New Build - The DumpsterDiver CNC

[louieatienza]

Nice job! I also need to know what town you live, because here there's nothing good in the dumpsters! And aluminum... the scrap metal places here are paying big bucks a pound for it, so I don't think I'll ever see scrap aluminum here!

[Logon_Bob]

Quoting from another thread, Logon Bob

Dan, your a genius.

ahhhh, that's what my Mom used to say.. :-)

Glad it turned on the lights. This thread is where I thought I was before, oh well.

So whatcha going do? A stout endplate in your Dumpster? Or a piece of angle so the eye bolt is inline. Or maybe run the nut out to the eye , add a washer and piece of larger diameter pipe or tube or even a block, then tighten the nut so it's all in compression against the I-beam might be enough to solve the bending.

I scored enough 8020 ( 1530 ) to do the gantry at a decent price today. Now I have to go back and pick up the 6x8.5 lb steel channel I had my hands on yesterday. I was looking at same 6" channel you used in alum but even at surplus prices of $3 a pound , that's like $ 90 just for the 2 ea X rails. Steel should be about $ 35 at .50c a pound.

Getting excited to get started. Just need a deal on a Gecko G540. :-)

Dan

[PaulRowntree]

So whatcha going do? A stout endplate in your Dumpster? Or a piece of angle so the eye bolt is inline. Or maybe run the nut out to the eye , add a washer and piece of larger diameter pipe or tube or even a block, then tighten the nut so it's all in compression against the I-beam might be enough to solve the bending.

A stout 1" thick endplate is one option, but it would not have the best purchase onto the end of the I-beams. I think a solid post or block extending off the side of I beam would work (your second idea). It needs to project out by ~3" so it should be ~3" along the chain axis, and I am thinking 1" thick (in the Z direction). Two side bolts to hold to the I beam, and we should be in business. It would still take a bit of machining to add adjustment slots (parallel to Y) but that should n't be too much trouble.

I will have to check my Al 'store' and see what is in stock ...

[garyo1954]

Chain drive was appealing to me until I did a search here and discovered chain brings a loss of precision with the increase in speed. For most applications it will work fine unless one expects resolution to .001.

Perfecting a chain drive would be a leap in building a CNC. I am interested in seeing it.

Just for info:

http://rockcliffcnc.com/ReportonLine...cs.aspxhttp://

[PaulRowntree]

Chain drive was appealing to me until I did a search here and discovered chain brings a loss of precision with the increase in speed. For most applications it will work fine unless one expects resolution to .001.

Perfecting a chain drive would be a leap in building a CNC. I am interested in seeing it.

Agreed. The ServoBelt design by Everman comes pretty close though

ServoBelt + this site

I haven't seen any DIY implementations, and I have no idea of the cost.

[robe_uk]

Agreed. The ServoBelt design by Everman comes pretty close though

ServoBelt + this site

I haven't seen any DIY implementations, and I have no idea of the cost.

Have you seen this thread.

[PaulRowntree]

Have you seen this thread.

Thanks Robe_uk! I hadn't visited that thread since Christmas and had not seen ger21's work. It is good of Mike to be helping people understand the details too.
Cheers!

[Logon_Bob]

Chain drive was appealing to me until I did a search here and discovered chain brings a loss of precision with the increase in speed. For most applications it will work fine unless one expects resolution to .001.

Perfecting a chain drive would be a leap in building a CNC. I am interested in seeing it.

Just for info:

http://rockcliffcnc.com/ReportonLine...cs.aspxhttp://

Agreed but the same applies to Rack & Pinion hence the 3:1 reduction. In fact a typical Pinion is 1" OD so 3.14" of travel per turn or more than the 9 tooth gear.

Been reading like all weekend, still no expert. :-) Some stuff I found.

Two sprocket manufactures state " minimum tooth of 17 unless RPM is less than 100. Though 15 can be successful". I assume this to reduce or eliminate Cogging but I didn't find that out.

To get to .001 with a 9 tooth you need 10x micro stepping. At 15 teeth, resolution just got worse.

I still like the single motor with chain drive for my 27"x40" . So at this point I am still pursuing#25 chain but with a 3:1 reducer. Starting with a 15 tooth sprocket on the chain. Looking for 20t & 60t Belt pulleys and a 60t belt.

Other facts: You shouldn't rely on Micro Stepping for accuracy but for motor smoothness at low speeds. So you want to build in a much resolution as you can and still have some speed. Micro Stepping is not good at high RPM's, Good drives like the Gecko 540 can Auto switch to Half or Full steps at high RPM's. Now what happens to the resolution?

Also, can you build stiff enough to hold .001??? :-)

I will attach a simple Excel sheet where you can plug in "yellow cells " different combos and look at the charts to see where your are and what you need to do to get there. I sure hope the math is correct, there's my disclaimer. :-)

Regards
Dan

[PaulRowntree]

I think that 0.001" is a bit unrealistic for my wood router. I would wonder if wood is in and of itself stable to that level. Humidity, grain, final sanding etc will all give much greater losses of precision. And then there are coffee spills. My goal is to be able to lay out pieces in CorelDraw and have them materialize on the cutting table so they fit together as expected. With the machine as-is my V-carved pieces are pretty good, but with room for improvement.

The rubber hits the road tonight. I just finished 2 end plates for X and will mount them up ASAP. I expect to see the tensioning be much more robust, in which case I will make 2 more X plates and start thinking about Y.
Cheers!

[Logon_Bob]

I would agree. Rigidity and lack of backlash for repeatability should be more import. Your eye won't pick up a hole is .005 over or undersize but it might if it's +- .010 out of round it might.

I have a full size mill ( 2500 lbs of steel ) in the shop eerrrrr garage with digital readout. I used it to make all the parts for a cnc hotwire foam cutter from MDF and Melamine. I found out just with varying vise pressure would compress the wood, I drill on the money and when released my hole is .005 over there. :-(

Hope the more robust chain tensioning system works out for you.

Dan

[PaulRowntree]

Why is it that experimental results are never as clear cut as they should be ..?

1) The plates are mounted, although it is a bit tough drilling into the end of the I-beam with a hand drill.
FIX : Print accurate paper templates with CorelDraw, use a good centre drill to get things started.

2) The chain mount plates now flex ~0.001"-0.002" inwards when the chain is deflected by hand. This is a 5-10 fold improvement over my original eye-bolt design (Thanks again Dan!). There is still some room for improvement in the way I secure the end plates.

but, the chain still allows some perpendicular motion, and the tensioning did not hit a 'hard stop' when taut as I had expected or hoped. That must still be some slop somewhere. The best news is that I can see no gantry motion as I deflect the chain by hand.

The eye-bolt shafts have flats on them to prevent rotation while tightening, and they are in a slot to give some lateral (parallel to the gantry axis) adjustment. This is why I had gone to eye-bolts in the first place. I welded the eye-bolt loop closed for strength They are 3/8" steel, so I expect they are stronger than the chain in tension.
So I will complete the other chain mounts and see how it runs.

PS : the dent in the bumper is not my fault.

[voltsandbolts]

Paul,

I'd say that looks plenty beefy!

I have only put a little thought into how I am going to mount / tension & adjust my chain. Glad your getting it all worked out for me.

Rob.