UmbralRaptor changed the topic of #kspacademia to: https://gist.github.com/pdn4kd/164b9b85435d87afbec0c3a7e69d3e6d | Dogs are cats. Spiders are cat interferometers. | Космизм сегодня! | Document well, for tomorrow you may get mauled by a ネコバス. | <UmbralRaptor> egg|nomz|egg: generally if your eyes are dewing over, that's not the weather. | <ferram4> I shall beat my problems to death with an engineer. | We can haz pdf
e_14159_ has joined #kspacademia
e_14159 has quit [Ping timeout: 378 seconds]
<_whitenotifier-d13c>
[Principia] eggrobin opened pull request #2584: Point to the planned initial Frenet frame before ignition - https://git.io/JfzR3
<_whitenotifier-d13c>
[Principia] eggrobin opened issue #2585: principia.journal.serialization.Status should have a message - https://git.io/Jfz0v
<_whitenotifier-d13c>
[Principia] eggrobin labeled issue #2585: principia.journal.serialization.Status should have a message - https://git.io/Jfz0v
<_whitenotifier-d13c>
[Principia] eggrobin assigned issue #2585: principia.journal.serialization.Status should have a message - https://git.io/Jfz0v
<egg|laptop|egg_>
!wpn
* galois
gives egg|laptop|egg_ a continuous wing
* egg|laptop|egg_
pets a cat
<raptop>
!wpn ANBOcat
* galois
gives ANBOcat an instruction set
<raptop>
!wpn kotya
* galois
gives kotya a torch
* raptop
apparently needs to know the names of more peoples' cats
<raptop>
!wpn -add:wpn coffee
<galois>
Added wpn 'coffee'
<raptop>
!wpn -add:wpn tea
<galois>
Added wpn 'tea'
<egg|laptop|egg_>
the cat is purring at me
<egg|laptop|egg_>
!wpn whitequark
* galois
gives whitequark a field
<whitequark>
hi
<whitequark>
!wpn egg
* galois
gives egg a tellurium fuselage with a bazooka attachment
<whitequark>
is that rq-9
<egg|laptop|egg_>
tellurium fuselage? that seems odd
<whitequark>
speaking of which i want to build a drone. like a big drone. without any weapons systems but with an internal combustion engine and many hour loitering capability
<whitequark>
fixed wing, hybrid drivetrain, ideally 100% autopilot including takeoff and landing if i manage to pull it off
<whitequark>
made mostly out of laser cut plastic to make it cheaper and easier to rebuild in case of a loss
<whitequark>
well, more or less
<raptop>
Big drone with long loiter sounds like it would drift towards being a turboprop
<raptop>
Well, unless you want to do something silly like centurion or helios?
<whitequark>
my roommate gets all excited about building a turboprop
<whitequark>
i have less confidence about the ability to pull it off
<raptop>
heh
<egg|laptop|egg_>
well, unless you can source a surplus turboprop engine from something this sounds messy
<whitequark>
the current plan is to use an uh, 10cc petrol engine
<whitequark>
to give you a sense of scale
<raptop>
Obviously this is Russia so you can just go to the local boneyard and get one off of an old An-12
<whitequark>
"big" as in "big compared to quadcopters that can fly for four minutes"
<raptop>
Ah
<whitequark>
not as in "ten meter wingspan and you have to use an actual airstrip to fly it"
<egg|laptop|egg_>
ah, not big as in "yoink an Ильюшин and plonk some electronics in the cockpit"
<whitequark>
christ you folks have no sense of scale :p
<whitequark>
i just want to fly a camera and a radio relay link so i don't have to get my butt off the chair if i want to know what happens fifty kilometers from here
<egg|laptop|egg_>
well, my aerospace eggsperience is KSP so,
<whitequark>
the payload is like a kilo. the *drivetrain* seems like it'd weigh about a kilo at most
<egg|laptop|egg_>
fixed-wing at those scales sounds like you would need some sort of a smol strip though? or a ramp/catapult thingy?
