[1.9-1.10] Throttle Controlled Avionics

[Peter Balholm]

Here is an up-to-date log file. It is using the 64-bit version.  I am using the most recent version of TCA, I believe. Thank you so much for your help.

This time I had the same blank TCA window, and the game eventually crashed. I have an 32-bit version of KSP installed, as well. Could that be causing trouble?

https://drive.google.com/file/d/0ByFshSPbart8X0xwZkMxc2o4NWs/view?usp=sharing

[Peter Balholm]

Here is an up-to-date log file. It is using the 64-bit version.  I am using the most recent version of TCA, I believe. Thank you so much for your help.

This time I had the same blank TCA window, and the game eventually crashed. I have an 32-bit version of KSP installed, as well. Could that be causing trouble?

https://drive.google.com/file/d/0ByFshSPbart8X0xwZkMxc2o4NWs/view?usp=sharing

[allista]

1 hour ago, Peter Balholm said:

Here is an up-to-date log file. It is using the 64-bit version.  I am using the most recent version of TCA, I believe. Thank you so much for your help.

This time I had the same blank TCA window, and the game eventually crashed. I have an 32-bit version of KSP installed, as well. Could that be causing trouble?

https://drive.google.com/file/d/0ByFshSPbart8X0xwZkMxc2o4NWs/view?usp=sharing

I'm sorry, but you can't use TCA versions above 3.1.0 with KSP-1.1.2.

TCA-3.2.* are only compatible with KSP-1.1.3. Here, see the Changlog page at SpaceDock: http://spacedock.info/mod/198/Throttle Controlled Avionics

[Peter Balholm]

Thank you very much! I am downloading the newer version of the game right now! I MUST get this working. :-)

[allista]

On 9/26/2016 at 6:51 PM, volcanicshrimp said:

That is a possible cause of the problem, however it still loses control in the same manner if it's hovering stationary 50m above the ground (in stop mode) and I tap w (this was in a TAC+dependencies-only install) with my control input only accelerating it to 1-2 m/s.

I've played a little with this problem. It turned out I was wrong with my guess: the RAPIERs are not at fault (they respond faster than jets for that matter).

What happens is that Attitude Control of TCA cannot cope with the combination of slow thrust change and large torque. It takes into account engine's response speed, but it is not an AI (yet), so there's limits: when you place slow engines this far from the CoM, even the slightest variation/oscillation of their thrust due to control loop hysteresis results in enough unwanted torque to cause what we see. I will try, as I said, to tune the Attitude Control a little more, but this will take time and the result is not guarantied. But you can already guess a simpler solution: just place those engines closer to the CoM. At some point the instability will be gone.

[Denko666]

For my VTOLs i tend to use these engines

Edited September 30, 2016 by Denko666
found engines on spacedock

[volcanicshrimp]

On 9/27/2016 at 5:58 PM, allista said:

just place those engines closer to the CoM

Thanks for looking into it, and thanks for the tip. I never thought of torque affecting it like that.

Now back to ill-advised craft designs!

I did try some atmospheric ballistic guidance with a rocket, and it didn't go very well. Think the rocket design was to blame though.

[ss8913]

So.. things are working better with KSPIE than they were.. now I'm questioning if I'm doing something wrong... if I just want my VTOL spaceplane (with 2 engines situated for and aft of the CoM) to balance the thrust so that I can fly like a helicopter and have it balance the thrust around the CoM... what settings do I use?  ie, how the "old" version of this mod used to work... what are the proper settings?  I don't want it to do anything else.. I don't want it to follow terrain, I don't want it to control altitude, I don't want it to hold position over the ground, just balance thrust around CoM on a spaceplane designed as described...?

[allista]

6 hours ago, ss8913 said:

So.. things are working better with KSPIE than they were.. now I'm questioning if I'm doing something wrong... if I just want my VTOL spaceplane (with 2 engines situated for and aft of the CoM) to balance the thrust so that I can fly like a helicopter and have it balance the thrust around the CoM... what settings do I use?  ie, how the "old" version of this mod used to work... what are the proper settings?  I don't want it to do anything else.. I don't want it to follow terrain, I don't want it to control altitude, I don't want it to hold position over the ground, just balance thrust around CoM on a spaceplane designed as described...?

Set those two engines to "Thrust" mode; disable VTOL Assist and Flight Stabilizer; do not enable anything you don't want to.

[Peter Balholm]

I have installed KSP 1.1.3, and now TCA works! Thank you!

However, I have one problem. I turn on "To Orbit" and use the default settings. Liftoff proceeds smoothly. (the fact that "TO Orbit is shown in red is worrisome, though). at about 1500 meters elevation, the autopilot throttles down to less that 1.0 TWR. That can't be good (see picture). It gradually recovers, and begins accelerating again. Finally, at about 5000 meters, it throttles back to nothing and the engines shut off completely. This is no way to achieve orbit!

