1/16 Scale FrankenDiff

[ddmckee54]

Regarding the Sherline carriage lead-screw issue, when I started taking it apart last night I discovered that the issue went away when I took the hand-wheel off. Upon closer examination of the hand-wheel and the lead-screw shaft I found a nick on the face of the hand-wheel and a burr on the lead-screw. The burr on the lead-screw was caused by the grub screw, but I've got no idea what could have caused the nick on the hand-wheel. I stoned off the raised areas and when I reassembled things that particular problem was gone.

I've been plottin' & conivin' for a while as to how I can adapt the Frankendiff into a powered steer axle. Blender's build of a 1/14 scale TH407C telehandler gave me the idea I needed. He used Losi CVD front axles to build his steer axles. Since I've already got a butt-load of the WLToys 12428 replacement rear axles I decided to get a set of replacement front axles to see if they'd be usable. After doing a little (very rough) 3D modeling, it looks like things will work out. They both use a ball and socket joint in the knuckle, but the Losi uses a CVD joint while WLToys uses a CVA joint. A CVD joint has the ball on the axle with the socket on the half-shaft. A CVA joint flips that around and has the ball on the half-shaft with the socket on the axle. Here's a picture of the WLToys CVA joints.

At the top is the assembled CVA joint, in the middle is the axle with the socket, and on the bottom is the half-shaft. At the left is the dog-bone that would normally attach to the front diff. This 14mm section on the left is totally useless to me, I only need about the last 25mm on the right. The center section of the half-shaft currently has a 3.18mm OD. My current plan is to turn part of the center section and part of the tapered area just to the right of it down to a 3mm OD. This will then be Loctited into the outboard end of the modified donor axle, just like I did when shortening the donor axles for the rear diffs.

Edit: Due to the black finish I was worried at first that the axles might be hardened. However when I checked them with a file I found out they aren't. The dog-bone section was easily removed with a hacksaw.

[Zabco]

Lord only knows what happened to the site this time. Was completely gone this time, not even any html code showing up when I tried to inspect the site. Glad to see it's back though and hoping it holds together for awhile yet. Still the best site for real information about building trucks and equipment.

Your axle really looks good Don, great job! Also sounds like you've got a good handle on the steerable axle as well. Can't wait to see how it comes out.

As to the Sherline, I'm glad to hear that removing the burr on the hand wheel fixed the problem. Thought you might have a bent lead screw when I started reading. Don't know about the bearings. Sherline claims they are lubed for life. Course no accounting for what the previous owner did to it. May want to drop Sherline a not asking about it.

[ddmckee54]

I was afraid I had a bent lead-screw too. Yes the bearings are supposed tp be lubricated for life, but, they are shielded bearings - not sealed. If they sit for several years without being used, the grease will dry out and harden. According to the part number they are standard precision bearings. Once upon a time in the west I used to repair machine tools for a living, so swapping spindle bearings is not outside my comfort zone. At least now the parts are a LOT lighter, and they're not covered in nasty smelling coolant.

I just got done modifying the first CVA half-shaft.

The modified half-shaft is just slipped into the brass tube for now, I'm just getting started on the working 3D model that will be used to print the parts. I've got bushings and shoulder bolts on order for the knuckle pivots, so I can accurately model things now. I model things in mm because that's how most 3D printers think. I really wonder sometimes when reverse engineering this stuff, some of the measurements make no sense. Like why is part of the half-shaft turned to 3.18mm, and why is the socket bored to 7.62mm. Earlier today it finally dawned on me, 3.18mm is 0.125", and that odd 7,62mm bore is 0.3". When I convert all the odd Metric measurements to Imperial units they come out nice and neat.

[ddmckee54]

I spent most of the day yesterday fighting with Alibre locking up. I'm trying to do the final 3D design for the powered steer diff parts. The design I showed a couple of posts ago was quickly thrown together to see if I could even make it work. NOW I need to cross the I's and dot the T's so that it will all work together, and that means accurately modeling all the individual parts. That's actually a good thing as I found a couple of gotchya's while doing it.

