Sunday, June 11, 2006

 

Frankenmotors running all three axes on Godzilla...

I got the little x-axis platform that will carry the Mk II extruder built this morning.

The x-axis is translating at about 28 mm/sec right now. I expect that that velocity will pick up substantially as the drive collar coupling nut and the threaded drive rod wear in a bit.

With the z-axis translating at 32 mm/sec and the x-axis moving at 28 Godzilla has a maximum extrusion speed of about 42-43 mm/sec right now. That's using 12v.

I ran the test over again using 5.45v and got the z-axis moving reliably at 10 mm/sec and the x-axis moving at 7 mm/sec. Hah! I may be able to start working at improving the MK II even before I get the axes' speed controller boards finished. :-D

Here are a couple of videos of the z and x axis running at the same time.

Short clip

Longer, more uncoordinated clip

Comments:
Nice stuff Plaas! Have you waxed the contact surfaces on Godzilla? Most greases and petroleum jelly evaporate over time. Even cheapo wax candles rubbed vigorously along the wooden rails are a step up from raw wood, and it stays on the surface where it's needed. I've also used common home canning parafin,(even better, and it's in handy blocks) but I'm sure there is something out there that is better than that.

BTW, seems brass screws are not what they used to be, seems the demand is low for high quality and expensive brass, so we are stuck with weak easy strip stuff now. The old boat builders gripe about it. ;)
 
This comment has been removed by a blog administrator.
 
This comment has been removed by a blog administrator.
 
The overwhelming source of friction in the system to this point so far has been between the coupling nut that I use in the thrust collar and the threaded drive rod. One of the reasons that I went for the Frankenmotors was precisely so I wouldn't have to worry too much right this minute about alignment and friction in my terrible example of the woodbutcher's art.

I have been looking around for some "Dance Wax", though. :-D

It's important to remember Godzilla is a bootstrap (repstrap to use Simon's nomenclature) system on the way to a much more efficient replicator somewhere down the way, not an end product of any kind. What I want now is something that can make parts for me in a reasonable time frame. If I can get Godzilla to do that for me I'll have succeeded.

Just to kind of put you in the picture I did a sensitivity study yesterday on Godzilla's performance if I manage to amp up the Mk II extruder performance to about 9x what it is now, something that I think is quite possible.

Godzilla already provides considerably more than the operating speeds required for that, viz, ~35 mm/sec. To do that draws about 70 watts of power. Amping up the Mk II extruder to perform at the same level would draw about 52 watts of power. That is enough power to light a few lightbulbs and considerably less power that is required to run the PC needed to feed instructions to Godzilla to actually make someths, so from where I stand there's not much point in trying to make it use a lot less power than it already does... for now anyway.

Running that enhanced Godzilla 24x7, something I don't think is anywhere near reasonable, and a 67% effectiveness figure, which I do think is reasonable, would let you extrude about 34 lbs of polymer/month. On a simple volumetric replacement calculation it takes about 15 lbs of polymer to replace the poplar structure in Godzilla which is to say... make another Godzilla. The time needed to do that is (15/34)(365.25/12)(24) or just over 13 days or 322 hours.

At roughly $5/lb Godzilla would replicate it's structure for $75 dollars. The electricity that it would take to run Godzilla to make the structure for another Godzilla would be ((70+52)/1000)(322)(0.21) {and yes, my marginal rate for electricity right now is a shade over $0.21/kw-hr this being California, the home of NIMBY} works out to right at $8.25.

That's a total cost of about $84.

Poplar in large pieces costs about $3/board-foot. Godzilla has about 9 board-feet of poplar in it or about $30 worth of wood. That, of course, assumes that you have the woodworking tools to turn nine board feet of poplar into Godzilla. I paid considerably over $100 for poplar precut in modular pieces, since I had a handsaw, rasp, mitre box and electric drill. Fastening technology, viz, screws, glue and the like is not a large, but still a nontrivial expense as well. The big bit of making a Godzilla by hand is the cost of your time. I wouldn't even guess at how many hours I put doing the woodwork on Godzilla, but then I was making up the design as I went along. I suspect with a set of plans I could have done the job in a man-day. You can price that in or not depending on how you think about such things.

You figure electronics for Godzilla to do again would probably cost about $100 and motors would cost cost another $25 or so while the remaining fixings would be maybe another $50.

That leaves us $84 for structure and $175 for the rest or about $260 total. That's not big bucks, but you'll also need another machine to turn polymer powder into filament which should cost another $100 and if you intend to recycle something like HDPE instead of the CAPA we're using you're also going to need a shredder/grinder to turn the recycled HDPE into something the filament maker can eat. That's likely to cost another $250. You're talking about a bit over $600 for a whole system, not including the PC which you can get for another $400.

Figure $1000 for everything.

From where I stand what you've got to worry about isn't the cost of the system. Godzilla plus filament maker is less than $400 and I already have several PC's so never mind about that.

