New Build - The DumpsterDiver CNC

[Logon_Bob]

Big improvement, congrats.

Pic one: I see a gap between the new end plate and the Ibeam. Pretty tough to get dead square without a mill or good saw. IF it is moving there, maybe some shims, even soda can aluminum to fill the gap.

Pic Two: Can't tell if the chain is tightened yet in this pic. If so, it appears to be climbing as and will get tighter as the Idler pulley gets closer to the end. The eye bolt is on center line and the chain going around puts it above. And or, you went wrapped the opposite way from the other side. Might line up better if the chain wrap went from bottom to top. Or if the eye bolt is in fact inline to where it needs to be, the chain would now be low. Only thing I can think of then would be a shackle around the eyebolt, once pinned the chain would be at least inline with the eyebolt. This of course to keep from slotting the eye bolt hole by hand.

As far as not tightening like a stop when the chain gets taught, I think it would be like stringing a bow. The farther ( leverage ) the chain is away from the Ibeam... Might be trying to bend the Ibeam.

I spent the day drilling and tapping steel. My X rails are 6" steel C-channel. Even at only 40" long and held in the vise near one end I was surprised how much it would bend with just hand pressure. I may be having some of the same issues you are, we will see soon. I will try to keep the chain as close to "X" as possible.

Dan

[PaulRowntree]

Here are some pics of my dust shoe. The fixed piece is cut from 2x6 stock, and is attached to the router with a hose clamp. The bottom of this piece projects ~1" below the bottom of the router, about as far as the rotor lock key way lies. The removable part lies below the fixed piece as well as under the router. The hole for the tool is as small as the nut will allow.
Cooling air from the router is mostly on the outer diameter of the router, not near the bit itself. This cooling air gets deflected outwards, parallel but above the work to minimize chip-storms. I know there is some debate about doing this, but it seemed that since this cooling air is not helping the vacuum do its job, why not get rid of it?
The skirt is 0.25" thick neoprene picked up at a local textile store, and it staples well using T50's. Two lenghts of skirt for different tool lengths; colour coded by my daughter. The lower piece of the shoe is aligned using 2 1/4"-20 studs projecting downwards from the fixed piece, and it is held in place with three strong magnets scavenged from hard-disk drives (used to move the read/write heads). On each lower piece there are bolts heads flush to the top surface that line up with these magnets and make for quick changes but solid mounts.

BTW : The reason the fixed piece is attched using a hose clamp is because this shoe projects out far enough to interfere with the 'elephant feet' that support my rails. Depending which axis is causing problems, the shoe can be rotated in 30 seconds.

[PaulRowntree]

Big improvement, congrats.

Pic one: I see a gap between the new end plate and the Ibeam. Pretty tough to get dead square without a mill or good saw. IF it is moving there, maybe some shims, even soda can aluminum to fill the gap.

Pic Two: Can't tell if the chain is tightened yet in this pic. If so, it appears to be climbing as and will get tighter as the Idler pulley gets closer to the end. The eye bolt is on center line and the chain going around puts it above. And or, you went wrapped the opposite way from the other side. Might line up better if the chain wrap went from bottom to top. Or if the eye bolt is in fact inline to where it needs to be, the chain would now be low. Only thing I can think of then would be a shackle around the eyebolt, once pinned the chain would be at least inline with the eyebolt. This of course to keep from slotting the eye bolt hole by hand.

As far as not tightening like a stop when the chain gets taught, I think it would be like stringing a bow. The farther ( leverage ) the chain is away from the Ibeam... Might be trying to bend the Ibeam.

I spent the day drilling and tapping steel. My X rails are 6" steel C-channel. Even at only 40" long and held in the vise near one end I was surprised how much it would bend with just hand pressure. I may be having some of the same issues you are, we will see soon. I will try to keep the chain as close to "X" as possible.

