Will there ever be an advanced ascent computer(for RSS)?

[wb99999999]

I play a lot of RSS, both with and without RO. One thing that has been bothering me for a while is that the auto ascent trajectory used by MechJeb is rather primitive: it dumbly follows the parabola preset, and never correct for those low TWR upper stages frequently used in RSS. So would there ever be a "smart" ascent program, whether from MechJeb or not, do the leg work of fine tuning pitch trajectory for a RSS launch? For example a smart ascent would automatically adjust pitch angle to maintain a constant 1G downward thrust once it reaches a set Apogee; it would also "depresses" the final Perigee of the upper stage so that after separation of the payload, the spent stage will fall back into atmosphere. Or maybe something would do a dogleg to reach desired inclination since MJ is terrible at high latitude launches...

I might be using the MJ wrong... in that case feel free to point me out and laugh at me but still, any ideas?

[Wadusher0]

GravityTurn is my go-to ascent autopilot. It should work with RSS but it will have problems with ignitions if you're using the rest of Realism Overhaul (RealFuels, Engine ignitor, etc.). It is however vastly superior to MechJeb and can handle all of what's mentioned in your post - except for deorbiting stages. You have to do that yourself no matter what you use as GravityTurn deactivates when both your desired apoapsis is reached and you're out of the atmosphere.

 

[sarbian]

3 hours ago, wadusher1 said:

It is however vastly superior to MechJeb

It is not for RSS. 

I am tired of reading the same propaganda in half the thread. It is so superior that it needs MJ to have full functionality...

Edited April 14, 2017 by sarbian

[Wadusher0]

4 minutes ago, sarbian said:

It is not for RSS. 

I am tired of reading the same propaganda in half the thread. It is so superior that it needs MJ to have full functionality...

It requires practically every setting to be changed from the default and experimented with when using RSS but that does not mean it is "not for" RSS.

By superior I mean it does the actual ascent with less DeltaV expenditure, with the right settings. If by full functionality you mean the circularization burn, then so what if it doesn't automatically create a maneuver node for it? The way it goes about that is to rotate your ship such that it faces prograde at apoapsis, making the already tiny (relatively speaking) burn even more trivial.

Besides, it's the only alternative to MechJeb that I know of short of flying 100% manually or using kOS - which is a whole other thread. Let OP judge whether it'll do a better job in RSS.

[sarbian]

10 hours ago, wadusher1 said:

it does the actual ascent with less DeltaV expenditure, with the right settings.

Which is what I call propaganda. Each time I tested it I could no see a significant difference in expanded dV. 

[StahnAileron]

I prefer GT because after seeing how it works, it's a bit more intuitive with what I might want/need to change in the settings to get the launch profile I want. I always found it hard to get the right settings/profile with MJ's system. (Setting the right shape.) I can now fly a GT-style (throttle-controlled) launch manually far more consistently than I ever could a MJ-style (always full throttle) launch. It's finesse versus brute force in my eyes. I haven't tested for any dV expenditure differences though. I still value MJ for everything else it offers.

GT should work for RSS (if I recall, it was coded to use whatever parameters were present in the game for the body you're launching from; no hard-coding.)

On ‎4‎/‎14‎/‎2017 at 4:55 AM, wb99999999 said:

[...] For example a smart ascent would automatically adjust pitch angle to maintain a constant 1G downward thrust once it reaches a set Apogee [...]

GT actively adjusts the throttle to maintain a Time-to-Apoapsis. Ideally, it doesn't need to actively to muck with your attitude controls. The only time I've had it do that is if there was a big difference in TWR between stages.

On ‎4‎/‎14‎/‎2017 at 4:55 AM, wb99999999 said:

[...] it would also "depresses" the final Perigee of the upper stage so that after separation of the payload, the spent stage will fall back into atmosphere. [...]

Honestly, that's just putting retros on your upper stage and having them fire at the same time as you decouple it (and decoupling right before the payload's Apoapsis circularization burn). I use the decouplers from FuelTankPlus all the time because of their built-in retro-SRBs. However, that does require some manual control. (I kinda wish the current GT had an option to make the MJ circularization burn optional because of that.)