Cost of polymer is what bites. Run full time at a 67% effectiveness ratio can get rid of 34 lbs of polymer a month, $170 dollars worth. THAT's the number what you should keep an eye on, in my opinion. Run Godzilla flat out for a year using CAPA and you've burned up $2,000. That's serious money. We've for sure got to recycle and it would be better if we could use more conventional polymers like HDPE. I don't know if we're going to be able to do that last one, but for sure we've either got to do the first or we have to look a LOT harder at polymer/filler formulations.
 
Speaking of, it might be a bit early, but...
Are there plans for making a godzilla? That is, the sort of thing you mention in paragraph 9?
 
Oh! You mean like dimensions and the like? Blueprints?

I could certainly whip up a set, I suppose. The design has been kind of a moving target, though. For example, I started out to use paired stepper motors to run the z and y axis positioning stages, but since then have gone over to high torque, shaft encoded DC motors and added a bunch of pulleys and the like to transfer forces.

Since Da Witch has got so near to producing useful polymer parts I've evolved Godzilla into a high risk development platform to test the possibility of extruding at a rates of 10-50 mm/sec.

Mind, you're welcome to duplicate it and I'll help you with plans and the like.

You're in the UK now, though, aren't you? All of the dimensions and modules for poplar are in US inches and lumber inches at that most of the time.

Probably what I ought to do is to do an Art of Illusion solids model of Godzilla as it stands and make that available. Then anyone wanting to either build it as it stands or use parts of it as an input into their own designs can get at it with no problems.
 
Thanks for putting up all the pricing estimates for the machines. I'm mulling over getting into some sort "pre-repstraping" mode myself. I was researching von Neumann machines when I found you guys.

I just knew someone out there had to be trying to build a universal constructor!

I've finished building a workshop, and am now setting up the ventilation system for soldering.

BTW, electricity rates in Quebec are 5 cents a kilowatt, and here in Ontario they are 14 cents! And thats in Canadian dollars.
 
This comment has been removed by a blog administrator.
 
Take those prices that I gave you with a BIG grain of salt. Those are "if I did it again and knew what I was doing right off" prices. It gets a LOT more expensive than that if you're feeling around trying to find a design solution that works.

As I said, I could build Godzilla over again now for a song. I doubt I'd build Godzilla over again, though. There are too many things in the design that could have been done a LOT better. We live and learn. :-)

I'm not surprised about the electricity prices up your way. Sadly, California is a place where people think that meat comes from the supermarket in film covered foam trays and electricity comes to them via wires from somewhere far, far away from their neighborhood, preferably out of state. I'm fascinated by this place. It is, taken by itself, about the 4th largest economy in the world with only about 30 million people. On the other hand half the population is functionally illiterate and the other half seems to have difficulty working out cause and effect half the time and still... the damned economy seems to work.
 
I am indeed in the UK. The last time I went to B&Q (think a smaller version of Home Depot), the 2x4s were still 1.75x3.75 inches, the packaging just converted that to mm.
It feels to me like the godzilla design is probably simpiler to build, and it feels like it's stabilizing, but if you think the daWitch design would be easier to replicate, I'd do that instead... I'm largely thinking that Godzilla would be easier because it's bigger, but I suppose that's quite likely a false assumption.
I'm afraid my woodworking skills (and equipment) aren't that great.
 
I am indeed in the UK. The last time I went to B&Q (think a smaller version of Home Depot), the 2x4s were still 1.75x3.75 inches, the packaging just converted that to mm.

Here 2x4 refers to the size of the uncured, unplaned wood. In reality 2x4 is 1.5x3.5. Lumber mills here have always been crooks in that way.

It feels to me like the godzilla design is probably simpiler to build, and it feels like it's stabilizing,

It seems to be, but that may be just me. The problem with building something like Godzilla is that its success depends on being able to run the Mk II extruder head, or some variation on it at a considerably higher rate than it has been tested for previously. I've corresponded with Dr. Bowyer who created it and he says that he simply didn't test it for higher extrusion rates in that 4 mm was adequate to his needs at that point. I suspect that it will be able to run at just about 8 mm/sec in its current state and maybe double that if the heating capacity of the extruder barrel is increased. I can't prove any of that right now, though.

The point is that there is a risk in building Godzilla that it won't work, ultimately. Be warned.

but if you think the daWitch design would be easier to replicate, I'd do that instead...

Da Witch would definitely be easier to replicate, if you have access to a 3D prototyping machine.

I'm largely thinking that Godzilla would be easier because it's bigger, but I suppose that's quite likely a false assumption.

If it is, it's the same false assumption that I'm operating under. I overdesigned Godzilla simply because I wanted elbow room in the design in case I made mistakes. I made mistakes. So far, though, that elbow room has let me correct them by reducing the initially very large working volume. So far the working volume I initially designed for has decreased by mayve 30-50%. It's still huge, though. :-)

I'm afraid my woodworking skills (and equipment) aren't that great.
LOL! Neither are mine. I'm certainly no cabinet or furniture maker. :-)
 
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]