OO Planet - Development

[DavidBigcheeseplant]

I use AutoCAD everyday for my day job, Fusion 360 even though it is an Autodeck product, it works or looks. Although there are similar commands such as mirror array etc.

[What About The Bee]

Hi Rana

Quick note on the 32 bit. That would be version 0.18 which is available at Github, and the version I am using.

https://github.com/FreeCAD/FreeCAD/releases/tag/0.18.3

The tool at this level has been adequate to get OO Planet underway. I have yet to encounter an issue that prevents development.

OO Planet Update

I've broken through the Gordian Knot of the gear train, valve gear motion bits and motor arrangement. It must all fit in to a OO Planet shell and not interfere with itself. I have just finished installing the eccentric rods and the eccentrics into the gear train. The eccentric sheaves will be along shortly; with the motion on the front of the smokebox afterwards. My goodness I was chasing my tail, but the epiphany came yesterday. A bunch of hours of FreeCAD jockeying and a solution forward was proven. Wow that was tricky!

Bee

[What About The Bee]

Hello Big Cheese

Do yourself a favor and google "fusion 360 paywall issue".

The sense I get is that they can lock up your models behind their paywall, if they arbitrarily deem you to not be a "hobbyist". There are plenty of complaints online about this.

That cannot happen with FreeCAD, as everything resides on your computer and you absolutely control all your files.

I originally wanted Fusion 360 but had some issues. When I did some poking around to find the solution, I encountered chatter about the paywall issue. I was done. Bye Bye Fusion 360. I intend to invest quite a bit of time in my models and I am not going to be held for ransom

Bee

[DavidBigcheeseplant]

I have never had an issue with Fusion 360 as I am only using it for my own use, the paid for version has things I don't really need, although export of DXF files would be useful.

[What About The Bee]

I've successfully integrated the gears, eccentrics, eccentric rods and motor. It all fits inside the shell and clears all the axles.  

Before I begin, a quick review of my design objectives

1) I have the Hanazono Motor Bogie on hand. It has been tested with a 3D printed LMR tender, also on hand. It has fantastic tractive effort.  

2) My OO Planet must speed match the tender. The easiest way to do this under DC is to start with another identical Hanazono motor. A gear train was devised to match speeds, such that OO Planet is 3 parts per thousand faster than the tender. The additional Hanazono motor is on hand.  

2a) I've not given up on the belt concept, that idea has merit. For now, I will continue with the geared solution.

3) Make the visible valve gear be in motion. What is a steam engine with static valve gear? The valve gear must move!

4) Practical manufacturing considerations. In the first pass, manufacturing the components was only a second thought. Ultra fine parts look great, but cannot be made in an economic sense.

5) The look and feel of Stephenson's Planet. OO Planet will have some compromises, but these should not detract from the presentation.

The first part of the integration was quite simple. The motor was in the firebox, but it was simply too wide to let the valve gear pass between the rear wheels and the firebox. As such, the Hanazono motor was moved into the smoke box. To insure that the motor would fit, the shell of OO Planet has a 1mm wall thickness, and the motor was placed such that it does not interfere the upper boiler shell. The gears were re-arranged to fit.  

Next, the eccentrics were added to the crank axle. These provide proper quartering to the valve gear and work essentially as does Stephenson's Planet, albeit not on the main axle. The red axle has a 1:1 gear to the wheel axle. Thus, the valve gear and oscillating handles will absolutely be in time with the wheels. You can see the eccentrics in the following image.

The eccentric rods must clear under the motor. Those bends in the eccentric rods are to clear the front axle

The Walking Rods hold the Eccentric Rods up in front of the smoke box. The Walking Rods pivot on the Eccentric rods and will appear to walk when in motion. Scale Walking Rods connectors are simply too fine to be practically manufactured. My current concept is to etch copper plate 0.4 mm thick.

The red/white scale is 1 mm colored, to give a sense of size. The copper plate will be bent. The etched material accounts for the bend through the neutral axis. The axle shown will likely become a shoulder screw.

The arrangement of the Walking Rods and Eccentric Rods is shown. This is drawn properly quartered, with all relevant angles.

At the top of the smokebox, observe the blue connectors which are from phase 1 and compare to the new etched connectors. This mechanism on Stephenson's Planet permitted the Engineman to lift the eccentric rods, to essentially put Stephenson's Planet into neutral. The purple mechanism will be static, as there is no need for functionality.