Anyway, as far as I know, GT is the only other launch autopilot I-... Oh, I think ThrottleControlledAvionics has a launch mode as well, now that I think about it. So yeah, 2-3 choices you have. Take your pick: MJ, GT, or TCA. (I have no idea how TCA might work as I don't use it.)

[toric5]

1 hour ago, StahnAileron said:

I prefer GT because after seeing how it works, it's a bit more intuitive with what I might want/need to change in the settings to get the launch profile I want. I always found it hard to get the right settings/profile with MJ's system. (Setting the right shape.) I can now fly a GT-style (throttle-controlled) launch manually far more consistently than I ever could a MJ-style (always full throttle) launch. It's finesse versus brute force in my eyes. I haven't tested for any dV expenditure differences though. I still value MJ for everything else it offers.

GT should work for RSS (if I recall, it was coded to use whatever parameters were present in the game for the body you're launching from; no hard-coding.)

GT actively adjusts the throttle to maintain a Time-to-Apoapsis. Ideally, it doesn't need to actively to muck with your attitude controls. The only time I've had it do that is if there was a big difference in TWR between stages.

Honestly, that's just putting retros on your upper stage and having them fire at the same time as you decouple it (and decoupling right before the payload's Apoapsis circularization burn). I use the decouplers from FuelTankPlus all the time because of their built-in retro-SRBs. However, that does require some manual control. (I kinda wish the current GT had an option to make the MJ circularization burn optional because of that.)

Anyway, as far as I know, GT is the only other launch autopilot I-... Oh, I think ThrottleControlledAvionics has a launch mode as well, now that I think about it. So yeah, 2-3 choices you have. Take your pick: MJ, GT, or TCA. (I have no idea how TCA might work as I don't use it.)

TCAs launch autopilot it currently full of bees. (great mod, but its autopilots need some work. a lot of work)

[Wadusher0]

12 hours ago, sarbian said:

Which is what I call propaganda. Each time I tested it I could no see a significant difference in expanded dV. 

Oh ok. Would be nice if you said that beforehand - as is stands I understood that to mean my entire post was propaganda... But oh well. 

Still the points I and StahnAileron brought up make GT worth using for launches over MechJeb. And the OP is clearly dissatisfied with MJ's ascent guidance, and needs whatever alternatives are available.

[wb99999999]

Thanks for the advises! GT sounds like a worthy alternative, but the fact it uses throttle rather than attitude to control the trajectory leave me a bit concerned: as you guys might know many engines in RO don't throttle AT ALL... and it feels rather unauthentic to throttle my engine cluster to like 5% during the launch (at least when I'm using RO or RF in RSS). Still I will try GT.

12 hours ago, StahnAileron said:

Honestly, that's just putting retros on your upper stage and having them fire at the same time as you decouple it

Nice idea, but I found it adding a bit of dead weight, and is not very necessary when say, manually adjusting pitch angle during the final insertion could achieve an ideal apogee at 200+ km and perigee at something like 100 km (which is high enough so that the payload or upper stage will only need to burn for 40~120 m/s to circularize, but low enough so the second or third stage will fall back and burn up). This is entirely doable manually, but involves staring painfully at orbital parameters and velocity displays and carefully tapping W and A for about 3 minute. It's just a huge hassle where an automated system would do 100x better.

[wb99999999]

13 hours ago, StahnAileron said:

GT actively adjusts the throttle to maintain a Time-to-Apoapsis. Ideally, it doesn't need to actively to muck with your attitude controls. The only time I've had it do that is if there was a big difference in TWR between stages.

 and this is exactly what happens in RSS... typically you'll end up with a lower stage that burn out at 4G and a upper stage that only do half a G or so at the start and burn for 10 minutes up.

[StahnAileron]

2 hours ago, wb99999999 said:

 and this is exactly what happens in RSS... typically you'll end up with a lower stage that burn out at 4G and a upper stage that only do half a G or so at the start and burn for 10 minutes up.

First, disclaimer: I only play in a stock system (but with plenty of other mods).

