The need for a Moon direct *2* - ...continue here.

GCNRevenger · 2004-10-31 15:09:12

Operating a machine teleroboticly with a millisecond time lag and a multi-second time lag is completly different Grypd, an entirely and wholey different thing. Remote-control human surgery or NASA's Robotnaut are possible because of near-zero "real time" control, but putting them a quarter million miles away and the story changes completly.

Think about it, having to perform delicate work like electrical wiring, fine welding, that sort of thing without real time feed back of any kind. It wouldn't be working in slow motion because action and reaction would be disjointed, which is a problem that is unavoidable. Try rewiring somthing with only opening your eyes every few seconds, and with no feeling in your fingertips. Computers can't predict that well, and are only about good enough to drive rovers around on flat ground. The MERs can't even drive straight on slopes because the computer can't predict accuratly. So, even if you do build a robot that is dexterous enough to rival humans, you still can't use it without a human controller nearby.

I also don't think that robots in general are good enough yet to actually build things, like taking metal bars or plates or pipes and bolt/weld them together with high accuracy. You can do this on an assembly line on Earth because everything is easy to align perfectly, which you can't do on the Moon. It only takes a small error to ruin any structure or machine.

John Creighton · 2004-10-31 15:46:28

Here is the approach I envision. The persons arm is attached to some feed back mechanism. The greater the distance he/she moves his arm from that predicted by the robot arm the greater the force he will feel. Acceleration will depend on the force the person feels but there will be a saturation in the desired velocity.

The velocity saturation will be dependent on the safety level and the predicted proximity to various objects. So for instance if the robot arm is very close to delicate components it may only be able to move a few millimeters or less per seconded. The operator will feel small forces or predicted forces in the robotic arm by temperature and will feel larger forces by vibrations or even changes in roughness.

There will also be controls in place so that the external forces in the robotic arm do not exceed those which we predicted when giving the operator his sensory feedback. If the robotic arm is resilient then the force it feels may not change much when it is deflected from its desired position. Consequently there will be a strong correlation between the sensory feedback in the operators control device and the force in the robotic arm. Clearly if the robotic arm was to drop something while operating in a high safety mode the operator would not have time to react. However if the operator is very careful he/she should not drop anything and should be able to do such things as change wires on a bread board and replace integrated circuits.

GCNRevenger · 2004-10-31 16:24:01

If it were like that you'd be reduced to flailing your arms around like the robot on Lost In Space (the show)... The only practical method for real teleoperation is for the arm on the machine end to reccord the motion that actually happened and send that data back to the operator and move his prosthetic arm to what the position actually is (within limits).

This STILL does not get around the time lag problem. Even if you move very very slowly you still have to wait to see what actually happens, and that will kill the whole proposition right there. It simply precludes real force feedback except for a computers' prediction, and the operators arm is going to get moved around all over the place of the operator end. It will be like working with no feeling in your hands except for the computers' guess.

And when i'm talking about delicate work, I'm not thinking only about fixing circut boards or redoing damaged wiring, I'm think moving multi-meter multi-kilo metal pipes to within a fraction of a millimeter and welding them, with no support from the ground except the rough Lunar surface. Increased sensitivity controls by making large operator movements make small robot movements will help, but you are still limited by the time lag. Put some gloves on and try to thread a needle with your eyes closed every other four seconds, except you can only move your hands with your eyes closed.

John Creighton · 2004-10-31 16:55:21

Even if you move very very slowly you still have to wait to see what actually happens, and that will kill the whole proposition right there. It simply precludes real force feedback except for a computers' prediction, and the operators arm is going to get moved around all over the place of the operator end.

If you move slow enough not much changes every 11 seconds so if you are careful you don’t have to wait to see what happens. Therefore it doesn’t preclude force feedback and the force that is feed back will be close to the actual feedback. The principle is over sampling. If you sample fast enough relative to the rate the signal changes then the previous estimate is a good predictor of the future estimate. So what if the operators arm gets moved around. The key is to have good enough mechanics and control systems so the robots arm doesn’t get moved around much.

