Wednesday, May 7, 2014

737NG Simulator Project

Update 8/30/2015: Some news in the progress, scroll all the way down...

By Mike Sherick


This letter is in response to Ian's reply to me, but I thought it best to copy you Thanos, as I wanted to communicate the same information to you as well.  I truly thank you both for your emails and objective comments.  

I fully understand that it is difficult for to comment specifically on the "general information" that I provided.  This is especially challenging since I am not a mechanical engineer and I do not have formal knowledge concerning the proper use of engineering terms to describe things - something I mentioned to you on the phone when we talked.  You have been kind to reply, and I respect the general nature of your comments, considering the limited set and somewhat disjointed of information I have provided.

Given that the nature of tying to build a full-scale 737 Sim Motion Platform is in and of itself a huge project, I am sure that most folks would pass this kind of project by, and rather spend their time playing golf or reading a good book.  All that is fine, but that has never been my personal mode of dealing with interesting challenges - like building  full-motion platform, or building my full-scale 737 Sim, which is now flying nicely.  Many of the projects I have done over the years have all turned out to be good learning experiences, and quite rewarding.  I expect that the 6dof Motion Platform falls into this category.

That is mainly why I would like to build a 6dof motion platform - to challenge myself and learn something new.  For almost all of the projects I tackle, I find myself having to learn and acquire new sets of technical knowledge and skills.  This certainly stretches my limited capabilities, but I have found it rewarding at times - driven by inspiration and dreams that in the end, were simply learning experiences.  With respect to the Motion Platform, I guess it is no different.  I simple see the possibilities, and like the range of complexity for this particular challenge - despite the so-called obstacles.  Anyway, I will take my best shot at all this with the information at hand, as it looks like I am going down the road where I will actually build a full size 6dof motion platform.

Regarding my decision to use a specific motor/gearbox...   I can use a 220 VAC Motor in the range of 1.5 to 2.0 HP because six of these represents the limit for the 220 VAC 50 A power circuit that is available in my Sim Room.  Since the NEMA mounting for these motor types is 140TC, each motor type will mate properly with the same Nord Gearbox, which can handle the extra HP.

The Gearbox unit seems to be the main issue for me at this time, as it needs adequate torque to do the job.  As you most likely know, each gearbox needs to be matched properly within the power range for the motor being used.  Likewise, the greater the motor HP, the larger and more expensive the gearbox becomes.  The a lightweight single seat 6dof motion platform, gthe cost factors are not too big of a concern for such a project.  But the cost considerations for the kind of motion platform required to handle the size and mass of a full -scale 737 Sim is a more serious issue - not to mention general safety concerns.  For the "home-builder" like me, cost is an issue, so this is one of the decision factors related to my proceeding with the Motion Platform Project.  In this regard, I have somewhat set the price limit on each gearbox at $2,000.  I believe I can get the performance I need from a gearbox in this price range for a 6dof motion platform.  I seemed to have found this with the Nord unit.

As you eluded to, the question becomes how much torque is required of the output shaft of each motor/gearbox in the 6dof platform system to accomplish the requirements for reliable Stewart Platform functions, supporting and moving 900Kg of mass.  Perhaps more specifically, how much additional torque is available to overcome the static loads, plus handle the additional load dynamics placed upon each motor/gearbox in the system.

In this regard, I have observed operation of the 6dof motion platform model I recently built.  It seems that almost always, at least two or more motors are moving the push rods in the same direction (or a vector thereof).  Many times, two or more motors are moving Up while at the same time two or more motors are pulling Down (all at different rates).  I suppose worse case, although I have not seen it yet, is with movement cues actuating only one push rod - essentially where only one motor/gearbox would be doing all the work.  

When I think of the mechanics related to the work to be done by the motor/gearboxes in a Stewart Platform system, and the work required for a single motor/gearbox unit in the 6dof system to lift the upper platform...  A simple static test can be done to help illustrate how much force is required by a single motor/gearbox unit that contributes to the movement of the upper platform.

Static test:  Consider lifting a 1 meter square box (cube) that has a mass of 30 Kg (CG is central, as in the case with the Stewart Platform).  If this mass is lifted from the edge of the middle of one side of the box (opposite side constrained, essentially tilting it upward), it only requires roughly 50% of the force that would be required to begin lifting the total mass of the 30 Kg box upward off the ground.  This is of course a bit over-simplified, as in reality, the various points about the Stewart Platform are somewhat constrained by Push Rods, and as the mass is lifted, it is also being "pushed horizontally" - and demonstrates a kind of "yaw effect" as it were.  