I am very excited. This has moved from being a tentative design into a design that can be manufactured. Going forward, I will lay in the remainder of the valve gear, to include the oscillating handles on the footplate.  

Drawbacks

A) you can see the motor under the boiler. This will be in a black shell. I am mindful that most of this is below the level of the sandwich frames. It is unfortunate, but speed matching to the Tender, under DC, requires the same Hanazono motor. It may be a tiny motor, but so is OO Planet!

B) The connectors are ridiculously huge when inversely scaled. In fact, most of the valve gear will be as well. Yet if I was to scale a 1" rod, I end up with 0.013" (0.333 mm).  I need to be able to attach parts with a rod or shoulder screw, this requires some compromise. I will have to live with it, or spend a fortune manufacturing tiny non-functional parts.

Onwards 

Bee

[What About The Bee]

I left off with the eccentric rods integrated with the gear train. The eccentric rods thrust forwards and pull backwards. This force acts on a lever which is mounted to an axle at the front of OO Planet.

Shown is a fuscia colored axle near the eccentric rods. Each side of the valve gear functions independently. All of the timing is integrated at the eccentrics, shown in an earlier post.

Torque

The eccentric rod linear motion is converted to rotational motion. Naturally this creates torque and for OO Planet, several dilemmas.

Firstly, the rotation will be impeded by friction. If the bearing surfaces are rough, high friction will be encountered. Possible breakage.

Secondly, torque and lever length can create high forces. The lever driven by the eccentric rod will be long compared to the radius of the shaft. Possible breakage.

Thirdly, gripping a shaft, without slipping, is non trivial. All of these parts are tiny! That shaft you see up above is ~¼" (actually 6.625 mm) long. 

Fourthly There will be three levers on that shaft. One to be driven by the eccentric rods. One to operate the slide valve over the piston. One to cause an oscillating lever on the footplate. In ¼".

Manufacturing

The dilemmas posed preclude 3D printed plastics. The dot size of the printer (0.4mm) compared to the feature size of the 3 levers is simply too large. Plastic layers can shear apart under load.  

Metal seems like a good choice. It can be smooth, lowering friction. Metal will likely not shear apart under load. But how to create tiny metal parts? The parts could be cast, but how to create the tiny mold?

Copper can be etched, but the thickest practical sheet is 0.5mm. 0.5 mm by itself is quite thin.

Solution Set

All parts are metal. The screws are 0000-160 × ¹/₁₆", commercially available. Some will be used as set screws, others will serve as the underlayment for a sleeved axle. 

There are three levers on the shaft.  

I begin with the lever that is driven by the eccentric rod.

[edit: the arrow should point to the other notch on the other eccentric rod. Whoops!!]

It consists of three parts. The outer sheet is etched and folded into a U shape. The two inner sheets will be tapped (commercially available taps 0000-160) to accept the screw. All will be 0.3 mm sheets, etched. The bottom screw will engage the notch on the eccentric rod. 

Note that the U is sized such that, when the axle is in place, the two inner layers are tightly bound, because the axle holes, around the U, are positioned to tight tolerance. Each U assembly will be epoxy glued on a temporary shaft and will likely require reaming. Tapping to occur post reaming.

In the middle is the scottish yoke. The lever consists of two etched inner sheets, tapped, held by a U surround.

The outer lever will drive the control rods, which in turn will drive the oscillating levers in the footplate. Another U shape surrounding two inner flat sheets, configured appropriately. More etched copper for the control rods. Note the sleeve in this case, over the screw. This lesson will be back fitted onto the lever driven by the eccentric rods when imported into the main model.

Onwards!

Bee

[What About The Bee]

A Progress Update

The left hand side of the Stephenson Valve Gear is in place. All relevant moving parts are now installed in the model. With the exception of the control rods (in green), the moving parts are copper colored, representative of the material they will be made from. The control rods will also be copper, but I wanted to emphasize the control rods, as they caused this re-work

The 0000-160 screws are all in place, with the appropriate sized holes for tapping.

The handles are finally on the model, the inspiration for this entire process. These curved handles are located on the valve gear axle, above where the footplate will be, and are shown properly quartered The left hand handle is drawn at mid-phase, whilst the right hand handle is drawn at the furthest forward point of rotation.

Those pesky green control rods now clear the gap between the firebox and the wheels!  You can observe the control rods behind the wheels and in front of OO Planet's shell. These are connected to the control levers in front and back. As the front Valve Gear axle rotates, it drives the front control levers. These simultaneously push on one control rod while pulling on the other. This will cause the control lever on the Valve Gear axle above the footplate to rotate. In turn, the causes the curved handle to wiggle back and forth.