One of those gotchya's was determining how much I needed to shrink/stretch the ends of the diff front and rear housings to accommodate the single front wheel instead of the dual rear wheels, and still maintain the same wheel track. It turns out I needed to take 13.5mm off each side. I know that doesn't make sense that I need to move the end of the diff housing inboard when going from duals to singles, but I spent a couple hours convincing myself that moving things 13.5mm inboard was actually right.

The second gotchya was when I realized that the screws holding what I'm calling the diff pivot in place (that's the part where the steer axle pivots in order to steer) would be hitting the outboard bearing - that would not be good. So I had to change the design of the diff pivot, it's now about 6mm longer. THIS change interfered with the outboard screws holding the front and rear diff halves together. Since the diff pivot slides over the ends of the front and rear diff halves holding them together, those screws aren't really needed any more. Problem solved, but I did have to change the 3D models for both halves to reflect this.

The third gotchya is what does the length of the brass tubing splice need to be? My modified rear axle measures 147mm from end to end. My modified axles were 70mm nominal in length, so that means the spider gears were occupying 7mm of space in the middle. My 3D printed non-powered front axle needs to be 172mm from end to end. If I subtract the 7mm for the spider gears, and divide by 2, the stack-up for my front axle length should be 82.5mm? (Why do I feel like Rodney Dangerfield in "Back to School" right now?)

Plus I realized that there's nothing keeping the outboard bearings of the steer axle from sliding outwards. On the rear axles the hex driver holds the bearing in place, on the front axle there's nothing but air in that location. Oh well, that's a problem for future me to solve.

Don

[Zabco]

Don't you love it. Think you got one simple change to make and it cascades into an avalanche. Good thing this is a hobby, I'd hate to have to charge someone for the time involved in some of these changes.

As to your bearing problem. Would it be possible to cut a small groove in the axle shaft just outboard of the bearing and install a C or E clip to act as a stop. This is how the bearings in Tamiya trailer axles are held in place.

[ddmckee54]

Somehow today I did something stupid. I must have changed the font size or something for this site because the text on this website is now huge. But it's only on this website, any other website appears normal, and I've got no clue how to get this back to normal.

[ddmckee54]

Never-mind, I figured it out. Somehow I had bumped the Zoom up to 200%.

I'm still fighting with Alibre's assemblies, trying to figure out what I can and cannot do with it. (I think part of the problem is that I'm NOT modeling any mechanical fasteners. This just might be what's driving Alibre up a wall.) This is a shot of the suspension assembly up to this point with the front axle just hanging out in mid air - things are shown at the approximate ride height.


I've got the front diff mostly complete internally, I've got the flexible CVA axles & half-shafts modeled. I haven't completely modeled the brass tubes and inboard parts of the axle shafts yet. I do have the brass tubes modeled to give me a nominal over-all axle length of 82.5mm. I did this as a sanity check to see if I had about 7mm between the axle ends like there was with the rear diffs. The distance between the axle ends measures 7.143mm - as Mr. Miagi would say "CROSE ENOUGH!"

I'm working my way out now, I've got to design the steering knuckle/steering arm widget. I've got a rough idea of what will work, but I need to refine it. It needs to fit the tie-rod in so that it clears both the top on the pinion cover, and the bottom of the frame rails - with some clearance to allow for suspension travel. Now I need to find the specs on the bushings and shoulder bolts I ordered - so I'll know what modifications they will need.

Don

[frizzen]

Ugh, this sounds like a whole lot of fun.

How often are you stopping to take a break to push that Bruder truck around while making engine noises?

[ddmckee54]

That phase will probably start when I actually begin bolting pieces to the frame rails. Instead of just making making pretty pictures on the monitor.

[ddmckee54]

I think that I just might be starting to get a handle on this Alibre assembly crap.


You are looking at a functional 3D assembly model. I can spin the wheel on the axle. When I turn the steering knuckle the wheel moves with it, and the tie-rod brings the knuckle on the other side along for the ride.

