<awygle>
do you mean your DSO's SI or the signal's SI
<awygle>
that waveform looks like it would work out to ... somethingl ike 35/40 MHz, based on the images you sent
<awygle>
which could be DSO-limited
<awygle>
you might get some EMI though. you look under-damped
<maikmerten>
I assume that the image of the 21 MHz waveform already is not a reliable representation of reality
<maikmerten>
(due to limits of my DSO, and the fact my ground wire is... a wire)
<maikmerten>
I'm certain that my approach to attaching the probe to ground is... bad in almost every conceivable way: https://pasteboard.co/HA4mawN.jpg
m_t has quit [Quit: Leaving]
<maikmerten>
(and sorry for the solder job - yes, I did this by hand, and yes, with that probably oversized tip)
<ZipCPU>
shapr: Sorry I missed your questions yesterday. I tend to take Sunday's off.
<shapr>
no worries
<shapr>
I had much fun, learned many things
<shapr>
and of course, broke everything this morning
<ZipCPU>
;)
<shapr>
I decided to teach a weekly FPGA class at work, so now I'm highly motivated to learn enough to teach
<ZipCPU>
Heheh. Let me know if you run into any trouble.
digshadow has quit [Ping timeout: 244 seconds]
dys has joined #yosys
<edmoore>
shapr: that's a really good idea
<shapr>
teach a class to learn? or what?
<edmoore>
yes
<edmoore>
an fpga class
<shapr>
I convinced two of my coworkers to buy a beaglewire, since it can run the yosys suite itself
<shapr>
so we'll see
digshadow has joined #yosys
<edmoore>
i bought a few icesticks for work just today
<edmoore>
arriving tomorrow
<edmoore>
i might do the same
<shapr>
I teach a Haskell class on Tuesdays at lunch, I used to teach Python on Mondays, but I want to do verilog now
<shapr>
another coworker teaches Rust on Thursdays
<edmoore>
we're doing rust at work
<shapr>
oh yeah? where do you work?
<edmoore>
and informally julia
<shapr>
Ah, I haven't tried Julia yet
<ZipCPU>
cr1901_modern: Ok, I'll buy that.
<edmoore>
i work for a tiny engineering r&d place in the UK
<shapr>
so, not myrtle software?
<edmoore>
just 6 of us
<edmoore>
nope
<shapr>
I know myrtle is using clash-lang for FPGA
<edmoore>
mechanical/thermal-related engineering
<edmoore>
though we design some electronics too
dys has quit [Ping timeout: 240 seconds]
<shapr>
Work code is 50-75% Golang, I'm not really a fan. The rest is mostly Python, but we have some C++ and C as well.
<edmoore>
we're mostly python for number bashing and c for electronics and some systemsy stuff
<edmoore>
i would be surprised if those two gradually shift to julia and rust respectively
<edmoore>
wouldnt*
<cr1901_modern>
ZipCPU: Guess I should prob actually test that tho :P
* cr1901_modern
just woke up/didn't sleep well
<edmoore>
shapr: our interest in FPGAs is for dsp and control in bits of equipment, and also we're looking to replace ethernet in some of our dataloggers with an in-house thing that does hard realtime bidirectional comms
<edmoore>
with time, data and power all down one cable
<edmoore>
and an FPGA could handle all that quite nicely, i think
<shapr>
yeah, it could
<shapr>
my barrier to entry was the pain induced by xilinx tools on linux
<edmoore>
yes same
<ZipCPU>
Painful, huh? Me too.
<shapr>
so when the yosys suite meant I could just git clone and make install, I was a much happier person.
<edmoore>
i had a flirt with this stuff about 4 years ago, got nervous about baking proprietary and flakey dev tools into our otherwise FOSS company, and it fizzled out
<edmoore>
and yosys changes everything
<shapr>
yeah, our research dept uses a truly massive amount of GPUs to train machine learning things
<shapr>
there's some vague interest in FPGA options, but zero internal expertise
<shapr>
would be nice to see if that's a good option
dys has joined #yosys
<shapr>
edmoore: does the icestick come with pmod headers? or grove?
<shapr>
Are those the two most common ways to connect other hardware to an FPGA?