Note the uneven length of the footplate Valve Gear axles. This places both handles to the left hand side of Planet's footplate, where the OO LMR Engineman could use use them.  

Plan Going Forward

Get the right hand side of the Valve Gear installed. This is fundamentally a mirror image of the left hand side, but needs to be drawn. It is a tedious process of grinding out the parts, all of the inventive nature is complete.

Design the bracketry to support the Valve Gear. All of the moving bits just float in space right now. I needed to understand the requirements of the moving bits before designing the static bits of the Valve Gear. With the moving parts in situ, the static parts fit around them. As per usual, I will overthink every detail!

Is there any doubt?

I started this process with the thought that it would be a lengthy process. That may have been a ridiculous under estimate!  Of course, for a first time effort, I expected to make mistakes and create re-work. But that has not been the major consumer of time. If this is what Hornby goes through, is there any doubt now why new products take so gosh darn long? I need only please myself. Hornby is in the unenviable position of trying to please everybody. Each detail will be flogged to death by the crowd, causing Hornby mountains of work in an effort to get everything just so. Is there any doubt?

Bee

[threelink]

Hi Bee. I am in awe of the detail work going into your Planet. Given the minute size of some of the components I do not envy you the task of producing and assembling them!

[What About The Bee]

Hi Three Link

Those 0000-160 screws terrify me. Once started they work like any other screw. It is the getting them started part that is insanely difficult.

Thank heavens I only have approximately 30 of them in the valve gear, at latest count. I shall ask you to prepare my straight jacket in anticipation of the event. Please and Thank you!

Bee

[What About The Bee]

I left off with the left side valve gear in place. I indicated, at that time, that the right side valve gear should be relatively easy, as it was a mirror image. My goodness was I wrong!

The left side valve gear was drawn at top dead center. This meant that the right side valve gear, which is quartered, is 90° out of phase to the left side valve gear. In and of itself, drawing the right side meant copying the left side dimensions, albeit in a new orientation. How could that be hard?

Unfortunately, the quartering uncovered a major interference within the valve gear mechanism. If I was to attempt to run that valve gear, it would bind the system. The motor would stop rotating, unable to force the valve gear forward, likely burning it out.

The interference was between the eccentric rod follower and the eccentric rod.  

After checking the dimensions several times, I realized the dimensions were correct. The root cause of the problem lay elsewhere.  

At the front of Planet, the eccentric rods are connected to the walking rods and to the eccentric rod followers. As the eccentric rods thrust back and forth, the front of the eccentric rod is constrained in motion to an arc, the radius of which is the length of the walking rod; the length of arc defined by the eccentric rod stroke. Similarly the end of the eccentric follower, is also constrained to an arc, the radius of which is the length of the eccentric follower, the arc length defined by the self same eccentric rod stroke. These two arcs are at a fixed distance along the eccentric rod and indicate, for a given thrust of the eccentric rod, the vertical rise of each arc. Those rises are different!! Root cause found. 

While this may be obvious in retrospect, that took quite a bit of thinking and studying to realize.

As can be observed, the ends of the arc for the walking rod do not rise as high as the ends of the arc of the eccentric follower, for a given eccentric rod linear travel.

A parallelogram can freely rotate, but a trapezoid is locked. The control rods on the side of OO Planet form a parallelogram. The walking rod, eccentric rod, eccentric follower and smokebox face form a trapezoid, hidden in plain view.

The solution is to elongate the slot in which the eccentric follower rides, permitting compliance.  

So why did I not see this? Armengaud shows us many things, but obscure details are not brought forward with reasons and explanations. They are left as an exercise for the reader to discover. This was one of those exercises.  

Bi-Stable Mechanism

I added all of the clutch mechanism, shown in green. 

 I have no intention of making the clutch functional. It will be part of the static decoration of OO Planet, an integral component of appearance. Once again, they will be etched copper parts. To provide access to OO Planet internals, the clutch can be taken off, via those tiny 0000-160 screws.

The clutch control handles (2) are on top of the firebox: a left side and right side handle. The long control rods, on top of the boiler, connect through a pair of levers to the walking rods. When the clutch handles are drawn towards the engineman, the walking rods are lifted, which lifts the eccentric rods.  This disengages the eccentric rods from the eccentric followers, in turn freeing the engineman to manipulate the curved handles and thus manually control the slide valves.