It is also understood that to accelerate the mass (our box) upward to some terminal velocity, it will require more force to overcome inertia of the mass  in order to get it moving in the desired direction - either in a static state or dynamic condition, as will be experienced in typical Stewart Platform operations.  In conclusion, it seems that "combinations of cues" compliment each other, thereby reducing torque demands on the motor/gearboxes - ALL pushing UP, sharing the load, with some pushing Up, while others pull Down - and combinations of such movements.  Again, worse case seems to be a single motor/gearbox having to do ALL the work - though I suspect that this never occurs.

You most certainly will be able to comment on this with more authority than my simple mind can comprehend, but this is from my lay-person's perspective. *:) happy  I greatly appreciate your on all this.

Although this is a very simple static example, it does speak somewhat to the static torque requirement, and does give a sense as to the torque requirement needed to handle additional dynamic loads that essentially become part of moving the mass, eg. acceleration issues and inertial loading.  Since in the case where my 737 Sim has an approx mass of 900 Kg, it seems that a single motor/gearbox only needs to act on (lift) 450 Kg of static load.   In using a 1.5 HP/Gearbox combination that produces 980 Kg of lifting torque (assuming using a 4" Lever to act on each Push Rod), then there is approx. 530 Kg of excess torque available to manage the various dynamic loads induced onto the Stewart Platform by a given amount of mass at a given CG (the Sim), by by factors such as start/stop, acceleration, etc. - in addition to duty cycle considerations for such activity.

I know that this is over-simplified, but I believe that although the 1.5 HP Motor/Gearbox mentioned above may be "over-kill", it seems capable of doing the job, without mechanically or thermally stressing the system components or assemblies.  In addition, if the torque is not enough, the same model of Nord Gearbox will also be able to accommodate a 2 HP motor to yield 1,363 Kg of static lifting force at the Push Rod using a 4" Lever.  Also, I will order this gearbox with high capacity heavy duty bearings at the Output Shaft that will better able to stand the overhung loads and force vectors commonly applied to the Output Shaft with the Stewart Platform.

Having stated the above...   I think I will proceed and plan on using the Nord 11 RPM Gearbox, that has an output shaft torque of 8600 in.lbs.  This gearbox will accept either the 1.5 HP or 2.0 HP 220 VAC Motor types for use with the VFD approach that Thanos' AMC1280 supports.  I will purchase a single gearbox this week and do some initial testing with the 1.5 HP motor under the full control of FSX/motion software/motion hardware.  I will initially use a 100 ohm 300 Watt Braking Resistor for initial testing.  I will make a fixture to test this combination under various dead and live weight loads, with various Lever lengths to determine response time, and torque/thermal limitations - if any.  In effect, this represents a bit of empirical testing, but I think it may be the best approach for me at this phase of the project exploration.  I'll keep you posted on status and results.

Regarding the acceleration values I had previously indicated...  You are right that what I specified was a velocity.  Since Thanos calls for the fastest acceleration time from the VFD, the 5(plus) in/sec rate should accomplish what is required to give a sense of motion.  What I am not sure of is whether or not there will be an issue with the short acceleration time demanded of the VFD to actually reverse the full mass required to move for a particular set of cues.  This equates to a "high duty cycle".  A problem here could result in overheating of the VFD and/or its Braking Resistor, or even the motor.  I am not too concerned about stalling of the system, as the VFD will deliver 150% of torque for the fast response times needed.   To balance all this, I think I can get away with reducing the magnitude of cue parameters in the 6dof motion software as needed to accomplish the required amount of movement, yet still induce a "real sense of motion" in the Pilot(s) in the cockpit flying the 737 flight model.  

On an encouraging note...  I should say that some of my hope for this is based on and related to an existing full-scale 737-800 Sim (virtually exactly the same as mine, but a bit heavier) that is being used by a friend.  His motion platform is a commercial servo-driven 3dof system, central spring, hinged back.  The interesting thing I observe is, that the mass of his Sim is a bit more than what I will have, yet the motor/gearboxes (right angle) are slightly smaller than the motor/gearbox units I am considering - although I don't know the exact output torque of his setup.   He has had issues with overheating.  But that was due to a CG issues, which he has now corrected, and his motion platform is now working nicely.  Anyway, that to me represents a bit of "real-world" comparison that helps to demonstrate that I am not too far off in considering the motor/gearbox mentioned above for my 6dof Motion Platform system.

