#garfield Logs

May 08 2020

12:17 AM rue_mohr: and I successfully 3d printed a transformer spool
12:23 AM rue_mohr: I got an idea to split it in half, and that made it printable
12:23 AM rue_mohr: winding wire on it holds it togethor
02:43 AM polprog: rue_mohr: nice https://twitter.com/RueNahcMohr/status/1258623403092635648?s=09
02:43 AM polprog: why is ch3 this weird?
09:34 AM rue_mohr: the diode contributes differently than the fet invterter
09:35 AM rue_mohr: its all enought to be a 1 tho
10:54 AM polprog: interesting
10:54 AM polprog: good morning also
10:54 AM rue_mohr: hey
10:55 AM rue_mohr: I'm a late start work day
10:55 AM polprog: i was thinking this local morning, what did you stuff into your teamaker to make it remotely controlled?
10:55 AM polprog: iirc it uses ssh? or am i talking crazy
10:55 AM rue_mohr: esp32
10:55 AM rue_mohr: no its running arduino-ness a little web server
10:55 AM polprog: nice.. mine is still unused
10:55 AM polprog: ah
10:56 AM rue_mohr: less the adapter circuit, it was a quick hack
10:56 AM rue_mohr: ota updates
10:56 AM polprog: looks like the 500G barracuda hdd in this nixdorf box is wiped clean
10:56 AM rue_mohr: so I can sit in my bedroom and rewrite it
10:56 AM polprog: i wonder what can i do with it now
10:57 AM rue_mohr: what doesn't fit on 500G?
10:57 AM polprog: idk, i was thinking a small PXE server would come in handy
10:57 AM rue_mohr: my shop fileserver is only 200
10:57 AM rue_mohr: house server is 1G
10:58 AM rue_mohr: er
10:58 AM rue_mohr: 1T
10:58 AM rue_mohr: haha
10:58 AM polprog: oh, reminds me i enabled old samba in my win 10 laptop
10:58 AM polprog: haha, 1G would be enough for a server that doesnt need space :P
10:59 AM rue_mohr: I'm not sure which design to wind on this core
10:59 AM rue_mohr: I'm all excited about the custom spool
11:00 AM polprog: the one you posted?
11:00 AM rue_mohr: yea
11:00 AM rue_mohr: well, the followup, thats redesigned to sit on its end so It'll fit a 14 pin pcb
11:01 AM polprog: i wonder if it makes sense to solder the uln2003 DCDC onto the rasterboard
11:01 AM rue_mohr: na, use a socket
11:01 AM polprog: if i found enough enamel wire i could easily make the inductor more inductive, just rewind it
11:01 AM polprog: no, i mean, whether to leave it like this or add the caps
11:01 AM polprog: and i think im out of narrow sockets.. might have to cut up a wide one
11:01 AM rue_mohr: did you see the thread I did there on making the litz wire?
11:02 AM polprog: no
11:02 AM rue_mohr: oh, I take apart solinoids from printers for wire
11:02 AM polprog: frankly, this thing, even though it has only 2g of ram and a pentium E5600 makes a pretty good workstation.. the onboard video can do 1920x1080 easily
11:03 AM polprog: ive got a broken relay with lots of thin wire
11:03 AM polprog: but i feel its too thin
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150939.jpg
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150940.jpg
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150941.jpg
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150942.jpg
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150943.jpg
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150944.jpg
11:03 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150945.jpg
11:04 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150947.jpg
11:04 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150949.jpg
11:04 AM polprog: for a file server id prefer raid
11:05 AM rue_mohr: yea I do software mirrored raid
11:05 AM rue_mohr: dont use hardware raid, the controllers fail and then your screwed
11:06 AM polprog: ill keep that in mind..
11:06 AM rue_mohr: so I realized, with the power converters
11:06 AM rue_mohr: you need a larger core to do flyback then pushpull
11:06 AM polprog: hmm, is that wire gauge enough for my application>
11:06 AM polprog: ?
11:06 AM rue_mohr: the flyback has to store the energy, the pushpull just conducts it
11:06 AM polprog: that might mean the busted relay is a good source
11:06 AM rue_mohr: I only needed 48mA
11:06 AM polprog: but thats one winding? or two
11:07 AM polprog: yeah mine doesnt take much more anyway.. <100mA
11:07 AM rue_mohr: pushpull is more windings, but you make parallel wire and wind it all at once
11:08 AM polprog: and then connect the wires in series or paralell?