<whitequark>
pick any road without cars
<egg|laptop|egg_>
or RATO,
<whitequark>
if we get far away from moscow that it becomes legal to fly it (coordinated with ATC and everything) there are lots of roads we could use
<whitequark>
far enough*
<whitequark>
apparently the process for legally flying a mid-sized UAV in russia is that you have to snail mail ATC
<whitequark>
that's it
<egg|laptop|egg_>
hah
<whitequark>
anyway the current plan is to design an airframe, get a dynamics model, plug it into jsbsim, write an autopilot that can do everything we need, and once that works, actually start building it seriously
<egg|laptop|egg_>
this sounds cool
<whitequark>
i have no idea what i'm doing
<whitequark>
well, mostly. i have a small amount of idea what i'm doing
<egg|laptop|egg_>
smol idea
<whitequark>
do you uh have any idea how one would go about designing an airframe
<whitequark>
like a very boring one conceptually
<whitequark>
it's mostly this is an optimization problem that has a circular dependency in it
<egg|laptop|egg_>
maybe WeylandsWings has some input
<whitequark>
8 mach
<whitequark>
poor cow
<egg|laptop|egg_>
or ferram but ferram4 does not appear to be here
<WeylandsWings>
wat
<egg|laptop|egg_>
whitequark: note that I suspect that the modeling here is very much spherical cow, the cow would not hold attitude or be rigid
<WeylandsWings>
there should be shocks on the front legs
<SnoopJeDi>
I would guess that designers do a lot of finite-element simulations as part of that iterative loop
<egg|laptop|egg_>
WeylandsWings: you are looking at a slice down the middle of the cow
<egg|laptop|egg_>
so you see the shock downwind of the legs
<WeylandsWings>
that would make sense
<WeylandsWings>
but then slice the model as well to make it more clear
<egg|laptop|egg_>
but you see the snout shock where it is
<whitequark>
poor cow ^2
<whitequark>
snout shock
<egg|laptop|egg_>
also udders,
<SnoopJeDi>
Like, I would be surprised if someone hasn't done airfoils in OpenFOAM, and it's a natural extension from that to design?
<whitequark>
SnoopJeDi: how do i avoid it becoming garbage in, garbage out
<egg|laptop|egg_>
transonic udders
<SnoopJeDi>
whitequark, I would personally make sure I understood enough of the analytical tools to check myself, and probably be able to work out a toy model.
<SnoopJeDi>
in terms of tying the two realms together I mean. Design philosophy itself...? *whistles* who knows
<whitequark>
hm
<SnoopJeDi>
Magnetostatics design gets off easy, here.
<whitequark>
how did people do this before FEM
<egg|laptop|egg_>
WeylandsWings: how does one into plane
<SnoopJeDi>
whitequark, other numerical methods
<whitequark>
point
<egg|laptop|egg_>
wood files and wind tunnels
<SnoopJeDi>
And popsicle-stick models
<SnoopJeDi>
Also yes, NACA did a lot of "go measure it, dummy" design
<SnoopJeDi>
and uh, so did most of early aviation history
<SnoopJeDi>
in maybe not the best sense of the word measure
<SnoopJeDi>
whitequark, NASA has some good CFD material IIRC, if you end up getting that far into the weeds.