What am I doing wrong? How can I fix this?

 

[allista]

46 minutes ago, Peter Balholm said:

I have installed KSP 1.1.3, and now TCA works! Thank you!

However, I have one problem. I turn on "To Orbit" and use the default settings. Liftoff proceeds smoothly. (the fact that "TO Orbit is shown in red is worrisome, though). at about 1500 meters elevation, the autopilot throttles down to less that 1.0 TWR. That can't be good (see picture). It gradually recovers, and begins accelerating again. Finally, at about 5000 meters, it throttles back to nothing and the engines shut off completely. This is no way to achieve orbit!

What am I doing wrong? How can I fix this?

Picture?

Anyway, I suppose you're trying to launch a classic rocket with engines aligned with the CoM? Then you should set those axial engines to UnBallanced Thrust mode.

I think I'll try to make TCA guess the most probable mode by engine's location/orientation, but for now you have to assign them manually.

Please, read the Manual about engine mods.

[Stone Blue]

On 9/27/2016 at 2:09 PM, Denko666 said:

For my VTOLs i tend to use these engines: https://www.dropbox.com/s/3w9beelss7yrlvi/TGOLgroup.tgz?dl=0 

I'm linking to my dropbox because i cannot find them anymore, they have vanished from Curse, KSP Forum and Spacedock..

Ummm... those are still up on SpaceDock... just not updated... not that they needed it for 1.1.3???

Edited September 30, 2016 by Stone Blue

[Peter Balholm]

Sorry about the photo. Hope this works. I will (re)read the manual.

https://www.dropbox.com/s/o5c06nv6rz1dpbk/TCA screenshot.jpg?dl=0

Edited September 30, 2016 by Peter Balholm
Picture didn't take first time.

[Peter Balholm]

I have reread the manual.

Once again, the "To Orbit" autopilot throttled down at about 1500m, and totally shuts down about 5000m. And the "To Orbit" button is shown in red, which I thought would mean it can't work. However, it displays info like "gravity turn" at the bottom, so It seems like it is working at least a little.

The first stage on my rocket has well over a hundred clustered engines, and if modeled as such, would lead to an unworkable part count. I custom built the bulk of these engines together in Blender, etc. to act a one part. It has the combined thrust of all the engines, and multiple thrust points. This bundle is Engine Group G5. After your last message, I have switched it from "Thrust" to "Unbalanced Thrust" since it is on the axis of the rocket, thrusting in the direction of the axis, and is symmetrical about it. (Although I really don't understand this terminology; it seems to me that this thrust IS balanced.) My intention is that this central cluster would thrust at full power the whole time.

There are also 18 individual engines clustered around the central block, which are intended as the steering engines. I have them as group G4. They are set to "Thrust and control."

https://www.dropbox.com/s/8sh5hu9snqzoaaa/TCA screenshot 2.jpg?dl=0

Thank you for your patience and thank you for writing a manual! I am sorry I haven't quite got this yet.

[allista]

27 minutes ago, Peter Balholm said:

Sorry about the photo. Hope this works. I will (re)read the manual.

https://www.dropbox.com/s/o5c06nv6rz1dpbk/TCA screenshot.jpg?dl=0

 

8 minutes ago, Peter Balholm said:

I have reread the manual.

Once again, the "To Orbit" autopilot throttled down at about 1500m, and totally shuts down about 5000m. And the "To Orbit" button is shown in red, which I thought would mean it can't work. However, it displays info like "gravity turn" at the bottom, so It seems like it is working at least a little.

The first stage on my rocket has well over a hundred clustered engines, and if modeled as such, would lead to an unworkable part count. I custom built the bulk of these engines together in Blender, etc. to act a one part. It has the combined thrust of all the engines, and multiple thrust points. This bundle is Engine Group G5. After your last message, I have switched it from "Thrust" to "Unbalanced Thrust" since it is on the axis of the rocket, thrusting in the direction of the axis, and is symmetrical about it. (Although I really don't understand this terminology; it seems to me that this thrust IS balanced.) My intention is that this central cluster would thrust at full power the whole time.

There are also 18 individual engines clustered around the central block, which are intended as the steering engines. I have them as group G4. They are set to "Thrust and control."

https://www.dropbox.com/s/8sh5hu9snqzoaaa/TCA screenshot 2.jpg?dl=0

Thank you for your patience and thank you for writing a manual! I am sorry I haven't quite got this yet.

Wow... just wow!