That out of the way: This has only been a concern for me during the main boost phase(s) of a launch. I.e. the part were maintaining vertical velocity is more of a concern than applying horizontal velocity. If you've gotten high enough that you're flying more horizontal than vertical, a lower TWR isn't usually too much of a concern unless you're still too low (and relatively slow) in the atmosphere.

I mentioned this elsewhere, but TWR after a certain point (enough to fight and win over gravity) is less of a concern and the dV spread across the various stages/TWRs of your rocket becomes more important. As an example in Stock system (dunno if you can extrapolate this to apply to RSS), it takes ~3400dV average to get to an 80km parking orbit. Generally, I prefer at least 2000-2500 of that to be in high TWR usage. The last 1000dV or so can be applied via a more efficient, lower TWR stage. 2000-2500dV is usually enough to get me to the point were atmospheric drag and the vertical component of my trajectory aren't my primary concern; it's gaining horizontal velocity.

As long as my upper stage's max TWR isn't too far below the TWR of the lower stage(s) when it's ditched (remember, GT has been decreasing the throttle all this time to maintain a specific Apoapsis ETA), the launch should be smooth.

Again, not sure if this is analogous with RSS, but I imagine the same principles apply.

EDIT: I'm a dumbass. Didn't realize you play with RO as well, so engine throttling is out... So yeah, just ignore the above

2 hours ago, wb99999999 said:

Nice idea, but I found it adding a bit of dead weight, and is not very necessary when say, manually adjusting pitch angle during the final insertion could achieve an ideal apogee at 200+ km and perigee at something like 100 km (which is high enough so that the payload or upper stage will only need to burn for 40~120 m/s to circularize, but low enough so the second or third stage will fall back and burn up). This is entirely doable manually, but involves staring painfully at orbital parameters and velocity displays and carefully tapping W and A for about 3 minute. It's just a huge hassle where an automated system would do 100x better.

This is why I stage the upper stage at the Apoapsis before the payload's circularization burn. It takes either no dV (the periapsis still is below the cut-off) or only a little (peri is above the cut-off). Granted, again this is in stock. I know the dV requirements in RSS are larger, so perhaps this would bite into your dV.

Edited April 16, 2017 by StahnAileron

[wb99999999]

4 minutes ago, StahnAileron said:

I mentioned this elsewhere, but TWR after a certain point (enough to fight and win over gravity) is less of a concern and the dV spread across the various stages/TWRs of your rocket becomes more important. As an example in Stock system (dunno if you can extrapolate this to apply to RSS), it takes ~3400dV average to get to an 80km parking orbit. Generally, I prefer at least 2000-2500 of that to be in high TWR usage. The last 1000dV or so can be applied via a more efficient, lower TWR stage. 2000-2500dV is usually enough to get me to the point were atmospheric drag and the vertical component of my trajectory aren't my primary concern; it's gaining horizontal velocity.

The thing about RSS launch is that the vast majority of time was spent out of atmosphere and gaining horizontal velocity. An orbital launch from Earth takes about 9.6 km/s, and when you're at 120 km altitude where atmosphere is basically nothing, you usually only spent something like 3 to 4 km/s and are at about 2 to 3 km/s speed. The gravity turn part is mostly done at this point, and what one should do is to burn almost perpendicular to the surface for the rest of the launch, while pitch up a little according to TWR to keep vertical velocity at around 0 or slightly positive. This phase of launch takes so much more time and fuel in RSS when compared to stock...

[Maxsimal]

It would be really nice if mechjeb supported something like switching over to GT's system at a preset altitude.  For my launches, I usually do this semi-manually - I figure out what a good MJ profile settings are to get to about 70 kilometers with a nice smooth prograde-facing gravity turn, and then just adjust MJ's pitch angle to manually control vertical speed the rest of the way up, knowing roughly what sort of vertical speed/altitude I need to be at to hit my typical desired apoapsis of 150km.

I've generally found that I can't get MJ to give me a good gravity turn but also an efficient burn-to-apoapsis otherwise, due to the very different TWR and burn times of the stages.

Of course, it gets much easier once you unlock booster engines that have more than one restart and you're ok with MJ hitting your desired apoapsis before your booster stage has run out of fuel, but for most of the difficult portion of RSS/RO that's not the case.