<whitequark>
hm
<SnoopJeDi>
Although OpenFOAM lets you hold your nose a little bit from what I recall
<egg|laptop|egg_>
I mean, even not-so-early aviation, lots of nontrivial feedback comes from test flights with actual humans
<SnoopJeDi>
I dunno, I was spoiled for years with access to COMSOL
<SnoopJeDi>
But I remember being very impressed with OpenFOAM's domain specificity and how much better it felt than roll-your-own FEA
<whitequark>
oh i can just get COMSOL
<whitequark>
roommate mentioned ANSYS Workbench and -that- i cannot get, i think
<SnoopJeDi>
I would recommend reading through their documentation if you can also get that
<SnoopJeDi>
Ah, it's the same sort of stuff under the hood
<SnoopJeDi>
ANSYS is AFAIK a lot more specific in terms of its numerics
<whitequark>
ok but like i'm not building the next f-16
<SnoopJeDi>
COMSOL has some reaaally neat stuff under the hood because it's designed explicitly for multiphysics
<SnoopJeDi>
So like, CFD yea sure, but okay throw heat transfer in there now too
<SnoopJeDi>
anyway sorry, uhhh it's good and is the best interface I can think of for that kind of work
<SnoopJeDi>
and their documentation is very mathematical, they show you the variational forms they use and everything.
<whitequark>
cool, noted
<SnoopJeDi>
if you care, anyway
<egg|laptop|egg_>
whitequark: tbh I guess design a thing that vaguely looks like a plane, check that its basic dynamical properties are OK, do a few rounds of that, and get something in the air and there you will see how well the modeling modelled?
<SnoopJeDi>
Y'know, I bet this is the title of someone's monograph
<whitequark>
egg|laptop|egg_: mm, probably
<SnoopJeDi>
"Designing Flyingbois for Fun and Profit"
<whitequark>
so i guess my main priority is validating my chosen construction method
<whitequark>
because i'm going to be building quite a few of them
<SnoopJeDi>
Validating for what?
<whitequark>
as in making sure what i want to use is actually good
<SnoopJeDi>
yea but I mean good in what sense
<whitequark>
oh
<whitequark>
cheap and good mechanical properties
<SnoopJeDi>
okay, so lift and stress sims simultaneously, potentially?
<whitequark>
like every recent version of comsol is on rutracker
<SnoopJeDi>
I didn't know which "get" you meant :)
<whitequark>
i'm "production engineering group"
<SnoopJeDi>
So yea, COMSOL is definitely the shortcut here. The interface is...idiosyncratic, but it's got logic to it
<egg|laptop|egg_>
yeah and if you spend too much time on the CFD your design will be the aerodynamics group one,
<whitequark>
egg|laptop|egg_: look, i don't want to excel at cutting material, i want the thing to fly. and the thing will fly into the nearest swamp more than once
<SnoopJeDi>
I would guess if they have a article about it, there's probably aero folks on the forums, too
<SnoopJeDi>
There is a powerful wizard who lives there. Give him my regards, should you meet him. I never spoke with him, but tales of his magic rescued me from many a dark pit.
<whitequark>
so the airframe is going to look like "i could get this thing assembled by getting it all cnc cut"
<SnoopJeDi>
What's the material?
<whitequark>
very much straight wing for example
<whitequark>
not sure yet, but some kind of tough sheet plastic
<whitequark>
nylon, perhaps
<whitequark>
bulkhead construction with ribs that slot into each other
<SnoopJeDi>
Got a density in mind?
<egg|laptop|egg_>
yeah I suspect the aerodynamic constraints aren’t going to be the main problem
<SnoopJeDi>
Well actually
<SnoopJeDi>
Just look in COMSOL's materials library lol
<whitequark>
oh
<whitequark>
no, that's not how it works heh
<whitequark>
i'm going to call a few shops that laser cut material and figure out what they have
<whitequark>
most of it is going to be too brittle to use or too expensve
<SnoopJeDi>
Hm, good point. I'm a little too used to the magnet design world where this is...not the leading cost driver...
<egg|laptop|egg_>
it’s not like you are doing aerodynamically super fancy things (I suspect you don’t want to do transonic flight with your drone,) so it very much sounds like a case of « give it a wing, a tailplane, control surfaces, nudge the centre of mass so it doesn’t fall out of the sky »
<whitequark>
it's not so much cost specifically, though cost too, as my refusal to treat it like an unique project as opposed to something that is fundamentally designed to be a small series
<SnoopJeDi>
I think that's precisely what she wants to do, guided by numerical experiment.