I did have to give up on the idea of putting the tie-rod above the drive shaft though, there just wasn't enough room. Where it is currently located the tie-rod is about 4-5mm below the driveshaft and won't touch any thing trough its' full range of travel. The bottom of the tie-rod is still above the bottom of the differential though so the tie-rod shouldn't get caught on any thing.

[Zabco]

Wanted to reply to your post on the 28th Don but I was at the Lafayette RC truck show this weekend and only had access to open public wifi and I don't like submitting passwords over those.

Alibre's assemblies can be a huge PITA. Most times they fall right in place and other times they just aaaarrrrrggghh. A few things I've learned about them over the years.

You can only have 3 constraints between any two objects or assembly.

Sometimes just changing the order you create the constraints will make a difference.

Under each part listed on the work history on the left side of the screen will be a list of each constraint you created for that part as well as a list of all the constraints created for the assembly as a whole. If a constraint has failed for some reason it will be shown in red. I usually just delete these as they can cause confusion when trying to figure out how things went together.

Get one part in place, usually 0,0,0 and anchor it. If you don't you run into the possibility of parts suddenly flip, flopping around and you can't figure out how the heck that happened. And usually the only way to correct it is to delete parts and start over.

It's often easier with complex systems to make parts into sub-assemblies and then assemble the sub-assemblies in a final assembly. But there is a catch to this method. In your last post as an example, you explained how you had the steering knuckles and tie rods able to move back and forth. If you were to make that drawing a sub-assembly and then add it to an assembly of say, the entire frame, those knuckles and tie rod will no longer be able to move. They will be frozen in whatever position they were in when you last saved the file. Only parts added in the current assembly will be able to move.

I don't know to how many decimal places Alibre takes their precision but it must be quite a few. If you try to constrain two parts and the difference in distance or angle is off by the slightest amount they will not go together. For example if I want to assemble a cross beam to a frame and they both have a square 4-hole layout that bolts would be put through to bolt the parts together I would probably assemble them in Alibre by first mating the two flat faces together, then make a coincident constraint between two of the holes in the frame and the bear. But the parts would still be free to move about each other around that bolt hole. To fully fix the beam in place I need to add a third coincident constraint between another pair of bolt holes. Three point of constraint and the two parts can't move in relationship to each other. But when I designed the beam I got the distance between the two pairs of holes .0001mm shorter than the same pair of holes in the frame. Such a tiny amount of length that you couldn't see it in the real world and a bolt would still easily fit, Alibre won't accept it. Round-off errors if you do your math outside of Alibre can cause a real problem. I don't know of any setting that you can change to allow a certain amount of 'play' between parts. The same is true when trying to mate parts with angled surfaces that are already constrained to other parts.

The nice thing about the assembly window is that you can create new parts or edit existing parts within the assembly. In my example above where I can't make the third constraint because of the slight difference in distance I could edit the beam by deleting the holes in the beam, place new holes on the face of the beam and make them individually coincident with the holes in the frame and then do an extrude cut on the beam. Perfect match, constraint made and when I save the assembly file the part file for the beam will be updated.

If all else fails, make a plane and use it to constrain a part. Planes have to be created relative to another plane, face or line of an object. So Alibre knows exactly where the plane is located. You can then do a direct of offset constraint between the plane and the part or sub-assembly you are having problems with.

[ddmckee54]

Constraining things to a plane is what I do - sometimes.

I placed an axis in the centerline of the axle bores, and I've got an axis on the pinion shaft centerline for each axle. The origin for my suspension assembly is the tail end of the center of the frame rails at ground level. I've got offset planes where the center of each axle shaft should be located. I've also placed a plane at the ride height. I've constrained the axles so they can move vertically along the locating plane and the pinion axis is constrained to the YZ plane, this keeps the axle where it belongs front to back and side to side.

I constrain the axle shaft axis to the ride height plane when I initially place the parts/sub-assemblies. When I want to check suspension clearances I suppress that constraint. When working in the various sub-assemblies I will constrain the various parts to get them properly oriented, anchor the part, then delete the effected constraint/s.