11:08 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1150938.jpg
11:08 AM rue_mohr: thats 6c
11:08 AM rue_mohr: I seperate out the conductors after
11:09 AM polprog: i see a spool of wire, what do you do with it later? wind on and then cut?
11:11 AM rue_mohr: https://twitter.com/RueNahcMohr/status/1256828419242778626
11:16 AM polprog: ahh, i see so you just make a thicker wire
11:16 AM polprog: clever
11:20 AM polprog: interesting, seems like 10-odd turns on the graphite-black cores give me way more voltage than the same count of turns on the yellow cores
11:20 AM polprog: must be different material
11:20 AM polprog: i have a small spare bobbin, maybe that could be used
11:20 AM polprog: i kindof wanna make it a small module
11:26 AM rue_mohr: ferrite
11:26 AM rue_mohr: yea, the yellow ones suck
11:26 AM polprog: i think the yellow ones are metal inside?
11:26 AM rue_mohr: the common chocked have a good permeability
11:26 AM rue_mohr: Ur
11:26 AM polprog: and the ferrite is some pressed stuff
11:27 AM rue_mohr: Iron oxide
11:27 AM rue_mohr: ceramic
11:27 AM rue_mohr: so they break like glass
11:27 AM rue_mohr: remember when Iwas trying to take apart all those transformers?
11:27 AM rue_mohr: and the ferrite was all broken on most of them?
11:28 AM rue_mohr: some of them are cintered iron
11:28 AM polprog: yeah i think i do
11:28 AM rue_mohr: but its rare to find those
11:29 AM rue_mohr: often their in power converters for automotive amps
11:29 AM polprog: ive got a tiny bobbin, h=8mm, r=2.5, R=6
11:29 AM rue_mohr: for the flyback supply, rememeber that the core has to hold all the energy as a magnetic field for that cycle
11:30 AM rue_mohr: like a capacitor, its limited
11:30 AM polprog: and the litz wire from the dead relay is... 0.1mm
11:30 AM rue_mohr: litz means multi-strand
11:30 AM polprog: yeah, but how much energy is that i have no idea
11:30 AM polprog: ah
11:30 AM polprog: so i probably have to make it multi-strand
11:31 AM rue_mohr: yea, if you try to pump too much energy into the core, it'll saturate, and then the resistnace between the windings goes to zero
11:31 AM rue_mohr: and the transistor is a dead short against the supply
11:31 AM polprog: theres an inline ferrite here which did 12->20 easily. It has 4 windings total
11:31 AM rue_mohr: I do the multi-strand thing to put multiple windings on a transformer at once
11:31 AM rue_mohr: so, if I need 1:1:1:1 I'd make a 4 conductor litz and wind it all on at once
11:31 AM polprog: do you think 0.1 mm is enough?
11:32 AM rue_mohr: depends on current,
11:32 AM rue_mohr: I'm not sure
11:32 AM polprog: 200mA max..
11:32 AM polprog: i guess ill make 3 strands..
11:32 AM polprog: to be on the safe side
11:32 AM rue_mohr: yea, you can parallel up strands
11:32 AM rue_mohr: if the ferrite saturates, it will get hot
11:33 AM polprog: yeah, eddy currents?
