#robotics Logs
Jul 05 2018
#robotics Calendar
12:01 AM rue_: the old toys ran at 28Mhz didn't they?
12:05 AM rue_: oh but they were crystal
12:05 AM mrdata: crystals help a lot
12:07 AM mrdata: a really fine LC meter is usually constructed fom a pecular bridge circuit. i'll try to remember which...
12:08 AM mrdata: Wien bridge i think; checking...
12:09 AM mrdata: plus the truely awesome meters have shields on each probe i think
12:11 AM rue_: no I need 1-10Mhz
12:11 AM mrdata: yes
12:11 AM rue_: crystals wont help me
12:12 AM rue_: Its probably a breadbaord, problem, I tried to space this one out a lot, it oscillated at 80Mhz nomatter what capacitor or inductor I used
12:13 AM mrdata: https://en.wikipedia.org/wiki/Wien_bridge
12:13 AM rue_: not for 1-10Mhz
12:13 AM mrdata: breadboards have horrible parasitics
12:13 AM rue_: yea
12:13 AM rue_: I'll heave to deadbug a generic oscillator somewhow
12:14 AM mrdata: for RF i recommend using manhattan construction
12:14 AM mrdata: it's deadbug-like but controls the parasitics
12:14 AM rue_: http://ruemohr.org/~ircjunk/images/canosc2.jpg
12:14 AM rue_: http://ruemohr.org/~ircjunk/images/canosc1.jpg
12:14 AM rue_: any part off ground is a standoff
12:15 AM mrdata: using a continuous ground plane and pads standing off by a precise distance
12:15 AM rue_: THO, most parts on a lot of osc' seem to go to +, so maybe making the base + would be a better idea
12:15 AM rue_: I stood off with 10M resistors
12:16 AM rue_: and I didn't waste any pcb ;)
12:17 AM mrdata: a uniform, continuous ground plane isnt a waste in RF
12:17 AM rue_: ^ can
12:17 AM mrdata: http://www.unixnut.net/files/manart.pdf
12:17 AM rue_: I made another one in the inside of the lid for an RF tuner, but it didn't work
12:19 AM mrdata: i like putting little RF circuits inside tuna cans
12:19 AM rue_: how do you clean them out?
12:20 AM rue_: (the linig)
12:20 AM mrdata: soap and water
12:20 AM mrdata: dont remopve the lining
12:20 AM rue_: OH you didn't mean soldering them to the inside
12:20 AM mrdata: just mount the board so that the ground plane comes close to the can and the components and traces are inside
12:21 AM mrdata: you can punch holes for coax
12:21 AM mrdata: etc
12:21 AM mrdata: this helps control interference because it's a faraday cage
12:23 AM rue_: huh, those pcb pads have a lot less capacitance than I'd thought
12:24 AM mrdata: depends on diameter of the pad, dielectric constant of the pad's insulation, and thickness
12:25 AM rue_: nice pdf
12:25 AM rue_: hmmmmm
12:25 AM mrdata: cheers
12:25 AM rue_: I wonder how much a square foot of copper roof flashing is worth
12:26 AM rue_: wild guess ~$23
12:26 AM mrdata: a few bucks
12:26 AM mrdata: you can get scrap cheaper
12:26 AM rue_: not here
12:26 AM rue_: all recycling are unidirectional
12:28 AM mrdata: hardware stores here carry thin copper sheet in sizes similar to 8.5x11 paper
12:28 AM mrdata: thin enough that it can be folded
12:28 AM mrdata: easily
12:28 AM mrdata: and soldered
12:30 AM rue_: oh, come to think of it, there is a brass roof flange that might not go fro much more
12:31 AM rue_: i was amused how ugly the 10mhz from the stm32 was
12:32 AM mrdata: bouncy transitions?
12:33 AM rue_: ding ding ding ding
12:33 AM rue_: about 3 or 4 cycles
12:33 AM mrdata: i add lots of RC filters to make those settle down
12:33 AM rue_: at 10Mhz?
