Another Vacuum Table Question

[canman77]

UPDATE:

I ordered the motor, got it wired up and connected to a vacuum box that bolts to the bed of my machine. I have a vacuum gauge coming in the mail so I'll update later with how much vacuum I am pulling. I am thrilled with the suction this vacuum provides. My machine is a CNCRP 4x2 standard, and when I throw a 2x4 sheet of baltic birch on the vacuum table, I can't move it with all the force I can exert. Very cool.

However...I am having some trouble once I start cutting parts. I am leaving a 0.02" onion skin to maintain vacuum as long as possible, but when it comes time to cut full depth, I get a few parts that move. And not just small parts, some parts well over 100sq.in. It seems to happen only in certain regions of the bed. I'm not sure if this is because of a bad vacuum box design, or if there is warpage in some of the parts that causes vacuum loss once they get cut out.

Considering the success everyone else has with plywood on a vacuum table, I'm guessing my issue is coming from a bad design. The base of the box is 1/2" MDF. Glued onto the base are the frame pieces, measuring 2.5" tall. I started with a rectangle frame with two cross pieces in the center. When I put the 1/2" MDF top on and turned on the vacuum, the top is held in place very securely and doesn't go anywhere. BUT...the vacuum is so strong that it causes the MDF to bend between each cross piece. That was mind blowing. As a result, no sheet good would really make good vacuum with the table. So I added more 2.5" tall pieces to hold up those sections that were getting bent down and tried again. Still a tiny bit of pulling. So I glued the board down, and I believe it stays fairly flat with the vacuum on.

I ran a quick couple of cuts before surfacing the board and had 2 of 14 parts move. Not bad, but not good. I figured it was due to not being flat relative to the machine, so I surfaced the top and ran another little cut, with two parts that are only 50sq.in. One part stayed down very well and wouldn't move when cutting the onion skin, but the other part jumped around and lost vacuum. I'm at a loss. My spindle isn't perfectly trammed, but I only have a little ridge between each pass. Perhaps that is enough to cause vacuum issues, I don't know. Maybe the 1/2" MDF top is still getting pulled by the vacuum and I need to use a thicker top?

Do you guys have any suggestions?

[MARV]

Hi Canman77,
A couple of observations; Baltic birch can be fairly stiff stuff (you didn't say how thick a piece you were using) and not as flat as most people would like to think. If you aren't getting tight contact everywhere, then you are loosing hold down force. Second; the deflection you're seeing in the 1/2" MDF top is resistance to air flow. That resistance equates to a loss of vacuum (not total loss, but a % loss).

As a side note; most shop vac blowers are measured in inches of water column, where as vacuum pumps are often measured in inches of HG. The conversion factor is; 1" of WC = 0.0734824" of HG. My shop vac blower (purchased 30 years ago and never used) was rated for 160" of WC. Thats equal to approx. 12" HG (which works out to a nominal 6PSI). Again, its not what the pump output measures, its what the vacuum is at the interface between spoil board and work piece. Plug your new vacuum gage into a gasketed panel and place that on your spoil board to check your system. You may be surprised at what you find.

Stay with your effort here; its all part of the learning curve and you'll get there soon enough. Its the details that'll make the difference.

[ger21]

Are the box bottom and sides sealed? If not, you're possibly losing a lot of vacuum through them.
If the boards or sheets you're cutting are warped or bowed, and the concave side is up, it'll be nearly impossible to hold them down.

If the top of your box is still bowing even a little, you'll lose a lot of vacuum.

In my world, 100 sq in is still a very small part.

When I cut a lot of small parts, I cut them in rows across the sheet. When I cut the bottom row, I start my cut at the top of the part, so that the part is attached to the full sheet until the very end. After the first row, move to the next row.
This works very well, until you get to the last row.
Also, try to cut the smallest parts first, and the largest parts last.

[ger21]

Hi Canman77,
Second; the deflection you're seeing in the 1/2" MDF top is resistance to air flow. That resistance equates to a loss of vacuum (not total loss, but a % loss).

