Bt32/48 vacuum fixtures

[dana_swift]

Here is one of my rigs for vacuum holddown on a BT32.

To start with there is the base board, all mounting holes are drilled, as well as C rails installed. All of my vacuum fixtures bolt to the table using the 1/4 20 mounting holes. Below the table you can see the vacuum accumulator.


29613.jpg

The vacuum accumulator allows the part to suck down and make a good seal when the vacuum valve is first opened. It would be hard for a vacuum pump/blower to move enough CFM to do the initial evacuation. The pump is not currently running, when it is it moves over to 28-30in Hg pretty quickly.


29614.jpg

The accumulator is an empty freon bottle I was given by a friend of mine. It fits nicely under the moving table parts and is right where I can reach it while I am holding the part I am about to secure while I hit the vacuum valve to start the evacuation.

The first layer of the vacuum fixture is a channel that routes the input fitting to the individual fixtures. This jig has two hold down areas where I intend to use them one at a time. This fixture was made very quickly, so I didn't worry about machining the outside perimeter of the material. Everything is referenced to the holes for the hold down bolts so the layers line up correctly. The gasket tape is from all-star.


29615.jpg

The top layer is bolted through the bottom layer and the base board with bolts long enough to engage the tapped holes there.


29616.jpg

Now I connect the 3/8 vacuum line to which ever fixture I am using. I have used 1/4 inch vacuum line without any problem, but moved to 3/8 so the initial evacuation will happen quicker.


29617.jpg

I let the hose slide back and forth with the table, it has never pulled out or bound on anything. Just simple.

When my part is on the appropriate fixture half, I open the right hand valve on the accumulator. The part either seals in about a tenth of a second, or something is wrong.

I have made quite a few fixtures like this of various sizes to hold different parts. Making one of these fixtures takes about a half hour to an hour including the time with partworks.

Large wooden parts with rough edges tend to leak a bit around the edges. I still get a minimum 15" hg if will seal at all. I am using a standard AC vacuum pump rated at 5cfm.

Hope that helps-

Dana

[erik_f]

I am still waiting for my BT and have been thinking about ways to set up a nice system. I like your idea...like you said...simple. I have been thinking just coming in from the front or back and creating a regular grid like on the bigger machines. One other question...kind of off topic...but it seems almost that the BT could be even more accurate than some of the bigger machines because of the fixed gantry and aluminum table. What have any of the BT owners notice in terms of tolerances from piece to piece. My old PRT I would say I was around + or - 0.01 inches on average. Sometimes better...but sometimes worse. Really hoping my new BT-48 will hold closer to +/- .005.

[Brady Watson]

Hey guys,
I'll try to post some pics of the vacuum system that I fabbed up for the BT line of tools in the next week or so. I'm on my 5th version and I think that it should be the final one. I'm going to install a setup on my own BT48 and verify everything. Prior versions were done for other customers with BT tools...Hang tight. Pics to follow soon...

-B

[dave_k]

Thanks Dana... this is the kind of post I've been looking for. My BT32 is in the shop but not hooked up and working yet. (still another week or so of other work before I can "play"). However, I've been brainstorming how to do this, and you have given me some GREAT ideas. Thanks.

[dana_swift]

My experience with accuracy and precision (two different values) is that it depends on how you do your setup, and if more than one setup is needed.

Just for grins, I measured the two hold down areas in the template shown above. The material is 1/4" HDPE. The orientation is 90deg between the two fixtures, but the artwork was identical. That allowed me to measure the consistency between X and Y. I used one of the pressure distribution channels as a test case as it was designed to be 1.000 inches.

The real channel width was about 0.995 for channels in one axis and 0.992 for channels in the other direction. The variation along the channels was about +/- 0.002 for either horizontal or vertical channels. Since in this case the exact dimension was not important I made no effort to make it precise. I have not checked the partwork generated cut files to see if they are exact for that channel width or not.

In other cases it matters greatly to me, and by making careful setups and adjusting dimensions to correct for router runout I am getting a reliable 0.003 to 0.002 for finished part dimensions in X and Y. Where I need this level of accuracy I use C# to generate custom cut files. Then I can correct for as many factors as I want to.

Z is the most un-repeatable axis in my system. The zzero plate isn't completely flat and it is held against the material by hand during the zeroing process. I have a hunch some of my Z axis precision can be traced to that procedure. I changed the ZZero routine to get more precision in the Z which has worked, but I still would like to get a reliable .002 from the ZZero routine. So far that has not happened without hand adjusting the Z using a 0.500 calibrated metal block and test fitting it under the bit.

From one point to another the Z is less accurate than X and Y. Add another .003 for Z consistency. There has been a lot of discussion on here about Z variations. My total Z variations are around 0.004 over 20" when the cuts are performed using a single MX or MY command at a constant Z. The accuracy of Z being highly dependent on doing the 0.500 tweak after the ZZero.

I think if you will get what you are after if you are willing to make a lot of test cuts and zero in your machine. Some time spent setting Z to specific values and checking it with feeler gages is quite instructive also, it sure has been for me.

D

[sawdust535]

Having trouble posting pics, I'll try again.

29625.jpg

[sawdust535]

Hey, it worked, just need to learn how to resize my pics so you don't need binoculars. This is one of my holddowns for the T-Track table. The lower right corner where the two guides meet at a 90 deg angle is my 0,0 mark. The table is oversized so that it will hold a 24 x 32 workpeice.

29628.jpg

[dana_swift]

As an afterthought: I changed my X and Y definitions so the part on the table has the same orientation as my drawings. So in the pictures X is horizontal, and Y is the Table motion. If you use this definition the XY zero routings from SB dont work as they check the wrong limit switches. I have a version that works with either definition.

George I like the Rail system you have setup. It looks like you have two layers of plywood there. I have the two C/T rails in my base board also (Thanks to Gary Beckwith for the suggestion) I use them to hold down things occasionally. Most of my hold downs use vacuum only.

Also when Z precision matters I use more holddown bolts to minimize flex in the HDPE. All of the vacuum fixtures I use mount to the table using bolts. The rails only get used about once a month.

In the pictures you can see the wheel assembly under the BT32. That is handier than pockets on shirts. I made it from wood out of a palette. The whole assembly pushes down to lift the bot up so I can move it around. Normally it sits solidly on the floor.

D

[sawdust535]

Actually, there is one sheet of 3/4 ply bolted to the metal table. It looks like 2 because in the lower pic you can see the front strip, or guide made from 1/2" baltic birch and the other guide to the right and perpendicular. Both also have T-Tracks. I made those guides 1/2" high so I can hold down a 3/4" mdf spoil board when thru routing. Thats when I need a vac, since then I have to use screws, or some other way to secure the workpeice.

[sawdust535]

Dana, do you have anymore info on how that 1/4" HDPE is machined and how it works, I'm unfamiliar with vacuum hold downs and have only seen Brady's set-up that he has posted.