Casting Process Successful!

I received a nice shipment from Curto containing 3 full sets of usable castings!  It is usually pretty cool when parts for a client's project come in but when the project is your own baby, the excitement is greatly magnified.  Imagine seeing a new dirt bike in front of a Christmas tree and you are in the right ballpark.

The combination of gating assistance from Curto and the 3D printed sand molds from Humtown resulted in parts with crisply defined features and just a bit of surface roughness, a traditional characteristic of the 3D sand printing process.  Though a bit rough, the finish was very consistent.  Think of it as a mini heatsink!

Both the pour and the post processing of the parts by Curto was excellent.  The T6 heat treat left no appreciable distortion at the gasket surface, and if I did not do the mold CAD myself, would not be able to tell where the feeders were attached.

 The high damping properties of magnesium and the internal and external ribs on the oil sump and left side cover (along with the hydrodynamic crank/counterbalance bearing system) will help keep transmission of internal mechanical engine sounds to a minimum and allow the intake howl and the exhaust growl to dominate the aural frequencies.

The clutch/generator cover's deep cylindrical shapes are themselves resistant to vibrational excitation, so less ribbing is needed here.  Internal features on the generator mount allow use of an unmodified Panigale stator with its wiring, crankcase seal, and waterproof connector.  It is a minor detail, but the stator is an electrical part subject to failure, so it is nice to be able to drop in an off-the-shelf replacement part.  The same goes for the voltage regulator, which is also an OEM part and easily accessible.

Having not worked directly with sand cast AZ-91-T6 Magnesium before,  I requested a piece of the runner/feeder system also be heat treated and sent along with the parts.  I will use this piece to do some machining speeds/feeds optimization, but more importantly to do some pull-out testing for the few M6 threaded inserts and for the oil drain plug in the oil sump and the M5 inserts for the clutch cover.  Magnesium is quite a soft material so I would rather not depend on threads created in it.  I will test both E-Z Loc and TimeSert inserts to see which works best.  TimeSerts are very low profile so maybe use them first and if there is a failure, there will still be room for an E-Z Loc.

Due to casting being a slightly unpredictable process, I was not sure if all the parts would be fully filled and usable but it turns out my fears were unfounded.  Curto's design assistance was top notch and with this completely successful pour, I now have 3 sets of castings.  They mesh up nicely with the 3 sets of billet blocks for crankcases, and enough 4340 VM steel for 3 sets of crank and counterbalance shafts.  A bit of ebay shopping later and it is possible to have 3 engines, and then a bit further, 3 bikes!

One last, unrelated item: last week I made a significant change to my original plan of using a hacked OEM Panigale ECU to do the engine and chassis management.  This did not go as planned and required some regrouping and rethinking, and in a week or so I will be able to go into some more detail on the new solution.

All right, time to stop typing and start doing!

Casting the Engine Side Covers and Oil Sump

I am really starting to gain some momentum on the engine part of the project.  Now that all the vendors for the various outsourcing needs have been contacted, parts and specifications discussed, quotes given, and deposits sent, the design really becomes frozen.

Crankcase and crankshaft posts will be coming shortly, this post deals with the engine side covers and oil sump for the semi-dry sump oiling system design.  The parts themselves are pretty straightforward and will be sand cast in AZ91 Magnesium alloy and heat treated to a T6 specification.

 

Why magnesium and why cast, instead of the billet aluminum I do so often?  Magnesium for weight and a nod to tradition, as the dark green Dow17 (or sometimes with gold paint) color is a traditional sign of a factory race engine.  I am a factory and this is a race engine!  Though it is not that strong, magnesium is very light, and for lightly loaded applications like engine covers, is a great, through pricey solution.  Why cast and not billet?  Pricing of magnesium billets for these parts is prohibitively expensive.  In this case, casting is the 'cost effective' approach!

The process I am following is a very modern take on the ages-old sand casting technique.  Most of the recent improvements in sand cast prototyping have been using 3D CAD to generate accurate models, then use 3D printers to quickly create low cost patterns that would be used to create the sand molds.  The approach I am using goes one step further, the 3D CAD model is used to create a digital mold, that can then be printed directly in casting sand via very specialized (read expensive) 3d printing equipment.