John Creighton · 2004-10-31 16:59:49

And when i'm talking about delicate work, I'm not thinking only about fixing circut boards or redoing damaged wiring, I'm think moving multi-meter multi-kilo metal pipes to within a fraction of a millimeter and welding them, with no support from the ground except the rough Lunar surface.

Fine you can talk about whatever task you want and deliberately choose a task you think will be two difficult for a telleoperated robot to make your case. I don’t think that positioning a pipe within a fraction of a millimeter is an unrealistic task. However, the key is not to find the most difficult tasks for the robots. The key is to figure out what areas they will be most helpful in. Then the engineering problem can be tackeled in two ways. The first is to find a way to make the task easier. The second is to improve the performance of the robot.

GCNRevenger · 2004-10-31 17:52:29

Even if you move very very slowly you still have to wait to see what actually happens, and that will kill the whole proposition right there. It simply precludes real force feedback except for a computers' prediction, and the operators arm is going to get moved around all over the place of the operator end.
If you move slow enough not much changes every 11 seconds so if you are careful you don’t have to wait to see what happens. Therefore it doesn’t preclude force feedback and the force that is feed back will be close to the actual feedback. The principle is over sampling. If you sample fast enough relative to the rate the signal changes then the previous estimate is a good predictor of the future estimate. So what if the operators arm gets moved around. The key is to have good enough mechanics and control systems so the robots arm doesn’t get moved around much.

Frankly, I don't think that you can move slowly enough that you could overcome a 3-4+ second gap even with oversampling when you are dealing with delicate tasks, the human hand is only so steady even if you do magnify the motion. If the operator end of the arms is always moving to update with reality, the arms are going to be moving all over the place and the operator will have a rough time keeping up. Kind of like having a twitch in your hands when you are doing surgery, it just won't do.

No matter what computational trick you pull to make the gap "feel" smaller, you aren't ever ever going to have fast enough updates with a multi-second delay to make it like real time. If it were a quarter of a second then maybe, but its just not, and thats the way it is.

comstar03 · 2004-10-31 18:00:16

John Creighton,

Any lag time would be adjusted for in the Virtual Landscape environment. The first test of developments using telerobotics would be in earth orbit assembling structures in orbit.

We would have all the bugs out of the system before going to the moon with the droid system. Remember these droids are for the large work not the precision work that humans can do. The main functions include assembly of buildings, open cut mining operations, mass driver assembly, Landing platforms and movement of cargo just to name a few.

another method of robotic controls is semi-telerobotics where you work with the droids avatar and controls the positioning not the actions ( similar way a foreman tells a trademan were to go to commence , but the trademan has the knowledege foir the task at hand.

John Creighton · 2004-10-31 18:18:12

Any lag time would be adjusted for in the Virtual Landscape environment. The first test of developments using telerobotics would be in earth orbit assembling structures in orbit.

If you have enough knowledge about your environment then this is possible. However it is a considerable software challenge. Clearly such techniques are in use and will get more sophisticated. However if the job is critical enough and the components delicate enough there will be a point where people may decide this method is not worth the risk and they will instead use a slower safer method.

Frankly, I don't think that you can move slowly enough that you could overcome a 3-4+ second gap even with oversampling when you are dealing with delicate tasks, the human hand is only so steady even if you do magnify the motion. If the operator end of the arms is always moving to update with reality, the arms are going to be moving all over the place and the operator will have a rough time keeping up. Kind of like having a twitch in your hands when you are doing surgery, it just won't do.

The operator’s device would be designed to hold the operator in place where the robot arm is supposed to be. Therefore, it won’t matter how steady the hand of the operator is because the operators control will keep his hand and arm steady. Clearly the operators arm will deviate from this position and the control will exert a restoring force. However, this deviation will not change the motion or position of the robots arm unless this restoring force exceeds some threshold level. The level should be set high enough so we know this is a desired motion and not the inability of the operator to keep his arm or hand stead. Clearly this will get tiring so at any point the operator can with his other arm push a disengage button and his sub can take his previous position.

GCNRevenger · 2004-10-31 18:21:01

Any lag time would be adjusted for in the Virtual Landscape environment. The first test of developments using telerobotics would be in earth orbit assembling structures in orbit.

Trouble is, you just can't do that Commstar. You can't make the time lag just go away through a computational trick. Reality will never mesh entirely with what the computer predicts will happen in the VR world.