To help out with this issue, use a 3/4" spoilboard, and surface 1/8" from both sides. This will allow more air to flow through then 1/2" MDF, as the core of the MDF is much more porous.

[canman77]

Thanks for the input and help guys. It is much appreciated. I would be very lost without your help. So, let me address your posts below:

1) The baltic birch I am cutting is 1/2". Stiff for sure, and certainly not flat. I can see the bow when I put it on the machine. I make sure the bow is bending up, so the sides of the plywood are touching the vacuum. Gerry mentioned this earlier in the thread.

2) The vacuum I purchased is rated at 153" of water, which works out to 11.25inHG, which then works out to roughly 5.5psi. When the gauge shows up I will compare these numbers to the real life numbers and update.

3) Everything in the vacuum box is sealed with the exception of the edges of the top surface. The sides, cross pieces, supports, everything is glued together. I sealed the outside of the box with a few coats of water/glue mix. The bottom of the box is sealed with a couple coats of shellac.

4) The top of the box...I can't tell if/how much it is bowing when the vacuum is on. I only have construction levels, which aren't notoriously straight or flat. But according to the levels I have, the top surface does not *appear* to bow when the vacuum is on. Worth noting, I surfaced the board when the vacuum was running.

5) Nesting/cutting order: I have seen several of you recommend cutting smallest first and biggest last, which I have been doing in my tests. I'm not opposed to tabs I can break with my fingers, 0.02" thick is no biggie. But what I found in one test is that even with tabs, I lost enough vacuum that part of the entire sheet was moving with the cutter (some of the waste in the sheet was maybe 1/2" wide, so pretty flexible) and still lost a part.

I know how capable the suction can be with this pump/box setup. I've taken some of the smallest parts I make (one at a time, so lots of vacuum wasted) and placed them in various spots on the box. In some spots I get great suction and would feel confident that the cutting force to remove the onion skin would NOT move the part. In other spots I look at the part wrong and it slides around. So I'm thinking the issue lies in my box somewhere. Which brings me to some questions:

1) How imperfect can the top surface be? I have tooling marks and ridges from surfacing the top. They aren't horrible, but a fingernail can definitely detect them. Is this enough to cause a problem?

2) Let's assume the top is bowing. How much bow is acceptable, or is this an "any bow = no go" situation? I think if this is the case, the best solution is to use 3/4" ultralite MDF, skimming off just the sealed surface. 3/4" is pretty dang stiff.

3) How big a difference will 1/2" vs 3/4" (both skimmed) MDF make in the vacuum? I know it is tough to get numbers, but if anyone has anecdotal information that would be great.

4) Vacuum gauge location: I like the idea of having a gasketed piece to toss on the vacuum bed to see what I am getting, but I'd also like to monitor the vacuum real-time. I want to test the vacuum when nothing is loaded vs when a sheet is loaded, when the vacuum is completely sealed off, and when a nested sheet is finished cutting. Basically get as much info to pinpoint the problem. Can I install the gauge into my PVC plumbing or do I need to take readings directly from the spoilboard?

[MARV]

You should know that with industrial applications, many of the issues you are chasing are overcome by overpowering system losses with high HP pumps that deliver high CFM. We're talking about 40 to 50 HP pumps in certain instances. Remove much of the power and most of the CFM and your left with the prospect chasing minute gains thru seemingly endless exploration.

I don't necessarily have answers, but I can offer some direction to consider;
1) If your baltic birch isn't pulling down reasonably well, you may want to consider developing some generic gasketed fixtures that can be made to work with a separate fractional HP pump. Keep in mind that there may be some work pieces that simply require an alternate method to secure.
3) My experience with MDF is that it breaths more thru the edge then thru the faces. Any surface that isn't intended to "breath" should be throughly sealed. And shellac may not do it.

Regarding your questions; 1) A small ridge left from surfacing your board shouldn't matter, but keep in mind you're concerned with fine tuning your system. Scuff any rag left from surfacing your spoilboard that may prevent tight surface to surface alignment.
3) With industrial systems the spoil board is typically supported by a rigid underlaying framework, and quite often a heavy gasketed phenolic panel. The thickness of the spoil board isn't important as long as the support is there.
4) Location of the vac gage should be as close to the spoil board as possible, but wants to be readily visible.