The length of the eccentric rod forward of the eccentics are held up by the walking rods with the entire weight of the walking rods sitting on the very front of the eccentric rods. That is a significant mass to keep elevated.

Did the engineman have to hold the clutch levers up? Didn't all that weight draw them back down?

No!

In the forward position of the clutch handles, the weight is supported on the eccentric follower. Indeed, the weight locks the eccentric rod onto the eccentric follower. It is stable in this position.

When the clutch handle rotates towards the footplate, that curve in the clutch rod comes into play. It brings the near end of the clutch rod beyond dead center. Therefore, the pull of the weight wants to draw the control rod through the axle of the clutch handle. It cannot, of course.

Therefore, the clutch is stable in the up position as well as the down position. The engineman need not hold the clutch handle. Elsewhere in travel, the clutch handle is unstable, and will be drawn by the weight forward until the eccentric rod rests on the follower. The eccentric rod will engage the pin of the eccentric follower, becoming stable.

A mechanism that is stable in two positions, but no other, is called a bi-stable mechanism.  

Stephenson was well aware of a bi-stable mechanism. We can see this clutch rod curve in his drawing.

On the shoulders of giants.

Bee

Some images of the current state of the CAD, OO Planet

Close to the front

The green clutch control rods pull on the green levers, which lift the copper colored walking rods up. There are two independent controls, the left (fuscia) axle and right (blue) axles work independently. The clutch will be decorative (non-functional) for OO Planet

The lower portion of the smokebox face is filled with valve gear components. This image attempts to annotate my nomenclature.

[EDIT: Hysterically, I managed to leave out (D) Walking Rods.]

It should be noted that, once again, there are two halves, fuscia left, blue right.

The view from the footplate shows the four control levers. 

[threelink]

Wow again, Bee. A modelling tour de force if ever I saw one. I presume that on the prototype the walking rods were removeable for access to the smokebox, and the eccentric assembly for access to the valves and cylinders. The shed staff must have hated Stephenson!

[What About The Bee]

Hi Three Link

There are a few things to consider in your questions, which frankly, I do not know the answers to.

1) OO Planet has comparatively huge, clunky rods and levers when compared to the prototype. The prototype, as drawn by Armengaud, has comparatively much finer details. Certainly beyond the modeling realm with rod diameters approaching zero if properly scaled. There would be much better access on the prototype than on OO Planet. Look at the walking rods on Stephenson's drawing for an example.

2) The smoke box door is just a plate and is not hinged. I do suppose they could remove the nuts/bolts and slide that plate laterally off of Planet. Access would be permitted to internals without removing the walking rods.

3) All of the fixings are indeed bolted and threaded. So yes, the walking rods could be readily removed.

What surprised me about many of these fixings was the mixture of square nuts and hexagonal nuts. I truly expected 100% square nuts. Yet the Armengaud drawings show a mix.

4) About the shed staff. I don't know if you ever worked for a true genius. I did. The man had over 50 (fifty) US patents, predominantly in electric motors. You could discuss engineering points with him, but in the end if he wanted it a certain way, that was the way it was going to be. Even if it meant I was going to struggle with Manufacturing, who had choice words about practicality. Stephenson, in my eyes, was the locomotive genius of his day. The shed staff must have been in awe of their genius, as we were of ours.

Improvements in serviceability would come. It is indeed very hard to accurately predict what is required to service a novel mechanism.

Bee

[What About The Bee]

Development Continues

I've added it quite a number of controls and details. Steam Regulator. Sight glass. Firebox door. Clinker Cleanout door. Blowdown valves. A Timothy Hackworth Safety Valve and a Salter Safety valve, as per my research, detailed in my post "Under Pressure". Man way into boiler, so it can be riveted together.

The Steam Dome was cast and therefore had details unavailable to period machining.

Bee

[wairoalbany]

Looks brilliant. Well done so far.

[What About The Bee]

Thanks Wairo!

Major update coming...

[LTSR_NSE]

I do wish (but don’t hold out a lot of hope) that someone in Hornby marketing reads this thread, and sees just how fantastic their Engine Shed blog (or newsletter) could be, if they gave regular updates on progress of all models in varying stages of production.

@bee please take this as a massive compliment of your excellent & fascinating thread. 👍

[What About The Bee]

Hi LT&SR_NSE

A journey of 1000 miles starts with one step. I'm on my way, but heavens, this is challenging.

Thank you for the encouragement! Hopefully, all this work yields something worthwhile.