These molds are being printed by Humtown Products, who provide pattern and moldmaking services using a variety of traditional and cutting edge technology.  They use an ExOne 3D sand printer that has two refrigerator-sized printing tanks that can print sand molds over 6 feet long!  Brandon from Humtown did the hand-holding with me to ensure the molds were prepared exactly as needed.

Imagine an inket printer crossed with a sandbox and you are not far away from this machine!  A layer of sand is applied by a precision scraper, the printhead sweeps past and sprays a specialized adhesive on the areas that are to become the mold.  It builds the part layer by layer (sound familiar?) and at the end the un-adhered sand is brushed away.   A good video can be seen at here.

Since these molds will be poured with magnesium, they need a special treatment to prevent the reactive liquid metal with the surface sand of the mold.  Potassium fluoroborate is mixed with the sand before printing to accomplish this.

These completed, treated 3D printed sand molds will be shipped to Curto-Ligonier Foundries Co. for the parts to be poured and heat treated to T6 condition.  Curto has extensive experience with both aluminum and magnesium sand casting for both the aerospace and motorsports industries.  They are not aware of it yet, but after this sub-project is successfully completed, I will spring my one-piece sand cast swingarm project on them!  Curto have been very helpful with overall gating design assistance, not to mention dealing with someone that has as many questions as I do!

Now on to the meat of the post:




The Generator Side Engine Cover:

-covers the generator rotor, counterbalance shaft gear, and clutch
-mating surface for clutch pack access
-mount for generator stator and wire routing

Starter Side Engine Cover:

-covers starter geartrain, cam chain drive, and shift lever mechanism
-access port for crank
-oil fill
-outboard seal support for shift lever shaft

Oil Sump:

-main oil reservoir
-secondary support for electric water pump

The cope and drag halves of the molds were manually created by referencing the existing part geometry.  The molds have pouring and riser openings in the top surface to ensure the parts fill and cool correctly without sink marks.

The molds are assembled like a Tetris game in the printer so that they can be made in the smallest vertical height to save both sand and printing time.

When printing is complete, the molds are cleaned, packed, and shipped off by truck to the foundry.  I will ask Brandon form Humtown and Jovan from Curto to take many pictures of the process for the follow-up blog post.  The printing should take a couple of weeks, and then a few more for pouring and heat treat, so hopefully in under 2 months we will all see a lot of pictures of perfectly filled cast magnesium parts!

Enough for 3 engines!

That's it for now.

All Parties Heard From

Came in today to receive a nice email from the crankshaft vendor of choice, David at Marine Crankshaft, Inc.  He had reviewed my drawings and requests and agreed to take on the process of turning a rough-machined crankshaft blank into a ready-to-use part.

The plan is to use 5 1/2" diameter 4340 vacuum melt material from Yarde Metals in a normalized and tempered state (Rc28-34 for you techies) which will be machined in the lathe and mill, leaving adequate stock on all bearing surfaces.  The process will be very much like the process used for the V4 engine detailed here, here, here, here, and here, but with bigger, more rigid machine tools.  I'll then send the crank to Marine where they will heat treat, finish grind, detail oil holes, superfinish, then plasma nitride as the final step.  This will produce a crankshaft with extremely fine bearing journal finish, strong and ductile core section properties, harder/stronger shell properties, and an extremely hard and lubricious plasma nitrided final surface.  The last truly beautiful crank I saw was a Rick Schell Stage 4 crank for a TZ250.  It was so gorgeous you didn't want to put it inside crankcases.  I am hoping the Hypermono crank will elicit similar feelings.

The main dimensions of the crank (journals diameters, stroke) were retained from the Ducati part to make bearing sourcing a simple proposition but the other proportions were made to match the new single cylinder application.  The resulting design is short and rigid, two nice adjectives to use for a crankshaft.

I used as many of the design tricks for a high performance crankshaft as possible, including hollow rod journal, large bearing fillet radii, and tungsten slugs for balancing.

Determining the optimum engine balance factor was not straightforward.  The offset cylinder complicated the situation enough that a simple piston primary and secondary force analysis was not sufficient.  Again I called on the help of Tony Foale to determine the best approach.  One of his papers on engine balance  and his basic engine balance software were a good starting place.