But now I understand the problem. Your rocket has low control authority: it's too heavy, control surfaces are small and few and so on. So despite the correct engine configuration, TCA is unable to steer the beast as quickly as it should (see the "Err: 22.5°" display?). In that case TCA lowers the throttle to decrease overall error of the maneuver. This is arguably not the best strategy for all scenarios, but for most it gives more accuracy and fuel efficiency.

The best way to handle this, as I see it, is to add several radial RCS thrusters at the top of the rocket (as far from the CoM as possible). This will allow you to have more leverage independent of the throttle. Or, if RCS are too weak, add some LFO thrusters in Maneuver mode (they will still be limited by the throttle, no way around it, but otherwise they will act like RCS).

I'll think about how to improve the autopilot to work around such cases.

Thank you for the report!

[Peter Balholm]

Okay, thanks! I'll see what more-meaty thrusters I can add. Your explanation sounds good, although I am almost positive that the engines throttle down while the ship is still on course (errors of just a few degrees and "system nominal"). It drifts off after shutdown.

By the way, you have no idea what an improvement your mod has made for me. I have been trying to get this beast into orbit with no gimbling and nothing but RCS thrusters. It turns like frozen molasses.

I have made the design choices I did, not because I didn't know any better, but because I am working on a story about travel to Mars with 1917 technology. They are using a hybrid rocket with paraffin and nitrous oxide. A hybrid cannot be scaled up indefinitely without extending its burn time. So, in order to get off the ground, massive clustering is REQUIRED. And gimbling might still be possible, but it would cause serious technical headaches.

[allista]

1 hour ago, Peter Balholm said:

I am almost positive that the engines throttle down while the ship is still on course (errors of just a few degrees and "system nominal"). It drifts off after shutdown.

The equation is simple:

throttle = 1 if error < 1, 0 if error > 10, else 1-(error-1)/9

So error of 6 deg results in 4/9 of the throttle.

Quote

I have made the design choices I did, not because I didn't know any better, but because I am working on a story about travel to Mars with 1917 technology. They are using a hybrid rocket with paraffin and nitrous oxide. A hybrid cannot be scaled up indefinitely without extending its burn time. So, in order to get off the ground, massive clustering is REQUIRED. And gimbling might still be possible, but it would cause serious technical headaches.

That's amazing! 

[Denko666]

3 hours ago, Stone Blue said:

Ummm... those are still up on SpaceDock... just not updated... not that they needed it for 1.1.3???

And you are absolutly right. I Just didn't find them because i was looking for T.G.O.L. or VTOL. 

I'll update my post

[Peter Balholm]

8 hours ago, allista said:

The best way to handle this, as I see it, is to add several radial RCS thrusters at the top of the rocket (as far from the CoM as possible). This will allow you to have more leverage independent of the throttle. Or, if RCS are too weak, add some LFO thrusters in Maneuver mode (they will still be limited by the throttle, no way around it, but otherwise they will act like RCS).

I'll think about how to improve the autopilot to work around such cases.

I added four BIG thrusters (c. 75 kN) and put them right up near the top. It flies straight up beautifully, and doesn't deviate more than a degree or two. (of course, other than the RCS, I have no roll control). But when I set the orbital autopilot, It still does the wobble at 1500 meters, and goes completely off the rails at 6000 meters, shutting down entirely. Thank you for the formula, by the way. It makes it all much less mysterious!

So, it would appear that the gravity turn is kicking in too fast for even my enhanced craft, and forcing it to shut down.

I tried to fly it nearly straight up, basically beyond the atmosphere, and then hit the Prograde control. I had hoped that this would force it to thrust over at the horizon. But It did not seem to try, and continued at about a 60 degree angle.

I don't know what I should do, but let me explain how this rocket's throttle steering should work, in my fictional real-life.

The cluster of 127 rockets would all fire at launch. The rocket would lift off. If it got a little off course, all of the rockets on one half of the rocket (roughly 60) would throttle back. In an extreme situation, they would shut down entirely for a moment, perhaps. Thrust would be lost, of course. But under no circumstances could it drop below 50%.

To make this work with the first stage (for example), I would probably rebuild my part files. Instead of having one giant cluster with 18 strapped-on control rockets, I would have one central rocket (required by symmetry) and six-pie shaped wedges of clustered rockets. (It is sixfold-symmetry because that is how cylinders pack the best--the reason for honeycomb.) Each of these wedges would be one engine as far as the game was concerned. Therefore they could be throttled up and down and control the ship with a great deal of authority, and the part count would only be 7 for the entire first stage!

My question is, can TCA control such a craft? Could I get past the shutdown gremlin?

Thank you for your help!

[allista]

1 hour ago, Peter Balholm said:

I added four BIG thrusters (c. 75 kN) and put them right up near the top. It flies straight up beautifully, and doesn't deviate more than a degree or two. (of course, other than the RCS, I have no roll control). But when I set the orbital autopilot, It still does the wobble at 1500 meters, and goes completely off the rails at 6000 meters, shutting down entirely. Thank you for the formula, by the way. It makes it all much less mysterious!

So, it would appear that the gravity turn is kicking in too fast for even my enhanced craft, and forcing it to shut down.

I tried to fly it nearly straight up, basically beyond the atmosphere, and then hit the Prograde control. I had hoped that this would force it to thrust over at the horizon. But It did not seem to try, and continued at about a 60 degree angle.

I don't know what I should do, but let me explain how this rocket's throttle steering should work, in my fictional real-life.

The cluster of 127 rockets would all fire at launch. The rocket would lift off. If it got a little off course, all of the rockets on one half of the rocket (roughly 60) would throttle back. In an extreme situation, they would shut down entirely for a moment, perhaps. Thrust would be lost, of course. But under no circumstances could it drop below 50%.

To make this work with the first stage (for example), I would probably rebuild my part files. Instead of having one giant cluster with 18 strapped-on control rockets, I would have one central rocket (required by symmetry) and six-pie shaped wedges of clustered rockets. (It is sixfold-symmetry because that is how cylinders pack the best--the reason for honeycomb.) Each of these wedges would be one engine as far as the game was concerned. Therefore they could be throttled up and down and control the ship with a great deal of authority, and the part count would only be 7 for the entire first stage!

My question is, can TCA control such a craft? Could I get past the shutdown gremlin?

Thank you for your help!

What you describe about how it should work is exactly what TCA does. There's just not enough torque.

(Just in case, your 75kN thrusters at the top are pointing perpendicular to the main axis, right?)

You may try to use MJ's Ascend Autopilot instead of TCA's one, leaving for TCA the engine-balancing work only.

Otherwise, I'll have to do some programming to overcome the issue. Thank you for the challenge  

[Peter Balholm]

9 hours ago, allista said:

What you describe about how it should work is exactly what TCA does. There's just not enough torque.

(Just in case, your 75kN thrusters at the top are pointing perpendicular to the main axis, right?)

You may try to use MJ's Ascend Autopilot instead of TCA's one, leaving for TCA the engine-balancing work only.

Otherwise, I'll have to do some programming to overcome the issue. Thank you for the challenge  

I'll try to increase the torque by rearranging my engines. I'll also try MechJeb. I didn't know that I could still use it. I suppose I just keep the Stability button clicked on TCA, and nothing else?

[allista]

22 minutes ago, Peter Balholm said:

I'll try to increase the torque by rearranging my engines. I'll also try MechJeb. I didn't know that I could still use it. I suppose I just keep the Stability button clicked on TCA, and nothing else?

It's better to disable both VTOL Assist and Flight Stabilizer as they're designed more for VTOLs than rockets.

Meanwhile I'm trying to fix things on my end and getting interesting results. I'll post a video later...

[DStaal]

2 hours ago, Peter Balholm said:

I'll try to increase the torque by rearranging my engines. I'll also try MechJeb. I didn't know that I could still use it. I suppose I just keep the Stability button clicked on TCA, and nothing else?

It might also be worth trying GravityTurn - it often requires less torque to perform the ascent in the first place, in my experience.  Though I expect using it and TCA together at the same time won't work to well, as the main control it worries itself with is the throttle.

(There's a community-supplied build for 1.1.3 near the end of the original thread.)

[allista]

16 hours ago, DStaal said:

It might also be worth trying GravityTurn - it often requires less torque to perform the ascent in the first place, in my experience.  Though I expect using it and TCA together at the same time won't work to well, as the main control it worries itself with is the throttle.

(There's a community-supplied build for 1.1.3 near the end of the original thread.)

I don't see why TCA would not work with that if it works with MJ. Just use the engine modes and basic balancing functionality with this.

[Denko666]

Hey @allista You're probably busy updating TCA to KSP 1.2. 

But When u have some time/interest for adding something to TCA: I got a little challenge for ya. . 

I use TCA on my spaceplanes. the idea is: an incomplete interplanetary vessel is built in orbit. A spaceplane then takes off from the KSC just like a regular plane, but climbs all the way up to space and meets with the interplanetary ship. It docks, completing the setup by serving as its command bridge. once arrived at destination, the spaceplane undocks again and serves as lander on the new world. 

But

Lets say the destination is a planet with an atmosphere, like say Duna. Landing a spaceplane o, a, uninhabited world (and so: no runway) is next to impossible. So once the spaceplane has descended enough it would need to switch from plane mode to VTOL mode. I haven't found a way to do that with TCA. by the time i got it switched over (switching engine config, activating proper modes) something disastrous either has happened or has become inevitable. Maybe TCA can already do that, but then i don't see how.