The next update has something exciting and spectacular. Do not miss it! 🙂

Bee

[What About The Bee]

I am making an effort to keep the word descriptions smaller and use more illustrations. In doing so, I recognize that there are details which I do not discuss and may slip by the attentive reader. If there is something you’d like to investigate or ask about, please feel free to do so.  

Stephenson’s Planet

Period drawings of Stephenson’s Planet show the support structure. The wheels sit on the track. The axles are supported by the wheels. The axles ride in horn blocks, which in turn ride in the horn guides (yellow). Between the axle horn blocks and the sandwich frame are springs. The mass of the remainder of Planet is fully supported by those leaf springs and the sandwich frame constitutes the base to which everything else is attached. 

Whilst the footplate is bolted directly to the sandwich frames, diagonal struts support the boiler, the smoke box and firebox.  

OO Planet

There are differences. Aside from the obvious size difference, OO Planet is supported and driven differently. An electric motor drives a gear train. The load of OO Planet is transferred to the axles via axle bearings. The exact scheme is shown in the following image. 

To clarify the rotating end of axle, this sketch is marked up

Many months ago, I sketched a gear train based upon a mathematical analysis. The analysis considered the Hanazono motor providing 1 unit rotation. OO Planet’s tender is driven by a Hanazono motor bogie.  When power is applied, both OO Planet and Tender should run at the same velocity. I sought and found the gears that will cause OO Planet to run at the same velocity as the tender, for that same unit motor rotation; even though OO Planet’s drive wheels are of a different size. The sketched gear train was found to be feasible and it was left as a feasibility sketch, as I moved on.

FULL CIRCLE

While I played around in Freecad, learning the tool, I created many parts and components. But in the background was looming a difficult task. The chassis and gearbox integration was lurking, waiting to pounce. It was always there, haunting me. I finally had to slay the dragon.

Here, I’ve placed the OO axle bearings (red) on OO Planet’s rear axle.  The very start of the chassis is shown in blue.

The first dilemma is how far apart the OO axle bearing flange faces must be. Instead of performing a mathematical analysis, as detailed by the Scale Four guys, I simply drew a second radius curve of two rails and placed the axles and wheels on that track. 

It should surprise no one, lateral play in axle mounting will NOT be required for a 2-2-0. Perhaps when I get to 2-2-2 Patentee, or any of the Bird class where a central axle requires lateral play to permit the locomotive to get around tight layout curves but for now, no excess lateral play is needed. Simple clearance is all that is necessary. That is, the flanges should not bind the wheels. The Scale Four guys claimed that simple clearance of 0.0635mm total, inclusive of both sides, is all that is needed. Thanks very much but no! That is just a bit too tight a tolerance for my preferences and practical machine tool experience. I selected 0.1mm per side or 0.2mm total. That is still less than I measure on my Hornby Rocket but should easily prove sufficient.

With that dilemma resolved, I examined the clearances of the gears to the axle supports, should they have some sort of chassis to hold them in place. Needless to say, there was conflict. Sigh. It was finally time to revisit the gear box. Thankfully, this time around, I have a better grasp of Freecad.

GEARBOX REQUIREMENTS

Any two meshing gears must be held at a specific distance apart. Too close and the gears bind, inhibiting rotation. Too far apart and gear lash occurs. Gears do not float in space, rather, they are mounted to axles. Each axle must be constrained vertically, longitudinally and laterally with respect to OO Planet.  These requirements must all be simultaneously met, without interference with any other components.  

GEARBOX DEVELOPMENT

I begin with vertical planes. These will constitute the beginning or end of an axle, or the lateral beginning of all gears that mesh.

Illustrative only, far more vertical planes exist!

From the planes, gears were placed and accurately sized to fit within the allowable space. The axles also begin or end at the respective planes, now turned off so as to see the gears

To constrain the axles vertically and longitudinally, plates (blue) with through holes are placed in the gearbox, again with the aid of the invisible vertical planes. Note the clearance around each axle in those plates, permitting simultaneous free rotation. With the blue plates in place, both ends of all axles are supported and the radial distance between gears is fixed.

Additional plates are added (green) to constrain lateral motion of the axles.

A top view of the gear box, with plates, axles and gears is presented.  It is very tight, but indeed, it all fits. 

The gearbox designed meets all requirements so far, except one: interference with other components. I will table that until later in this discussion.  

ECCENTRICS

Up until now, I have not shown the interface between the eccentrics and gear box. Here they are!

 The next image shows the eccentric endcap removed; the keyed eccentric disk and the eccentric rods extending forward. The hole is for an axle shown above. Note there is a gear on this axle driving a gear on the rear wheel axle, see above, with a one to one ratio. Thus the eccentric motion is mechanically bound to wheel rotation, meaning the “chuffs per rotation” or in my case, “handle wiggles per rotation” is matched perfectly to velocity. Yes, it is properly quartered!

VALVE GEAR

I’ve made every effort to make Stephenson’s valve gear function on OO Planet. I’ve written about that extensively, there won’t be repetition. Yet here is the valve gear with the eccentrics. The chassis will need to support all of this kit with the body off.

The motor must also be supported and constrained by the chassis

The exact separation and mesh between the Hanazono worm and worm wheel is still being investigated, yet you may still see that mesh. Once fully defined, all internal objects will be on location, permitting the chassis to be resolved.

WHAT ABOUT INTERFERENCES?

The critical interference check is the body shell vs the gear box.  As such, the gear box is shown top center, with three views of the gear box with the shell on, less any decoration that may distract the eye.  From this, we can observe it fits, without interference!! 

DECORATED

All the other tiny bits and bobs are now added on. OO Planet will have all of the requisite controls.  Its details that make a model pop. If you look towards the back, under OO Planet, some gear faces are still peeking out. The sandwich frame and accoutrements shown in the first image are still missing. They will go a long way to hiding the mechanism still visible under the boiler. Yet, in the end, there will remain some peeking through. This is the price to pay for speed matching the tender and the great tractive effort already demonstrated for that tender.  In essence, OO Planet need only drive itself and wiggle its handles. Any excess will simply contribute to the tender’s tractive effort.

The gearbox was a difficult challenge, taking most of this update period to get right. While this write up may take you a few minutes to review, it took hour upon hour upon hour of struggle to make everything fit and work properly. It is virtually 100% parameterized, so in the event something needs to shift, everything will move programmatically with that shift, including plates, axles and gears.  And shift it may, as the mesh of the Hanazono worm and worm wheel still eludes me.

Bee

[Fazy]

why not just go for a smaller motor, you could also dump the gearbox go for a micro motor with a pulley with a band drive, others have done such when driving small models.

[What About The Bee]

Hello Fazy

Thank you for your excellent suggestion.

The short answer is "speed matching". I have a 3rd Generation LMR Tender which is motorized with a Hanazono motor bogie. Testing showed wonderful tractive effort with that motor boggie, with well over 20 pieces of rolling stock easily hauled. I posted about that earlier this year. OO Planet should speed match the Tender. As a comparison, Hornby's Rocket cannot handle anywhere close to 10 pieces of rolling stock. In real life, Stephenson's Planet was recorded pulling a consist of 20 wagons, the tonnage hauled and the time between Liverpool and Manchester.¹

The longer answer of speed matching must account for how a DC motor operates and the velocity it provides. The urging force of a DC motor is the magnetic field of the coils attracted to the magnetic field of the rare earth magnets. The magnetic field of the coils is a function of the wire turns and how densely the turns are packed. The magnetic field of the rare earth magnet is a function of the density of the neodymium product. The primary retarding forces are the inertial mass of the rotor as it relates to acceleration and the frictional forces of the rotor stator arrangement. There are other forces, of course, like the back EMF & etc, but we just need consider the primary forces. The torque produced is generally stated as a torque constant (torque/amp) but even this is must consider the temperature of the coils, higher temps change the resistance, which reduces the torque. Even by this highly simplified paragraph, we can observe that for a given DC voltage and current, some arbitrary motor will NOT have the same rotational velocity as the Hanazono motor. You can see this on your DC layout any time you please. Plunk two locomotives on one track and turn up the controller. You know as well as I do, they will not run at the same velocity.

So if I wish to have any chance at speed matching the tender, OO Planet must use the same motor. Even that is fraught with speed mismatch. The two Hanazono motors will be very similar, but not identical. For example, the gap between the two fields affects the force produced, small changes in gap produce large changes in force (1/r²). Yet the same type of motor will still be far superior to an arbitrary motor, if I wish to speed match.

I recognize your suggestion as an excellent one, and it has been recommended to me before. It is a valid strategy and one I may pursue should the gearbox prove outrageously expensive or not feasible. I would still need to use the Hanazono, but I would substitute the band for the gears.