I did something similar with the front axle assembly. When I was initially placing the CVA axle and the CVA half-shafts their axis were constrained to be coaxial. Once I had everything placed properly I suppressed the coaxial constraint, and constrained the axle so that it could pivot around what are effectively the king-pin bushings.

I stumbled across a "Flexible" check-box and hoped I'd found the mother-lode regarding the "moving" sub-assembly parts no longer moving issue. Alas, there was no joy in Mudville that night since it made no difference. I've been doing a little digging on-line and it seems that I was on the right track.
That option IS supposed to fix the no longer flexible issue - but it doesn't always work.

I was able to use the assemblies to check clearances, what broke the rear suspension assembly was trying to get the walking beam to play nicely. I'm pretty sure the full scale rubber bushings are there to absorb any minor alignment problems, and I've got silicon O-rings to substitute for those bushings in the model. I just don't think Alibre assemblies were designed to simulate something like that.

[Zabco]

Sounds like you have a real good handle on the assembly process in Alibre. Just wish it all would work. Where is this 'flexible' check box? I don't recall ever seeing such a thing. Or is it only in the full blown version and not in Atom?

[ddmckee54]

I've got Design Pro V22, so I don't know about Atom. In Design Pro if you right click on the assembly in the Design Explorer, you get a drop down menu - Make Flexible is one of the options in that menu.


Don't know if that helps.

[Zabco]

It does indeed, last place I had thought to look. Had gone through all the preference menus and couldn't find anything. Will have to do some experimentation as I have never tried it. Weird in that I use the functions on either side of it all the time and never really noticed it.

FYI - I talked with Travis Biddle while at the Lafayette RC show this past weekend. Travis is the guy who owns this web site. Asked him how much longer the site will remain active. He didn't say when it would go dark but that it's days were probably numbered. Lack of use, number of known bugs in the program and no one wanting to hose forums anymore. If you want to save anything you've posted on here you better back it up somewhere. I've always kept copies of my postings and thankful that I never used photobucket for my images. I'll keep coming in here till it goes down but I am considering options. I already am a member on Scale Builder's Guild and they have a section for trucks and equipment. And I know there are a couple of other similar sites. Wait and see what happens.

[ddmckee54]

I know that Blender, Frizzen and I have all posted over on the Semi Truck and Construction Equipment forum of RCGroups. I'd never heard of the Scale Builders Guild until now, I just registered there. I'll have to take some time and go through it later. I never used Photobucket either, but they sure did screw up a LOT of forums.

[Zabco]

The two programs are pretty much the same. I've still got a copy of Cubify Design (from when 3DSytems owned the software) and it works the same way. I played around some today with the 'flexible' function. It does work but is somewhat persnickety about how you set up the assembly. I drew up a basic engine using 3 cylinders with wrist pins and connecting rods. A piston with connecting rod and wrist pin was created as one assembly. A crankshaft with 3 - 120 degree connecting shafts as a part, and a basic engine block as the third component. Started a new assembly by bringing in the crank and constraining it to the Z axis and centered on 0,0. Then imported three copies of the piston assembly. First I connected each piston without enabling flexible to the crankshaft with a vertical constraint. As expected, the assembly locked up with the second piston and would not rotate. Next I enabled flexibility. Connected first piston assembly to crankshaft and constrained to vertical position directly over center of crank. Piston would move up and down with crank rotation. Added second piston same way and crank would rotate. Thought that this was going to work ok. Added third piston and the assembly locked up and would no longer rotate. All three pistons were in alignment and connecting rods were free to swing back and forth. Happened to glance at left hand column and noticed I had a number of constraints shown in red meaning they were bad.

Thought about what was going on and came to realization that I was over constraining the piston assemblies. What I forgot is that they already had some constraints carrying over from their original construction. Adding the three additional constraints placed on them in the new assembly and they froze. Deleted the piston assemblies and all their related constraints. Imported the engine block and centered it in X and Y over the crankshaft and anchored it in place. The engine block would serve to align each piston directly over the crank, properly space each piston along the crank and serve as the vertical guide for the pistons. Add each piston assembly back, enabled flexible, put concentric constraint between each piston and cylinder bore and a concentric constraint between crank and connecting rod. That's all, no other constraints. Added other two piston assemblies in same manner. "Grabbed" the crank and gave it a twist and it spun with all pistons going up and down as the should. So the 'flexible' function does work but you just have to be very careful in how additional constraints are applied. And, I think, not constrain part that will move to planes or axis.

