|
Post by vegemeister on Oct 15, 2018 20:58:51 GMT
Super-accurate guns are only a problem if the target has high enough acceleration to consistently dodge your fire, but scatterguns rarely perform satisfactorily in such cases. I disagree. In point defense use, the AI likes to target the engines of missiles. If the missile is accelerating forward (say, in terminal phase), every single shot misses if the gun is too accurate.
Against capital ships, spread allows guns to attack the entire front face of the ship at once, which can immediately disable a large fraction of its turrets. Low-spread guns have to engage one turret at a time, and the rate of turret destruction is limited by the time-of-flight of the bullets.
|
|
|
Post by vegemeister on Oct 15, 2018 4:18:29 GMT
Here's a capacitor coil that throws 15.0 grams @ 8.14 km/s @ 30 hz with a 1.5 meter turret diameter and it doesn't overheat. It's a little heavy though, if the game had the option of fitting a water jacket and a radiator and a small pump instead of only using passive cooling on high cyclic rate guns this would be a lot lighter and a lot smaller. But naa, lets just use a solid 1 meter cylinder of diamond because that sounds easier to build. Better economy of moving parts this way, fewer things that could go wrong. You didn't need that delta v right? Good.
That thing is laser accurate. Or better. I'm pretty sure that spread's less than the laser wobble.
A tip: if you click that "compare" button in the top right, and you haven't made any changes since opening the editor, you can get all the statistics in one screenshot.
Edit: one thing to look at is the overheat time. Overheating isn't a problem if it only happens after 5 minutes of continuous fire.
|
|
|
Post by vegemeister on Oct 15, 2018 4:12:09 GMT
On the one hand, player-optimized reactors totally clobber chemical batteries in both power and energy density. ( Wiki says 20 kW/kg for state-of-the-art Li-ion. Player-optimized reactors are >60 kW/kg including radiators.) You'd be better off building for Moar Power in the first place than storing it. And capacitors are already in the game for pulsed-power. On the other hand, that *is* ridiculous. Reasonably sized missiles and non-capship-sized drones can be deployed with 100 kW or less. Try gadolinium for the stator and calcium for the forcer. For very small missiles, blast launchers may be lighter overall because of their smaller crew requirement, and they don't need power at all. Optimal material is boron filament, if you care about cost, or UHMWPE fiber, if you don't. Optimal propellant is nitroglycerin if you're minimizing mass and don't care about launch velocity, or either nitrocellulose or octogen if you want high launch velocity.
|
|
|
Post by vegemeister on Oct 14, 2018 1:50:28 GMT
You have to type in the "Cavity Semiminor Axis" parameter instead of clicking the buttons, which can't get close enough to circular. You're trying to maximize "Cavity Shape" efficiency, which is affected by cavity shape (obviously; set the semimajor axis then approach from below with the semiminor axis), and by the radii of the lamp (leave it at the minimum) and the lasing rod (tweak as necessary). With proper tweaking, you can get it fairly close to 100%. All of the other components of efficiency are either entirely dependent on material choices, or can be adjusted independently. While you're tweaking, crank up the turbopump to avoid temperature errors.
After you've got the cavity shape and lasing rod radius correct, you can increase optical nodes either until M^2 = 3.00 (for short/light lasers), or until you get worried that the laser is too long or heavy to fit in the ship (for a small increase in pumping efficiency).
Then, add each frequency doubler. Increase the optic radius until the error goes away, and then tweak the length to get as close as possible to 100% efficiency (the sweet spot is pretty wide).
Finally, tweak the aperture size to taste. It controls the size of the laser spot at the target. Big aperture -> small laser spot/high intensity. But, like I said in the other post, lasers have wobble, and if you go much bigger than 50 cm aperture, the laser won't be able to hit extremely small targets at long range.
Here's an example of a 4.34% efficient laser:
(Ignore the outlet temperature and turbopump parameters. It's modified from a 1290 K laser for this post.)
Edit: very small values of Optical Nodes will reduce cavity shape efficiency somewhat, so if you start with a clean sheet design you may be frustrated by that. Also, if you change the power of a laser, the only parts that really have to be re-tuned are the frequency doublers and the turbopump.
|
|
|
Post by vegemeister on Oct 13, 2018 20:23:12 GMT
Well, I lose. But here are some high cyclic rate guns.