<edmoore>
shapr: yes it has a pmod
<shapr>
ah cool
<edmoore>
I made my own pcbs fit pmod the first time round so we’re good
<shapr>
I've never made PCBs, I can barely use kicad enough to open existing projects
<shapr>
there's so much to learn :-)
<shapr>
My work team specifically does packet processing, I'd love to see how FPGAs could be used to speed up packet processing.
<shapr>
I've seen FPGAs that do that, but I haven't seen an open source tcp/ip stack in verilog
<shapr>
edmoore: so you have an ethernet pmod and you'll try your in-house protocol on that?
<edmoore>
shapr: will you be around in about 1-1.5hrs? Just going for a point with my colleague (actually to discuss fpgas!)
<shapr>
sure!
<qu1j0t3>
shapr: maybe ask astrid about that
<qu1j0t3>
shapr: but i think she's more about hard cpu networking
<shapr>
hm, google says there are piles of tcp implementations in verilog, many on github
<ZipCPU>
There's also an implementation on OpenCores.
X-Scale has quit [Ping timeout: 268 seconds]
m_w has joined #yosys
develonepi3 has quit [Remote host closed the connection]
develonepi3 has joined #yosys
<edmoore>
shapr: sorry, i meant i was going for a pint, not a point
<shapr>
ha, no worries
<edmoore>
shapr: no ethernet pmod, currently it's using an arm cortex m4 microcontroller
<edmoore>
i would make a custom physical layer for this thing, we were thinking of going over coax
<edmoore>
but i have some other pmod things, like a gps RF front end
<shapr>
ah, I should get one of those
<edmoore>
this is one i made
<shapr>
I can plug in arduino shields and that kind of thing, that's my level of PCB design :-P
<edmoore>
i think like most things it's scary on the outside but once you start pulling at a thread it quickly unravels
<shapr>
yeah, that's how most things go
<edmoore>
slightly odd metaphor sorry
<shapr>
no, I get it
X-Scale has joined #yosys
<shapr>
a solid ten hours of FPGA dev yesterday got me enough of a handle that I can now build basic stuff
<shapr>
at least, from the code side
<edmoore>
i like pcb design too because it's quite relaxing
<edmoore>
i can listed to music with lyrics, or podcasts, or whatever
<edmoore>
vs coding of maths where i need silence
<edmoore>
or maths*
<shapr>
I can code most things without having enough sleep, but I've been writing code for money for near 25 years
<shapr>
not gonna say I prefer writing code that way
<shapr>
math is hard work for me
<shapr>
though taking existing math and turning it into code isn't so difficult
<shapr>
I learned a bunch of crazy stuff from awygle's twitch streams doing PCB design
<shapr>
didn't know stackup was even a thing before that
<shapr>
Like, the traces might need to be different widths on different layers because... field effects?
<shapr>
PCB design is not like programming
<edmoore>
yep
<awygle>
yep
<edmoore>
i do a non-zero amount of radio-freq stuff
<awygle>
i should do more of that
<edmoore>
and that's where it starts to become a bit woo
<awygle>
it was fun despite a slight self-consciousness
<shapr>
awygle: yeah, I've learned much from your live streams, I am appreciative!
<shapr>
also, cute kitty
<edmoore>
the tracks themselves become components
<awygle>
edmoore: recently i've started doing more high-speed digital which is frankly way scarier than all but the highest RF i've ever done (8GHz)
<edmoore>
well, they're always components, but it starts to really matter
<shapr>
I can't think of any case where I need to know about analog effects in programming
<shapr>
well, maybe cosmic rays when you have more than 128GB of RAM
<edmoore>
awygle: oh fun!
<awygle>
spectre and meltdown? :p
m_t has joined #yosys
<shapr>
awygle: still not analog effects
<shapr>
that's just side channel
<awygle>
okay, rowhammer
<shapr>
ok, good point
<edmoore>
i've not done a huge amount of high speed digital, but i've done RF and instrumentation stuff with ultra high impedance sources
<edmoore>
where things like moisture content on the pcb surface ruins your signals through leakage currents
<shapr>
wow
<edmoore>
or would, if you didn;t design around it
<shapr>
I wish I'd gotten ECC ram in my laptop, I'm curious how many cosmic ray errors I'd get with 64GB
<awygle>
yeah ultra high impedance stuff is also scary
<edmoore>
but fun
<awygle>
did you have to do air wires?
m_t has quit [Excess Flood]
m_t has joined #yosys
<edmoore>
yes in a couple of places
<awygle>
or use a fancy dielectric?
