#robotics Logs

Jul 06 2018

12:18 AM rue_shop4: whats 17db in normal
12:56 AM z64556 is now known as z64555
01:19 AM mrdata: rue_shop4, using this calculator, http://www.dicks-website.eu/coilcalculator/index.html
01:19 AM mrdata: i put the following parameters in,
01:21 AM mrdata: Bmax 0.4 tesla, Ae = 78.5 mm^2, μr = 3000, le = 45 mm, gap = 0.4 mm; and effective permeability is calculated to 108
01:22 AM mrdata: this means the sliding core idea can work, if you control the gap to at most 0.4 mm and have a high enough μr
01:23 AM mrdata: that is, gap is at most 0.4 mm when fully inserted
01:23 AM mrdata: seems to me, it's possible to get the gap smaller than that, ven
01:23 AM mrdata: *even
01:24 AM mrdata: and if you can, then effective permeability is higher
01:25 AM mrdata: so your 100:1 range from fully inserted to removed will achieve your 1:10 MHz
01:42 AM mrdata: ok, i looked at inductance factor for gap 15 mm (fully removed) to 0.15 mm (fully inserted) and the ratio is 100
01:50 AM mrdata: so with 5 turns, this is 1.642e-7 henry fully removed, and 1.494e-5 henry fully inserted
01:50 AM mrdata: oh
01:50 AM mrdata: that's not quite enough... fully inserted needs smaller gap i guess...
01:52 AM mrdata: ok, with max 0.1 mm gap, 5 turns is 2.14e-5 H fully inserted; and 1.64e-7 H fully removed
01:52 AM mrdata: more than your needed 100x
02:08 AM mrdata: 0.130 mm total gap is still okay; so divide this by two, to get 0 .. 65 μm as your tolerance for both sides
02:09 AM mrdata: i'll draw a sketch of what i think this looks like
02:19 AM rue_shop3: ok I been thru almost every IF chip I have, dont seem to be any amps that are gonna be a quick fit
02:20 AM rue_shop3: mrdata, and there is the downconverter idea
02:20 AM mrdata: how does that work
02:20 AM rue_shop3: 6:1 @ 20Mhz, is enough delta F at 10Mhz to 10:1
02:20 AM rue_shop3: you mix it to down-convert it
02:21 AM mrdata: interesting idea
02:21 AM rue_shop3: so if you mix, for example, 19-21Mhz with 18Mhz you get 1-3Mhz
02:21 AM rue_shop3: I think
02:21 AM mrdata: but that might get a big ugly
02:21 AM mrdata: how clean do you want your wave to be
02:21 AM rue_shop3: yes, so I'm going for a transistor based wien bridge
02:21 AM rue_shop3: I want square, so I dont care
02:22 AM mrdata: ok
02:22 AM rue_shop3: I'll come back around to the LC if the RC thing bails on me again
02:22 AM rue_shop3: fff its 12:30
02:22 AM rue_shop3: this project is killing me at work
02:23 AM rue_shop3: I dont even REALLY need it, its just got me hooked now
02:23 AM rue_shop3: ):|
02:27 AM rue_: my finger is hovering over the button to buy 10x NE5539 for $15
02:28 AM rue_shop3: any day of the week I'd say someone using a op-amp in an RF circuit is tooo green out of school
02:30 AM rue_shop3: http://www.nutsvolts.com/uploads/wygwam/NV_0901_Marston_FIG7.jpg
02:30 AM rue_shop3: rrrr
02:33 AM rue_shop3: ok, maybe this is the test then
02:38 AM mrdata: the PN2222 seems to be okay?
02:39 AM * rue_shop3 pukes
02:39 AM rue_shop3: how about...
02:39 AM rue_shop3: 2N3858 ?
02:39 AM rue_shop3: 9014?
02:40 AM * rue_shop3 starts an elaborate incantation
02:40 AM rue_shop3: 2SC2026
02:40 AM rue_shop3: }:)
02:41 AM rue_shop3: YOU WANT BANDWIDTH!?!?!
02:43 AM rue_shop3: oh, hehe, the 2222 is better than the 3858
02:44 AM rue_shop3: well 2sc1856
02:45 AM rue_shop3: ok wait, I should just get soemthing working first
02:45 AM rue_shop3: arg
02:46 AM rue_shop3: I ahve 2 2sc2026, I'll start with 2222 tho
02:47 AM rue_shop3: there are lots of nice transistors in these UHF tuners that I cant identify
02:47 AM rue_shop3: :/
02:50 AM rue_shop4: ouch, transistor amp gain,
02:50 AM rue_shop4: RE/RC ?
02:51 AM mrdata: find data sheet; look at gain bandwidth product
02:51 AM rue_shop4: for the 2026 its 2Ghz
02:51 AM rue_shop4: }:)
02:51 AM mrdata: cool
02:52 AM rue_shop4: I have two of them
02:52 AM rue_shop4: maybe other stuff but {480 square feet of workshop filled to the 8' mark}
02:53 AM rue_shop4: I want to rejig the amp I see online for self biasing...
02:53 AM rue_shop4: Hfe = fn(Re,Rc)
02:53 AM rue_shop4: ...
02:53 AM rue_shop4: aha
02:53 AM rue_shop4: https://backyardbrains.com/experiments/img/GainCalc0.jpg
02:53 AM rue_shop4: that sounds clear
03:12 AM mrdata: your workshop is larger than my residence
03:12 AM mrdata: 400 sq ft two bedroom
03:13 AM mrdata: and the 2nd room is occupied
03:13 AM mrdata: i originally planned for it to be a work shop but things didnt work out financially
03:13 AM rue_shop4: :(
03:15 AM rue_shop4: I better be carefull of the current on these
03:16 AM rue_shop4: operating point 1mA for the transistor...
03:17 AM rue_shop4: 1.5k 10pF, 10Mhz, whats the current?
03:17 AM rue_shop4: z of a 10pf cap at 10Mhz?
03:18 AM mrdata: X = 1/ ( 2 pi f C )
03:19 AM rue_shop4: 1.6k
03:19 AM mrdata: yep
03:19 AM rue_shop4: so, hfe = 0.5 @ 10Mhz, cause its up against... no...
03:20 AM rue_shop4: oh, actually the wien bridge will have a lot of attentuation
03:21 AM rue_shop4: the low side of the divider of it is about 750R
03:21 AM rue_shop4: the top side is 3k
03:21 AM rue_shop4: 0.2
03:21 AM rue_shop4: and the amp gain is set to 5
03:22 AM rue_shop4: oo just on the margin
03:22 AM rue_shop4: but it gets worse, at 1Mhz
03:23 AM mrdata: worse?
03:23 AM mrdata: at 1MHz the transistor has an easier time of it i would have thought
03:24 AM rue_shop4: well, the bride has 1k||10pf on the bottom, and 10pf||(1k5-15k)
03:24 AM rue_shop4: at 1k5 its running 10Mhz, at 15k its 1Mhz
03:26 AM rue_shop4: bottom at 941ohm, top at 31k
03:26 AM rue_shop4: 0.029
03:26 AM rue_shop4: I'm gonna need more than Av=5
03:27 AM rue_shop4: gain of 50 would do...
03:27 AM rue_shop4: :/
03:27 AM rue_shop4: I think this is the design crunk
03:28 AM rue_shop4: gbp=200, gain=50...
03:28 AM rue_shop4: 4Mhz
03:28 AM rue_shop4: gbp = 2000Mhz, gain=50
03:29 AM rue_shop4: 40Mhz
03:29 AM rue_shop4: huh
03:29 AM rue_shop4: GBPmin=
03:29 AM rue_shop4: 500Mhz
03:29 AM rue_shop4: iiinteresting
03:30 AM mrdata: ok! now you can select your transistor based on that
03:30 AM rue_shop4: so, its a 2 stage amp, maybe I can do gains of 5 and 10
03:31 AM rue_shop4: or make it a 4 stage amp, how many degrees of phase am I allowed?
03:31 AM rue_shop4: ugh
03:32 AM mrdata: phase does count
03:32 AM mrdata: idk
03:32 AM rue_shop4: so, feedback is a few cycles behind is ok... :)
03:33 AM rue_shop4: haha, uh
03:34 AM rue_shop4: the pn2222 amp I have here, with a gain of 5, starts over its gbp@80Khz
03:34 AM rue_shop4: or its self resonating...
03:35 AM rue_shop4: its got a gian peak at 80Khz
03:36 AM rue_shop4: huh, 400khz is the start of the -3db
03:37 AM rue_shop4: hmm pretty good tho, I have the freq gen set to min. and its outputing 11Vp-p
03:38 AM rue_shop4: input is shy of 150mv p-p
03:38 AM rue_shop4: so, its gain jumped to 73
03:38 AM rue_shop4: huh
03:38 AM rue_shop4: supposed to be 5, damn math
03:42 AM mrdata: hmm
03:42 AM mrdata: this is why i model with ltspice
03:43 AM rue_shop4: ok, so, I'll call it a night, I established that greasy tacos suck, that I dont have any if-amps that by the book seem suitable, and that a transistor wien bridge osc is plasable
03:44 AM mrdata: k
03:45 AM mrdata: i made a quick sketch of a gapped coil
03:46 AM mrdata: uploading...
03:48 AM mrdata: ugh. browser crashed
03:48 AM mrdata: this will take another few minutes
03:50 AM rue_: that last coil was pretty tight
03:53 AM mrdata: https://imgur.com/a/9ROLTnJ
03:53 AM mrdata: so the flat parts have to get as close as possible
03:53 AM mrdata: when it is fully inserted
03:54 AM mrdata: total gap is the figure that matters
03:54 AM mrdata: each wing of this is g/2
03:54 AM rue_: yes, but then I ahve to take the other ferrite right out to get the freq up
03:55 AM rue_: the lower peice only contributes about 25% the low end
03:55 AM mrdata: so the piece inside the coil slides in or out fully
03:56 AM rue_: I was, that took it from 20Mhz to ...5?