Digging into his big bag of software tricks he was able to modify an existing program to take into account not only the basic piston primary and secondary forces, but the relative positions of the crank, piston axis, and counterbalance shaft.  He ran a few optimization studies and the results started flowing.  For a basic analysis a 50% balance factor on the crank and 50% on the counterbalance shaft produces the lowest overall engine vibration levels.  The optimized result was a few percent different than the basic simulation and, just as important, the balance weights are optimally not 180 from the throw, again off by a few degrees.  Even though 50/50 and 180/180 would have been good enough and resulted in a smooth engine, if I am going to go through all this trouble, why not make the parts to the optimal values instead of approximate ones?  There is no good answer not to!



The crank assembly is configured a little different than most crankshafts due to the idler gear being between the crankshaft and the clutch.  This gives me the same radial room on both ends of the crank for components.  As the design progressed and I shuffled components back and forth, the best overall layout ended up with the primary drive gear and generator on one end, and the timing wheel/starter clutch, starter gear, and cam drive sprocket on the other.

Most of these parts are Ducati OEM, no reason to reinvent the starter clutch or generator rotor.

Now that i have a definite path forward on all the outsourced components the last bits of material will be ordered, CAD files finalized and frozen, and toolpath generation started.  Lots of clicking ahead for me in the next several weeks but after that, fabrication starts in earnest.  That will produce much more exciting pictures and videos!

Details, Details, Details.

There seems to be plenty of empty time these days so there is plenty of time to obsess over some bike design details!  I'm sort of going a little stir crazy and am finding it hard to be really productive, so please bear with a scattershot blog post.  I'll be covering a bunch of stuff, from parts status to design details, to fabrication prep.

The last, last minute change is that we have decided to go from a 1299 piston/cylinder assembly which is a 116mm piston in a thin steel liner, to the 1199 piston cylinder assembly, which is a 112mm piston in a thicker aluminum liner.  Seems that the 1199 is the basis for all the race development, like 2 ring pistons for less frictional drag.  The lost 50cc of displacement will be more than made up for in less friction and better breathing.

Now that the important dimensions have now been set in stone, purchase orders have been issued and the bank account is cringing in submission:

• custom length connecting rods from Carrillo 
  

•   custom length high performance roller cam chain from IWIS

 

• 6061-T6511 aluminum blocks for crankcases 

The engine CAD model is undergoing subtle refinements as time allows.  Current state is good enough to freeze and start making chips but there seems to be no sense rushing things so I am revisiting details that, while satisfactory from a pure engineering perspective, just don't have that elegant feel to the design solution.  These areas are, of course, few and far in between.  ;)

I have contracted with Curto-Ligonier Foundries to sand cast the two side covers and the oil sump in AZ91 magnesium.  Curto will be providing all the feeding, gating and sprue design information which I will use to create cope and drag parts of the molds for sand casting. 

The sand molds themselves will be 3D printed sand that the metal will be poured directly into.  The sand molds will be 3D printed by Humtown Products.  Brandon Lamoncha there has been a big help in gathering the various details needed for the process.

As the engine nears design completion, I can focus on the rest of the bike.  Since the chassis and suspension are already fully detailed, I can progress to details like air intake routing, airbox, and electrical components.

The dual air intake path threads through both small triangles formed at the crest of the frame then merges, has a filter mount, then enters the airbox.  Its volume is approximately 9.5l, plenty big to tune the Helmholtz resonance frequency to help boost the mid-range power.

For the air runners, I determined the minimum air intake size, doubled it, then used Creo's advanced 3d surfacing features to create duct with a curvature-continuous walls and a linearly increasing cross-sectional area. This helps slow the air from road speed to 0 with minimal losses, ensuring max static pressure.  Even though ram air benefit is pretty small for a bike that can only do about 160mph, it makes sense to try to get what you can if it does not get in the way of something else that is more important.

The air intake system leaves plenty of room for a nice size filter at the front of the airbox housing.  The various electronics modules (ECU, water pump control, coil, IMU, etc.) will live on to and around the base of the ram air ducts.  I am not sure if there will be brackets molded into the ducts or a complex 3D printed housing that has a few discrete mounting points.  Those details have a minor enough impact on the surronding parts that I can push them back until the first test assembly phase, where details like this are best resolved.

That's about it for now.  Forthcoming posts will cover the sand mold design process and the resultant castings, and the start of crankcase and crankshaft machining.

Hope you are all staying sane during this craziness.