<edmoore>
and isolation slots
<edmoore>
and guard rails
<awygle>
huh i didn't even think about iso slots for this but that makes sense
<edmoore>
i didn;t go mega fancy dielectrics but i did go to a slightly better fab where i knew exactly which dielectric i was getting
<edmoore>
and where the weave was consistent in x and y
<awygle>
like an FR408/RO-4350 kind of thing?
<awygle>
"FR4 but like... with process control"
<edmoore>
yes exactly
<edmoore>
not a fingers-crossed chinese fab
<edmoore>
i mean i know there are chinese fabs that do it all properly
<edmoore>
but i went with eurocircuits who gave me good guarantees
<edmoore>
but i've never tried to do something really high speed digital
<edmoore>
there is a nice youtube series called The Signal Path, and he recently took apart a cheapo automotive radar
<edmoore>
was very neat to see some of the ways they can make acceptable radar on a budget of cents
<awygle>
yeah signal path is great
<awygle>
somewhat more technical than eevblog ime
<edmoore>
yes
<edmoore>
he is an expert
<shapr>
not sure I'm an expert in anything
<edmoore>
and presents things much more like a well designed lab session you might have on a university course
<awygle>
Dave's no slouch either, that's just not what he's there for
<shapr>
wait, I probably am an expert in some very small niches, like refactoring browsers
<edmoore>
but i think shariar really is an authority on a lot of the stuff he does
<edmoore>
not that it's required to make the videos
<edmoore>
i am not an expert at anything either
<edmoore>
i have ended up falling into the hole you often do at small companies
<shapr>
where you do a bunch of things better than everyone else, but you're still not a complete expert?
<edmoore>
whereas some of my customers are just carrying on with their PhDs, but even more so, at their company
<shapr>
Our ops team tracked me down last week to talk about hyperthreading and interprocessor interrupts
<edmoore>
looking for half a percent here and there on combustion simulations
<awygle>
oo sim stuff is cool. What kind of sims? FEM?
<edmoore>
i'm not sure if i do a bunch of things better than other people
<edmoore>
but i know enough about enough things to be able to cost together a research programme quite quickly, which seems to be useful
<awygle>
I have no idea how combustion sims work actually
<shapr>
edmoore: rotating detonation engines?
<edmoore>
shapr: no
<shapr>
as far as I know that's the only combustion tech that can beat the limitations of the carnot cycle
<edmoore>
but close enough
<shapr>
oh you have my interest now
<edmoore>
well, it's not operating beyond the limits of the carnot cycle because i'm not convinced you can
<edmoore>
but we are trying to push chemical rocket engines beyond where they are now
<shapr>
huh, interesting
<shapr>
I assume you've already read the book "IGNITION!", right?
<qu1j0t3>
edmoore: Ah i watched that Signal Path vid, was good
<shapr>
cause if you haven't, you are in for a treat
<edmoore>
shapr: i helped push for the reprint :)
<shapr>
w00
<edmoore>
awygle: so there are a bunch of methods, all pretty rubbish
<shapr>
my stomach muscles were sore after reading that book, I laughed so much
<edmoore>
it's a hard problem
<edmoore>
usually large eddy simulation + chemistry
<awygle>
my father is a rocket test engineer
<edmoore>
but i really like the adaptive mesh + direct navier stokes stuff that's becoming more widespread
<edmoore>
it seems sensible
<edmoore>
awygle: oh cool!
<edmoore>
where?
<awygle>
I just skimmed the CFD wiki page and my head hurts :-P
<awygle>
edmoore: aerojet
<edmoore>
well CFD is really hard for a reason
<edmoore>
which is memory
<edmoore>
it's actually super simple
<edmoore>
you just divide up your 2D or 3D space into a grid/mesh, and track all the properties of the fluid in each cells (average temp, momentum, pressure, whatever), and iterate it by some time t - so just solving ODEs on a grid
<shapr>
cfd?