Bee

¹ I checked the reference. "Penny Magazine" 1833. The locomotive: Planet. The number of wagons: 18 (whoops, my memory failed me!) The tonnage: 80 tons. The time between Liverpool and Manchester, 2hrs 54mins. [including two stops for water/fuel at 5 minutes each]. The distance: 31 miles. This works out to just over 10 mph average speed. 2 mph over the limit for luggage!!

[What About The Bee]

ADDENDUM

With the gearbox nearly complete, I indicated I needed to define the separation distance between the worm wheel and the worm gear that is on the end of the Hanazono motor.

I start with the worm gear on the end of the motor. I measured the major diameter and pitch and sketched those in FreeCAD. Was it correct? How to check that?

I used the optical comparator trick, commonly used in Incoming Inspection. Use the image of the actual worm gear and optically compare it to the image from CAD of the same part.

Here, the result of successive iterations has occurred. Pressure angle is 20° and the pitch is indeed 1 mm.

Again, successive trial an error yields a precise length, 4.5 mm.

Confirmation. With the CAD image partially overlapping, it looks very nearly perfect!

With the characteristics known, it is merely an exercise to create the worm wheel in CAD. The only tricky bit is that the teeth aren't straight across, they are at an angle that matches the pitch advancement of the worm.

The worm wheel gear and the worm gear are non enveloped. This will reduce my cost and further, the efficiency simply isn't required.

PROBLEM SOLVED

The separation distance between the axles is now defined, the motor placed firmly in model, the gear box adjusted and I will be moving on to the chassis.

Bee

[What About The Bee]

The Chassis Conundrum

The chassis is the component that ties all the floating bits together. The motor will be held on position to the gear. The gearbox will be supported and made rigid, transferring torque to the wheels. In the instance of OO Planet, it must support all the valve gear without the body shell on. Additionally, the chassis must not interfere with any parts, particularly rotating gears. Without further ado, here is my current solution. Naturally, it is subject to change and adaptation as I go along.

 

Of note are all the notches, openings and cutouts. I was forced to add and cut out various bits and bobs as each feature was added. This is to be one cast metal part (hopefully). I took my clues from the various Era 1 Hornby models I possess.

Here are the bearings, bearing supports and wheels, to give you a sense of orientatation.

 

The gear box was far more challenging. There are four main vertical plates which support several axles. The center plates have a tab which notch into the chassis. Each main plate has subordinate cap plates which restrict the lateral motion of the axles. Note the presence of 0000-160 screws. These are provided to permit assembly and disassembly, but mind, the screws must not interfere with any gears. I must also be able to actually assemble the gearbox, no use designing something that cannot be put together!  

 

Add in the gears

 

Add the outer plates. Note the outer plates have screws affixing the plates to the chassis

 

Add a top plate and some 00-90 screws. The gearbox is now rigidly attached to the chassis. note the blue tabs from the main gearbox plates slotted up and into the top plate.

The motor nestles into the chassis, with a notch in the top plate such that the worm gear clears as the motor goes straight down

 

I was hopeful of being able to get flashes of the eccentric rods working under the locomotive. I sculpted away quite a bit of material, but I think I am safe

 

Whilst the valve gear isn’t yet attached, I do hope you can see the reason for the fore and aft vertical plates

 

Of course, all of this is useless if it doesn’t fit in a body shell. All 290 components turned on. The vertical plates will take the place of the body shell in those regions, so right now you can see both blue and grey at those locations.  

 

The reader should understand that I did not start out with a shape in mind. Rather, the shape evolved as requirements were met. This was a major challenge, now sorted…I think. A few tweaks perhaps!

Bee

[threelink]

Stunning work, Bee.

[What About The Bee]

Thank you Three Link.

I must admit, this is not my first locomotive. Here is an image of some others I have worked on. It is an unusual scale for the enthusiast!

What scale is that? 1 😉

In the foreground, 4 Evolution AC4400 Locomotives, ready for shipment to the client. While they freewheeled to the client, 4 would provide ~670,000 lbf of continuous tractive effort. Up to 7 locomotives can be distributed throughout any one consist.

Extra points if you can identify the locomotive in the background and why it is mounted on a flat car!

Bee

[What About The Bee]

Welcome back to another OO Planet update.  Since the time of the last update, I have focused on two areas. The first area is the motor itself, and the second area is the support for the valve gear.