My worry about building large structures is that without human help to allign the pieces, that robots won't be able to position pieces accuratly enough to join them together properly since spacing/angle/etc isn't easy to get right on the rough Lunar surface.

comstar03 · 2004-10-31 19:39:14

GCNRevenger,

You don't believe that our technology is not advanced enough to build a droid that could assembly large structures on the moon without humans being there.

I don't think you have got the VR worlds have been constructed eg.: for software developer in german to reprogram the entire railway sstem they needed a way to speed large complex processes up by assemblying blocks of software code in VR land just click them together.

The VR world for the droid assembly team works from the view of the droids and created into a graphical landscape, or we could do a radar mapping of the landing site and build a 3D virtual landscape, then use that 3D world as the basis of the VR Landscape to work with the droid avatar. ( All current technology )

Remember, The preparation droids would first clear and survey and mark the site out from development as per the maps devised on earth construction office.

Once the site is prepared then the assembly droids are used to set the assembly process for the moonbase / mining site / any other type of site suggested.

I understand that technology hardware is there, and the software will be there to run the hardware for the droids. It just needs resources to do that process.

Grypd · 2004-10-31 20:07:57

Any lag time would be adjusted for in the Virtual Landscape environment. The first test of developments using telerobotics would be in earth orbit assembling structures in orbit.
Trouble is, you just can't do that Commstar. You can't make the time lag just go away through a computational trick. Reality will never mesh entirely with what the computer predicts will happen in the VR world.
My worry about building large structures is that without human help to allign the pieces, that robots won't be able to position pieces accuratly enough to join them together properly since spacing/angle/etc isn't easy to get right on the rough Lunar surface.

Then if you cant get rid of the time lag and you are worried about a delicate operation you need a robot to do, it becomes a software program. A system the nuclear industry use is to have the pieces you wish to manipulate mapped by use of a minimum of three cameras. It is then recreated in a computer model and is then manipulated by the operator and since the actions of the operator are exactly recorded it is downloaded to the robot which exactly mimicks the action performed, except this time in reality. They have been able to operate as good as any human could unprotected in an enviroment that is considered very unhealthy to man. This is the answer to your problem.

But for the most part we wont have this problem we will be using telerobotics initially to move and help deploy our initial landings of equipment. They will then clear a space and provide and secure a beacon so that the rest of the craft coming down will land in place. It will be then that they start to actually harvest the Lunar Regolith for the factories. And I agree with John Creighton the first would be oxygen and solar cell production factories.

GCNRevenger · 2004-10-31 21:00:53

Thats correct Commstar, we don't have the technology. Computers must be able to have some understanding of the task in order to deal with unplanned issues, which they cannot do that yet, nor can robots coordinate their motion without a fixed, rigid reference point (concrete floor, not springy tires on soft dust & gravel).

You are missing an important issue with your virtual reality idea, is that what really happens is not what you predict it will in your virtual model. Trouble is, you don't know what really happens for about two or three seconds that it takes a signal to reach Earth and be processed, and you have to wait to tell robots on the Moon what to do for the signal to reach the Moon and be processed. In either event, you have unavoidable problem that it takes time for signals from Earth to reach the Moon and vice versa, which is not going to go away. Period.

Mapping out everything requires exact knowledge of where the cameras are I imagine so that probobly wouldn't work on the Moon. Even if it did, I don't think that the robot's motion could be coordinated to the cameras' location very well because of differences in terrain on the Moon versus on Earth at mission control on flat cement (or VR).

comstar03 · 2004-11-01 00:34:49

GCNRevenger,

Yes, you need to provide a reference to the droid to understand where it is , just like humans. You have little understanding how robotics work and CAD/CAM/CAE Systems and Software have advanced over the past few years. Using topographics mapping sytems, laser marking systems and other ground / site preparation systems we could have a droid understanding where it is, what it is doing, using graphic sight software that can pixelize images into a computer mapped graphic environment in real-time that can be used within the droid programming to navigate and work.

Sub-systems monitoring its functions and connected into a local area network of droids (team members) the virtual reality environment can be set identical to the lunar surface.