<shapr>
oh, computational fluid dynamics?
<awygle>
computational fluid dynamics
<awygle>
yes
<awygle>
"solve ODEs on a grid" represents like 90% of all problems, with the remaining 10% being "solve PDEs on a grid" :p
<edmoore>
HOWEVER - for practical problems, the size of the mesh to track all the eddies you need to actually capture the physics is absolutely tiny, so let's say you could capture all the properties of a grid cell in say 1Kb of memory, which sounds reasonable, and then you work out how many cells you need, you quickly find you need likea yottabyte of memory to just simulate, say, a compbustion chamber in a car
<edmoore>
so all of CFD is finding bodges and approximations, usually specific to the problem, to avoid having to have such tiny meshes
<edmoore>
LES which i mentioned above is one-such
<shapr>
or convincing someone to let you spend an amazing amount of money on hardware?
<shapr>
We have a box at work with a terabyte of ram and .. eight? of nvidia's largest GPUs
<edmoore>
but, really fast wayes of having adaptive grids, which are fine where they need to be and sparse when they can be, and then just doing brute-force but correct navier stokes, will be the future i think
<edmoore>
so the physics becomes obvious and dumb and the mesh generation becomes hard and important
<edmoore>
rather than the other way round as has been prevelent
<edmoore>
shapr: even that probably wouldn't be enough to brute-force-homogenous-mesh solve any practical problem
<shapr>
I'd believe it
<awygle>
yeah adaptive meshing makes tons of sense to me in other domains
<sorear>
as opposed to phenomenological models of turbulence etc
<edmoore>
yes
<edmoore>
i'm convinced, as a signal processing guy with a bit of thermodynamics, that this is the right way and 90% of CFD PhD knowledge will become redundant when we're proved right
<awygle>
lol there's an xkcd for that
<edmoore>
but they obviously don't like hearing that and I'd never be so rude as to say it to them
<edmoore>
well i don't work directly for them, i'm in a company of 6 that has worked with them since it was just alan, doing a lot of the r+d on the nozzle end
<awygle>
ah okay
<edmoore>
it does all work, but i do not represent the new management who come from Professional Aerospace
<awygle>
lol uh oh
<awygle>
well i wish them all the best, that would be a real breakthrough
<edmoore>
and should the project end up failing, i can say it's not because of the tech
<awygle>
unlike most of the "breakthroughs" that industry has
* awygle
loves IRC
<edmoore>
yes, it's exciting
<shapr>
yeah, so many interesting people on IRC
<edmoore>
i will be over in your part of the world in a couple of months
<edmoore>
we are testing an important part of it over there, because lots of people seem to be interested
<awygle>
well if you're in the PNW ping me and i'll buy you $FOOD | $DRINK
<edmoore>
well, we're actually in colorado
<awygle>
err that's pacific northwest for our friends across the pond
<edmoore>
the desert bit
<shapr>
if you're in atlanta
<shapr>
oh too bad
<edmoore>
presumably so as not to upset too many people with noise
<edmoore>
i guessed
<edmoore>
awygle: having said all the stuff about sims, i too am a testing guy at heart
<awygle>
"nobody believes a sim except the one who ran it, and everybody believes a test except the one who ran it"
<edmoore>
and think that if you have good inhouse machining and testing, it's often quicker and obviously more truthful to just build something and put it on the rig
<edmoore>
and cheaper*
<awygle>
that same thought is why i've been thinking about building out some home pcb manufacturing capability
<edmoore>
yes, making test data trustworthy is a fulltime job
<awygle>
i haven't done it because a small poorly ventilated apartment with cats is a bad place for HCl and whatnot
<edmoore>
i cannot tell you the number of pathological interactions you have with random industrial control gear in the loop with your stuff and someone else's stuff and network latency and and and
<edmoore>
this is why we make all our own datalogging and control stuff
<awygle>
yeah i believe it
<awygle>
slower but much surer
<awygle>
and ultimately probably not much slower
<edmoore>
just this week we've been properly chasing ghosts in the machine in a very very very expensive swiss brushless motor control system (of order hundred kW) that is playing the role of a turbopump in an experiment
<edmoore>
turns out despite being swiss, they don't necessarily understand nyquist
<edmoore>
oh it's almost always fater in the end
<edmoore>
faster*
<awygle>
shoulda done it yourself, then you could formally verify it with yosys :p
<edmoore>
i am totally sold on the power of a small number of good people in one building
<awygle>
oof sampling rate problems. my favorite one of those was a vendor who told me their crash accelerometer sampled at 10 Hza
<awygle>
*Hz
<edmoore>
well so long as your impact with your steering wheel nice and slowly you'll be ok
<awygle>
"uh, it's a *crash*. it will definitely be faster than 10 Hz"
<awygle>
this particular crash was of the variety "balloon pops and test equipment falls out of the sky"
<edmoore>
balloon pops?
