Dave+Lopez

The internal teeth of the lock blocks apparently don't sit quite as far down as I had thought, so the spiral feature of the planetary gear carrier that it sits on can eat into the teeth (almost like a tap) and then force the block upwards under certain conditions. This puts an axial load on the e-clip, eventually forcing it off.
 * //Problem://

//Why was this only a problem on one hub?// The block I had made for the failing hub sat a little tiny bit higher (few hundreths of an inch is my best guess) than the other hub. This makes the teeth more susceptible to failure, since the very tops of the inner teeth are more easily sheared away than a point further down in the meat of the tooth.

//Solution:// Now that I have an approximate value for the point at which the inner teeth interface with their mating surfaces, I can make the teeth longer, which should alleviate this problem (assuming that the aluminum is strong enough to hold up to this kind of abuse; I'll know the answer to that question when I open up the other hub). This isn't really a big deal and won't add much time to the production process if we can keep using aluminum. If the material ends up being an issue, I can make the block out of steel. This will definitely be strong enough, but will add some weight to the hub (it will actually be eliminating our weight savings obtained by using aluminum and adding a small amount on top of that) and will take longer to machine.

//Now, for the good news:// The e-clip is actually very strong - the lock block has a fairly impressive looking ~0.01" deep groove cut into it by the e-clip/washer before it popped off. This means that we don't need to turn to some other way to secure the inner mechanisms in place as long as the lock block doesn't fail.****


 * Lock Block - Manufacturing**

The shop guys and I have decided that the best way to make this piece is to take a 3" x 3" x 1" rectangular aluminum plate and waterjet the teeth and OD out of it. Once we do this, we can stick it on a lathe to create the profile, ream out the center hole so it is the correct size with no water-jet-induced taper, and then we can use an endmill to carve out the detail of the mating surface in the inner diameter.

The piece takes ~17 minutes to water jet and ~1:30 on the mill. I haven't turned down the profile yet, because I'm not sure if the teeth are strong enough to withstand the forces put on them by the hub casing if they are shortened down that much. We'll see next week.


 * Lock Block Solid Model**

- Version with longer internal teeth to solve the e-clip problem. - Further modified outer teeth, should be stronger than v2.3 - This one has stronger outer teeth than v2.2 does but should still fit (testing that will be required before we start producing more than one at a time).
 * [[file:Lock Block v2.2.SLDPRT]]** - Basically the design we currently have on our chair. The only difference is the shape of the internal teeth - I'll update this once Joe makes it.
 * [[file:Lock Block - Water Jet.SLDPRT]]** - Outline of v2.2. We could also use the v2.2 or v2.3 files directly, but the inner diameter must be shrunk so it will be undercut - this will let us bore it out to size so we're sure it will fit around the shaft and not have water jet taper give us problems.

//To water jet something:// The water jet only takes 2-D models. So, make a drawing of a part in SolidWorks that includes only the outline of the part to be created - **no** center mark in the hole or Title/Description stuff on the bottom. Save this drawing as a .DXF file (File-> Save As... choose .DXF) and get it to the water jet computer. You need one dimension of the part so you can make sure the drawing wasn't scaled up or down during the transfer, then let Bob do the rest.

We figured out how shifting works, and now have a good understanding of how the whole hub works in regards to what we're going to do with it. We can fairly easily have a high-gear only hub.
 * 11/08 Lab Notes:**


 * 5 Speed to 3 Speed Fixed Conversion:**

This is a simple process that consists of two parts:
 * 1) Removing the pawls
 * 2) Fixing the outer skin of the hub (called “case” in the rest of this description) to the internal gear mechanism


 * 1.) Removing the Pawls:**
 * 1) Remove the two nuts on the shaft
 * 2) Pop off the bearing/plastic cover
 * 3) Slide out the internals from the hub
 * 4) Remove the pawls (finger-like things that make the clicking noise when freewheeling) and wire spring from both pawl carriers (yellowish one right next to the e-clip, larger, silver one on the other end of the device)


 * 2.) Fixing the hub to the Internals:**

There are several ways that we could go about this. Two simple ones:


 * 1) __Replace__ the “front” (yellowish, next to the e-clip) carrier with a fixed block that is shaped to fit snugly around the bumps on the inner surface of the case, fixing the hub speed to the internal speed at this point. This is how the conversion kits that we’ve seen online work, and is also the basic principle under which the pawls work (though these only work in one direction), so we know that it will work and is structurally sound.

Note: Italicized items are only needed if we make our own · **Grease:** The part will need to be greased up like the rest of the hub for it to work · **//Metal://** A chunk of metal (McMaster?) approximately the size of the inner diameter of the case is needed · **//Machine Time://** Shouldn’t be too difficult to get, but this means we have to work around their schedules instead of independently setting our own.
 * Materials Needed:**


 * Possible drawbacks:**


 * **Time:** It takes time to either order the block or to order the material, figure out how to make it (CNC mill? Broaching? Water jet*?), and then actually make it.
 * Our best bet


 * **Tolerances:** This piece must have somewhat tight tolerances, so it doesn’t rattle around in the case or appreciably increase the backlash present.


 * **Cost:** Not really an issue, but this will probably be the most expensive route to take.


 * 1) __Attach__ the pawl carrier to the case. This can be done using a few screws or pins that pierce both the case and carrier and do not allow them to individually rotate, though they must allow them to axially move relative to each other.


 * Materials Needed:**

· **Screws/Pins:** Easy to find, cheap, not difficult in any way · **Machine Time:** We’ll probably need some time on a machine to solve the issues raised by the **Shifting** drawback below.


 * Possible drawbacks:**


 * **Case/Screws/Pins/Carrier Structural Integrity:** We are fairly certain that the case, carrier, and screws/pins that we use will be able to take the stresses, but we can’t definitively say whether or not they’ll be able to.


 * **Bearing/Clip Structural Integrity:** The carrier has an oddly shaped mating surface where it attaches to the rest of the internal mechanism. This mating surface lets it rotate rigidly in the forward direction, but when spinning in reverse attempts to push the pawl carrier outwards, the same way that the coaster brake worked in the Sturmey-Archer hub. Since this hub does not have a coaster brake, there is an e-clip to prevent the carrier from actually moving outwards.

This works in the existing design because the carrier is not loaded when spinning in the opposite direction; the pawls simply slip over the bumps on the inner surface of the case and make that clicking sound we hear. Our modified hub will be driven in the opposite direction, however, which will put a fair amount of force on the e-clip and the thrust bearing that they were not designed to hold. Will they be fine with this new load?


 * **Foreign/Domestic Object Damage:** The internals of the geared hub are delicate and precise. We must be careful when cutting holes for screws or pins that we do not let any debris enter the casing, and we must seal these holes well to prevent sand and dust from getting in over the life of the product. We also must be absolutely certain that the screws or pins will not shear. This mode of failure could introduce fairly large chunks of debris that will very quickly jam the mechanism. Once this happens, the wheelchair will not be able to move.


 * **Shifting:** The hub shifts by moving the internals back and forth axially inside a stationary casing. This means that any method of attachment would also need to allow this axial movement while remaining rigid in the rotational direction.


 * Thoughts/Comments:**

At this point in time, I think that purchasing a block (or machining our own, if we can’t get it in time) is the best way to go. Here are a few reasons why:
 * Fixing**


 * 1) We know it will work, and we know that it’s possible to implement once we have the materials we need.
 * 2) If we try to attach the carrier to the casing and we find that we cannot do it satisfactorily, we will have to fall back to this plan anyway. We may not have time to try two plans, so I’d rather go with the plan that’s sure to work from the start and not run the risk of running out of time.

As we have seen, there are two pawl carriers in the hub – a yellow one near the front, and a larger, silver one near the back. When we modify the hub, we are essentially making the yellow carrier’s shaft the only output and disconnecting the silver carrier’s shaft from the casing. This nets us two low gears. If I’m thinking correctly (and I may not be, this is just a thought I had last night) we may be able to build a wheelchair with two high gears by making the silver carrier’s shaft the only output and disabling the yellow carrier’s, which is the exact opposite of what was done above. Instead of disabling the high gears and netting us two low gears, this may disable the low gears and net us two high gears instead. I’ll have to take a closer look at an assembled hub to see if this is the case.
 * High Gears**