it pulls 25mA
11:03 PM rue_shop1: It'll be due to my class A 8Mhz amp
11:04 PM rue_shop1: hahaha
11:04 PM aandrew: rue_shop1: just drive it from 48V, that'll get you down to 4mA :-)
11:04 PM rue_shop1: without that amp enabled, it pulls 4.4mA
11:04 PM rue_shop1: ok
11:05 PM rue_shop1: so, I should shoot for serial interface
11:05 PM rue_shop1: hmm
11:05 PM rue_shop1: I still want to not have to have 1 ADC chip per device
11:06 PM rue_shop1: I can rake a bunch otogethor with a tiny26 but
11:06 PM rue_shop1: the signals dont carry well
11:06 PM rue_shop1: one of them being an 8Mhz square wave and the other being a 1.4V DC voltage
11:07 PM rue_shop1: maybe if I convert the 1.4V to a freq
11:07 PM aandrew: can you do v-F
11:08 PM aandrew: hell you can do that with a 555 I think
11:08 PM rue_shop1: but thats hard to do without picking up drift
11:08 PM aandrew: fair
11:08 PM aandrew: you can do a better v/f without a 555 and pick up less drift
11:08 PM aandrew: I wonder if you can do some 1-bit ADC
11:08 PM aandrew: yes
11:09 PM rue_shop1: and I already know there are no single chip ADCs with the price range I'm trying to hit
11:09 PM aandrew: you can feed a ref voltage or PWM *to* the sensors that drive one input of a comparator
11:09 PM aandrew: the ohter input is the output of the LVDT
11:09 PM rue_shop1: hmm
11:09 PM aandrew: the ref voltage is constantly sawtoothing, the source knows what vref is and can look at the 1 bit out
11:10 PM aandrew: and you can even multiplex those 1-bit outputs
11:10 PM rue_shop1: its the analog travel that disturbs be
11:10 PM rue_shop1: me
11:10 PM aandrew: there is no analog travel
11:10 PM rue_shop1: if i'm feeding it a ref wave tho
11:10 PM aandrew: the vref can be transmitted as PWM
11:10 PM rue_shop1: I'm disturbed by the drp is sensor bandwidth tho
11:10 PM aandrew: and besides, you're in charge of vref, you can drive that with a very stiff source or even a current level
11:11 PM aandrew: take your 4-20mA and use that to create Iref for all the sensors; I like that less because a single break in the wire kills all sensors
11:11 PM rue_shop1: one of the ponderings I have is a seperate current loop
11:11 PM rue_shop1: do it 4-20 with an external supply
11:12 PM rue_shop1: just make it a 4-20 sourcing output
11:12 PM aandrew: I think a PWM vref and a 1-bit output would work quite well; you could even make that sawtooth vref 1MHz or something, it's up to how fancy you want to get with the comparators
11:12 PM aandrew: but 1MHz really isn't anything fast
11:12 PM rue_shop1: the congrigator has a 250ohm sink
11:12 PM aandrew: you could drive the PWM from a timer, and the LVDT outputs to another timer with the same source
11:13 PM aandrew: it's all handled in hardware then and you get an interrupt that tells you the PWM timer count of when the sensor "tripped"
11:13 PM rue_shop1: hmm
11:13 PM aandrew: which means there's no polling, and depending on the MCU you could even DMA the results and have almost no CPU load for this
11:14 PM aandrew: dammit rue now I kind of want to build something using this
11:14 PM rue_shop1: if I target a signal distance of 12"
11:14 PM rue_shop1: I'm doing the capacitor one eh?
11:15 PM aandrew: capacitor one?
11:15 PM rue_shop1: yea
11:15 PM rue_shop1: WAY simpler
11:15 PM rue_shop1: check my thread there
11:15 PM aandrew: thread where
11:16 PM rue_shop1: https://twitter.com/i/status/1338319991368679424
11:16 PM rue_shop1: $0.89 in parts
11:16 PM rue_shop1: I added it up
11:16 PM rue_shop1: driver circuit and all
11:17 PM aandrew: I have no idea what that is
11:17 PM aandrew: it looks like a foil based LVD..C?
