#robotics Logs
Jul 09 2018
#robotics Calendar
12:10 AM rue_shop3: I'm sure its $5 somewhere
12:14 AM rue_shop3: I cant digikey, can you tell me if they have it?
12:20 AM rue_: !!!!!!!!!! its avialable from china!
12:20 AM rue_: aliepress
12:22 AM rue_: https://www.aliexpress.com/item/free-shipping-used-ADE-1-24-ADE-1-Surface-Mount-Frequency-Mixer-SOP-6/32829158139.html
12:23 AM rue_: suppose I can order two and work on all the other ideas over the 2 months it'll take to arrive
12:30 AM mrdata: \o/
12:36 AM rue_: NOW THAT I'V PUT $4.40 INTO THIS PROJECT IT BETTER WORK!
12:37 AM rue_: I really didn't think it would be on aliexpress
12:40 AM mrdata: cute little thing
12:41 AM mrdata: but it's a thing that stuff on the ITAR export restricted list ends up being produced abroad, so that manufacturers can get around the export controls
12:43 AM rue_: ITAR export restricted list ?
12:43 AM mrdata: yeah, supposedly frequency mixers are on the list; idk why
12:44 AM mrdata: some dinosaur wrote it i suppose
12:44 AM rue_: whats itar?
12:44 AM rue_: zhanx, ?
12:44 AM mrdata: it's a treaty; it's about items that have both civilian and military use
12:45 AM mrdata: https://gov-relations.com/itar/
12:45 AM rue_: ok, military stuff, thought so
12:46 AM mrdata: for giggles, i wrote some crypto code awhile back and exported it to USA, so according to that treaty i'm an arms dealer
12:47 AM mrdata: consequently my poetry is classed as a weapon and may not be exported to listed countries
12:50 AM rue_: Vogon eh?
12:51 AM logicmoo is now known as dmiles
12:51 AM rue_: so, yesterday I learned about SAW oscillators
12:51 AM mrdata: i have to put labels on it.. "warning: this poetry may contain steganography, peanuts or other nuts"
12:52 AM rue_: AND that those cheap little chineese rf transmitters are just SAW oscillators
12:53 AM rue_: https://en.wikipedia.org/wiki/Vogon see poetry
12:57 AM rue_: the 315Mhz saw units are quite available, but they are only 75Khz bandwidth
12:59 AM mrdata: not ideal for your use
01:00 AM mrdata: whereas that little mixer is extremely simple
01:00 AM rue_: there is a 10.7, buts its only +-30Khz
01:00 AM rue_: I got the mixers
01:00 AM rue_: the saw things fascinates me tho
01:00 AM rue_: I DID find a 10Mhz saw filter in a vcr
01:01 AM rue_: but iirc its like 419Mhz
01:01 AM rue_: (10Mhz bandwidth)
01:02 AM rue_: atleast at <30Mhz, I dont have to go insane with shielding
01:02 AM rue_: but it does look like I'll end up having to etch a board
01:02 AM mrdata: you should put the oscillators and mixer in cans anyway
01:03 AM rue_: yes
01:03 AM rue_: the plan is a module with the 1-10Mhz out
01:03 AM rue_: well, I'm still trying to find an electronic tuner
01:03 AM mrdata: follow all the rules of thumb of RF
01:04 AM rue_: mrdata, right "A: electricity goes thru air. B: electricity does not go thru wires."
01:04 AM rue_: I'm gonna need help getting the signal levels in the right order of magnitude
01:04 AM mrdata: in microwave range that's precisely what happens, because parasitics dominate
01:05 AM mrdata: but down < 30 MHz, it's the more usual stuff: keep signal lines as straight and as short as possible
01:05 AM rue_: I gave up the day the rf output of a tv modulator was a wire from the middle of the coax to the ground plane
01:05 AM mrdata: connect modules with 50 ohm coax
01:05 AM rue_: no intermodules
01:05 AM rue_: one module
01:06 AM mrdata: continuous ground plane
01:06 AM mrdata: pay attention to skin depth
01:06 AM mrdata: put the module in a can
01:06 AM mrdata: SMA or BNC out
01:06 AM rue_: did you see the variac stuff I found today?
01:07 AM mrdata: no?
01:07 AM rue_: er
01:07 AM rue_: varactor
01:07 AM rue_: really cool!
01:07 AM rue_: using generic diodes and transistors (reverse biased BC junctions) as varactors
01:08 AM rue_: the power transistors 2SC4020 vary down to about 30% of their capacitance at ~17V
01:09 AM mrdata: other RF rules: place inductors so as to minimize coupling
01:09 AM mrdata: control parasitic capacitance
01:09 AM rue_: ^^ see the fun I had
01:09 AM mrdata: checking...
01:12 AM mrdata: wrong page; that only shows me the last page
01:12 AM rue_: I didn't get capacitance values, I just ran them in an RC osc and compared the freq differences
01:12 AM mrdata: i have scrollback tho
01:12 AM mrdata: how far back
01:12 AM rue_: http://tom-itx.no-ip.biz:81/~tom-itx/irc/logs/%23robotics/2018-07-08.html
01:12 AM rue_: "5:41"
01:17 AM rue_: https://www.aliexpress.com/item/2300-2500MHZ-VCO-VCO-crystal-2-4G-original-new-preferential/32714918670.html
01:17 AM rue_: two of those, against each other, could do this easy, I just cant make the board
01:19 AM rue_: https://www.aliexpress.com/item/VCO-voltage-controlled-oscillator-SM450-400-470MHZ/32844525891.html
01:20 AM rue_: I'm almost wondering if thats doable, I think that mixer can handle them
01:25 AM mrdata: what is the oscillator's Q-factor
01:25 AM mrdata: ie how stable is the frequency
01:26 AM rue_: no idea, they use them for things like wifi tho
01:26 AM rue_: https://www.aliexpress.com/item/Wideband-VCO-10dBm-signal-generator-for-RF-signal-source-of-70-200MHz-VCO-voltage-controlled-oscillator/32857343385.html
01:26 AM mrdata: the ltspice circuit i showed you has Q > 1500 but more likely about 2500. so stability might be 6 to 10 kHz variation
01:26 AM rue_: ^^ that ones interesting, they use a colpitts, with all the capacitors varistors
01:26 AM mrdata: what do you need
01:27 AM rue_: well, the HP signal gen I have now drifts by about 10hz, which seems pretty good
01:27 AM rue_: I can pll it if I need
01:27 AM rue_: wel.. +-
01:27 AM mrdata: 10 Hz drift is excellent
01:28 AM rue_: it maxes out at 1Mhz
01:28 AM mrdata: a guy wanted 5.7 GHz with 0.1 Hz stability for an active SETI project
01:28 AM rue_: hmm
01:30 AM mrdata: and very low phase distortion
01:30 AM mrdata: the plan was to compensate for earth's motion w.r.t the sun
01:31 AM mrdata: so to provide a receiver at 100 light years away a strong spectral peak at a magic frequency
01:31 AM mrdata: it is a beacon
01:33 AM rue_: everyone knows where the watering hole is, nobody wants to go get stuck out there.
01:33 AM mrdata: to be eaten by crocodiles, yeah
01:35 AM mrdata: this frequency was related to the cesium and rubidium atomic clocks: f(Cs) - 0.5 f(Rb), iirc
01:36 AM mrdata: an unmistakable signature
01:36 AM mrdata: that just happens to fall in ISM and Ham bands
01:37 AM rue_: heh
01:38 AM mrdata: 5775290464.548 Hz
01:38 AM rue_: slightly specific
01:38 AM mrdata: ya think
01:38 AM rue_: do you know a good guideline for impedence of an RFC?
01:39 AM mrdata: RFC?
01:39 AM rue_: RF choke
01:39 AM mrdata: air core?
01:39 AM rue_: not usually
01:40 AM mrdata: i use two calculators: for air core, http://hamwaves.com/antennas/inductance.html ; and for ferrite core, http://www.dicks-website.eu/coilcalculator/index.html
01:41 AM mrdata: impedance, Z is sum of reactance and resistance; reactance is given by X[L] = 2 pi f L
01:42 AM mrdata: resistance depends on skin effect
01:42 AM mrdata: for which i have another calculator,
01:43 AM mrdata: http://www.mantaro.com/resources/impedance_calculator.htm
01:44 AM mrdata: when these calculators do not quite agree (which is often) i know there will be some tuning required
01:44 AM rue_: lookign for a target value tho
01:44 AM mrdata: ...
01:44 AM rue_: for the impedence
01:45 AM rue_: I know its "high" I just dont know whats a good high
01:45 AM rue_: in the old days, people would just throw a 'lot' of turns on something, often a resistors body
01:46 AM mrdata: 50 ohm interconnections are considered 'a la mode'
01:46 AM rue_: its like a DC blocking cap, ...
01:46 AM mrdata: so that coax can be used
01:46 AM rue_: what value is good for a dc blocking cap? 1k? 1M?
01:46 AM mrdata: what cutoff frequency?
01:47 AM mrdata: the 3pF cap in the circuit i showed you does DC blocking, but it is also kind of a high-pass filter
01:48 AM rue_: yes, well, thats just it
01:48 AM mrdata: i was just wondering if i should tweak its value a bit
01:48 AM mrdata: because there is some attenuation presently at high frequencies
01:49 AM mrdata: so X[C] = 1 / ( 2 pi f C )
01:49 AM rue_: if I use a oscillator with a varactor, and its min freq is 1Mhz, and I want a dc blocking cap, its going to have some impendence at 1Mhz, but lots more at 0.1Hz
01:49 AM mrdata: ^^
01:50 AM rue_: but what value is a god level for the dc impedence, 1000ohms? 1Mohms?
01:51 AM mrdata: if the 3 dB cutoff of a varactor is around 1 MHz, then at 0.1 Hz it will be something like 10 Megohms
01:52 AM rue_: OH!, thats cool!
01:52 AM mrdata: give or take a 2 pi factor and stuff
01:54 AM rue_: "votlage controlled phase shifter" huh...
01:54 AM rue_: I'm learning about too many new things on this project
01:55 AM rue_: the stuff to learn about is expanding like a trinary tree
01:55 AM mrdata: take an example: 50 ohm impedance at 1 MHz, then C = 1 /( 2 pi f X[C] ) = 3.18 nF; at 0.1 Hz then X[C] = 500 Megohm
01:58 AM mrdata: by the way that last link, the mantaro calculators, has a strip line calculator. so you can find the impedance of a signal trace on a PCB
02:04 AM rue_: arg, its a worknight
02:04 AM rue_: its 12:04am
02:04 AM mrdata: yeah i should sleep
02:04 AM mrdata: nn
02:04 AM rue_: and I REALLY HAVE to try this ONE MORE circuit
02:05 AM rue_: my mondays are HELL
02:05 AM rue_: "If you have a problem with it, prop me up with a stick!"
02:10 AM veverak: lol
02:15 AM mrdata: aha! yes, that 3pF should be 2pF; and the transistors should be 2n3904
02:15 AM mrdata: and add a 150 pF cap on output
02:20 AM rue_shop4: dude, I can only mentally handle so many changes
02:24 AM mrdata: this brings the high-end attenuation down quite a lot
02:24 AM mrdata: i'll post another schematic; not right now, i'm going to sleep
02:31 AM rue_shop4: ok, I have a voltage controlled delay, 156.25 - 104.6ns
03:37 AM rue_: mrdata, 8-| 42Mhz on breadboard with 74ls04, 1uH from input to output, with 0-160pF capacitor to ground on input side
03:54 AM Jak_o_Shadows: ohhh. This library hasn't included the actual code to do the i2c writing
03:54 AM Jak_o_Shadows: that makes sense
03:57 AM Jak_o_Shadows: literally second paragraph in the docs
03:58 AM rue_bed: ghaha
03:59 AM rue_bed: I need to varry the timing 900ns
03:59 AM rue_bed: over 8 delays
04:00 AM rue_bed: so 0-112.5ns ea
04:00 AM rue_bed: er, no cant work like that
04:00 AM rue_bed: 12.5 to 125ns ea stage
04:01 AM rue_bed: delta 112ns ouch
04:01 AM rue_bed: I got 57ns so far
04:01 AM rue_bed: if each stage is 2 delays I might pull it off
04:01 AM rue_bed: thats 16 delay circuits
04:01 AM rue_bed: hmm
04:02 AM Jak_o_Shadows: Have you thought about just using a big length of coax?
04:02 AM rue_bed: for a vco?
04:02 AM Jak_o_Shadows: hmm, yeah, the best you could do with that is discrete steps
04:02 AM Jak_o_Shadows: if you had multiple taps
04:04 AM rue_bed: no, the delays can be rc or lc, and use varactors made of transistors
04:04 AM Jak_o_Shadows: But if you had the space, you could have a huge range
04:04 AM rue_bed: 1-10Mhz
04:05 AM Jak_o_Shadows: You were talking about 60ns a moment ago. Thats like 20metres of coax
04:05 AM Jak_o_Shadows: Or fibre.
04:11 AM rue_bed: the goal is a 1Mhz-10Mhz variable oscillator
04:12 AM rue_bed: apparently 1:10 is insane range
04:15 AM veverak: so
04:17 AM veverak: got a job, will have to create a temp. sensor
04:17 AM veverak: or, board to measure a temperature of 3d printers developed by the company
04:17 AM veverak: (hotend temp.)
04:17 AM Jak_o_Shadows: I thought a large part of the problem was getting the 10? Couldn't you divide down from that without as much difficulty?
04:19 AM veverak: I am thinking of using thermocouple tbh
04:22 AM veverak: yup, thermocouple with proper design should be enough
04:25 AM Jak_o_Shadows: don't they already have a sensor?
04:26 AM Jak_o_Shadows: You want one that is independent?
04:32 AM veverak: yup, they want something for testing
04:32 AM veverak: but I don't have temp. ranges yet, so just thinking lout what are the possibilities
04:32 AM veverak: RTDs for example are also a potion ofc
08:26 AM rue_bed: the 100:1 range on the parts you need to get the 10:1 on the freq is the challange, esp to keep it stable
09:14 AM rue_: mrdata, huh, MAX2605
11:03 AM mrdata: i think i spoke too soon; with 2pF, Q drops factor of five
11:03 AM mrdata: compared to 3pF
05:24 PM swampmonster is now known as justanotheruser
06:24 PM z64556 is now known as z64555
07:10 PM rue_: mrdata, today I learned about Digital Locked Loops, and Ring Oscillators
07:10 PM rue_: (while on lunch)
09:19 PM z64556 is now known as z64555
11:38 PM mrdata: rue_, yes, good