Dan

Yes, when I cut the beam I had no idea (of course!) that I would need the ends trued later on. One of 4 ends is reasonably flush. I will do some playing with the dial to see what is moving that last 0.001"-0.002". Wherever it is, I do not think it is my biggest problem now.
It is hard to appreciate how much leverage a 13 tpi bolt gives as a tensioner. 'Finger tight' looses its meaning. Still, the gantry moves smoothly with one side modified and under more tension than I had before, so I haven't messed anything up yet. Just using the manual control of M3 I was getting >400 ipm as is.
Cheers!

[PaulRowntree]

Big improvement, congrats.

Pic Two: Can't tell if the chain is tightened yet in this pic. If so, it appears to be climbing as and will get tighter as the Idler pulley gets closer to the end. The eye bolt is on center line and the chain going around puts it above. And or, you went wrapped the opposite way from the other side. Might line up better if the chain wrap went from bottom to top. Or if the eye bolt is in fact inline to where it needs to be, the chain would now be low. Only thing I can think of then would be a shackle around the eyebolt, once pinned the chain would be at least inline with the eyebolt. This of course to keep from slotting the eye bolt hole by hand.

Dan

THe chain at the rear of the machine (pic2 ) is about 1/16" higher than the other end, but the chain axes line up perfectly with the horizontal axes of the eye-bolts (the chain loops are free to rotate over the eye-bolt). You are right that the chain comes up on one end and down on other, but beyond the region where the chain is doubled over this has no effect. It does look dumb though, so it will be changed. I can get 1/16" from the vertical play in the flattened eye-bolt wrt the slots in the end plates.
Cheers!

[Logon_Bob]

THe chain at the rear of the machine (pic2 ) is about 1/16" higher than the other end, but the chain axes line up perfectly with the horizontal axes of the eye-bolts (the chain loops are free to rotate over the eye-bolt). You are right that the chain comes up on one end and down on other, but beyond the region where the chain is doubled over this has no effect. It does look dumb though, so it will be changed. I can get 1/16" from the vertical play in the flattened eye-bolt wrt the slots in the end plates.
Cheers!

Probably the camera angle, you know best as you are there. Sure looks like if you routed the chain from the bottom to top of eye bolt it would line up better. Like the other side.

Dan

[PaulRowntree]

Today the DumpsterDiver CNC made a major leap forward.
1) I tweaked the chain terminations and found a reliable way to ensure that both sides are evenly tensioned. I hang a std mass (1 kg) from the chain in the middle while the gantry is at the extreme end adjust the tension until I see a vertical deflection of 12 mm. This is very reproducible. This leads to an error (slop) at the tool of about 0.002". Plus, I reinforced the two idlers on each chain with a connecting brace; now when tightening there is no outward deflection of the two idler shafts.
2) I replaced the Z coupler with a rigid brass piece, and put flats on the stepper motor shaft. What a difference! My V-Carve corners are now crisp and essentially flawless, whereas before they showed considerable excess depth. It turns out the Z was lagging behind and so it was cutting the corners too deeply. Getting rid of this hysteresis makes such a difference that I am close to declaring Victory over the machine. Time to check the accuracy and LockTite the various couplers.

Now we are having fun.
Thanks to everyone who has contributed to get it this far!

[Logon_Bob]

Today the DumpsterDiver CNC made a major leap forward.
1) I tweaked the chain terminations and found a reliable way to ensure that both sides are evenly tensioned. I hang a std mass (1 kg) from the chain in the middle while the gantry is at the extreme end adjust the tension until I see a vertical deflection of 12 mm. This is very reproducible. This leads to an error (slop) at the tool of about 0.002". Plus, I reinforced the two idlers on each chain with a connecting brace; now when tightening there is no outward deflection of the two idler shafts.
2) I replaced the Z coupler with a rigid brass piece, and put flats on the stepper motor shaft. What a difference! My V-Carve corners are now crisp and essentially flawless, whereas before they showed considerable excess depth. It turns out the Z was lagging behind and so it was cutting the corners too deeply. Getting rid of this hysteresis makes such a difference that I am close to declaring Victory over the machine. Time to check the accuracy and LockTite the various couplers.