[canman77]

What kind of feedrates/depth of cuts are being taken in industrial applications? For comparison, I am cutting 150IPM at 0.3" DOC. I'd be interested to see how that compares to industrial cutting. I'm guessing I am on the VERY conservative side. And to be fair, I'm not expecting industrial grade machining from a 2hp pump. Based on results others have seen with these vacuums, I don't think my expectations are unreasonable. Especially when a small part is held down by vacuum in some areas of the table but not others! I know the potential is there, I just need to identify the solution.

That said, my "full sheet" (2x4 for my machine) of baltic birch definitely gets held down. Very much so. The problem arises as I take the finishing pass through the 0.02" onion skin. I will seal the edges of the spoilboard tomorrow, but not until after I install the vacuum gauge. I'd like to measure the difference. It's a cheap gauge so I'm not expecting insane accuracy, but I'm hoping it will at least show changes in vacuum as different solutions are tested.

What you say about the industrial machines with phenolic panel is interesting. I did not know that, but after seeing what the vacuum can do to 1/2" MDF, I see the rationale behind it. I don't know how to calculate how much support MDF needs in order to stay rigid while the vacuum is running (not a mechanical engineer!) but I suspect 3/4" MDF might help. What I lose in vacuum I would probably gain in rigidity and flatness...and thereby increasing the chance that small parts would stay put.

As I've noted before, my smallest part is like 16sq.in. or something. When I tested the first MDF vacuum box with a shopvac, I was able to cut these small parts without anything moving. That motivated me to buy the better vacuum and build a real box. Realistically, the only thing that changed is my box. I used 1/2" instead of 3/4" and I'm thinking that played a part. I was trying to minimize my Z axis sacrifice but now I'm wishing I had just used 3/4"!

[peteeng]

Hi Canman - To maybe explain some of your variables:
1) MDF has variable density and porosity across it , this is why some areas pull better then others
2) MDF has a "skin on its surface that is not very porous. So it needs to be skimmed to allow it to breath better. I think you have done that by levelling the board
3) A high volume pump used to generate "vacuum" requires airflow to generate a dynamic pressure. This approach is not like a static vacuum approach with a proper vacuum pump. So in fact if the air gap at the faying surfaces of your job to vac table top is zero then it can loose vac as there is no place for the air to flow. Try a thin piece of cloth or tissue or something under the part to allow some airflow
4) Yes you will need to seal the edges of the MDF as these are more porous then the through thickness direction. Air will flow via the easiest route (path of least resistance) and if you have no flow you have no pressure.
5) Can anybody comment on cutting a small grid across the surface to improve airflow? Like a 25x25 grid 1mm deep?
Cheers Peter

[MARV]

Canman,
Your feed speed is at the low end but certainly reasonable. You didn't say what the dia. of cutter was. Again, its lateral cutting force thats moving the part. You may want to reduce the cutter dia. and your depth of cut. At 16 sq. in. of part size, you are on the ragged edge using only vacuum for part hold down. All the variables will have to be perfect for your system to work.

Regarding your vac gage; 15 bucks should get you a quality gage. Use it on a small panel first to test your system. While the pump is running, use something like sandwich wrap on all exposed surfaces of your vacuum box and watch the gage like a hawk. A jump in value as your covering the box reveals leaks in your system.

Another test you might try; mount you gage on your manifold close to the vacuum table. Cover the whole top working surface with cling warp. Run the vac pump and read the gage. The value your reading should be your baseline reading for a "whole sheet" part. Anything less means you may have a problem.

[ger21]

What kind of feedrates/depth of cuts are being taken in industrial applications? For comparison, I am cutting 150IPM at 0.3" DOC. I'd be interested to see how that compares to industrial cutting.

Depends on tool size.

I was cutting some baltic birch Friday. I was in a hurry, and due to a typo, cut .70 deep at 350ipm with a 1/4" downcut spiral.

