Engine Startup Time

[BekfastDerp13]

Have you wanted to make a cinematic launch for a KSP video, but the engines start up too fast? Vote for timed engine (at launch and/or maybe in space as well) startups or normal engine startups from KSP 1.

Edited October 12, 2021 by BekfastDerp13

[AtomicTech]

Also engines not flicking around and slowly-ish gimbaling into position.

[jimmymcgoochie]

If you want some fancy ignition effects, throttle up just a little bit at first and then give it full power later. Slow start-up is fine when you're actually playing with realism mods, but outside of that very specific condition leave it as is.

[SciMan]

For the sake of making it easier to plan maneuver burns and landing craft, I think that rocket engine throttle response both on start-up and when "active" but not producing any thrust should be instantaneous.

That whole thing with all the jet engines taking an age to spool up and down in KSP 1 is why I never even bother trying to make jet-powered VTOL aircraft despite that being the best way to get a great performing lander in an atmosphere with oxygen in it. IRL only the largest jet engines take any appreciable time to actually "spool up" from "flight idle" to "max thrust", but with the way that KSP 1 does it you can sit there for half a minute and the engine's still gaining thrust just standing still at the runway (perhaps held back by a launch clamp or two as wheel brakes are not that useful when considering the thrust of a couple Goliath enignes).
The bigger issue when building VTOLs is that they take a whole lot of time to spool DOWN (which is unrealistic), which lands you in this odd situation where you need a TWR > 1 to slow down to a safe vertical speed when landing, so you have to have a fair bit of throttle, but when you touch down you need the thrust to become as close to zero as possible, as quickly as possible, and KSP jet engines just don't do that unless you actually go and set an action group for "shutdown the engines".

[mattinoz]

On 10/13/2021 at 2:18 AM, AtomicTech said:

Also engines not flicking around and slowly-ish gimbaling into position.

Yes a layer of ease in ease out curves is needed in the game generally.  I think a little smoothness would make things more controllable not less. 

[Brofessional]

Adding an engine startup delay would just make the game harder and less approachable to beginners. Maybe you could make it a difficulty option but it's probably better left to mods.

Gimbal animations could be fine as long as they're still quick enough to feel responsive.

[Thundy]

I believe this is a good idea, each engine will have some form of response, some are faster than others in throttling and gimbal control. This will give some engines more popularity due to their superior gimbal speed even if they're not as powerful or as efficient as others.

 

3 hours ago, Brofessional said:

Gimbal animations could be fine as long as they're still quick enough to feel responsive.

I'd argue that Gimbal animations will make the game easier to control using keyboard.
in KSP 1 there's already an option to do this by clicking caps lock, your controls (engine gimbal and control surfaces) will have a bit of a delay to reach full power allowing you to create precise movements.

 

 

 

[AtomicTech]

3 hours ago, Brofessional said:

Adding an engine startup delay would just make the game harder and less approachable to beginners. Maybe you could make it a difficulty option but it's probably better left to mods.

Gimbal animations could be fine as long as they're still quick enough to feel responsive.

 

9 hours ago, mattinoz said:

Yes a layer of ease in ease out curves is needed in the game generally.  I think a little smoothness would make things more controllable not less. 

Yeah! I really hate the spaz gimbals of KSP.

[Staticalliam7]

Simplerockets 2 does a wonderful job of realistic startups (which makes nuclear take painfully long to throttle up and down), However, for ksp, 

On 10/12/2021 at 11:32 AM, jimmymcgoochie said:

If you want some fancy ignition effects, throttle up just a little bit at first and then give it full power later. Slow start-up is fine when you're actually playing with realism mods, but outside of that very specific condition leave it as is.

Just do this. Much easier

[mattinoz]

20 hours ago, Brofessional said:

Adding an engine startup delay would just make the game harder and less approachable to beginners. Maybe you could make it a difficulty option but it's probably better left to mods.

Gimbal animations could be fine as long as they're still quick enough to feel responsive.

When I first started playing I didn't know there was a full throttle button so would hit space to activate the first stage then throttle up because well that seemed like the way it would work. Not until a watched Scott Manley (I think) till I learnt of a better way. 

To me nothing that mirrors the player initial expectations of how the world works makes their learning curve harder or the game harder. Bit of build up and excitement doesn't go astray either. Maybe an automated camera that zooms in shows the sparklers and pulls back as engines roar in to life would be good for immersion. just need to work in a launch clamp retraction. 

[t_v]

I’m all for startup times but I want them to be buffed in a similar manner to how ion engines have their thrust buffed to make the game fun without waiting forever for burns. Now that you can timewarp under thrust, the thrust can maybe be reverted but specifically for atmospheric nuclear engines, I don’t want to wait three minutes on the launchpad/runway to get thrust going. Perhaps this would work on the non atmospheric nuclear engines because their main purpose is to push a spaceship through a long burn, but anything that I am landing wit needs to have a quick response time. 

[SciMan]

One though I just had about engine start-up times:

Engine "start-up" times are probably fine to have in the game, with the exception of hypergolic pressure fed engines like RCS thrusters and perhaps the OMS engine (which we need larger versions of, hopefully with better ISP because IMO "monopropellant" covers all types of monopropellants AND hypergolic pressure fed bipropellants).

The issue I have is with engine SPOOL DOWN times. For gameplay reasons, those should be as close to instant as possible (maybe exponentially decaying thrust, so it decays FAST to start but it won't totally stop producing thrust for maybe 10 seconds, which is fine for an engine you're using to land on a planet or moon).

Ion engines and any other electric propulsion method included in KSP 2 should obviously have instant throttle response, as throttling it is as easy as controlling how much electrical power you send to the engine.
Likewise, any engine where the fuel is simply pressure-fed and admission to the engine itself is controlled by solenoid valves (such as RCS obviously, but also ANY pressure fed hypergolic engine), should be not only as close as the game allows to instant as possible, but it should also not have your typical throttle response!
These kinds of hypergolic and monopropellant engines used in RCS and small satellite engines are usually NOT able to be throttled at all. They're either fully on, or fully off, and operation anywhere but those two regimes should be avoided unless transitioning quickly thru them to get to either of the two main operation states (startup and shutdown is fine only because it happens quickly, don't let these kinds of engines "loiter" at like half the propellant flow rate or the engine might blow up because of a "hard start" or more likely just provide unpredictable levels of thrust which will throw off your guidance).
However, there IS a way to use Time to result in a "sort of variable" thrust engine that has quick response but can't throttle. You just need to use coarse Pulse Width Modulation on the flow of propellants, and therefore on the solenoid valves which are typically used to control the flow of propellants to hypergolic pressure fed engines. Pulse Width Modulation is a pretty simple concept actually. You decide on a frequency you want to actuate the control devices at (in this case solenoid valves), and then you simply vary the ratio between time on and time off (with the total still adding up to 100% of the time, if you don't end up with the total being 100% of the time, the only real thing that happens is the operational frequency will change, you'll still get your "smoothly" varying control output, but because it makes the math more complex I won't be analyzing it or using it as an example). So let's say you want the simplest case, 50% of the maximum thrust out of the engine. That's easy, just send the solenoid valves a square-wave signal with the on-time 50% and the off-time also 50%. For 25% throttle, the on-time would be 25% and the off time 75%. For 75% throttle you'd have the signal on 75% of the time and off 25% of the time. Full throttle would no longer be a square wave, instead it would just be a steady on signal, but to put it in the terms of the ratio, it would be 100% on and 0% off. Similarly, for zero thrust, you'd have the signal on for 0% of the time and off for 100% of the time.
As you can see, you just need to have the on-time be the same as whatever throttle setting you want.
Pulse Width Modulation is also used on the fuel injectors of car engines with electronic fuel injection, with the help of a "fuel air ratio table" programmed into the ECU from the factory, along with the additional "fine-tuning" feedback signal of the oxygen sensors both before and after the catalytic converter (this is how the car can tell if the catalytic converter is doing its job, if both the 'before converter' and 'after converter' sensors are reading the same thing at the same time for a given duration under a set of programmed in test criteria (usually highway driving), the car will turn on the check engine light and spit out an OBD2 "P0420" code, which is universal across all auto manufacturers for "the computer thinks something's wrong with the catalytic converter". Now if there's actually something wrong with the converter is another matter, and that takes an auto mechanic to figure out, but usually the computer's right given the signals it has available to it).

The whole point of "why does it matter that a hypergolic pressure fed engine has to use pulse width modulation to get variable thrust levels out of it" is in the sound it puts out.
It would only put out a constant "rocket blast" or "Hiss" sound at 100% throttle, anywhere from 1% to 99% and it would be putting out a fixed-frequency "buzzing" sound and it would sound different depending on the throttle level (not different in frequency but different in "texture" I guess, because your ear can tell the difference between a sound that's mostly on and a sound that's mostly off, even if they're at the same frequency from the same device).
Because these are rocket engines and even the fastest responding ones have a "maximum operating frequency" determined by their (rapid) startup and shutdown times, you'd probably have the engine controller pulsing the fuel supply to them at maybe 40-60 hz. Nice crackly buzz sound too, because it's supposed to be a square wave. But in reality if you took a frequency analyzer to it you'd find that it's actually the same rocket blast or hiss sound you'd hear at 100% throttle, but interrupted with periods of silence or just the engine bell resonating.
What I'm trying to say is that throttling an RCS thruster or something like the LV-1 Ant would be quite a different sound from your typical rocket engine.

 

And if they want to simulate the induced vibration of the vehicle from that unsteady source of thrust, they could do that too. Would give these RCS-only craft quite a bit of character that they're sadly lacking in KSP 1, and hopefully they've developed code to handle Orion and other pulsed-thrust type drives nicely, which might be able to be adapted to these pulse-width-modulated RCS and small satellite thrusters. It would be an interesting challenge to have to deal with to have to design your craft to NOT resonate at the same frequency as the propulsion system operates at, and that's something that has to be done in IRL rocket design as well (ever hear of "pogo oscillations"? Go look it up if you haven't, it's a very real and very problematic thing when making a liquid fueled rocket).