<awygle>
high altitude balloon tests
<edmoore>
OH!
<edmoore>
talk to me
<edmoore>
why were you doing them
<edmoore>
shapr: sorry, it's the SABRE engine
<edmoore>
have a goog
<awygle>
validating avionics for our cubesat. not convinced we got much useful out of it
* shapr
googs
<awygle>
except we proved our ground station pointing, i guess
<edmoore>
i have done quite a lot of HAB stuff
<shapr>
high altitude balloon?
<edmoore>
yep
<edmoore>
awygle: did you use, for example, any tools to predict the balloon flight path?
<shapr>
huh, sabre is a really cool design
<awygle>
edmoore: we had a subcontractor do _most_ of that stuff. Near Space Corporation out of Tillamook Oregon
<edmoore>
i know of them
<awygle>
not to talk out of school but i don't recommend them. although it's possible they've improved in ~4 years
<edmoore>
what's wrong with them?
<awygle>
shapr: to give you a little bit of context, for decades "I have a design for a single-stage-to-orbit rocket" was synonymous with "I am a con man". if SABRE does what it claims it can do (see above re: endorsements and reputable investors), that will no longer be true
<shapr>
well, it covers all the bases that I can think of
<edmoore>
however, i am learning the hard way about doing ambitious engineering projects in the UK
<edmoore>
we tend to invent stuff then have another country with actual money commercialise it
<edmoore>
there is just not the billions sloshing around here that there is in the US
<awygle>
i gotta get some of that sloshed on me...
<edmoore>
don't we all
<sorear>
Meanwhile SpaceX claims to have a working SSTO that they’re just not using
<awygle>
edmoore: they were very disorganized. there were 2 or 3 competent people and a bunch of ... less competent people
<edmoore>
sorear: some historical rockets could have been ssto too, just
<edmoore>
just with no useful payload
<edmoore>
it's the latter that makes it really really hard
<awygle>
also i can't believe the size of the payload they let us fly. it could _easily_ have killed somebody. oversight seemed incredibly lax
<edmoore>
awygle: yes, i wondered about that too
<edmoore>
we never flew more than like 2lbs on uk ones
<awygle>
tillamook is gorgeous though (although it smells like manure), and the hangar is amazing. so... go take the tour lol
<edmoore>
well that's not true, i was a bit naughty when i was 19 and flying bigger stuff
<awygle>
we flew like... i want to say 50 kg? but it might have been lbs
<awygle>
50-something
<edmoore>
i remember they had a glider or something
<edmoore>
but i never saw it fly
<awygle>
no it was def. kg
<awygle>
anyway, not important
* awygle
has to get back to work despite fun rocket talk
<edmoore>
enought to ruin your day if it lands on you either way
<sorear>
So you use atmospheric O2 for the first 20% of the velocity/first 4% of the energy. Surprised saving 4% matters
<edmoore>
intuition with KE can go out of the window with the rocket equation
<sorear>
Mm
maikmerten has quit [Quit: Ex-Chat]
<edmoore>
i mean just as a rough datapoint for calibrating KE intuition, if you consider the K.E you transfer to the satellite, the second stage of a falcon 9 does about 10x the work of the first stage
<edmoore>
and yet the 1st stage is about 10x the size/weight
<edmoore>
(very back of the envelope, 1st stage B.O is about 2km/s (so 0-2km/s = 2km/s delta-V), 2nd stage B.O about 8km/s (2-8km/s = 6km/s delta-v)
<edmoore>
(6^2)/(2^2) is very approximately 10:1
<sorear>
BO=?
