Got the mechanics for the Z belt drive on the Ingentis done tonight 
https://www.youtube.com/watch?v=20agR5ksUwU&feature=youtube_gdata_player
Awesome! Question, have you though of printing directly onto the aluminum? It’s thick enough that is should have good flatness. Just throw some kapton tape on it. It’ll save you both weight and warmup time.
Good idea on the counterweight, I had thought about the need for constant torque in the z belts but never thought about catastrophic failure of the drive system. Adding the counterweight will also lower the required holding torque in the z motor, but it will add considerable inertia to the z movement. This really shouldn’t be much of an issue though. Even if you do the auto bed leveling in Marlin (which is pretty badass) your z wouldn’t be accelerating too fast so it shouldn’t be an issue.
Looking good. Counterweights and/or counterbalance springs seem like a good idea.
Another possibility might be simply protecting the machine against the damage caused by a fall. Something like a spring at each corner to catch the platform instead of letting it smash into things. A fall is only bad if it breaks things, right? What if there’s a hand under it when it goes? Could that cause injury
Edit: On second viewing, I see you already had that idea.
Adding a momentum limiting brake like a flyball governor would be geeky cool, but probably not the most effective thing. Hmmm. Maybe there’s an arrangement of transistors that could turn the Z motor into a brake (short the coils) when the stepper enable pin loses power. I’m not sure what that might do to a driver if it was accidentally disabled by the controller instead of by a loss of power. Probably the controller would have to be cut off from the shorted motor. Relays might be the thing to use, instead of transistors.
A worm gear instead of herringbone gears should prevent a fall. You can’t backdrive those except for very low ratio cases.
I think I’m done spitballing ideas for that.
I hope you can stay with the linear bearings. A bushing inherently has clearance, and even 0.01mm clearance in one of your bearings could add up to a layer’s height of slop near the far edges of the platform.
@Eric_Moy yeah I print on Alu directly now. I’m thinking glass for this as the Alu tends to get a bit foxed after a while and it would be useful to swap out surfaces. If I can work out a way of managing the weight, then I’d prefer to use toughened glass cos I’m careless 
. I don’t think the Z ever needs to move very fast, usually the accel and jerk settings are much lower than xy and for head moves it’s only tweaking up and down by 0.1mm. The principal challenge with a counterweight seems to be where to put it so it doesn’t foul against those big brackets as they pass each other and doesn’t vibrate like crazy during prints.
@Dale_Dunn I’m thinking some big springs around the 10mm shafts to cushion a fall. Makes sense as I’ll only need 2 that way. I was considering a induction brake like they use in elevators but I’m at least attempting to keep it simple so maybe I just let it drop and catch it softly. Worm gear is intriguing, have you seen any printed ones?
Thanks both for your input!
2 springs on the shafts makes perfect sense. Nice and simple, with no more design work to do than to size them. Is the sleeping child ever allowed near the machine while it’s powered up?
I’ve never tried a printed worm gear, but there have been a few on Thingiverse. I’m sure some have source files, but I can’t remember which CAD system you’re using. I saw at least one with a .scad file attached, and I’ve modeled one or two myself in SolidWorks. But not with proper tooth forms. Great, now I’ll be thinking about what that would be all afternoon.
@Dale_Dunn Yep, the boy will get near it but not unsupervised. I’ll have some covers over the gears etc before then. I use rhino for the most part but also have inventor when I want to get all parametric without resorting to openscad. I have access to solid works too. Hmm. I think inventor has a gear wizard built in…
Just checked. It does. Hmmm.
@Dale_Dunn hmmm, I really like the worm gear idea, good thinking. I’m thinking of my childhood now and the worm gears I had in my capsela toy kit
@Tim_Rastall I’d be interested to see how fast your z can go on this setup. I have some mad schemes that I think will be killed by the slow Z axes on my machines.
Also, you could use a very weak electrical device like a solenoid or a 9g servo to pull a ratchet out of the way of the belt, so as soon as power is cut the ratchet falls in place and keeps Z from moving downward.
@Nick_Parker Clever, that’s similar to the Otis elevator failsafe, the brakes clamp when in a free fall. Just be sure to use a 100% duty cycle solenoid
@Nick_Parker Yeah, I’ve been thinking about something along those lines too. Except I was thinking about a friction brake rather than a ratchet.Ideally I’d like a stand alone printed mechanical solution as it makes things nice and accessible. I have about 5 possible solutions in my head at the moment. in the short term I’m going to just let the damn thing drop and use springs around the shafts to cushion the fall
Standalone printed mechanical solution as in no electronics to it? Have you got any in mind that wouldn’t cost you in performance (speed/accel/lash)?
I ask because I’ve been toying with a slicer that makes nonplanar layers. Making the layers slightly parabolic or even more complex shapes to work around the weak layer interfaces in FFM printers.
I just mean as a breaking mechanism. And I’d be comfortable with something that slows the descent of the platform rather than arrests it entirely. I have a slightly crazy thought that I could integrate the brake into an idler based belt tensioner -something that rides up as the platform drops while introducing further belt tension.
@Nick_Parker Interesting concept, I’ve seen the idea kicked around a bit in other forums. I’d been considering using nylon for infill as it’s layer bond is so much stronger.
@Tim_Rastall Yet another brake idea might be a clock’s escapement. With a light enough pendulum (radically lighter than in a clock), it wouldn’t interfere with normal movements, but would still impose a speed limit on how fast the platform could fall.
@Tim_Rastall Check out these camlock clamps for linear guide rails by Misumi. Perhaps you could do something like. The lever is attached to an eccentric pivot that cams into the rail. You can attach the head of a servo (if it has enough torque) to the eccentric cam that can be shaped to wrap around you linear rod. If you really got fancy, you could create a mechanism that squeezed the rod from both sides.
@Eric_Moy link?
crud, sorry, didn’t notice that my ctrl+V didn’t do squat:
http://us.misumi-ec.com/vona2/detail/110302272140/?Inch=0
Credit to Mike Carpenter in IRC for pointing this out to me:
If you short the coils of a stepper motor, the induction makes it way harder to turn. You could just make the Z motor default to shorted when unpowered, and it wouldn’t be able to move very fast without power.
@Nick_Parker good call. Much like a screw gun which shorts the motor when you release the trigger. If you drop a load or short across the motor, it has no choice but to resist