Now we are having fun.
Thanks to everyone who has contributed to get it this far!

Congrats again. Sounds like you got a handle on it. I spotted the hose coupler a while ago and was going to bring it up.... :-) That is my preferred method on a Hotwire Foamcutter and it lines right up but there's no real load with a wire cutter.

I know my foamcutter coupler can move a few degrees but it is acceptable. I got rid of the clear hose, auto fuel injection with the blue liner works OK. Plus a drop of Superglue, motor shaft must be right up the lead screw and at least 3 clamps with the middle clamp on the joint. With the motors locked, thumb and finger on the lead screw and it will still twist a bit.

Pic of my build so far. Many hours and many more to go.

Oh, when are we going to get a youtube video? :-) No music, I would rather hear it run thank you very much.

regards
Dan

[PaulRowntree]

Now you want me to master another technology! I have no idea how Youtube works, but it should be possilbe.
I missed my chance; I was just out cutting a large Celtic knot in MDF, but didn't think to film it. It looks good! Next time out.

I like the look of your machine. If/when I rebuild I think the 8020 will play a bigger part of it all, with some steel track reinforcements. First I should get something out of this one so the Ministry of Finance sees some return on the investment.

Cheers!

[PaulRowntree]

Dan, is that steel or Al channel supporting the X tracks?

[Logon_Bob]

Dan, is that steel or Al channel supporting the X tracks?

That would be 6 x 8.2 lb. Steel Channel. 96 lbs at .50/lb. :-) $ 5 for 3 cuts. Had my eye on the Alum equivalent much lighter but $3 per lb. I think it would have been about 3 times the $$$

It is rock solid. I laid on the gantry and pulled myself back and forth.. :-) Hey, got do sumpin until I get some motors attached.

I am very pleased so with the steel but it's a pain. Drilling and tapping 40 1/4-20 holes even with a mill is tough going since it's so long. I ran a dial indicator length wise, starting at zero I got -.002 to one spot of +.001.

Problem found later is the legs are parallel but not at 90 degs to the web. So back on the mill but I think shimming would have been a better way to go. I think alum normally is more accurate.

You can learn all about Youtube while your router is cutting for hours..:-)

Dan

[PaulRowntree]

I was amazed to see how far out some Al channel pieces were from being square. And of course I saw this after preparing the pieces. Similar results for some Al angle. One result of inexperience working with free metal. More and more I like the look of Ahren's roller designs running on cold-rolled steel, but I have not measured the locally available retail stock to see if it makes spec.

Drilling steel is one thing, but tapping an army of holes is quite another.

[Logon_Bob]

I was amazed to see how far out some Al channel pieces were from being square. And of course I saw this after preparing the pieces. Similar results for some Al angle. One result of inexperience working with free metal. More and more I like the look of Ahren's roller designs running on cold-rolled steel, but I have not measured the locally available retail stock to see if it makes spec.

Drilling steel is one thing, but tapping an army of holes is quite another.

Yup, same here. Alum angle, 3 x 5 x 1/4 is sitting on top on the cars and the gantry bolts to the angle, not square. I had to face mill it.

Anyone making a DIY without machines like a Mill and a Lathe ( heck, even with ) has to think ahead how can I adjust this. Shim stock, washers, slots instead of just a bolt hole. It's a ***** no doubt.

Yes, I like Ahren's roller method as well. I might do Y or Z that way. Same deal, when shopping for the steel plate, the sides need to be parallel or you will have issues either getting tight or loose. Might have to do some Flat Filing.

I ran across a Youtube vid of a guy who had Blackfoot, V bearing on alum angle, he wore it out in 20 hours. The Alum angle being the weak link where the V-bearings are hardened.

Best design principles web page I found was here. Discusses the why's and what's going on rather than just following a plan. Unfortunately it has not been updated and it's incomplete with many "under construction" pages.
CNC Router Plans: Download free CNC router plans

Dan