<edmoore>
and yet obviously the first stage is a really big deal and you're not going anywhere without it
<edmoore>
burnout
<edmoore>
this is all from the satellite's POV
<shapr>
really big slingshots ...
<shapr>
maglev track up the side of a mountain?
<shapr>
stage zero?
<sorear>
I guess if you’re filing patents that means your management actually believes you’ll have something by 2038
<edmoore>
well that's one reason to file patents
<edmoore>
there are others
<edmoore>
shapr: it all depends on what can form a business plan that adds up
<sorear>
shapr: such have been proposed
<shapr>
seems like a good use of money to me
<edmoore>
it's quite hard to build stuff up mountains, it turns out
<edmoore>
assuming you can find a friendly mountain in a friendly place with a friendly government
<shapr>
yeah, there is that
<shapr>
rocket friendly locations don't often have rocket friendly governments
<sorear>
If you’re not running a million amps through the entire world supply of niobium what even are you doing
<edmoore>
exactly
<edmoore>
chemistry is really quite good
<edmoore>
i mean, convenient
<edmoore>
from a power density pov
<shapr>
and the most efficient energy store
<shapr>
yeah
<edmoore>
compared to other methods
<shapr>
only better energy store is high tensile strength flywheels
<edmoore>
it's still super marginal for rockets
<edmoore>
if the earth's diameter was 10% greater we'd probably have never left
<edmoore>
we can only just do it with the oxidisers and fuels that we have
<sorear>
only because nuclear is so unpopular
<edmoore>
there is always nuclear, but that's tricky
<edmoore>
yes
<edmoore>
imagine doing all the early failures to learn how to do it, but with fallout every time
<edmoore>
it wouldn't be cool
<shapr>
yeah, I wish nuclear weren't so fraught with politics
<edmoore>
but you can get almost twice the ISP out of them, so i guess we'd have got off
<shapr>
that was my first major at university
<edmoore>
well i think the fear is pretty valid for rockets
<edmoore>
although their improved ISP would allow you the mass budget to make everything a good deal less marginal
<edmoore>
but that's really the same logic behind the sabre
<edmoore>
we can make an aeroplane with one that isn't only-just-not-exploding
<shapr>
that makes ICBMs look like a much saner option
<edmoore>
yes that nuclear ram tech is terrifying
<edmoore>
and also elegant and lovely and simple, wearing my engineer hat
<edmoore>
but, terrifying
<shapr>
true
<shapr>
but "fly around until someone angers you, then irradiate them"
<shapr>
that's extremely scary
<edmoore>
sorear: yes, absolutely reusable - that's the motivation for designing it. to make a reusable SSTO space plane
<edmoore>
shapr: i was actually researching those programs because they did some real cowboy engineering to test them
<edmoore>
to make a mach-lots wind tunnel to test them
<edmoore>
which would be very useful for lots of things in the present day
<shapr>
what other programs?
<shapr>
edmoore: anyway, I'm sure you can tell you've found kindred spirits on this irc channel :-)
<edmoore>
well ours for one! but in general ram and scramjet research
<edmoore>
supersonic wind tunnels are very expensive
<edmoore>
and difficult tomake
<shapr>
yeah, I've read about a few
<shapr>
past the speed of sound, prepare to spend big money
<edmoore>
for this they just, iirc, got loads of oil pipeline, several miles worth, and snaked it back and forth insome patch of dessert, slowly pumped it up to some very high pressure, they blew the cap off such that it discharged into the test vehicle
<sorear>
How much does cheaper access to space translate into cheaper testing for space hardware
<edmoore>
and it just blew down supersonically over some tens of seconds
<edmoore>
sorear: hmm, it's probably still at the expensive end
<shapr>
that's real cowboy engineering
<shapr>
but then I remember mercury delay lines...
