Reducing dynamic weight

I just received my kit and so I’ve been going through a weight reduction exercise, thinking where materials could be reduced WITHOUT going into exotics, hard to find, or expensive stuff.

Here’s what I measured (with some decent estimations for the plates and the rails, since I don’t have those yet):

Weight (kg) Percent overall
Rails 4.47 42.77%
Motors 1.85 17.69%
Nuts & bolts 1.21 11.57%
Spindle 0.87 8.32%
Plates 0.66 6.29%
Misc hardware 0.61 5.80%
Bearings 0.47 4.52%
Wheels 0.17 1.62%
Controller 0.15 1.41%
Total 10.46 kg

Screen Shot 2020-03-18 at 15.46.50

Applying the 80/20 rule, which is that 80% of the gains come from 20% of the work, it suggest we should focus on the rails, the nuts & bolts, and the plates. (We can’t change anything on the motors and spindle.)

Rails

The rails are the single biggest contributors, but they have been the subject of some discussion elsewhere, (A primer on stiffness), so I won’t go into it here except to assume that we could decrease weight without losing precision by a strategic swap of X and Z tubes locations, coupled with a reduction in X-tube wall thickness to 0.035.

Screen Shot 2020-03-18 at 15.46.56

Whether this increases price or not seems to be a matter of where you live and when you buy the material, since 0.035" wall tubing can strangely cost more than 0.065" tubing.

Nuts & bolts

The kit uses a significant amount of steel hardware. For the most part, the driving characteristic is outer diameter, not strength. The upshot is that 5/16" hardware-- which has a tensile load capacity of ~1,500kg-- is used as 8mm bearing axles. If we could optimize this design by using tubes for diameter (8mm) and spec’ing the threaded hardware for the design loads (1-10kg), we could save a lot of weight.

I have designed a new bearing roller which uses 8mm (5/16") brass tubes zip-tied in place. This would eliminate twelve 5/16" nuts and bolts, for a calculated savings of 0.235kg.

https://cad.onshape.com/documents/9bfd836abff9fc3b04fc7b23/w/8b821b524858853269344a9f/e/1f1612a96238da718f427de9?configuration=List_QPFgsXm53a08bY%3DDefault

I know this has been tried in the past with pins and dowels, but ran into issues with sourcing parts. AFAICT, brass tubes are novel in this context. 8mm brass tubes are cost-competitive with 5/16" hardware, and can be found worldwide at hobby shops.

I figured there’d probably be other opportunities for similar weight savings, so I just guesstimated a 50% reduction overall.

Screen Shot 2020-03-18 at 15.47.02

Plates

The plates have also been discussed (1/8th aluminum Y plate), so I won’t dive into that either except to say that a 50% reduction seems plausible by making a 1/2" thick tile-board/foam/tile-board sandwich.

Screen Shot 2020-03-18 at 15.47.08

Conclusion

Putting it all together:

Weight (kg) Percent overall
Rails 2.59 33.91%
Motors 1.85 24.21%
Spindle 0.87 11.39%
Misc hardware 0.61 7.94%
Nuts & bolts 0.61 7.92%
Bearings 0.47 6.19%
Plates 0.33 4.31%
Wheels 0.17 2.21%
Controller 0.15 1.92%
Total 7.64 kg

It looks like a 25% reduction is plausible, without increasing cost or decreasing performance. There’s no real-world testing to back this up, so it’s all purely hypothetical. Still, it points out some promising directions for enterprising souls to explore.


Appendix

Full doc here:

Motors

(Wait, I thought motors were off the table? Well…)

There are some elegant belt routing options which reduce the number of motors without reducing speed. CoreXY (like in the ZenXY), Hbot, and the decidedly strange MarkForged belt path all would result in fewer motors and quite possibly moving the motors off the entire gantry. I have no idea what practical downsides there are to this, as doubling (tripling?) belt length won’t be a trivial change. But if they could be made to work and honor the design intent of V1Engineering, it could be a win because it would shave 1.85kg (!) off the gantry. There’s a lot of prototyping and testing which has to happen before anyone could even begin to make a fair comparison.

4 Likes

I love the work and the presentation! Thanks for sharing and it will be interesting to see the results from any testing you can do!

Wonderful presentation!

We do need a bit of mass to keep the machine down, or we have to add wheels on the bottom to clamp it in place. The biggest gain from reducing weight would be slightly faster accelerations but I am pretty happy with them as they are some even suggest slower accels. If we could increase rigidity without increasing weight we would really have a winner!

That’s interesting! Do you have a notion of how much weight is necessary? And does it need to be evenly distributed, or could it all be on the Y-plates?

If we can strategically remove non-essential weight it can be added back in in the form of stiffening members. Not quite sure what this would look like, or if it would even be practical.

No, the first version had bottom wheels because I figured it would be best practice. turns out we didn’t need them. I assume we are actually pretty close to a lower limit though. you need a lot of fudge factor for bad cam. In the Z we have plenty but a dusty table with these wheels I bet along the X axis if we reduced to much guides would be necessary instead of optional. I haven’t gone too far down the new LR rabbit hole, still on the other one. I think I really want to have a better system for the router mounting more than anything. reducing the rail length can increase rigidity some and like you said, increasing diameter a bit.

So when I go to make a new version I really really beat up the old one. Since it has been in use for so long I can wiggle it and push on it and see what breaks or flexes…then beef that up.

3 Likes