03:56 AM rue_: the lower peice took from 5 to 3
03:56 AM rue_: at best
03:57 AM mrdata: your total gap must have been higher, or the relative permeability of your ferrite wasnt as high as 3000 ish
03:58 AM rue_: I only had two peices that fit in the coil, a loopstick, and the flyback TV core
03:58 AM mrdata: as i calculated above, g/2 has to be within 65 micrometers
03:59 AM mrdata: was the loopstick circular? or did it have flat sides
03:59 AM rue_: how thick is the shelack on the ferrite?
03:59 AM rue_: two flats
03:59 AM mrdata: oooooooooooooh shelac
04:00 AM mrdata: idk how thick 65 µm is like the thickness of copper on a pcb
04:00 AM mrdata: pretty thin
04:02 AM Triffid_Hunter: standard pcbs are 35µm
04:02 AM mrdata: ok
04:05 AM mrdata: i believe it's possible to engineer this; but, downconverting by mixing is interesting and could work too
04:06 AM rue_: the osc we made last night had enough range if it was downconverted
04:07 AM rue_: would need to come down by 10Mhz, which I think means mixing it with 10Mhz, which I could use a crystal osc
04:07 AM rue_: 20-10=10
04:08 AM rue_: I have LOTS of rf mixer chips with all sorts of jazz in them
04:08 AM rue_: tho, video IF is only 4.5Mhz
04:12 AM mrdata: mixing with 4.5 MHz will get both sum and difference, and you cant filter that
04:14 AM rue_: yes, but I need to mix the 20Mhzvar with 10Mhz
04:14 AM rue_: then put it thru a 1-10Mhz filter
04:14 AM rue_: (low Q?)
04:14 AM rue_: and there needs to be AGC, so the signals mix at the same levels
04:14 AM mrdata: this runs into filtering trouble if you dont have a really sharp low pass filter
04:14 AM rue_: the mixer chips usually do all that
04:15 AM rue_: well it could be a lowpass
04:15 AM rue_: kill anyhting under 20Mhz quite dead by 10Mhz
04:15 AM rue_: er, vise versa
04:15 AM rue_: there is that gap
04:16 AM mrdata: so ideally you want to begin at higher frequency, say 24 to 15 MHz and mix with 14 MHz, then low-pass filter 10 MHz cutoff
04:16 AM rue_: anything over 10Mhz can go, and the nearest noise is at 20Mhz, there will be a band up at 30Mhz, but who cares
04:16 AM rue_: if the 24Mhz osc will give enough range
04:17 AM rue_: is sounds doable tho
04:17 AM rue_: a wien bridge osc seems to much easier!
04:17 AM rue_: I'll try the wien
04:18 AM mrdata: ok
04:18 AM rue_: haha its 2am
04:18 AM rue_: DOOMED
04:18 AM mrdata: 5:18 am here;
04:18 AM mrdata: i'm going to sleep
04:18 AM mrdata: nn
04:18 AM rue_: I ahve to insall a phone system tommorow, I hope I dont screw it to the wall upside down or soemthing
04:18 AM mrdata: people will have to stand on their heads to use it
04:18 AM mrdata: no prob
04:18 AM mrdata: lol
04:19 AM rue_: ;)
09:19 AM weyland|yutani: https://youtu.be/QZ1DaQgg3lE
07:03 PM rue_: nice jump
07:09 PM rue_: oh, the gain of a wien bridge is always 1/3
07:09 PM rue_: huh
07:14 PM Tom_L: phone system upside down?
09:43 PM Jak_o_Shadows: Chaindriven knees is neat
09:47 PM rue_shop3: yea
09:48 PM rue_shop3: the replaement system turned out to be wrong, I had to scramble for something else
09:48 PM rue_shop3: *so far* it worked out in the end
11:03 PM Jak_o_Shadows: and the 3d printer is workiong again! For more than 30 minutes at a time!
11:04 PM Tom_L: that's a good thing since most prints are likely longer than that
11:04 PM Jak_o_Shadows: Yup1
11:05 PM Jak_o_Shadows: the problem was heat creep - so i changed to an insulated hollow bolt 9that joins the heat-sink to the hot-bit of the hot-end
11:10 PM rue_shop4: you have a fan on the hot end tho, right?
11:11 PM rue_shop4: you need one unless your running a quartz tip, trust me
11:11 PM rue_shop4: mrdata, ok, :) I have a wien bridge doing 1.17Mhz
11:11 PM rue_shop4: on a solderless breadboard!
11:11 PM Jak_o_Shadows: i have one on the heat-sink. No fan blowing air over the nozzle itself.
11:11 PM rue_shop4: with pn2222!
11:11 PM rue_shop4: yea heatsink fan
11:12 PM Jak_o_Shadows: Yup yup. It's one of those little 12V fans.
11:12 PM Jak_o_Shadows: i have a little bit of protoboard on the bottom of the cold-end, where the extruder mount plate is
11:12 PM Jak_o_Shadows: Has some LED's and breakouts for 12V, 5v, gnd
11:27 PM Tom_L: mower bot idea: http://www.evatech.net/PRODUCT.php?ID=6
11:49 PM mrdata: rue_shop4, ok!
11:49 PM mrdata: how about the upper end ~ 10 MHz?