11:33 AM rue_mohr: I tried a core that was too small and it heated up quikc
11:33 AM rue_mohr: not sure
11:33 AM rue_mohr: they aren't supposed to have much of anything for eddy currents cause there is no electrical conduction between the magnetic particles
11:37 AM rue_mohr: the flybacks voltage comes from the regulation system, there is a limit based on the duty its driven at tho
11:37 AM rue_mohr: the pushpull can double without regulation, but is subject to losses
11:38 AM rue_mohr: the core of the flyback has to be larger to store the energy
11:38 AM rue_mohr: the pushpull can be implemented with less parts
11:38 AM rue_mohr: the pushpull can also have a different turns ratio to deal with loss
11:38 AM polprog: but i only have 1 winding on 1 core
11:39 AM rue_mohr: I have not done a forward converter yet
11:39 AM polprog: not sure if you are talking about flybacks.. thats not a flyback i think
11:39 AM rue_mohr: pushpull is not flyback
11:39 AM rue_mohr: sorry, I'v lost my book or I'd post some images
11:40 AM rue_mohr: I need to find it
11:40 AM rue_mohr: its got a lot of math work for switching power supplies
11:40 AM rue_mohr: I want to make some modules for my resistor box today
11:40 AM polprog: mine is just a boost topology
11:41 AM polprog: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&ved=2ahUKEwiB6a-w2qTpAhXQo4sKHce1BFsQFjACegQIHxAF&url=https%3A%2F%2Fwww.st.com%2Fresource%2Fen%2Fapplication_note%2Fcd00003910-topologies-for-switch-mode-power-supplies-stmicroelectronics.pdf&usg=AOvVaw2ChiEryNQgAC3DYkNT9bDw
11:41 AM rue_mohr: that is flyback
11:41 AM polprog: ugh. google
11:41 AM rue_mohr: there is flyback and forward, it reffers to when you pull the energy from the inductor
11:42 AM rue_mohr: buck is just pwm with a filter
11:42 AM rue_mohr: pushpull is a type of forward
11:44 AM polprog: that appnote is full of equations
11:46 AM rue_mohr: yea, but their kinda wrong
11:47 AM rue_mohr: in their exclusions
11:47 AM rue_mohr: the good document was written by motorola
11:47 AM rue_mohr: (I found my book)
11:47 AM rue_mohr: did you see the EE.surrey page on making wound conponents?
11:48 AM polprog: no
11:50 AM rue_mohr: http://info.ee.surrey.ac.uk/Workshop/advice/coils/
11:50 AM rue_mohr: http://info.ee.surrey.ac.uk/Workshop/advice/coils/terms.html
11:50 AM rue_mohr: the document has a few forks that make it a bit hard to find all the good info
11:50 AM rue_mohr: all my main notes are from the terms page
11:52 AM rue_mohr: oh, the other document was from national
11:53 AM rue_mohr: oh, hmm
11:54 AM rue_mohr: http://www.ti.com/lit/SNVA038
11:54 AM rue_mohr: its only for buck designs tho...
11:56 AM polprog: ive made the coil, lets see if it blows up
11:57 AM polprog: 12->20V nice, at 49mA
11:57 AM rue_mohr: :) thats the way I test 'em
11:57 AM rue_mohr: do you have an inductance meter?
11:57 AM polprog: nope, thats the issue
11:57 AM polprog: ok.. we have 49mA in,
11:57 AM polprog: $ rclib -v 21 -r 1500
11:57 AM polprog: I = 14.000000 mA P = 294.000000 mW
11:58 AM polprog: 14mA out..
11:58 AM polprog: :(
11:58 AM polprog: wait
11:58 AM polprog: im calculating it wrong
11:58 AM polprog: 294mW out
11:59 AM polprog: 588mW in
11:59 AM polprog: thats bang on 50%
12:00 PM polprog: youre right, the diode in the ULN sucks
12:00 PM polprog: 21.50 Vout
12:00 PM polprog: wait, that caused the input current to drop to 38mA
12:01 PM polprog: 308mW / 456mW = 67.5%
12:01 PM polprog: not bad :D
12:01 PM polprog: i found this on the ground while urbex-ing last time..
12:02 PM polprog: the broken relay
12:05 PM polprog: at 680nF load on the ring osc stages, its running at 400kHz and does 12->29.3V
12:07 PM polprog: 572mW out / 720 mW in that is short of 80%
12:12 PM polprog: 6ms startup time..
12:13 PM rue_mohr: I should show you how to make a saturation tester
12:14 PM rue_mohr: zippo:/morfiles/programming/c/avr/Atmega8/resistorbox$ ohm -v 21 -r 1.5k
12:14 PM rue_mohr: Power is: 294.000mW
12:14 PM rue_mohr: Current is: 14.000mA
12:14 PM rue_mohr: Voltage is: 21.000V
12:14 PM rue_mohr: Resistance is: 1.500kOhms
12:15 PM rue_mohr: I was looking at the voltage drop on my scope, 1.8V
12:17 PM rue_mohr: it was literally adding 0.5w to the dissipation of the uln2803
12:18 PM rue_mohr: as soon as I used an external diode, the chip cooled down to nothing
12:18 PM polprog: 9V to 24V. this is good
12:18 PM rue_mohr: dispite lots of voltage drop on the transistor
12:19 PM rue_mohr: polprog, I cant find an image
12:19 PM rue_mohr: V*t in = V*t out, on the inductor
12:19 PM polprog: of what?
12:20 PM rue_mohr: so, 1ms@10V in, is 0.5ms@20V out
12:20 PM polprog: why would the time be different?
12:20 PM rue_mohr: duty
12:20 PM polprog: ah
12:21 PM rue_mohr: remember I said I had to adjust the duty on that 555 circuit with a resistor on pin 5?
12:21 PM polprog: yeah
12:22 PM rue_mohr: it couldn't achive 24V cause the duty wouldn't allow it
12:23 PM rue_mohr: if you have 100us@3V in, you can get 1us@300V out
12:23 PM polprog: why?
12:24 PM rue_mohr: thats how the math works
12:24 PM rue_mohr: so the duty would be 100/101
12:24 PM polprog: well, i know that.. its just what is the basis for the math
12:24 PM polprog: energy equality?
12:24 PM rue_mohr: its how the inductor stores the energy
12:25 PM polprog: makes sense
12:25 PM rue_mohr: its equiv to volt*seconds
12:25 PM polprog: i should be studying that asctually
12:25 PM polprog: got a test on 14th
12:25 PM rue_mohr: heh
12:25 PM rue_mohr: read the units.html
12:25 PM rue_mohr: er
12:25 PM rue_mohr: terms.html
12:25 PM polprog: the inductors part was like 2 classes.. its more of the general EM stuff
12:25 PM rue_mohr: http://info.ee.surrey.ac.uk/Workshop/advice/coils/terms.html
12:26 PM polprog: thanks
12:26 PM rue_mohr: we didn't have any inductor stuff in school, that document helped me understand everythign
12:27 PM rue_mohr: even in tech school, inductors were only covered as AC components, no switched DC stuff
12:27 PM polprog: what school did you finish?
12:29 PM rue_mohr: From tech school I got a dimpoma of technology for robotics and automation
12:29 PM rue_mohr: not sure how much I actually learned there tho
12:30 PM polprog: tech school like vocational college or?
12:30 PM polprog: i have no idea about us education
12:30 PM polprog: (well, a rough one)
12:31 PM rue_shop1: hmm I suppose
12:31 PM Tom_itx: the robotics they teach here are industrial
12:31 PM polprog: gtg
12:32 PM Tom_itx: by by
12:32 PM rue_shop1: https://www.bcit.ca/
12:35 PM rue_shop1: in the microcontrollers course I was usually doing paperwork for other courses, following with 1 ear
12:35 PM Tom_itx: https://wsutech.edu/admissions/programs-of-study/
12:35 PM rue_shop1: the instructor kept throwing questions at me and got frustrated cause I'd always answer correctly after a brief delay working out what he was talking about
12:36 PM rue_shop1: amazingly similar pages
12:36 PM Tom_itx: the classes i took, we'd talk about everything else
12:37 PM Tom_itx: i was done way early and did those 100+ extra problems
12:37 PM Tom_itx: including helping the guy next to me that was failing
12:37 PM rue_shop1: I wrote programs to do a lot of the homework
12:38 PM rue_shop1: once you have written code to do it, you dont forget how
12:38 PM Tom_itx: we've got the largest if not only aviation research center here
12:39 PM Tom_itx: dassault has a building on campus (SW & Catia)
12:39 PM Tom_itx: so does airbus
12:40 PM Tom_itx: if i had nothing else to do, i'd go back for fun
12:40 PM rue_shop1: :)
12:41 PM Tom_itx: they won't let you audit classes since they're work related technical classes
12:41 PM rue_shop1: at some point I realized not to get caught up in the stress of worrying about failing
12:41 PM rue_shop1: to not sorry about writing everything down madly
12:42 PM Tom_itx: the dean knows all 3 of us
12:42 PM rue_shop1: I'm going back to sparky school in oct, and I'm worried about it
12:42 PM Tom_itx: especially since the kids went to that skils thing
12:42 PM Tom_itx: and brought back iron
12:42 PM rue_shop1: I thought the plane parts were top secret?
12:42 PM Tom_itx: why are you worried?
12:43 PM Tom_itx: the one my kid was working on was
12:43 PM Tom_itx: it was a work study thing he did while there
12:43 PM rue_shop1: :)
12:43 PM Tom_itx: disassembling the fighter jet and scanning all the parts
12:46 PM Tom_itx: why are you worried about sparkie school?
12:46 PM rue_shop1: just stress
12:46 PM rue_shop1: I hate tests
12:46 PM Tom_itx: gonna be a lineman?
12:46 PM rue_shop1: joiurneyman
12:46 PM Tom_itx: meh, you got it
12:46 PM rue_shop1: I wonder what the tolerance is...
12:47 PM rue_shop1: red black, black, brown. (2k) but then purple yellow
01:06 PM rue_shop1: this is going to get complicated
01:07 PM rue_shop1: this resistor box
01:07 PM rue_shop1: so, I have a lot of resistors that are out by a bit
01:07 PM rue_shop1: and inerestingly enough, I can get the controller to compensate
01:07 PM rue_shop1: but its going to make the software a huge snotball
03:16 PM Tom_itx: smt or thruhole?
03:16 PM Tom_itx: smt should be pretty much right on
04:03 PM polprog: i wonder if i could make a charge pump on the ULN2003
04:04 PM polprog: https://upload.wikimedia.org/wikipedia/en/thumb/1/12/Dickson_voltage_multiplier.svg/1280px-Dickson_voltage_multiplier.svg.png
04:04 PM polprog: ofc i could !
04:04 PM polprog: heh, what if i fed the DCDC output to a charge pump.. i still hae 3 free stages
04:05 PM polprog: if i could make a relax osc on the ULN2003 i could have 5 free stages
04:23 PM Tom_itx is now known as Tom_L
05:32 PM Tom_L: rue_shop1, what file extension is a shell script?
05:33 PM polprog: any really. usually .sh
05:33 PM Tom_L: ok thanks
05:33 PM Tom_L: how to run it?
05:33 PM Tom_L: just call it from a terminal?
05:34 PM Tom_L: ./run.sh
05:35 PM polprog: yeah, if its executable
05:35 PM polprog: chomd +x file
05:35 PM polprog: if you want to make it executable
05:35 PM polprog: you can check with ls -l file
05:35 PM Tom_L: i got it thanks
05:44 PM polprog: http://tinyurl.com/y94puzdt
05:44 PM polprog: falstad sim, 3 stage charge pump
05:44 PM polprog: looks like it would work with open drain switcher
05:47 PM polprog: it would be more efficient with a push pull stage but if i disregard that i could make a compact boost + charge pump converter on the ULN
05:47 PM polprog: or just a charge pump
07:53 PM polprog: when did i mess wit that current controlled colpitts
07:53 PM Tom_L: polprog,
07:53 PM Tom_L: also
07:57 PM polprog: LA7876 Vertical Deflection Output Circuit.
07:58 PM polprog: havent played with that. but thats not what im looking for..
08:01 PM polprog: eh, must have posted the frequency range to twitter instead
08:01 PM polprog: i was thinking.. that colpitts design was using the yellow core inductor
08:01 PM polprog: what if i used the actual ferrite one
08:02 PM polprog: maybe i could get an even better range?
08:03 PM polprog: 12-44kHz, unstable
08:03 PM polprog: unstable later*
08:03 PM polprog: ok
08:05 PM polprog: looks like that sanyo LA7876 can do 3 Amps peak-peak easily for driving inductive loads
08:05 PM polprog: deflection yoke driver
08:06 PM polprog: should i build even more absurd converter? maybe a buck this time
11:32 PM rue_shop1: Tom_L, thru hole resistors, I'm making a decade box, so I'm trying to find resistors that are spot on
11:33 PM rue_shop1: polprog, charge pumps suck :)
11:33 PM rue_shop1: Tom_L, I'm told that, to be correct, you do not put an extension on a shell script, screw that tho, I use .sh
11:34 PM rue_shop1: polprog, doc it for me eh?
11:36 PM rue_shop1: :/ I'm gonna have to upgrade my circuits binder
11:37 PM rue_shop1: 2" binder and its getting hard to add pages