Since I last built up the 1/8th scale motion platform model (mini-servo driven), I have added a 1/8th scale 737 Sim model to sit on top of the upper platform - complete with its displays.  Although the little model is made up of cardboard, it is to scale and is a fairly accurate representation of what this setup will look like.  The Nose Section is from FlightDeckSolutions, and its geometry shows the correct form-factor.  As you might appreciate, supporting the three 70" LCD monitors (each monitor weighs 90 lbs.) on a "flying bridge" will be a challenge to minimize/eliminate resonant vibrations that could set up in this kind of extended load.  In addition to typical rear mounting, I plan to "cradle" the bottom of the monitors to further help support their weight, as-well-as stabilizing them from the top to minimize unwanted oscillations.

The enclosed video gives a good sense of how all this might work in concert to yield a nice motion result.  Your comments on this video (below) would be helpful.  I don't yet have all the "cue parameters" setup properly, but you can get a good idea from watching the Sim model moving about on the Platform how it might "feel" being inside the cockpit.  Also, keep in mind that I will induce "flight vibration" into the Seat/controls sub-platform inside the cockpit, which will help in "fooling the senses" for things like ground operations, where I don't want to rattle the Sim to pieces.

Also, I don't know if operations with the mini-servos will be significantly different from that of controlling the VFD.  Something I will soon discover...  Comments?

In the video, when on final approach to landing, the aircraft is under AutoPilot control in the AutoLand function, so it is a fairly smooth ride down the glideslope.  In addition, I am using realtime weather, which happened to be clear skies and winds calm.  I will do another very short video soon to show motion platform behavior under gusting wind conditions.

Also notice in the video how FSX has some issues with exagerated movements, eg. take off heave, and autopilot engagement.  I will really need to sort out those anomalies, as the movements are quite abrupt.  Any thoughts on this?

In this regard, your comments on the turbulence that FSX produces is certainly what I am experiencing.  Playing around with the cue parameters will be required to get an overall balanced motion result to achieve the best flight sensations for the 737 flight model.

Relative to the above comments, I hope you have some time to provide additional comments that might be helpful at this point in my 6dof Motion Platform project.  Thanks.

Best Regards,
Mike Sherick

I wanted to share with you some initial testing of my 1/8th scale model of a Motion Platform that might support my full-scale 737-800 Sim.  Here is the video links:

Some detail on the overhead panel of the 737NG simulator

Updates on the progress will be placed below:

Motor Gearbox Initial Test (5/22/2014)

Yesterday, I started testing of the VFD/Motor/Gearbox under the control of your AMC1280usb controller, BFF 6dof motion software, and FSX.  Everything seems to function correctly.  It is very helpful to see the 6" Lever move in "real-time" according to what FSX is feeding the BFF 6dof motion software.  The speed of rotation and response seems adequate for the 737 flight model I will be using.  I should also point out that the BFF motion cue parameters for scaling and acceleration are quite subdued in the configuration.  More testing and "tweaking" of these cue parameters will be necessary as I continue testing.

An interesting aspect of the Nord Gearbox is the fact that they use a Nylon spline coupler to mechanically interface the Motor's output shaft to the input side of the Gearbox.  This is a somewhat loose fit, and it tends to make a "clicking sound" as the Motor is commanded to start/stop - fwd/rev.  Not a big deal, as it is not a loud noise, and the engineers at Nord claim that it is capable of handling the full output torque of the Gearbox, which is 11,000 of torque.

Another noticeable thing is the noise that the WJ200 fan makes.  I expect that this can be reduced somewhat when the WJ200 units are placed in an electrical box with external fans helping to cool components.  Of course, nothing is getting hot at this point in the testing due to the fact that there is no load on the system.

In any case, this initial testing of the electrical/hardware/software interconnect confirms that things are interfaced and working correctly with FSX.  The more critical phase of testing will come next, as I place the Motor/Gearbox in a fixture that will allow testing this system under full load.  I plan to build a fixture in the coming weeks that will emulate mechanically the motion platform layout, with full-scale Push-Rods connected to the steel Lever on the Gearbox.  This test fixture will allow dead weight to be placed directly on top of the Motor/Gearbox actuator, which will provide the "load" for this system - essentially moving the load Up/Down as it would as part of a full-motion platform system with six actuators.  I will add weight 50 lbs. at a time, up to 700 lbs. to determine system capability.  In actual practice, I expect that a single Motor/Gearbox should be able to manage 500 lbs. given the dynamics of the 6dof motion platform in actual use.

