Jump to content

artwhaley

Members
  • Posts

    639
  • Joined

  • Last visited

Everything posted by artwhaley

  1. I haven't had any MAJOR trouble with the ascent autopilot - it sometimes overshoots the circularization burn, so I always keep an eye on that part of the launch... but otherwise, it does great. Things to watch out for - you must have 'control from here' set to a part that points towards the sky... if you're putting together a rocket that includes docking ports, probe cores, landers, etc. that face some direction other than 'rocket up' you need to verify before launch that the controls are being referenced from something that points the right way. You can tell easily by looking at the navball - if it's all sky before you launch, you're ready to go. If you see half ground or all ground, you need to right click a reaction wheel, docking port, or command pod that points the right way until you see all sky. If you don't, the autopilot will helpfully pitch you right over into the ground. Otherwise, if the rocket flies well - as in the loads are mostly straight over the centerline and there's not excessive wobble in between stages then mechjeb can launch it just fine. Does any of that help with your issue? Have you tried launching the rocket manually? Do you have any trouble controlling it at a certain point in flight? The docking autopilot frequently makes strange decisions- the only situational awareness it's got comes from looking at the math... so it doesn't see things the same way a human does. With two small ships that have well layed out RCS thrusters, it will do just fine. If you're trying to do anything more complicated... it has to make guesses about what is where based on the vessel size and such... so it will assign a far-off starting point and safe distance... and even when it doesn't get confused it can waste a lot of rcs fuel getting lined up the way it wants to. Here's a better way to dock with mechjeb - 1. On the less maneuverable vessel, you can select 'control from here' at the docking port you want to use, set the docking port on the OTHER vessel as the target, and then turn on Smart A.S.S. and select "Target +". This will make the vessel you're going to dock to point their docking port at you. You don't have to do this step, if it's a big station or something that doesn't like to rotate - but then you have to work a LITTLE harder because you have to get lined up yourself. 2. Switch back to the more maneuverable vessel. Target the docking port you're headed for. Select 'control from here' for the docking port on your vessel. Turn on Smart A.S.S. and select 'Parallel Minus.' 3. Now one vessel is keeping the ports parallel and the other vessel is keeping them facing at each other. Switch the navball to target mode, if it isn't already. Use the "H" key to get a little forward velocity towards the target, then use "IJKL" to put the yellow 'relative velocity plus' navball bug right in the middle of the pink navball 'target' circle. If you keep the yellow bug in the pink circle and keep moving forward at a reasonable rate- docking complete.
  2. I'm working on a hardware cockpit for kerballing... and discovering Telemachus today just saved me weeks of trying to learn to program in C# and figure out how to find my way around KSP and Mechjeb! I dug out the TI Connected Launchpad board that I bought when they came out, played with the demo app, and then put on the 'someday' shelf and told it that it's day has come.... and an hour later I had the two physical buttons on the launchpad commanding mechjeb to dance back and forth between prograde and retrograde! Along with the built-in ethernet support, the launchpad has over 60 GPIO pins, and with an impressive list of pwm, i2c, uart, and usb functionality - all for $20.00! And it can be programmed with Energia, which makes it accessible to anyone who's tinkered with arduino. And, because it's communicating with the simulator via network, there's no reason you couldn't expand it and have more than one... if you REALLY needed more than io (and didn't want to use an io expander). I think it's going to be perfect for getting myself a cockpit full of buttons switches and blinky lights.... and I'll of course share as I go. All that said - I've got a question for those who know more about Telemachus than I do... is there somewhere that documents the maximum and minimum value for all of the settable parameters? (other than the boolean ones... I think I can handle those! ) The API lists them... but I haven't found the range of values yet. I suppose it's the same range as what the query's report... so I'll just spend some time watching the values change and see if I can figure out what sorts of values it expects sent back... but I wanted to ask in case it's all written down somewhere I just haven't looked yet! And... one quick (and then one probably slow) feature request... Is there any chance Telemachus could get a mj.killrot api command to activate kill-rotate? And the (probably) slower one... I wish there was a way to query the mechjeb mode... or at least keep track of the last command that Telemachus sent to it... that way you could use lighted buttons that would show you what autopilot mode you currently had selected. This would be especially handy since the on-screen A.S.S. window just says "Auto" when it's responding to a Telemachus command. If that's overly tough... would it be substantially easier to query mechjeb to find out if ANY autopilot control is currently active, so you could activate a light that told you that mechjeb was controlling the ship, so you knew to press the 'off' button if you wanted it to let go? Thanks for any help anyone can provide! Art
×
×
  • Create New...