<whitequark>
i don't want lots of manual labor per drone
<whitequark>
if (when) it drowns in the nearest swamp i want to be sad and then spend a week making another one
<SnoopJeDi>
Makes sense if you're shopping it out anyway, might as well get a couple
<whitequark>
because i have a very realistic estimate of my skill as an autopilot programmer
<whitequark>
i remember how i had a tiny and silly model airplane as a child and it was kind of impossible to fly because it was so easy to break and so hard to control
<SnoopJeDi>
well, autopilot's hard enough when the aero's top notch eh, might as well get what you can
<whitequark>
and not even that cheap i think
<whitequark>
speaking of autopilot, i'll probably stuff everything i reasonably can into an FPGA so that i can't accidentally write a priority inversion or a segfault with unfortunate results
<SnoopJeDi>
Well, to be clear, I know absolutely nothing about aerodynamics except what I read on the back of a box of a box of PDFeeios
<SnoopJeDi>
But I can attest to COMSOL just absolutely ruling at "hey can you do up a simulation to check $THING" when you're doing iterative design and fab
<egg|laptop|egg_>
if you want stability maybe give it some dihedral?
<whitequark>
dihedral?
<SnoopJeDi>
This is like, literally the entire thing my lab did, albeit for magnets.
<galois>
[WIKIPEDIA] Dihedral (aeronautics) | "Dihedral angle is the upward angle from horizontal of the wings or tailplane of a fixed-wing aircraft. "Anhedral angle" is the name given to negative dihedral angle, that is, when there is a downward angle from horizontal of the wings or tailplane of a fixed-wing aircraft.Dihedral angle (or anhedral..."
<SnoopJeDi>
...and maybe a few CFD simulations of molten salt >_>
<whitequark>
egg|laptop|egg_: hmm
<egg|laptop|egg_>
whitequark: I mean if you just want an intuition of how moving wings around affects basic static analysis, toying with KSP+FAR is how I have any at all
<SnoopJeDi>
yea FAR is really good for aero, especially understanding the moving pieces of the picture
<SnoopJeDi>
I can't recall if the design tools did mechanical stuff, like pressure profiles?
<whitequark>
interesting
<SnoopJeDi>
I guess this gives you a sense for how big the span of design tools is haha
<egg|laptop|egg_>
Not sure what kind of flight characteristics you are looking for, but I suspect stable but slow-to-turn might closer to your needs than "fighter plane"
<egg|laptop|egg_>
(going by your range of 50 km)
<whitequark>
correct
<whitequark>
i have zero interest in aerobatics
<whitequark>
it has to well... move in the air. that's it
<SnoopJeDi>
what a delightful word
<SnoopJeDi>
Plane people are insane and I love it
<whitequark>
i feel like i have requirements that are not entirely aligned with the usual reasons people are interested in airplanes
<whitequark>
hobbyists at least
<whitequark>
seems pretty close to the reasons people build commercial drones
<egg|laptop|egg_>
btw you can achieve the same kind of effect that you would get from wings tilted up by having wings above the fuselage
<egg|laptop|egg_>
(Антонов-style)
<whitequark>
oh, is that why i've seen quite a few top mounted wings on models?
<egg|laptop|egg_>
that might be easier on your manufacturing
<whitequark>
bonus points: that makes it easy to manuf... yes
<SnoopJeDi>
One of the #1 reasons I like COMSOL here is because the models go pretty seamlessly between 2D and 3D
<egg|laptop|egg_>
yeah because if you put it below the fuselage with no dihedral angle you are going to have less dihedral effect
<whitequark>
does "dihedral effect" boil down to "you want your lift force applied above your center of mass"?
<whitequark>
well
<SnoopJeDi>
although 3D is...yea, not sounding likely here
<whitequark>
for stability
<whitequark>
SnoopJeDi: yeah i'm more or less looking at basic operations applied to arbitrary 2d shapes
<whitequark>
extrusions of closed contours
<SnoopJeDi>
cool, that makes everything simpler to build a toy model of
<whitequark>
it also probably has to disassemble somehow because i have got no space to store a pretty significantly large aircraft
<whitequark>
which precludes the use of clever techniques for making 3d structures
<SnoopJeDi>
probably loads more information about what a good cross-section looks like than there is about "okay let's talk about vorticity"
<SnoopJeDi>
Are you going to use fasteners?
<egg|laptop|egg_>
whitequark: I think it is less obvious than that; with the wings below the centre of mass but tilted up, you get that effect
<whitequark>
SnoopJeDi: not sure yet. maybe
<whitequark>
egg|laptop|egg_: oh interesting
<egg|laptop|egg_>
hm, do I have an install of KSP with FAR here
<whitequark>
SnoopJeDi: i mean there are definitely going to be fasteners involved but i'm not sure how central they'll be
<whitequark>
does it make sense to have the entire thing disassemble into a pile of plastic plates?
<SnoopJeDi>
I guess I probably should have just asked you directly how simple the (dis)assembly you were thinking of was
<whitequark>
or should i permanently glue the wings together
<SnoopJeDi>
makes sense to me if we're talking about something with a range of 50 km
<SnoopJeDi>
the fasteners I mean
<SnoopJeDi>
err, the disassembly
<whitequark>
50 km is like, the range below which i don't even want to bother getting into airplane building
<SnoopJeDi>
rad
<whitequark>
if i'm successful with this i'll want more
<whitequark>
but i'm starting (relatively) small
<whitequark>
hopefully makes it more tractable
<whitequark>
very long range means issues with radio link
<SnoopJeDi>
It's not exactly your first rodeo to "let's do a thing" so I would say that's a good instinct
<whitequark>
one of the reasons i want to eschew manual piloting is that if they fly themselves and loiter for a long time i can use several of them to establish a relay link
<SnoopJeDi>
I swear to god I was just going to make an OTA joke
<SnoopJeDi>
err, OTH
<whitequark>
with like actually enough bandwidth for HD video telemetry
<SnoopJeDi>
or over-the-horizon or whatever, I don't know that's actually a common-use acronym
<whitequark>
but that all starts with one
<SnoopJeDi>
I know enough about radio that I know I don't know radio
<whitequark>
the practical reality is that HD video requires satellite or line of sight and i can't afford satellite
<whitequark>
also that would definitely raise questions
<whitequark>
i mean this whole project is going to raise questions but hopefully i could keep their amount reasonably low
<egg|laptop|egg_>
whitequark: is this for staring at cats over the horizon,
<whitequark>
lmao
<SnoopJeDi>
I have a number of them myself
<SnoopJeDi>
But at this point I have decided that whitequark is too curious to be a James Bond villain
<SnoopJeDi>
Anyway, definitely you'll want to have your available materials in mind when you set out to lay down a model, if you do simulation de rigueur
<whitequark>
mhm
<SnoopJeDi>
And if you get into mechanical anything, you'll definitely want to be able to do some simple beam-loading estimates or whatever
<whitequark>
SnoopJeDi: realistically the materials are going to be some of PET, PC, ABS, probably not even PVC, PA is expensive but viable
<whitequark>
that's about it
<SnoopJeDi>
That seems same
<SnoopJeDi>
sane
<whitequark>
PC, PA, ABS would be my short list
<SnoopJeDi>
I'm a little too used to "what whackjob material are we designing around this time"
<whitequark>
hah
<SnoopJeDi>
Because when you're paying for a big ol' bunch of superconductor, materials prices start to shift around in your spreadsheets
<whitequark>
PC and ABS can be glued fairly easily, PA less so
<SnoopJeDi>
and also "I'm going to take this down to 2 K and then I'm going to put it under enough pressure to be a physical safety hazard."