Here is the video, which has descriptions along the way, with Lever movements under various weather/flight conditions:

6dof MP 1st Motor Gearbox Test (5/27/2014)

Update 7/25/2014
Proper J Rail system for the pilots seat!!! Necessary to be able to get in and out the enclosed cockpit... :)

Update 11/24/2014

I was pretty excited to see Steve's new video of his race car Sim with his 6dof motion platform working so well.  Steve has done an awesome job with hi project - very professionally done!  I have spoken with him several times on the WJ-200 issues, and he has helped a great deal in clearing some of the setup confusion with the "Auto Tune" function.   Steve seems like a great guy, and his motion platform project speaks highly of his skill and determination.

Thanks again for working on getting my two AMC1280usb Controllers constructed.   I hope the parts you ordered have come in so the build and testing can continue.

A quick update on my 6dof motion platform project...   I have received my six (6) WJ-200 2.2KW VFDs, and I have also received five (5) more 1.5HP 3-Phase BlackMax motors.  I installed the new 2.2KW VFD and did a new Auto Tune setup along with tweaking some other WJ-200 setup functions.   The initial test results using FSX, BFF 6dof motion software, with your AMC1280usb controller are very encouraging.  The system is now performing very well and is highly responsive, including being able to achieve low acceleration/deceleration times - all without error.

What I do realize now is...  That the Auto Tune function in the WJ-200 is very important in order to achieve best performance.  In the Sensorless Vector drive mode, the WJ200 is capable of providing "up to" 200% output current to the motor.  This allow the motor to respond more quickly to the heavy demand of the AMC1280usb controller in response to the motion cues provided by the BFF 6dof motion software.

In testing this new configuration last night...   I found that short acceleration/deceleration times (although now possible), are not necessarily better for my requirement with the 737-800 flight model in accomplishing highly responsive actuator movement as commanded by the motion cues.  In addition, increasing the acceleration/deceleration times to 1 second seem to greatly reduce the current output by the WJ-200 to accomplish its work, yet produces almost the exact same response applied to the actuator.  I realize that for a race car Sim, the need for faster response times is needed, but for the 737 Sim motion requirement, slower acceleration/deceleration seems to be better.

In monitoring the WJ-200 display in the "Current mode Amps", it seems that this setup will draw typically less than 4 Amps, with short duration peaks up to 15 Amps. I can see this demand for current being greater under the greater loads applied trying to move 1,800 lbs. of 737 Sim on the Upper Platform.  As a result, I met with an electrician last night to add greater electrical power to my Sim Room.  I plan to have a new power circuit added that will provide 120 Amps @ 240 VAC single phase.  I expect that this will handle my power requirements with the six (6) WJ-200 VFDs/Motors for my 6dof motion platform.

My Grove gearboxes are on order, and should be delivered sometime in the next two weeks.  The 160:1 ratio will provide all the lifting capacity I need, yet testing shows that it is still very responsive to motion cues.  I can't wait to get all this setup for an initial no-load test of a light weight test platform arrangement.

Update 8/30/2015

It has been a while since I last reported on any significant status with my 6dof Motion Platform project.  I have been somewhat delayed on this project with other project around the house.  In any case, I have been making steady progress.   The photos below represent the current status.   This is the testing phase for the motion actuators to determine potential mechanical interference.  The Upper Platform shown in the photos is made from wood, so that if there is an interference, the wood structure will crack or break, instead of the expensive gearboxes.

The Actuator Levers are the final design, and are made from 1.50" thick steel plate, precision machined for attachment to the Gearbox Shaft and Rod Ends respectfully.   As you can see, I have added a 2.00" Spacer to the Lever for additional clearance, in addition to a "bevel" machined on the clamp part of the Lever to allow for more Actuator movement.

Statically, at the "Drive" neutral position of the motion platform, the height of the Upper Motion Platform is 32" above the floor.  Notice also, that the motor/gearboxes are mounted on wooden platforms, and are not anchored to the floor just yet.   This is to allow adjustment of the motor/gearbox units for best possible 6dof configuration.   Each Actuator has a 2.5 degree inward tilt toward the center of the Upper Platform, which is intended to allow for maximum clearance of the Lever with the Rod End/Actuator assembly.  Thus far, the configuration shown in the photos seems to be working nicely as interfaced with the 6dof BFF motion software and FSX in actual flights.

I am now in the process of cleaning up the main cabling from the VFD Controller box to the AMC1280usb box.  I hope to complete this wiring in a few days.   I plan on continuing this testing of this configuration over the next few weeks, which will also include some fine-tuning of the motion cues.  Once I am happy with these initial test results, I will make any necessary mechanical configuration modifications to the platform design.   After that, I will proceed to fabricate the Lower Platform anchor plates to be bolted to the concrete floor, and fabricate the Upper Platform structure - all this done in steel plate/tubing. 

Thanos, 2015

1 comment: