open-bionics / brunel Goto Github PK

The outer palm frame extends over the top of the inner palm. The back cover doesn't have a Boolean to accommodate this.

The M2 bolts bite into the plastic badly. Washers need to be used. It would be good to have locating features for them.

The prongs of the thumb ligament can be much wider. Making this change will help with lateral stability.

The ligament cover partially covers the M2 bolt holes for the back cover.

M2 bolts are hard to push through. Check clearance section diameters

As the PCB slots in from the palmar side of the back cover, the boolean for the protruding USB socket needs to extend all the way through in this direction.

The way the booleans intersect around the connectors means there are gaps. The PCB is clearly visible.

Could make a tight fitting Cheetah clip that seals around the ports.

The GALA cover can't be clipped on with the first thumb pivot dowel in place. It needs a Boolean to accommodate it.

The inner palm needs to provide some space for finger ligaments to clip into

Change design of locating features to prevent gala cover rising up and allowing movement of actuators.

The part needs to perform the following functions:

• Hold the PQ12s
  • Provide protection for the PQ12s
• Contain the GALA springs
• Provide mounting for the PCB
• Provide mounting for the digit ligaments
• Route the tendons to the appropriate digits with minimal friction
• Provide attachment points for the dorsal cover

Whilst meeting these constraints, the part volume should be minimised to reduce weight, print time, and material cost

The ligaments appear to bee too long between the joints. The area around the joint seems right. However, on the distal phalanx, the end of the ligament doesn't tuck in very well. On the proximal phalanx, there appears to be too much material between the joints.

Possibly a result of the hook modifiers not working.

@OllyMcBride suggested adding a ridge at the proximal end of the Ligament. It can fit in a matching trough on the inner palm and provide another point of attachment for the ligament.

This could potentially be replicated on the fingers.

The "castellations" that hold the ligament have too narrow a cross-section at their base (i.e. footprint). This part is also in the layer plane so the castellations are prone to snapping off.

Suggested to make the footprints longer by extending them proximally.

Using Cheetah as the mould material was a definite improvement. There is not bleeding of Urethane or dye where the outer surface of the PLA was pressed against the Cheetah mould.

However, the top of the mould is made of PLA (intended to add rigidity). When cast, the urethane and dye fill the mould, reach the top and bleed through the small gaps between the PLA part and top of the mould. This means the ligament trough, M2 bolt hole and (on the distal phalanx) the tendon channels are flooded with urethane.

The same seal achieved on the rest of the part needs to be seen on the top as well.

P.S. Mould release definitely required.

At the moment the hooks on the ligaments must stretch 1 mm either side of the knuckle to reach the hooks on the phalanges. This pre-tension helps with lateral stability. There are also ridges on the underside of the ligament that slot into trenches. They are half way between the joint and the hook. At the moment, they are 0.5 mm misaligned. The idea is that stretching the ligament to the hook aligns the trench with the ridge. However, this means the ridge has to be located under tension. It might be better to have all the hook and trench fit in a relaxed state, then to apply the tension.

The 2nd thumb pivot dowel cavity is too tight for the M3 dowel

The PCB wants to fit loosely into the back cover. Otherwise, it will be pulled off the GALA cover when the back is opened.

The outer palm should have a lip along the dorsal edge.

This will catch glue in the moulding process. It will also hide the exposed edge, helping to prevent the skin peeling off.

The displace modifiers on fingers 2, 3, and 4 look like they're using local coordinates rather than the UV map.

The cover is presently fastened by two M2 bolts at the proximal corners of the inner palm block. This has a number of problems:

1. Those corners protrude, badly distorting the outer shape of the hand. They can't be softened much due to the springs around the motor horns. However, any would help.
2. The M2 bolt Boolean used is too narrow on the clearance section. The bolt bites on the clearance hole.
3. One bolt would be sufficient given the wedges also holding the cover, and the low loads involved. This is much more preferable from an assembly point of view

There is too much length between the USB socket and connector Booleans. This leads to the wall between the circuit board and connector being too thick. As a result, the USB connector can't be plugged in fully.

The clip could cover the teeth on the thumb to improve the look.

Currently, the PCB just sits in the GALA cover. It will pop out if there are cables pushing up from underneath or fall out if the hand is inverted. It would be better if it clipped in.

Is it possible to make the back cover's form completely symmetric so the ports and look the same on either side?

The USB and possibly the headphone jack will need recesses cut in the GALA cover for the PCB to fit in securely.

The step in the thumb, to allow it to fit around the palm needs to match the contours of the palm. This should be done when the palm shape is finalised.

There are sharp edges missing from the Boolean on the proximal end around the PCB

Slots for PTFE tubing are too large. The Booleans are 3 mm across and the PTFE tubing is only 2 mm diameter. It would make assembly and maintenance faster if the tubes could be popped into place without glue.

Currently, it is impossible to use one length of M2 bolt on the hand. Some holes are too long for the new 5 mm ones. Whereas, some are too short for the older long bolts.

It fits well but the castellations aren't tall enough to reach the ligament

The back cover is too snug over the body of the DC jack. The hole for the jack is also too small.

When the fingers are pulled hard enough, the ligaments can be dislocated. The ridge is pulled out the trough and pops into the interphalangeal space.

The fingers can be dislocated by pulling on them directly, or by applying leverage using a lateral load.

The distal phalanges appear to hold the ligament better. Although both joints dislocate, it's always because of the proximal phalanx.

Can they be hidden somehow?

The slots are quite fiddly to use. They also don't hold quite enough cable. Flaring them after the mouth will help both of these

Although it's the correct shape, the tolerances on the thumb ligament appear to be off. It fits poorly in both the palm-side ligament cover and the thumb itself

The cable tidy slots don't hold quite enough cable without double folding. If they held slightly more, the cables could perfectly reach the connectors on the PCB
It would also be useful for assembly if the cable stayed in place after it was put there. and didn't spring out.

As the hand is now made of hard PLA rather than Ninjaflex, are pinch points on the fingers now an issue?

Need to figure out how to mount the thumb elastic. It's very tight on space.

The new PQ12 caps allow for the proximal corners of the inner palm to be rounded. However, as can be seen in the figure below, the conical spring Boolean meets the rounded end of the cap at its centre line.

The conical spring lines are therefore not tangent. The tendon line will have to run over the slight corner created as a result. It will probably be pinched between the wall of the inner palm and the cap.

Pinching could cause the tendon line to fray. Additionally, it will prevent the 3rd and 4th fingers shared tendon line from sliding.

Adjusting the conical spring Boolean so this is tangent should prevent pinching.

At the moment it just slots in, and so can fall out during assembly. Create a restriction in the slot entrance to keep it in.

The addition of a ridge on the ligament that tucks into a groove and then clamped improved the constrainment of the ligament. Can a similar feature be implemented along the length of the fingers?

The print settings for the mould need to be changed to prevent this.

The GALA cover doesn't quite sit right on the inner palm block. It appears the protrusion from the inner for the thumb pivot dowel is slightly too tall for the recess in the cover. The suspected reason is that the roof of the recess is horizontally across support. A larger tolerance than normal should be used in these situations.

The distal bolt hole has a thin wall all the way around it. This could be removed, it would then be similar to the GALA cover M2 bolt hole boolean

The thumb pivot-dowel in the palm-side, thumb ligament cover is designed to pop in. It does this very nicely. However, it is loose in its notch after it has popped through the constriction. It can, therefore, rattle when it is in place.

The current clip-on design requires the thin frame to bend across its layers. Therefore, it tends to delaminate.