<whitequark>
ouch
<SnoopJeDi>
it's fine, we just push back about that hard with preload, lock keys in place, and bam, problem solved
<SnoopJeDi>
Anyway, glue is probably not a bad idea either, modulo this constraint on being collapsible
<whitequark>
glues are pretty impressive as a concept
<SnoopJeDi>
The story of chemistry in the industrial revolution is underappreciated in a digital world where ephemeral software is "innovative"
<whitequark>
this is why i work on a CAD
<whitequark>
bridging that divide
<SnoopJeDi>
I'm shocked when I think about all the stuff I straight up do not know about how we went from "yea idk slap some tar on it" to "the glue in the macbook is so good that the easiest battery recycling just crowbars the battery out, destroying the screen" in like, less than 100 years?
<SnoopJeDi>
I mean I'm probably not doing 19th century chemistry justice there but, jeez.
<whitequark>
heh!
<SnoopJeDi>
whitequark, do you do podcasts and if so, do you listen to "Well, There's Your Problem?"
<SnoopJeDi>
Not exactly adhesives I guess but this conversation is reminding me of the sealing mechanism they cover in the episode about the "atmospheric railroad"
<SnoopJeDi>
(I have no idea why, it is...not at all related)
<whitequark>
don't do podcasts
<SnoopJeDi>
That scans
<SnoopJeDi>
19th century train design driven by differential pressure. Whole thing sits on a vacuum tube with a tallow seal. You melt the seal in front of the train, reseal it behind.
<whitequark>
i... what
<SnoopJeDi>
Someone was absolutely off their rocker about making seals to come up with that idea
<SnoopJeDi>
And like, it workedish in a few places?????
<_whitenotifier-d13c>
[Principia] pleroy opened pull request #2587: Remove the 2519 traces and give meaning to conserve_angular_momentum - https://git.io/JfgCu
<_whitenotifier-d13c>
[Principia] eggrobin labeled pull request #2587: Remove the 2519 traces and give meaning to conserve_angular_momentum - https://git.io/JfgCu
<_whitenotifier-d13c>
[Principia] eggrobin synchronize pull request #2584: Point to the planned initial Frenet frame before ignition - https://git.io/JfzR3
<_whitenotifier-d13c>
[Principia] pleroy closed pull request #2587: Remove the 2519 traces and give meaning to conserve_angular_momentum - https://git.io/JfgCu
<_whitenotifier-d13c>
[Principia] pleroy pushed 3 commits to master [+0/-0/±5] https://git.io/JfgW1
<_whitenotifier-d13c>
[Principia] pleroy e7f6211 - Remove the 2519 traces.
<_whitenotifier-d13c>
[Principia] pleroy 40adc30 - conserve_angular_momentum should do something.
<_whitenotifier-d13c>
[Principia] pleroy 5176916 - Merge pull request #2587 from pleroy/NoTraces Remove the 2519 traces and give meaning to conserve_angular_momentum
<_whitenotifier-d13c>
[Principia] eggrobin closed issue #2577: The manœuvre node marker should show the ignition orientation prior to ignition for guided burns - https://git.io/JfEOB
<_whitenotifier-d13c>
[Principia] eggrobin pushed 3 commits to master [+0/-0/±3] https://git.io/JfgWD
<_whitenotifier-d13c>
[Principia] eggrobin 3fdb554 - Point to the planned initial Frenet frame before ignition
<_whitenotifier-d13c>
[Principia] eggrobin 17b8ecd - Merge pull request #2584 from eggrobin/guidance Point to the planned initial Frenet frame before ignition
<_whitenotifier-d13c>
[Principia] eggrobin closed pull request #2584: Point to the planned initial Frenet frame before ignition - https://git.io/JfzR3
<egg|laptop|egg>
(doubled the wingspan, moved the wing back a bit because the heavier wing moved the centre of mass back closer to itself and thus reduced stability
<egg|laptop|egg>
)
<egg|laptop|egg>
I wonder how slow I can get it to fly