I've been told cyclic rate higher than 30 Hz doesn't actually do anything, because the game only allows one bullet per simulation tick per gun.
|
|
|
Post by vegemeister on Oct 13, 2018 20:07:51 GMT
All lasers should be unmounted, because it reduces the crew requirement (which competes with the radiators for heaviest part of a laser system), and because the optimal shape for the laser cavity is long and thin -- much longer than the width of reasonable turrets. (Although short cavities don't hurt efficiency *that* much.)
The best (non-modded) lasers are Nd:YAG with Krypton gas lamps. Ti:Saph with Xenon is sometimes used, but the greater efficiency of Nd:YAG allows a wider beam at the same intensity, which is good for small targets. There are also more temperature options with Nd:YAG: Cavity Wall/Mirror
| Output Coupler
| Radiator Temp
| Approx Eff.
| Silver
| Fused Quartz
| 1173 K
| 4.34 %
| Gold | Fused Quartz
| 1270 K | 4.10 %
| Copper | Fused Quartz | 1290 K
| 3.96 %
| Molybdenum
| Fused Quartz | 1841 K
| 2.16 %
| Molybdenum | Diamond
| 2131 K
| 1.91 %
|
I like copper at 1290 K. For large aperture lasers, the mirror material (*not* the "focusing mirror" material) affects the weight of the laser. I think that's why apophys uses copper for that and gold for the cavity wall at 1270 K. If you want to use Ti:Saph, gold and copper are no good. Arc lamp envelope is always diamond (strong and transparent), and coolant is always hydrogen (low density and transparent). Lasers should always be frequency-quadrupled, unless you don't care at all about turret size and want to squeeze out that last bit of efficiency with a silver primary mirror. The cavity should be very nearly round.
Efficiency is somewhat improved by reducing power density in the arc lamp gas, which shifts its power into the longer wavelengths. Increasing the radius of the lamp greatly reduces Cavity Shape efficiency, which must be counteracted by increasing the cavity radius (a lot), which requires re-tuning and bloats weight. So it is better to use the minimum lamp radius and make the cavity very long ("Optical Nodes"). This is a small effect at reasonably high powers, however, so if you want to make short lasers (perhaps directly attached to turrets), you don't lose much.
The takeaway from the ablation cap is that aramid and polyethylene are the best armor against high-intensity lasers. but if you're facing low-intensity lasers tuned against them, the best choice is amorphous carbon. The crossover intensity is 17.4 MW/m^2 for aramid, and 38.8 MW/m^2 for polyethylene. This means missiles should have an outer crust of amorphous carbon, and AC is also preferred for laser turret armor if you expect laser dueling. Furthermore, lots of low-power lasers are better than one high-power laser.
An interesting consideration for dueling, is that if you define survival_time * intensity_at_target as a figure of merit for laser turrets, it works out to be proportional to the mass of the turret and only the mass of the turret, under the assumptions that 1) the armor is thin enough to treat as locally flat, and 2) the mirror is large enough that all of your laser energy lands on the enemy's turret (no overspray). This is because the area of the armor and the area of the mirror are both proportional to the area of the turret.
That means you can put thick AC armor on your wide beam point defense lasers, and they are perfectly adequate for dueling laserstars -- or at least keeping them occupied long enough for your kinetics to hit. The laserstar will probably still win a lasers-only duel, given that they put all of their reactor power into lasers and you're probably using some of yours elsewhere.
If you want to kill the enemy's turrets quickly (important for kinetics), not just win laser duels, you need big apertures. But laser wobble makes big aperatures largely useless against small missiles. Here's a table relating aperture size to beam radius at 1000 km, for frequency-quadrupled Nd:YAG lasers: 1000 km Waist
| Aperture
| 30 cm
| 1.78 m
| 50 cm
| 1.07 m
| 1.0 m
| 53.5 cm
| 1.5 m
| 35.6 cm
| 2.0 m
| 26.7 cm
|
High time-to-kill is pretty much unavoidable, assuming the enemy has remotely effective armor. But the hope is that that you can lase off the enemy's guns from outside their effective range and dodge/tank any bullets they managed to get off before the guns were destroyed. Then slowly drift in and burn off their radiators over the course of minutes.
|
|
|
Post by vegemeister on Oct 13, 2018 18:16:06 GMT
But you can't slow down again once you get to your destination with a laser sail You can.
You need a two-part sail. Near the destination the sail separates and discarded part acts as a mirror reflecting the beam backwards to let you brake with the remaining part.
That'd require the discarded part to be an optically flat mirror with attitude control. It might end up a lot heavier than the basic aluminized kapton sheet you could use otherwise. Plus, depending on how long the braking takes and the ratio of the masses of the main ship and the discard mirror, it might travel very far from the main ship and have to be very large to control beam divergence.
Something else that might work would be to make a close pass to the destination star and capture with a solar thermal rocket, taking advantage of the Oberth effect and high intensity radiation. Then, use the planets in the destination system to gravity assist your periapsis out of the fire.
|
|
|
Post by vegemeister on Oct 9, 2018 17:31:23 GMT
Do stock flak missiles even work with the current standoff warhead bug?
|
|
|
Post by vegemeister on Oct 9, 2018 5:19:40 GMT
The fishtailing like a drunken sailor problem makes me suspect that the control law is using line-of-sight angular rate and its derivative instead of perpendicular velocity and acceleration. The same perpendicular velocity would equate to a larger and larger angular rate the closer the missile is to the target, but the response of perpendicular velocity to thrust is the same no matter the distance. So using angular rate would cause the control gain to effectively increase the closer the missile is to the target, which seems to match the behavior of the fishtailing. It's also consistent with being able to use lower damping factors in the earlier guidance phases without instability.
A possibly related issue is "adverse sway", caused by attitude control torque provided by gimballed engines or RCS behind the center of mass. If the missile wants to thrust right, it must first turn right, which (if torque comes from thrust behind the CoM), requires thrusting *left*. This can be avoided by using clustered-throtted-engine steering or RCS in front of the CoM.
I've also found that giving missiles any roll control at all tends to result in severe problems when combat-launched.
There's a potential problem with using APN in the boost phase for high-acceleration missiles. A target that thrusts with low acceleration during the boost phase, then immediately stops, can force the missile to expend quite a bit more Δv than it does, on the assumption that the target will maintain its acceleration through the entire coast. With PN, if the target wants to force the missile to expend Δv, it must itself expend that same Δv. In theory it's also a problem in midcourse, but the AI seems to do most of its attitude flailing at the very start of combat or right before impact, and most of my missiles are tuned to avoid midcourse corrections anyway, so as not to expose their soft squishy side armor to the enemy.
I've also run into quite a bit of trouble getting missiles to work reliably against fleets of several identical ships (or flights of drones). It seems like the missiles fail to lock on to any particular target and instead arbitrarily change what they're homing on based on tiny fluctuations of lumiance or something. A missile that works beautifully against a single evasive target -- all missiles through the hole made by the first -- is totally stumped by a fleet of 5 measly gunships. I have a couple designs that somewhat work around it, one by being capable of ridiculous terminal accelerations and the other by carrying a 100 kg warhead so that large miss distance doesn't matter, but both of those are highly vulnerable to point defense.
|
|
|
Post by vegemeister on Oct 7, 2018 13:31:00 GMT
On a serious note, has anyone else had any trouble with turrets moving around on designs when exiting and reentering the game? I keep having to put the Ring's center turret back in place because for some reason it shifts when I exit and reload, which is extremely annoying. It seems like the game is shifting the front turret back by a single unit. Same thing happened when I tried to load up the laser spider; I had to manually fix it. It might be my installation, but I am wondering if I need to report it as a bug. Same here. I suspect it's a floating-point decimal conversion issue. It seems like "front-mounted guns" are defined by having a position exactly at the farthest forward extreme of the ship. Their position gets converted to decimal for saving, when read back in the result is not bitwise-identical, and the the guns become side-mounted. I try to work around it by tweaking the overall length of the ship very slightly so that the gun positions randomly happen to round in the right direction. It seems to mostly work, although it is important to make sure a ship is save-load stable before trying to show it around.
|
|
|
Post by vegemeister on Oct 3, 2018 9:54:11 GMT
Excellent. I seem to have figured out the permissions issue then. With that out of the way, I present the absolutely terrifying "Rebuke" frigate. snip The Rebuke is, in every measurable sense, a very dangerous ship. Its armaments are simple but deadly for their size; Nine "Fury" 50 km/s turreted railguns and one "Requiem" 100 km/s spinal railgun provide the offensive weaponry with which to eliminate enemies. Defensively, the ship is equipped with 16 cm of graphite aerogel sandwiched between 1 mm of tin and 1.5 mm of amorphous carbon. Additionally, the front section of the ship boasts an impressive 98 cm's of graphite aerogel, providing an extremely tough barrier to laser elimination. Propulsion is provided by a duel system of MPD thrusters and the "Colossus" 32 MN manevering engine; a set of four "Vector" NTR's provide RCS control allowing for aiming the spinal railgun. The craft is also almost completely immune to attack by drone and missiles: The Fury railguns double as an incredibly potent point defense system. The entire package comes in at 730 tons fully load, and costs a mere 6.49 megacredits. It also has an MPD delta-v of 24.3 km, and a combat delta-v of 5 km. The Modus operandus of the rebuke is to close to just beyond the range of the enemies primary weapons, and then open fire using the fury railguns: snip Example of the Rebuke firing its Fury railguns. Enemy consisted of 5 deep fryers, a fleet carrier, 4 laser frigates, 5 corsairs, and a siloship. The Rebuke won. Although not incredibly accurate at ranges of up to one million kilometers, the sheer volume of fire is usually sufficient to shred any exposed weaponry and eliminate lightly armored ships entirely. Smart commanders attack laserstars first with the fury's, knowing that the turrets are somewhat vunerable to concentrated laser fire. However, enemy gunners usually focus fire on the spinal "requiem" weapon. Should the fury turrets be destroyed by enemy action, the Rebuke has a backup: the "Requiem" railgun. Nested safely behind 98 cm of graphite aerogel and further consisting of 2.44 m of graphite aerogel surrounding a railgun module 60 m long, the Requiem is almost indestructible by laser barrage. The Vector maneuvering thrusters allows a damaged Rebuke to align itself with enemy laserstars and send a barrage of shots downrange to eliminate the threat even without the fury turrets. Battle tests with the Rebuke have been extremely positive. Among other accomplishments, it has taken out a swarm of 21 deep fryers, eliminated an entire capital fleet (pictured above), and soloed Vesta Overkill. Some additional combat photos of it in action: snip A Rebuke facing off against 21 Deep Fryers. The battle is about to start. snip Dealing damage with the Fury railguns.
snip The aftermath ------------------------------------------- None of the AI's are quite perfect to run the Rebuke, however "Aggressive" comes close. I would recommend testing using this setting, or alternatively controlling it manually and setting broadside after opening fire. The Rebuke Frigate is attached. That is remarkably effective. The spinal gun almost never shoots, because the attitude control can't aim it correctly. The AI doesn't shut off the dodge engine, and it wobbles too much. Under manual control with RCS only and the "broadside" order, it seems the attitude error broadside is willing to tolerate is too large. It shoots sometimes against extremely-low-acceleration ships and when pre-aimed before combat, but when I tried taking the ship I'm working on now against it, it was the Furies that did the work.
I was able to kill it with missiles, but only under manual control or when the Rebuke's AI is set to something that doesn't charge in with capitals, like "Ranged". For some reason, with the Rebuke on "Aggressive", the other AI doesn't try to intercept it with missiles while it's charging in.
Combat-launched missiles, on the other hand, did not work so well, because my missiles' terminal-phase engines can't match the Rebuke's acceleration. They work when launched outside combat because they hit before the boost stage burns out. Realistically, though, you could trickle in flights of missiles and run the Rebuke out of Δv.
Two things I noticed:
1. The rear pair of radiators is rotated by 2°?
2. You've got titanium diboride control rods in your reactors. Hafnium carbide can take higher temperature, so you can get more Isp from your NTRs and push the radiators up to 2600 K.
|
|
|
Post by vegemeister on Sept 26, 2018 9:36:41 GMT
The only thing broken about laser stars is how ineffective they are. With pulsed lasers[1], fixed beam intensity[2], realistic efficiency[3] and a less limiting damage model[4], you could get one or two orders of magnitude more ablation out of your laser than you can do with the ones in game. Not to mention the unrealistic turret wobble and hard cap of 10 Mm on the lasing range. The ship might be significantly heavier and slower than what our laserstars are, but that doesn't really matter when you can kill any ship from a few dozen Mm's out. [1] Pulsed lasers can ablate armor by shattering it instead of vaporizing everything, which is far more energy efficient [2] The game has confused beam diameter and radius, which means all lasers have 1/4th of the intensity they should. [3] Real lasers can reach around 30-40 % efficiency instead of the max 4% we can do in game [4] The laser damage model in game is almost useless. It only simulates heating armor to melting point, and disregards any energy past the ablation cap of a material, so against well optimized targets, you just waste any intensity past around 4 MW/m². With enough power and big enough mirrors, you could lase targets until your effectiveness is limited by light lag instead of beam intensity falling off due to beam dispersion. Past a light minute or so of range, the target can start dodging your lasers effectively due to the travel time of the laser. I think, as AtomHeartDragon said below you, that the primary limitation on lasers would not be the overall design but the logistics of supporting them on a ship (I'm paraphrasing a bit though). The only thing questionable about lasers that I can think of is the efficiency of the frequency doubler, being able to pulse them would be a nice addition. As far as logistics go, having paper-thin radiators is a no go due to micrometeorites and excessive thrust blowing them off. Reactors should not run on the cusp of going super-critical. Another thing, lowering the efficiency of reactors as much as possible only makes sense from a gameplay perspective. "Lets get as little power out of this as possible" said no engineer ever. As such reactors would run much cooler and the radiators would dispense less heat. This would drive up the cost of super lasers because of the rarity of real estate needed to support all those radiators on the ship. I think lasers would still be seen but only on very large expensive ships especially due to the added cost of radiators. Even if you're optimizing for efficiency, heat engines are more efficient the greater the difference between the hot side and cold side temperatures, so you do want to run the reactor as hot as physically practical. But the limited resource for nuclear power on a spacecraft is mass, not fuel, so power systems would be engineered to minimize mass, not fuel burn. If the radiators make up most of the mass of the power system, their size is minimized (in spherical cow world) by setting the cold side temperature to 3/4 of the hot side temperature. See the derivation beginning on PDF page 9. But in the hexagonal cow world of CoaDE, we have an additional constraint that the hot and cold side temperatures of our thermocouples cannot differ by more than ~500 K. It seems like the optimal hexcow temperature ends up being ~2630 K radiators, which is as hot as possible without reducing the ΔT across the thermocouple.
But it is indeed correct that doom lasers are unrealistically advantaged by the ability reject heat with paper-thin radiators, without regard for flow rates, structural strength, or inlet-outlet temperature gradients. (Somehow, my reactor pumps sodium at 2630 K into the radiator loop and gets back sodium just above the melting point, even though the radiator works as if the whole surface is at 2630 K.) Now I'm not a laser expert, but I've never read of any real life laser with continuous power output of more than a few hundred kW, and those don't have to reject heat to space.
|
|
|
Post by vegemeister on Sept 21, 2018 23:25:21 GMT
The big win here isn't that the armor spacing makes a good fuel tank. It's that a fuel tank makes *fantastic* armor, because fuel is far denser than anything you could reasonably use as whipple stuffing without compromising the mass ratio of the ship. A meter of ethane has much greater sectional density than a meter of graphogel. And if you make it a self-sealing fuel tank, the fuel doesn't even leak away if you get shot up a bit.
|
|
|
Post by vegemeister on Sept 13, 2018 5:11:18 GMT
Working under Steam Play/Proton on 1.2.1.
There's a minor sound issue where the sound gradually becomes garbled over time, but it doesn't start until the game has been running for several hours. Other games have had audio problems with Proton which were fixed by a recent update, but I didn't try CoaDE before that update. Maybe it's trying to use alsa instead of pulseaudio, which has always been problematic for wine on my machine.
|
|
|
Post by vegemeister on Sept 11, 2018 16:15:32 GMT
Argh. The table was hosted on Dropbox. I don't suppose anyone has a copy?
|
|