12:33 AM rue_: a termination resistor I recon
12:34 AM rue_: but the idea died when 50Mhz/5 = 10Mhz but 50Mhz/6 = 8.3Mhz
12:34 AM rue_: bit course
12:34 AM rue_: heh 'bit' course
12:34 AM mrdata: tau = RC, so at 10 MHz you're looking at tau = 0.1 microsecond; that's 0.1 uF and 1 ohm, say
12:35 AM rue_: hahaha
12:35 AM rue_: even at 2V p-p
12:35 AM mrdata: or 10 nF and 10 ohm
12:36 AM rue_: 100pf?
12:36 AM rue_: actually how about 60-160pf
12:37 AM mrdata: 100 pF and 1 k ohm; but you want some signal to get through
12:37 AM rue_: hmm
12:38 AM mrdata: 50 ohm tends to be a sweet spot between power and attenuation in RF
12:38 AM rue_: I'm thinking a really lv osc and the "0-160pf" capacitor, with some amps on it before the 74hc14
12:38 AM mrdata: so 2nF and 50 ohm would give you tau = 0.1 us
12:38 AM rue_: oh, but
12:39 AM rue_: I also have these opto-fets
12:39 AM rue_: for light controlled resistors
12:39 AM mrdata: do you want a square wave or other oscillator
12:39 AM rue_: H11F1
12:39 AM rue_: at 10Mhz, it dosn't matter
12:39 AM rue_: all pretty round at 5V
12:39 AM mrdata: k
12:40 AM mrdata: pierce oscillator is good
12:42 AM mrdata: so stage your RF circuits putting each in a can and connect via 50 ohm coax such as SMA or BNC
12:42 AM rue_: its worse than that jim
12:43 AM rue_: I need the 1-10Mhz, as the high range
12:43 AM rue_: then I'll be dividing it down by 10 to get all the way down to a 0.1-1Hz range
12:43 AM rue_: array of 7490, see if I can use them up
12:45 AM rue_: the opto does 500R to 300M
12:45 AM mrdata: counting down is easy; you just dont have HF pollute LF by looping around it and shit
12:45 AM rue_: yea, once I have teo 1M-10Mhz, I'm golden
12:50 AM rue_: hmm
12:50 AM rue_: https://www.changpuak.ch/electronics/RC_HCT_Generator.php
12:50 AM rue_: says use 2 optos for resistance control, and a 200pF cap
12:50 AM rue_: maybe if I run it at like 3V
12:50 AM rue_: says by the time I hit 3k, its at 1Mhz
12:51 AM rue_: ok, its worth a quick try
12:53 AM mrdata: yes, a pierce thats not crystal uses an NPN, and a couple caps only, i think; jussec
12:54 AM rue_shop4: google is just giving me a mishmash and I'v not got the familiarity to tell
12:54 AM mrdata: yeah put your screwy variable inductor in place of that crystal
12:55 AM rue_shop4: hey if I can get an RC to pull off 10Mhz, I'm all for it
12:55 AM mrdata: https://en.wikipedia.org/wiki/Pierce_oscillator
12:58 AM mrdata: so for 10 MHz, sqrt( L C ) will be something like 1.6e-8 or so; if C = 2 nF, then L ~ 127 nH
12:59 AM mrdata: then have L vary from 127 nH to 12.7 uH maybe?
01:00 AM rue_shop4: if I have a core with a ur of 3000 that goes into whats otherwise an aircore
01:00 AM rue_shop4: I dont see it being a small component
01:01 AM mrdata: if it's a solenoid, the gap is probably large
01:01 AM rue_shop4: I wonder what ur is of a ferrite rod antenna
01:02 AM mrdata: the coil could be about the diameter of a penny maybe?