[ddmckee54]

For the last couple of days I've been working on Sherline mods. One of the mods I did earlier was to add some cheap Chinese scales with remote readouts to the carriage and cross slide, that was a game changer. But I learned some things about those scales that I wasn't happy with. First, they're battery powered and no AC adapter is available. My scales also do not have an Auto-Off feature. I don't know how many times I've gone down to the shop the next day only to discover I'd left the stupid things on. The most aggravating thing is that the carriage scale will periodically reset to zero.

I did some research and found out that the Shahe clone scales, which is what these are, can be sensitive to electrical noise - there's no shielding on these things. I did mention they were cheap didn't I? 'Cause I got two 8" scales for 36 bucks and change, including shipping. I am replacing them with Igaging EZ-View Plus scales which have better resolution, shielded cables, and an available AC adapter. I got a 6" scale, a 12" scale, and the AC adapter - all for about twice the price of the Shahe clones. As an added perk, Igaging actually gives you instructions on what scale functions are available and how to use them.

I was having trouble wrapping my head around what everything would look like, so I decided to 3D model it. For the last couple of days I've been modeling my Sherline's bed, the base I've got it mounted on, the carriage, the cross-slide, the new 6" scale and read-head, and the new 12" scale and read-head. I modeled the scales and read-heads as separate parts so they would work correctly in the assembly. Once I got the parts modeled it was time to start arguing with Alibre about what I can and cannot do with constraints in assembly drawings. I've learned some lessons from the other assemblies I've done about being methodical and renaming the constraints to something that will make sense to you in the future.

I added the base to the assembly and then I anchored it. I then added the Sherline bed, and used offset constraints to position it. Then I anchored it also, and deleted the constraints I broke when I anchored the parts. I added the carriage and used coincident constraints on the dovetails to keep the carriage in its' proper location. I then added an offset constraint between the carriage and the headstock end of the bed, and played with the limits until I had the carriage restricted to it's normal movement. I did the same for the cross-slide and got the carriage and cross-slide working together as they should.

Then I started adding scale bits that were just hanging out in mid-air; getting them to move like they should without breaking anything else. Once I had everything dancing about in a coordinated fashion I started designing the brackets that would connect Tab A to Slot B. This is where I'm at now.


I've still got to finish the design of the bracket that ties the cross-slide scale to the cross-slide. The cross-slide read-head is fixed and the scale moves through it. The carriage is the opposite with the scale fixed to the base and the read-head moving on it. I still have to design the brackets that will anchor the carriage scale to the base. I just noticed that I never modeled the two existing screw holes in the carriage. I'm going to use those screw holes to mount the read-head bracket. I'll need to reposition and redesign that bracket to use those screw holes. I guess I'll have to send a change order to the customer for the extra engineering time.

Don

[Zabco]

The iguaging scales are very popular in the home machinist hobby world. If you want an even better readout google 'yuris toys'. Yuri developed a, relatively, inexpensive way to combine the scales with an old android based tablet and some and some additional hardware to create a very good DRO for lathes or mills. Has his own forum on the www.hobby-machinist.com website.

[ddmckee54]

The new Igaging EZ-View Plus scales use a different protocol than their older DigiMag scales did. I know Yuri is working on it but I'm not sure if he's got his Touch DRO working with that protocol yet.

I got the 3D design done done for the lathe scales and printed the parts out last night.


The 2 parts that I couldn't design any adjustment into are the ones that don't fit. Turns out it was operator error, I didn't model them correctly. I fixed the models and the parts are re-printing now, they should be done in a couple of hours. Then I can start installing the new scales.

After that, about the only thing left for Sherline modifications are new headstock bearings, which arrived in the mail today.