The pixelization of images and landscape could be done earlier and then referenced to the droid via a mapping system, any changes to the landscape can be updated on site or remotely in the same way into the system and then updated through the sentinel droid providing the network interface.

We are talking about next generation droids that are development and can do alot of the functions and more. Some of the droids will have specialized functions such as crane / lifting duties, others are diggers / bulldozers etc. The assembly droids are their to put columns, beams, and bolts together.

I think you put more complex natures into the droids then they require and thus make a complex issue out of something simpler for the tasks associated with these droids.

Grypd · 2004-11-01 06:11:54

Thats correct Commstar, we don't have the technology. Computers must be able to have some understanding of the task in order to deal with unplanned issues, which they cannot do that yet, nor can robots coordinate their motion without a fixed, rigid reference point (concrete floor, not springy tires on soft dust & gravel).
You are missing an important issue with your virtual reality idea, is that what really happens is not what you predict it will in your virtual model. Trouble is, you don't know what really happens for about two or three seconds that it takes a signal to reach Earth and be processed, and you have to wait to tell robots on the Moon what to do for the signal to reach the Moon and be processed. In either event, you have unavoidable problem that it takes time for signals from Earth to reach the Moon and vice versa, which is not going to go away. Period.
Mapping out everything requires exact knowledge of where the cameras are I imagine so that probobly wouldn't work on the Moon. Even if it did, I don't think that the robot's motion could be coordinated to the cameras' location very well because of differences in terrain on the Moon versus on Earth at mission control on flat cement (or VR).

At first we will not be needing the robots to do any real complicated component construction. Time is not a real problem so we can set the robots to slow and methodical, its a good safety measure anyway. What is needed for the robots is to do the groundwork and beginning of Lunar harvesting of materials. We need the robots to place more solar cell banks and dig the cable trenches then infill them. We need the robots to dig out the trenches to put the first habs in, we need the robots to put the materials to the factories that they have placed to produce more solar cells and materials needed.

When the Human crews come to the base they will find a decent start and a place to stay ready for them. They will be guided in by the beacons that the robots have placed. All the time the robots are adding to the base and increasing its capability and if we find that we need the robots to do really complicated tasks that cannot be done by Earthside then we can use the Human crews to Telerobotic control the Robots. At a lot less degree of risk and a lot more improved work capacity in time and flexibility.

GCNRevenger I know you have difficulty in expecting anything to work but such a system can be fully tested with reasonable ease on the Earth including an artificial 4 second delay put into the system. It will be good practice to create such a system anyway if only to give the operators a good grounding in operations. We will use this system to work out our problem scenarios and to see what fixes are needed done.

SpaceNut · 2004-11-01 07:26:19

This is probably all a mute point (repair or broken droids) for once enough oxygen has been generated I would think we would have a manned presence on the moon and that any repairs could be done then. Yes not necessarily be man himself but possibly by robots at that time.

GCNRevenger · 2004-11-01 07:50:51

"Yes, you need to provide a reference to the droid to understand where it is"

Trouble is Commstar, you don't have this on the Moon to the high accuracy you need it. There are no reference points except the loose Lunar soil, which is uneven and would make setting up a stable platform extremely difficult if not impractical.

And, you can make your VR world snazzy until you run completly out of money, but the reality is that you cannot know what actually happens on the Moon until two seconds or so after it happens. So, you cannot do anything in real time, no force feedback, no seeing if your robot hand overshot the hole for the bolt, no nothing as it actually happens. Your VR world can try to make up for this by predicting what will happen, but no computer simulation can make up for this inherint physics problem.

I've got no problem with robot bulldozers, robot tractors, robot ore trucks, that sort of thing, because they don't need the extreme accuracy required for things like construction or repair. Plus, if humans are going to be on the Moon anyway, lets just use them for the detail work.

John Creighton · 2004-11-01 09:38:07

After some thought I think the best way to do it is this, a human maps out the action in a virtual reality world at the pace they would normally do it at. The virtual reality world would include virtual force feedback for such actions as gripping a glass. The simulation may not know exactly when the glass will break but a human should have a good idea of how tightly to grip it. Then the motion is mapped into a computer. Then each part of the motion is time scaled on the computer by how difficult each part of the motion is. For instance right before the robot has to grip something small the velocity of the motion is slowed down to a millimeter every 11 seconds. The system will have error correction that will identify the objects that are suppose to be near the appendages at each frame and adjust the motion accordingly. The operator can then watch the simulation and the updates will be fast enough given the speed of the manipulator to identify any potential problems. This way it may be possible for on operator to supervise multiple robots performing delicate tasks. It would be easy to set up an assembly line based on these principles. Moreover this has applications in industry because standerdized robots will perform a repetitive task with very little human supervision.

SpaceNut · 2004-11-01 09:54:46

Well it would appear that Nasa is spinning its wheels when it comes to planning future moon missions. The article references that the election results could influence the future of the space program. Like dah. we know that..

NASA's robotic moon mission spins wheels

John Creighton · 2004-11-01 10:09:17

NASA now plans to spend $5 billion between 2005 and 2020 to launch a dozen robotic missions to the moon, or one per year, beginning in 2008.
The idea is to have robots map the moon, search for water ice, survey potential landing sites, and test prototypes for oxygen production and electrical power plants, among other things.

From the above link. It looks like there is some consensus as to the value of producing oxygen on the moon.

SpaceNut · 2004-11-01 10:20:35

And in another article The Sober Realities of Manned Space Flight we get the broad picture of the vision and of Nasa not invented here bias on space cost.

The Sober Realities of Manned Space Flight

This is the picture of a federal agency immune to the competitive influences of the private sector. "You see this little metal loop? It's called a carabineer," California space entrepreneur Rick Tumlinson told a Senate hearing right after the President's announcement in January. "You could go to any sporting goods shop and buy it for $20. Yet NASA pays over $1,000 for the same object because of its procurement methods. It's the 'not-invented-here' mentality and distrust of the private sector that makes the cost of these projects so astronomical."
Each Space Shuttle (there are now three) must be virtually reconstructed after each flight. The process takes two months and 20,000 people. Some of the parts are so outdated that engineers troll eBay for replacements. "What we save on re-use, we throw away on maintenance of the aging fleet," says Alex Roland, professor of military history at Duke and former historian of NASA. The Shuttles are scheduled to be retired for good in 2010.

comstar03 · 2004-11-01 17:06:00

GCNRevenger,

I have never talked about the precision work unless it was under real-time human control that means humans on the lunar surface inside the lander controlling the droids outside.

Before all this, we need top have the infrastructure around the earth orbit and transfer vehicles from earth to moon for cargo and humans. Robotic Cranes and other site preparation vehicles can be controlled from earth ( heavy engineering work ), but wiring droids are preceision work and its best to functional in real-time thus onsite. All the work would be divided in categories that could be done via remote earth based robotic control and things that require real-time precision robotic control.

We could use the same process for mars, using site preparation and survey drodis first before the human -centric droids are send to work with humans to assembly a mars outpost. We need this practice to get the bugs worked out and the training personnel because humans can assembly a complete base within days and test the assembly without leaving their habitats.

But the Virutal environments are used to communicate and visual see what the droid sees and understands. To expand space for humanity we need to have extension systems such as probes, droids and other telepresence vehicles to explore and develop for humanity.

comstar03 · 2004-11-01 17:20:50

Also, GCNRevenger,

Repair Droids are different again,

The droid repair bay would have all the knowledge and scanning, repair and precision tools to automatically repair the droidss and test their work before sending the droid back to the field for more work.

and the repair bay would be able to disassemble the droid and change damage components and also re-assemble the droid into a fully-funtional droid.

You would require a repair bay facility with your team of droids as they land on the surface and commence work thus reduce loss of droids and loss of time and cost to send additional droids.

John Creighton · 2004-11-01 17:32:45

GCNRevenger,
I have never talked about the precision work unless it was under real-time human control that means humans on the lunar surface inside the lander controlling the droids outside.

Robotic technology must be pursued and will lower the cost of science, exploration and industrialization in space and on earth. It is clear that now we could have rovers to dig dirt to be analyzed and processed by a lunar oxygen producing machine. Initially these rovers will be built to be long lasting with redundant components and fault tolerance. Depending on when humans arrive on the moon robots that can repair other robots either by telleoperation or AI may arrive before during or after humans. These robots will greatly improve the scale of the future industry on the moon. The transition from telleoperatied robots to AI will be a smooth natural transition that will have applications on earth and in space.