<shapr>
what a way to save data
<edmoore>
sorear: but it will probably change the economics of making a satellite in the first place
<edmoore>
if you can launch your telecomms sat for $5M, it now longer has to be a triply-redundant gold plated $200M thing (to justify the $100M launch)
<edmoore>
instead you can just make 3 of them at $10M a piece, and launch the others if the first fails
<edmoore>
(maybe)
<edmoore>
shapr: i've only ever read about those
<shapr>
I'm not *that* old
<shapr>
though I did buy the DIP ram chips to upgrade my first PC to 512k
<edmoore>
sorear: i think probably cubesats, in conjection with the rather imperfect market for them that exists where you can find some agency to pay for your launch, is the best way to test something for now
<edmoore>
in the $100k range
<edmoore>
though it depends what you mean by space hardware
<edmoore>
if you want to test martian entry tech then maybe it become economical to put a test article in LEO and try it out for real
<sorear>
I mean testing Sabre 2 by dropping it out of Sabre 1
<edmoore>
oh sure
<edmoore>
well, i do think there will be a market for high alt hypersonic flights
<edmoore>
sounding rockets ++
<edmoore>
acting as flying wind tunnels
<edmoore>
that's probably the easiest way to get yourself into the most challenging conditions
<edmoore>
the space part of the sabre is really just a conventional rocket
<edmoore>
albeit a high performance one
<edmoore>
the really hard part is airbreathing and changeover
<edmoore>
but i think a nice business for someone, too small for an airbus of lockheed to be too bothered (i.e. not a 10B market), but great stuff for a small to medium company, will be sounding rockets and hypersonic wind tunnel flights
<edmoore>
make 5 and 20-30km, some good sideways flight time
<edmoore>
with a few hundred lbs strapped to the front
<edmoore>
mach 5*
<edmoore>
at 20-30km*
<edmoore>
sorry my typing is awful, laptop keyboard with stupid apple keys
<shapr>
would you like to buy a nice ergodox infinity?
<edmoore>
i have a nice mechanical keyboard already
<edmoore>
i like it
<shapr>
I need to get rid of some
<edmoore>
not split, but it's still never caused me any wrist pain
<edmoore>
they didn;t like my cherry blues one in the office
<shapr>
that's another fun idea for an FPGA design, an unreasonably low latency keyboard
<edmoore>
so it's a cherry brown one for there
<edmoore>
ha, yes that'd be fun
<edmoore>
so i could use vim even more hardcorely
<edmoore>
right, it's my bed time, early start as i need to change some tires on my car
emeb has joined #yosys
<shapr>
nice to chat with you edmoore
<edmoore>
yes, and you
<sorear>
=shapr
<shapr>
false?
<edmoore>
icestick arrives tomorrow so i can just ask lots of silly beginner questions
<shapr>
sorear /= shapr
<shapr>
:-P
* shapr
hugs sorear
m_w has quit [Quit: Leaving]
<awygle>
only the _very horrific_ nuclear propulsion methods actually make sense technically (and obviously they do not make sense socially)
<awygle>
nuclear thermal might as well be "worse ion" and Orion, Pluto, and NSWR are all varying degrees of "NOPE"
<awygle>
(google NSWR, it's truly horrifying)
<awygle>
oh wikipedia, i love you: "This is however of little use on the surface of a planet, where a NSWR would eject massive quantities of superheated steam, still containing fissioning nuclear salts. Terrestrial testing might be subject to reasonable objections."
<ZipCPU>
Ya think!
<ZipCPU>
I read about this in Popular Science some time ago (IIRC).
<ZipCPU>
An absolutely crazy idea, from an age when every problem was solved with nukes.
<awygle>
although the wiki page now links to "fission fragment" rockets, which seem way more reasonable
<awygle>
particularly the "dusty plasma" variant
<awygle>
which i'd never heard of before
<awygle>
love me some magnetic mirrors
<sorear>
Well the point of a nswr is you get good specific impulse AND good thrust/mass
<awygle>
plus lots of cancer
<awygle>
everybody wins
m_t has quit [Quit: Leaving]
<shapr>
anyone know if there's a list of pmod things I can buy?
<shapr>
I found a bunch on digilent, since they made the standard
<shapr>
and a few on tindie
<shapr>
but where else would I look?
develonepi3 has quit [Remote host closed the connection]