All targeting solutions for sublight speed are computable.
It takes a hell of a lot more than a single clock cycle to run a targeting algorithm.
Yep, around 200. The Missile meanwhile travels 0,2mm
Source? I’d estimate thousands to millions of cycles.
Not everything has to be computed in every computation cycle. Most things are already pre-computed and the operations to do the corrections would mostly be simple additions and multiplications.
Written in hand coded assembler language: Maybe 100 cycles. Actually Quake from 1997 used a very similar prediction for network coding and not joking, it took less than 60 cycles and the very same code has been used to guide the IRIS-T system - which by the way uses an 1.4Ghz ARM Quadcore.
Written in Dotnet: Your computer is to old. No matter which computer.
So they wrote Starfield in .net? /j
hypersonic during launch phase only, at the target it’s more like Mach 3.5 if we’re talking about Kinzhals downed by Patriot some months ago
TIL hypersonic is genuinely a distinct word from supersonic.
It’s usually over mach 5, but originally it was used to refer to when you start needing to deal with heating up the air.
Also around the point where jet engines kinda just stop working right.
It’s not about the targeting computers not being powerful enough. It’s about the counter missile not being fast enough and the hypersonic missile being able to zig-zag like a cruise missile, but with similar speeds as ballistic has in it’s decent phase.
They don’t need to be as fast. They’re not going in the same direction.
Then how does it react when the missile is going 3x as fast and sweeps right when the counter missile is sweeping to left towards the missiles previous position and misses, in that case it needs to be faster than the incoming missile itself. Yes, the speed would not be a problem towards a predictable trajectory, that’s how ballistic missiles are intercepted even if they go super fast. it’s basically a high school math problem in that case to calculate the point of interception in a firing solution. It’s also fine if missile can change course, that’s how cruise missiles are shot down, because the counter-missile can still race with them when they turn, but when the missile is fast and can change directions mid flight then it doesn’t much help how fast the computer calculates if the hardware can’t react fast enough. it’s basically like if your mind were able to move at superhuman speed but your body is still human and you get shot and only thing you can really do is to watch the bullet approaching but being unable to dodge fast enough.
The issue with hypersonic missiles is that they go so fast that the air around them heats up enough to become a plasma. This prevents radio, microwave, and infrared based comms and radar systems from working. So any proper hypersonic missile is stuck with using either optical sensors (which require a ton of processing power and is slow) or inertial sensors (which aren’t very accurate due to drift) in order to plot its course.
The Russians are the only ones with “hypersonic missiles” because they’re using a different definition. Theirs are just ballistic missiles that move at hypersonic speeds, which by definition can’t change course and have a huge IR and radar signature, so it’s not that difficult to set a slower patriot missile to be in it’s way to intercept.
Khinzal for example is meant to be both a ground targeting missile and a “carrier killer”. That alone should mean it’s indeed maneuverable in flight.
There’s always a significant difference in what the Russian MOD says their equipment is meant for and what it’s capabilities actually are.
Also, if Khinzal was really that maneuverable then how were the Ukrainians able to intercept multiple with the much slower Patriot missiles?
Also, if Khinzal was really that maneuverable then how were the Ukrainians able to intercept multiple with the much slower Patriot missiles?
One can put serious doubt on their claims. Ukrainain MoD has every incentive to lie and on top of that they did show off some empty soviet era bombshells as remnants of a shot down Kinzhal.
Ukrainain MoD has every incentive to lie
And Russia doesn’t? Besides, it’s already common knowledge that Khinzal is just a ballistic missile
Slower objects are more nimble. They don’t have to fight inertia.
I can throw a rock at a moving car. It can swerve, but I can easily throw another and it can’t swerve again quickly.
The car here is moving like 500km/h and it’s doing a zigzag in erratic pattern. Even if your rocks can change trajectory themselves mid flight they simply aren’t moving fast enough. Oh and you have to hit the car from 1km away.
You live in a world where the sport of baseball is impossible. I find that a bit silly, but you do you.
But please don’t start a career in defense or rocket science. You don’t seem suited for it.
Your scales are way off, and you’re oversimplifying and call-of-dutyfying the laws of physics. Right now you’re arguing people about real life phenomenons by talking looney tunes. You’re flat out just wrong, and I’m not even going to try explaining why because you just ignore every single explanation anybody else have made for you.
By the way, stop fanboying about a ballistic missile, it’s weird and honestly really cringe.
This thread is earily similar to one where I recently had to explain to someone on the ebike subreddit why a bike going 35 mph won’t be fast enough to keep him from getting bitten by a dog if it wants to bite him. “No dog can run 35 mph” …“they don’t have to be faster, they just have to plot an intercept course” It’s not fucking complicated at all.
> hypersonichu
> ziz zag
Kek
okay, but for a missile 30km away radar signal gets 0.2ms to get there and back, that’s like 40cm of missile movement
Just need to fire at a target 1.3 feet away from the launcher.
Checkmate Ukrainians
This 2hgz SHARC processor with redundancy, what is it’s name?
Just to mention: The current MARS rover is using a Raspberry Compute Module 3. Several modern space probe also use off-the-shelf technology. Most Ukrainian drones use it too. The IRIS-T missile uses a 1.0Ghz Quadcore ARM. The times when such systems required extreme technology are far gone.
What? This is one of the most ridiculous things I’ve read in awhile and people are up voting this misinformation. No rover that’s on Mars is using a rpi. Here are the actual specs: https://mars.nasa.gov/mars2020/spacecraft/rover/brains/
I googled around to see where you could have gotten this information from. You might be thinking of this educational rover: https://github.com/nasa-jpl/open-source-rover/blob/master/README.md
I tried looking for the Mars Rover that is using a compute module and couldn’t find any that were using one which makes sense because it’s not radiation hardened and they need to operate well outside of the temperature range that the modules actually tolerate. They seem to all be using a radiation hardened version of an IBM power PC 750 that is manufactured by a company name BAE. Ignoring that fact though the Mars Rover does not meet the same low latency processing requirements that this meme is referencing so we can move along further.
“Most Ukrainian drives use it too” while I can’t quantify these numbers it does appear that the Ukrainian military uses orange pi’s for their “kamikaze” drones which is good for that application as it’s easy to source and doesn’t require any heavy demand. As for everything else I was able to locate on the subject, it all appears to be typical DSP processors Analog Devices SHARC’s and the TI equivalents such as the TMS320F28335 As for the Iris-T I could not find any information one way or another there. It’s indicating high velocity and use of radar so I can’t imagine they are using anything other than a DSP processor as you are going to need to shave every single microsecond of response time to maximize your effectiveness.