MOTOR

To date, the motor has been a place holder. I reserved a volume for the motor and placed the worm gear on a pseudo shaft. Yet the motor has electrical contacts, as well as other internals and is decidedly not just a reserved envelope. Time to resolve that. All of the motor details were taken directly from the Hanazono motor that I intend to use. The magnets, rotor and stator were drawn in. The electrical contacts were sketched, as well as the insulators. When placing the detailed motor into the chassis, I discovered that the motor envelope previously drawn was in error. The envelope was too long. Bank error in my favor! This resulted in a modest change to the chassis. I also modified the chassis to give room for a wire to come out from under the motor.

How to clamp the motor in place? The stator of the motor should not move relative to the chassis. I installed a wedge clamp (triangular wedge) at the inside of the smokebox face. The angle of that triangular wedge forces the upper front of the motor to the lower rear corner. At the lower rear corner of the motor are two chassis faces. Thus, the motor is locked into position via a diagonal force vector, upper front (on the left) to the lower rear (on the right). 

VALVE GEAR SUPPORT

To date, all of the Stephenson Valve Gear has simply floated on location. Nothing supported it, it was never tied to OO Planet's structure. In my eye, it looked marvelous, but lurking there is a nasty problem. It is a mechanism that must be made to function, not just look pretty.

I begin with the valve gear axles. They are indeed tiny bits, brass, 0.5mm in diameter. They will be plenty stiff, as they are very short, just under 7mm They will be supported in 1mm OD, .225mm wall tube. That leaves 0.05mm clearance betwixt the axle and the tube, all the way around. a 0.002" slip fit. That is near perfect! Very good control of location whilst permitting rotation.

When you look at the image, note that there will be several axle supports across the front and across the back. They not only support the axle but keep the axle from floating side to side, as the valve gear bits on the axles are constrained. The first thing to know is that for any axle, the center line of these supports must be co-axial. That is fairly easy to accomodate, simply thread them onto a long axle when installing them.  

That also goes for making them. There are always 3 parts to each support. The tube itself, an arm that permits longitudinal adjustment and a base. Note the base for the two outside supports in front is different from the others, because the cylinders and nominal base conflict. Indeed, Stephenson also had this problem, and I have followed his solution! The three components for each support will be mounted into a jig, simultaneously with the other axle supports, such that the longitudinal, or front to back distance is uniform and controlled over the set of supports. The jig will insuring co-axiality in manufacture. Whew.

And then I came to the most mind boggling problem of OO Planet so far. The length of the control rods, those diagonal struts running from front to back, is fixed. The control rods ride on axle rings mounted to control arms, with a specified clearance from the rod hole to the axle ring OD. It must not bind. The issue? The axle supports must be on vertical location. If they are too far apart, the control rod is too short and it will bind. If the supports are too close together, the control rod is too long, and it will bind.

It may look marvelous in CAD. Reduction to practice says otherwise.

BUILD UP OF TOLERANCE

This is a problem that occurs when we have multiple components in an assembly. Each tolerance adds to the next. This accumulates until the assembly may not function at all, unless the build up is considered and resolved. Where to tackle this? I studied the issue for days, experimenting with this and that. I could try for a precision placement of the axle supports, but I swiftly realized this wasn't going to be workable. After all, how tightly can I place these tiny bits! I decided on a vertical placement of the axle supports that I could achieve. And then looked for how to keep the mechanism from binding. What took an incredibly long time to realize was that relief from binding was in the control rods themselves.  

LOOK TO STEPHENSON

Stephenson's Planet had threaded rods, such that the length of the rods could be adjusted. Stephenson had the same problem! I simply will not have the luxury of threaded control rods. It is all just too very small.

Approach 1 is to open up the inside bores on the ends of the control rods to account for the achievable vertical placement of BOTH front and back axle supports. That turns out to be a 0.11mm increase in the radius (0.004") or 0.008" total slop from the axle ring OD to the control rod ID. Not terrible, but the control rods may be a bit sloppy and as these control the oscillating handles, approach 1 is sub optimal. 

Approach 2 is to make several control rods of varying length (in sets of 4). In placing the axle supports, the control rod range of lengths is defined. Pick the control rods that fit the realized model best. I can keep the ring OD to control rod ID tighter and substitute in a better matched set of rods, effectively taking Stephenson's approach. Approach 1 is presented in the image, although you will be extremely hard pressed to see it.

In the end, I will likely combine the two approaches. Various sets of control rods, with a tighter ID. The rods won't flop around and the mechanism will have good fitments, keeping friction lower.

These updates may not be visually obvious. The motor is buried inside the shell. The build up of tolerance solution is virtually invisible. Yet both took quite a bit of thought to accomplish.

Until Next update

Bee