01:02 AM rue_shop4: need to be carefull I dont use soemthing thats gonna snub out the rf
01:02 AM mrdata: ferrite rods between that and diameter of a pencil are easy to find
01:02 AM mrdata: definitely use ferrite then
01:03 AM mrdata: theyre made for RF
01:03 AM rue_shop4: I have a lot of loopstick ferrite
01:04 AM mrdata: good
01:05 AM mrdata: so lets say it's 10 mm dia
01:05 AM mrdata: or tell me what dia it is
01:07 AM mrdata: so using this calculator http://hamwaves.com/antennas/inductance.html
01:07 AM mrdata: i put in 10 mm dia, 10 turns, length 15 mm, and 1 mm dia wire
01:07 AM mrdata: and 10 MHz; and got aboutr 450 nH
01:08 AM mrdata: that's air core
01:12 AM rue_shop4: wholy flipp that RC osc is unstable
01:12 AM rue_shop4: watching it walk up and down on my freq counter
01:13 AM rue_shop4: ok, so 127nH is easily obtainable
01:13 AM mrdata: try to increase the Q factor
01:13 AM mrdata: Q = sqrt( L / C ) / R
01:13 AM rue_shop4: I cant, the optos I'v got only go down to 100R
01:13 AM mrdata: the higher the Q the more stable the frequency
01:14 AM rue_shop4: putting my finger on the chip warms it up enough to drift by 0.3Mhz
01:15 AM mrdata: also Q = f / bandwidth, so bandwidth = f / Q
01:15 AM rue_shop4: hmm
01:16 AM rue_shop4: dialed up the votlage a bit and lost the HC04
01:16 AM mrdata: so increase L and reduce R
01:16 AM rue_shop4: it was an RC
01:16 AM rue_shop4: was at 8Mhz tho
01:16 AM rue_shop4: I could potentially PLL a RC osc
01:17 AM rue_shop4: looks like it would have to be on its toes tho
01:17 AM rue_shop4: heh, I'm dividing it down anyhow, I could use a 4046 :)
01:17 AM rue_shop4: ok, try an option with a stable osc tho
01:18 AM rue_shop4: so, you got 2nF
01:18 AM mrdata: idk? 1 nF is more common
01:19 AM rue_shop4: 10 turns, 10mm d, 15mm long, #18 wire
01:19 AM mrdata: so maybe try C = 1nF and L = 255 nH
01:19 AM mrdata: sure
01:19 AM rue_shop4: see if I can get the 10Mhz first, then work on bringing it down
01:19 AM rue_shop4: I need an osc
01:19 AM rue_shop4: where is my magnet wire...
01:20 AM mrdata: ok, maybe try 0.1 nF caps as well then
01:21 AM rue_shop4: #18.. ok
01:23 AM rue_shop4: sloppy as all hell, but pretty close
01:24 AM rue_shop4: peirce and not heartley or colpitts eh?
01:25 AM mrdata: pierce and colpitts are very similar
01:26 AM rue_shop4: so, capacitor divider
01:26 AM mrdata: yeah
01:27 AM rue_shop4: geeez, I should just save all the real schematics off google of all the rf oscs, there are like only 10 or 15 non-concept circuits
01:28 AM rue_shop4: so, is the 0.1nF over both or each?
01:28 AM rue_shop4: prolly both, so I need 0.2nf
01:28 AM rue_shop4: (to series 0.1nF)
01:28 AM mrdata: with the above coil i think a pair 0.2nF gets you closer to 10 MHz
01:29 AM rue_shop4: ok
01:29 AM mrdata: but, i'm not certain of the exact frequency formula; use what you have
01:29 AM rue_shop4: so, I wonder if a PN2222 should do it
01:29 AM mrdata: and measure and adjust
01:29 AM rue_shop4: its also real world, so math is just ballpark
01:29 AM mrdata: PN2222 probably should be okay
01:30 AM mrdata: whats its frequency gain product?
01:30 AM rue_shop4: https://en.wikipedia.org/wiki/File:NPN_Colpitts_oscillator_collector_coil.svg
01:30 AM rue_shop4: I think its about 200Mhz
01:30 AM mrdata: mkay
01:31 AM rue_shop4: ok, I'm starting to see this stuff, so the L and C(s) are in parallel cause the power supply is low impedence
01:31 AM rue_shop4: and from what I can tell, the bais current is REALLY forgiving on an RF osc
01:33 AM mrdata: that svg looks fine
01:34 AM rue_shop4: ceramic for now...
01:34 AM rue_shop4: :) I love small caps, is "220" 22pf or 220pf? "221" is 220...
01:35 AM mrdata: 221 is 200pF yeah
01:35 AM mrdata: er 220pF
01:36 AM mrdata: 220 is 22 pF, 221 is 220 pF; 222 is 2200 pF etc
01:37 AM rue_shop4: maybe, 47 isn't 40000000 pf, its just 47pf
01:37 AM rue_shop4: :/ how am I gonna build this...
01:37 AM mrdata: oh well
01:37 AM rue_shop4: ok, I can do it off a ground plane
01:37 AM mrdata: yes
01:38 AM rue_shop4: oh I putzed up the caps...
01:38 AM mrdata: ??
01:38 AM rue_shop4: I put them across the inductor directly, heatley style
01:39 AM rue_shop4: https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRcCTwYWwQ9iwmlRHplchjoYML11ftITeUGEt8BQr4n_dzp47MS
01:39 AM rue_shop4: or I do that and use the same values
01:39 AM mrdata: hartly oscillator uses a saturated core
01:39 AM rue_shop4: ? oh
01:39 AM rue_shop4: I dont think that matters to me
01:39 AM mrdata: i dont think you will have saturation on a solenoid
01:40 AM rue_shop4: oh, not an aircore
01:40 AM rue_shop4: right
01:40 AM rue_shop4: ok, back to plan X
01:48 AM rue_shop4: I"m sure there shoudl be a RF bypass around the supply on that thing
01:51 AM rue_shop4: ...
01:51 AM rue_shop4: ok?
01:53 AM rue_shop4: .. applying smoke... I mean power...
01:54 AM rue_shop4: oscillating
01:55 AM rue_shop4: plausable
01:55 AM rue_shop4: hah, your out, 20Mhz
01:55 AM rue_shop4: :) FAIL
01:55 AM rue_shop4: its drifting, but they are ceramic caps
01:55 AM rue_shop4: core...
01:56 AM rue_shop4: ferrite...
01:57 AM rue_shop4: that one takes it down to 9Mhz
01:57 AM rue_shop4: this one....
01:58 AM rue_shop4: 13Mhz
01:58 AM rue_shop4: a loopstick takes is down to 6
01:58 AM rue_shop4: 6Mhz
01:59 AM rue_shop4: if I can get down to 2Mhz, the problem it beat
01:59 AM mrdata: 20 MHz is about where i figured it would be
01:59 AM mrdata: use a bigger coil
02:00 AM mrdata: how many turns were there? 10?
02:00 AM rue_shop4: its not wanting to go down to much less than 6Mhz
02:00 AM rue_shop4: 10 turns
02:01 AM rue_shop4: 2 caps in series, result 110 pf
02:01 AM rue_shop4: 20Mhz is ok, I can throw in a /2
02:02 AM rue_shop4: but I need the range
02:02 AM mrdata: but for higher Q, increase the inductance
02:02 AM rue_shop4: to get a 10:1 freq ratio, I need a lower Q dont I?
02:02 AM mrdata: to 1/2 the frequency, 4x the inductance
02:03 AM rue_shop4: yea, and i need to 1/10 the freq
02:03 AM mrdata: 10 you have 10 turns, try 20 turns
02:03 AM rue_shop4: it'll go to 30mm long
02:04 AM rue_shop4: I'm breaking out the hard stuff, smps ferrite
02:04 AM mrdata: ok
02:05 AM rue_shop4: using evil tricks I'm down to 4Mhz
02:06 AM mrdata: with ferrite in?
02:07 AM rue_shop4: yea and around
02:07 AM mrdata: um,
02:07 AM mrdata: ew
02:07 AM mrdata: lol ok
02:08 AM mrdata: because you had to make the gap smaller to use get effective permeability > 1
02:08 AM mrdata: ok
02:08 AM mrdata: um,
02:08 AM * rue_shop4 burries it in peices of ferrite
02:08 AM rue_shop4: hey, what if I do like they do in welders
02:09 AM mrdata: 20 down to 4 is getting close to the range you wanted
02:09 AM mrdata: now if those caps were variable,
02:09 AM mrdata: hmm...
02:10 AM mrdata: so now i think it should be 20 turns at least
02:10 AM mrdata: so you dont have to increase the caps
02:10 AM mrdata: keep them small
02:10 AM mrdata: a pair of 100 pF trim caps would do i bet
02:11 AM mrdata: also this solenoid is single layer but as you scrunch the bundle of wires closer together, the inductance goes up
02:11 AM rue_shop4: 3.3Mhz....
02:11 AM mrdata: not bad
02:13 AM rue_shop4: ok wait, I bolt should snub it out, right?
02:13 AM mrdata: ?
02:14 AM mrdata: what
02:14 AM mrdata: oh a steel bolt?
02:14 AM rue_shop4: ? allen key makes it go up
02:14 AM mrdata: i'm not sure
02:14 AM mrdata: carbon steel isnt super permeability; soft iron is best for low frequencies
02:15 AM mrdata: silicon steel is used for transformer cores at 50 or 60 Hz often
02:15 AM mrdata: in thin layers, insulated from each other by a thin layer of glue
02:16 AM rue_shop4: eddy losses, kill the osc
02:16 AM mrdata: yes
02:17 AM rue_shop4: SO CLOSE
02:17 AM rue_shop4: I put a flyback core thru it, full loop, its not gonna go lower than 3.3Mhz
02:17 AM mrdata: we're into test and adjust now
02:18 AM rue_shop4: if it cant get to 2Mhz, its beat
02:18 AM mrdata: double up the turns
02:18 AM rue_shop4: delta Ur isn't high enough
02:18 AM mrdata: doesnt have to remain single layer
02:18 AM rue_shop4: on 15cm or 30cm?
02:18 AM rue_shop4: k, to 15cm then
02:19 AM mrdata: i hope you meant mm
02:19 AM rue_shop4: hah yea
02:19 AM rue_shop4: "yer coil got a bit long there..."
02:19 AM rue_shop4: oh ohohohohoh, I can use the transformer tape I got!
02:19 AM mrdata: whats that
02:23 AM rue_shop4: ugh, I failed coil winding class
02:23 AM rue_shop4: wait, there was no coil winding class
02:23 AM rue_shop4: its the tape they put between layers to make them orderly
02:25 AM mrdata: oh
02:25 AM mrdata: sounds useful
02:27 AM rue_shop4: nup, starts at 9.4Mhz now
02:27 AM rue_shop4: so, for test sake
02:28 AM rue_shop4: 940Khz
02:29 AM rue_shop4: nope, flyback core only takes it down to 1.2Mhz
02:29 AM rue_shop4: not enough range
02:29 AM mrdata: close, tho
02:30 AM mrdata: seems to be getting better
02:30 AM rue_shop4: the 20Mhz coil was only 1Mhz out
02:31 AM rue_shop4: 20Mhz/3Mhz is 66
02:31 AM rue_shop4: er 6.6
02:31 AM mrdata: 9.4 / 1.2 = 7.88
02:31 AM rue_shop4: ah, yea 9.4 to 1.2 is yea
02:32 AM mrdata: so, replace the 220 pF caps with trim caps?
02:32 AM rue_shop4: the exact freq wouldn't matter
02:32 AM rue_shop4: I need to have one control that takes the osc 10:1
02:33 AM mrdata: thing is, theyre variable
02:33 AM rue_shop4: yes, if I can get the 10:1 I can trim it to where I need it
02:33 AM mrdata: next try 100 pF caps instead of 220 pF ?
02:34 AM rue_shop4: hell, at this point I'm willing to use a mixer to move the freq if I have to
02:34 AM rue_shop4: back up to 20Mhz?
02:34 AM rue_shop4: think that would get 8:1 or so?
02:34 AM mrdata: this will bring frequency to i think 13.9 MHz to 1.8
02:35 AM mrdata: and try 3 layers of coil
02:35 AM rue_shop4: I'm gonna stare at an rc osc in a book they said could peak at 10Mhz for a bit
02:35 AM rue_shop4: besides, its a work night and its already 12:35am
02:36 AM mrdata: later here
02:36 AM mrdata: i should sleep
02:36 AM mrdata: so i'm saying a combo of variable inductor and variable caps will get you your range
02:36 AM rue_shop4: you know
02:36 AM rue_shop4: there is series tuning
02:36 AM rue_shop4: if one were to size the inductor such that most of it was canceled out between 0-160pf
02:37 AM rue_shop4: I cant combo them
02:37 AM rue_shop4: I kinda need one dial here
02:37 AM rue_shop4: whats on the opposite side of 160pf
02:38 AM rue_shop4: well, a pH isn't gonna help me
02:39 AM mrdata: i wonder: can you control your parasitic capacitance so that you have less than 5 pF of it total?
02:39 AM rue_shop3: its gonna be pretty low now, its dead bug
02:40 AM rue_shop3: they have a circuit here, using an op-amp, that, with RC, does 20Mhz
02:40 AM rue_shop3: got to be one hell of an op-amp
02:41 AM rue_shop3: HA2539
02:41 AM mrdata: then you might have a ganged variable cap to give 0 to 100 pF
02:41 AM mrdata: and wind that inductor just right
02:41 AM rue_shop3: I have a radio cap that my meter says is 0-160pf
02:41 AM mrdata: three terminals?
02:41 AM rue_shop3: yes, the other side is 0-90pf
02:42 AM rue_shop3: ugh, the ha2539 is a 600Mhz opamp
02:42 AM mrdata: try it in place of those 220 pF caps then
02:42 AM rue_shop3: its lopsided tho
02:43 AM rue_shop3: one side is 0-90 and the other 0-160
02:43 AM mrdata: but maybe thats okay
02:43 AM rue_shop3: hmm
02:43 AM rue_shop3: I can try tommorow
02:43 AM mrdata: if theyre unequal, then try both ways
02:43 AM rue_shop3: arg, hard life of a night-owl
02:43 AM mrdata: one direction will give better results
02:43 AM rue_shop3: it'd be the same, the series capacitance
02:44 AM rue_shop3: ?
02:44 AM rue_shop3: do I have a 600Mhz amp around?
02:44 AM mrdata: i found this with another oscillator; one cap should be smaller than the other
02:44 AM rue_shop3: whats a video amp?
02:44 AM mrdata: slightly
02:44 AM mrdata: but which one, matters
02:44 AM rue_shop3: wait... I do have a high freq amp
02:44 AM rue_shop3: a few
02:47 AM rue_shop3: 120Mhz
02:47 AM rue_shop3: hmmm
02:47 AM rue_shop3: can I do this with a 120Mhz op-amp?
02:48 AM rue_shop3: wtf is the bandwidth of a video amp?
02:49 AM mrdata: a couple megahertz methinks
02:49 AM rue_shop3: NE592...
02:49 AM mrdata: but that depends on NTSC or HD
02:49 AM mrdata: etc
02:50 AM mrdata: idk
02:50 AM rue_shop3: 90Mhz...
02:50 AM mrdata: i must sleep
02:50 AM rue_shop3: gnight
02:50 AM rue_shop3: uA733...
02:50 AM mrdata: your oscillator is almost where you want it to be
02:51 AM mrdata: a 160 pF and 90 pF in series are like a 57.6 pF overall
02:51 AM mrdata: so if total parasitics are below 5 pF
02:51 AM rue_shop3: 120Mhz...
02:51 AM mrdata: oh wait
02:51 AM mrdata: you need 100:1 capacitance range
02:51 AM mrdata: argh
02:51 AM rue_shop3: :)
02:52 AM rue_shop3: atleast this problem is keeping more than just me awake
02:53 AM rue_shop3: I wonder if a gain of 10 is enough
02:53 AM mrdata: maybe
02:54 AM mrdata: ok; nn
02:54 AM rue_shop3: ... can I just trick an IF amp into this?
03:06 AM rue_shop3: seems to me, that a 120Mhz op-amp, at 10Mhz, has a remaining gain of 12
03:06 AM rue_shop3: this circuit uses 2.5
03:06 AM rue_shop3: so, there should be enough room
04:32 AM RJ722 is now known as _RJ722_
04:33 AM _RJ722_ is now known as RJ722_
09:02 AM rue_: apparently a comparator dosn't ahve a GBP
09:02 AM rue_: cant think of the part numbers for any high speed amps I might have
09:02 AM veverak: aok
09:02 AM veverak: I need reasonable way to subdivide simplexes
09:02 AM veverak: (n dimensions)
09:02 AM rue_: a cable tuner must have an IF amp that I can use
09:03 AM rue_: veverak, triangles?
09:03 AM veverak: n-dimensional simplexes
09:03 AM veverak: rue_: so naswert "also triangles"
09:03 AM rue_: veverak, or circle/triagle
09:03 AM veverak: *answer
09:03 AM rue_: ah, 3d too eh?
09:04 AM veverak: 4d/5d/6d...
09:04 AM rue_: they have not completely worked out intersections in librecad yet, its been years
09:04 AM rue_: veverak, hah, mutlidimensional cad eh?
09:04 AM veverak: nope
09:04 AM veverak: motion planning
09:04 AM rue_: "WTF, WHY DID YOU STOP AT 3 DIMENSIONS!"
09:04 AM veverak: :)
09:05 AM rue_: so, your doing 6d hypercube intersections?
09:05 AM veverak: nope, just need to divide simplexes
09:06 AM rue_: how did you break past 3d doing motion planning?
09:06 AM veverak: the simplexes are mapped into 3D, the actual mapping (as what it is supposed to be), creates a objectin 3D that has curves as edges
09:06 AM veverak: *object in
09:06 AM veverak: that is problematic to work with
09:06 AM veverak: so, I want to use the approach "make the simplex small enough, so the curves are neglible and consider them lines"
09:07 AM rue_: not many people need to do timespace motion planning
09:07 AM veverak: rue_: you consider each degree of freedom of robot as one dimension
09:07 AM veverak: the mamping from N dimensions to 3D, is simply kinematics :)
09:08 AM rue_: I need to identify op-amps I might have with a GBP>25Mhz
09:09 AM rue_: I wonder what the amps for the vcr heads is
09:09 AM veverak: this does test my math knowledge tbh
09:09 AM veverak: so, barycentric subdivison seems reasonable
09:09 AM rue_: I'm not good with timespace, I am usually out of both of them
09:11 AM rue_: I was thinking that I cant use the if-amp in most of the radio chips, cause its voltage controlled, but ya know, agc on an oscillator is kinda the goal anyhow
09:13 AM veverak: let's KISS it
09:13 AM veverak: I think that I will simply divide the simplex into two on it's longest edge
09:14 AM rue_: AM IF amp wont work, 455Khz
09:15 AM rue_: the FM radio has no FM IF
09:15 AM rue_: maybe I'm thinking for video
09:15 AM rue_: iirc video is AM
09:15 AM rue_: uses a Mhz level IF
09:16 AM rue_: At the cable company's set top box, the signal is converted to a lower IF of 480 MHz for
09:16 AM rue_: yea, a 480Mhz IF amp would work wonders
07:07 PM rue_: NE5539
07:28 PM mrdata-: whats that
07:28 PM mrdata-: hmm... https://datasheet.octopart.com/THS4021-Texas-Instruments-datasheet-152340.pdf
07:30 PM mrdata-: LNA
07:36 PM mrdata-: rue_, i dont get it; are you thinking of frequency synthesis now, using 350 MHz and divide by N?
07:40 PM mrdata-: by the way, concerning the sliding core idea, the smaller the gap when fully inserted the better
07:41 PM mrdata-: try to get gap less than 0.1 mm
08:33 PM rue_: no
08:33 PM rue_: 350 amp to gnerate a 10Mhz rc osc
08:37 PM mrdata-: controlled how
08:38 PM mrdata-: variable R?
08:38 PM mrdata-: what's the Q of this
08:38 PM mrdata-: what stability do you need
09:20 PM rue_bed: idea
09:20 PM rue_bed: variable tap inductor
09:20 PM rue_bed: 1000:1?
09:25 PM mrdata-: that will not vary continuously; do you need continuous variation?
09:56 PM mrdata-: meanwhile, https://www.youtube.com/watch?v=WSKi8HfcxEk
11:41 PM rue_: yes, I need continious variation
11:41 PM rue_: I THOUGHT OF SOMTHING
11:41 PM rue_: so
11:42 PM rue_: the 20Mhz went from 20Mhz to 3Mhz, thats a difference of 13Mhz
11:42 PM rue_: if you DOWN CONVERT THAT to start at 10Mhz, it will take you to 7Mhz!
11:42 PM rue_: !!!!!!
11:42 PM rue_: ??
11:45 PM rue_shop4: I pulled some chips out of a cable tuner, so I'm just seeing what they have for an RC osc