SpaceNut · 2004-11-05 14:18:46

It appears that there is truely an indirect plan as piece mealed together form all planned mission into the future.

So far this is all I have found
Lunar-A - JAXA (Japan) Lunar Orbiter and Penetrator Mission (2005) http://nssdc.gsfc.nasa.gov/database/Mas … sc=LUNAR-A

SELENE - JAXA (Japan) Lunar Orbiter Mission (2006)
http://nssdc.gsfc.nasa.gov/database/Mas … ?sc=SELENE

http://nssdc.gsfc.nasa.gov/planetary/chronology.html

with the only Nasa mission being in a preliminary stage.
Moonrise - NASA New Frontiers sample return mission under study.

I wonder if this is the LRO for 2008.

India will launch its first lunar mission, Chandrayaan-1, in 2007 or 2008
http://www.chandrayaan-1.com/

China would like to send a probe to the Moon by 2006.
http://www.spacetoday.org/China/ChinaMoonflight.html

Also I believe there are a few thrown in from India as well.

Grypd · 2004-11-05 17:07:10

You are of course relying on information being shared between the countries.

It is interesting to note that both the Japanese and Chinese have plans to use robots to develop the Moon for commercial reasons. It is not a benefit for these countries to share this sort of info with other potential competitive nations.

Interesting to see both nations consider telerobotics as the means to develop the Moon. so maybe im not a complete Robotic Lunatic

New Mars Forums

Announcement

#76 2004-10-31 15:09:12

Re: The need for a Moon direct *2* - ...continue here.

#77 2004-10-31 15:46:28

Re: The need for a Moon direct *2* - ...continue here.

#78 2004-10-31 16:24:01

Re: The need for a Moon direct *2* - ...continue here.

#79 2004-10-31 16:55:21

Re: The need for a Moon direct *2* - ...continue here.

#80 2004-10-31 16:59:49

Re: The need for a Moon direct *2* - ...continue here.

#81 2004-10-31 17:52:29

Re: The need for a Moon direct *2* - ...continue here.

#82 2004-10-31 18:00:16

Re: The need for a Moon direct *2* - ...continue here.

#83 2004-10-31 18:18:12

Re: The need for a Moon direct *2* - ...continue here.

#84 2004-10-31 18:21:01

Re: The need for a Moon direct *2* - ...continue here.

#85 2004-10-31 19:39:14

Re: The need for a Moon direct *2* - ...continue here.

#86 2004-10-31 20:07:57

Re: The need for a Moon direct *2* - ...continue here.

#87 2004-10-31 21:00:53

Re: The need for a Moon direct *2* - ...continue here.

#88 2004-11-01 00:34:49

Re: The need for a Moon direct *2* - ...continue here.

#89 2004-11-01 06:11:54

Re: The need for a Moon direct *2* - ...continue here.

#90 2004-11-01 07:26:19

Re: The need for a Moon direct *2* - ...continue here.

#91 2004-11-01 07:50:51

Re: The need for a Moon direct *2* - ...continue here.

#92 2004-11-01 09:38:07

Re: The need for a Moon direct *2* - ...continue here.

#93 2004-11-01 09:54:46

Re: The need for a Moon direct *2* - ...continue here.

#94 2004-11-01 10:09:17

Re: The need for a Moon direct *2* - ...continue here.

#95 2004-11-01 10:20:35

Re: The need for a Moon direct *2* - ...continue here.

#96 2004-11-01 17:06:00

Re: The need for a Moon direct *2* - ...continue here.

#97 2004-11-01 17:20:50

Re: The need for a Moon direct *2* - ...continue here.

#98 2004-11-01 17:32:45

Re: The need for a Moon direct *2* - ...continue here.

#99 2004-11-05 14:18:46

Re: The need for a Moon direct *2* - ...continue here.

#100 2004-11-05 17:07:10

Re: The need for a Moon direct *2* - ...continue here.

Board footer

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.

Re: The need for a Moon direct 2 - ...continue here.