11:17 PM rue_shop1: strange, part of it is misisng
11:17 PM aandrew: aha there's the link
11:18 PM rue_shop1: https://twitter.com/RueNahcMohr/status/1339343953938710529/photo/1
11:18 PM rue_shop1: yes, capacitor
11:18 PM rue_shop1: WAY simpler
11:18 PM rue_shop1: 1.4VDC peak right out of the sensor
11:19 PM aandrew: this is pretty cool
11:20 PM aandrew: doesn't the LVDC need the positional element to be one leg of the cap?
11:21 PM rue_shop2: no its grounded
11:21 PM rue_shop2: differential
11:21 PM aandrew: that's still in circuit then
11:21 PM aandrew: it means your actuators all need a decent electrical connection to ground
11:21 PM rue_shop2: yes, see how I have the inner lug grounded?
11:22 PM rue_shop2: like a light spring
11:22 PM aandrew: yes, I'm saying that is less than ideal
11:22 PM rue_shop2: no, closed up device
11:22 PM aandrew: and if you're moving around that's gonna be a noisy ground unless it's solidly attached
11:22 PM rue_shop2: spring
11:22 PM aandrew: ah so you'd have the actual actuator push and pull on some sealed cylinder
11:23 PM aandrew: I like the simplicity of this but the LVDT seems way more robust
11:23 PM rue_shop2: yea, you bring it out on a non-conductive coonnection
11:23 PM rue_shop2: the cap is the same
11:24 PM rue_shop2: I got the lvdt down to a few dollars, but that whole lvdc is $0.89
11:25 PM rue_shop2: there are like, a few orders of magnitude to the complexity gap
11:26 PM aandrew: yes
11:26 PM aandrew: that is a really novel design
11:26 PM aandrew: robust too
11:27 PM rue_shop2: and, I could NOT find ANY reference, anywhere to using a triangle wave with an lvdt
11:27 PM rue_shop2: which I dont get
11:27 PM rue_shop2: it seems to make so much sense
11:29 PM rue_shop2: so...
11:30 PM rue_shop2: I like the full bandwidth of a current source
11:30 PM rue_shop2: even if I cant make it just a loop
11:31 PM rue_shop2: you know
11:31 PM rue_shop2: for this bot, I'm supposed to be selling its tech to make the money to build it
11:31 PM rue_shop2: *sigh*
11:32 PM rue_shop2: aandrew, I'm working on some conductive paint to make the lvdcs with
11:32 PM rue_shop2: for the outter electrodes anyhow
11:32 PM aandrew: well you have a great idea here, if you can 3d print an simple enclosure for the encoder bit that you can screw an actuator in to, a simple board that generates 8MHz and spits out +/-2... I think that's something lots of robot enthusiasts would be all over
11:33 PM aandrew: you'd have to translate it from +/-2 to 0-4 though
11:33 PM aandrew: nobody wants to deal with -ve voltages until you hit a certain threshold of proficiency
11:34 PM rue_shop2: that circuit DOES generate the 8Mhz
11:34 PM rue_shop2: and because its not critical, its just LC
11:35 PM rue_shop2: I can recenter it around anything
11:35 PM rue_shop2: a 2.5V center is as easy as a TL431 and a 1k pullup
11:37 PM rue_shop2: I also like the lvdc casue it doesnt' need a custom power transformer
11:38 PM rue_shop1: if I finish the powersupply, I can have 1 finished project
11:38 PM rue_shop1: I mean, winding transformers is awesome fun and all, but
11:40 PM rue_shop1: http://ruemohr.org/%7Eircjunk/tutorials/elex/ldvc/lvdc.html