I'll usually make that cut with a 1/2" downcut spiral at 700ipm. I only use the 1/4" bit when the 1/2" is too big, or when the part is small. The cutting forces are much higher with the 1/2" bit, and it will move small parts, where the 1/4" won't.

This is a 15HP spindle on a 5x12 table, with two 10HP Becker pumps. I can usually cut a 2' x 2' board, and leave the rest of the table uncovered. But it depends on how slippery the material is, and the condition of the spoilboard.

I'm thinking that one reason you have different amounts of vacuum in different places is possibly due to your support ribs? Just a guess.

What you say about the industrial machines with phenolic panel is interesting. I did not know that, but after seeing what the vacuum can do to 1/2" MDF, I see the rationale behind it.

Our machine has an aluminum grid table/ plenum, with 1/2" holes every 6"-8". There's a rubber gasket around the perimeter, which forms a seal with the spoilboard.

Most DIY version use an MDF plenum, with a grid routed into, then sealed thoroughly, with the MDF spoilboard on top of this. So your spoilboard sits on a rigid box that doesn't flex at all.

[ger21]

As I've noted before, my smallest part is like 16sq.in. or something

What does this part look like? A 4"x4" square is a lot different that a 16 sq in part with multiple holes in it, or small "features" that stick out.

I don't know if I would attempt cutting something that small without tabs. But it depends. Creative toolpathing can make a big difference.

Vacuum holding can be incredibly difficult by itself. Using it to cut really small parts makes it even more difficult.

[canman77]

Sorry for the late response guys...had to help a friend build up his road bike and then spent some time with the vacuum table. Before I go any further, I have been making all my cuts into baltic birch with a 1/4" compression bit. Unlikely that I'll ever run anything bigger than 1/4" on the vacuum table, simply because of the increased cost of the tools as the diameter increases. Also, Gerry...the smallest part I run is 16sq.in and looks like this: https://www.slotfurnishing.com/wp-co...upholder-1.png For sure a part I plan on cutting with tabs. Though I was able to cut it without tabs on my "test" box and a shopvac.

I got the gauge installed, I don't think it is an overly precise gauge but it gets me in the ballpark. The needle doesn't quite sit on zero, it sits about one notch below zero. Below are the results of the few tests I ran:

Vacuum completely sealed: Just over 9 inHg (I did this by covered the opening with my hand so I could feel suction and make sure there were no leaks. Not advised for long periods of time.)
Vacuum connected to box, pulling air through MDF, sides UNSEALED: approximately 7 inHg
Vacuum connected to box, pulling air through MDF, sides SEALED (with tape...not ideal but better than nothing): approximately 7.3-7.4 inHg
Vacuum connected to box, with "full" (2x4) sheet of 1/2" baltic birch (sides of spoilboard still taped): Right at 9 inHg

(For the sake of simplicity for those invested in this thread, 7inHg = 3.4psi, 9inHg = 4.4)

The vacuum is rated to max out at 11.25inHg so I have to assume I'm getting leakage somewhere. I haven't drilled a 3/8" relief hole for cooling yet but perhaps I don't need to if I'm not getting 11.25inHg.

I'm not sure how to go about testing different areas of the spoilboard, since I haven't zoned anything. But, I was testing a small part (~55sq.in) on the bed again to see where the vacuum table would grab it and where it wouldn't and I noticed some spots on the board making contact with the part, and as I slid the part around, other spots on the board wouldn't make contact, leaving probably a 1/32" or even 1/16" gap. The gaps roughly correspond to spaces in between my supports. I am beginning to think the 1/2" MDF is the problem. I don't know if the bottom of the box is making good contact with the base spoilboard but I have it bolted down at the edges so I hope it isn't bowing. Either way, I am pretty sure the vacuum is what has caused these peaks and valleys to appear.

I think I am going to cut off the 1/2" board and replace it with ultralite 3/4". I will know before I glue anything down if the vacuum is creating peaks and valleys. If they still show up then my box isn't rigid enough and perhaps I need to redesign.

[canman77]

I think I've identified the source of the issue. It is my belief that the 1/2" board was not sufficiently supported, as others have already suspected. I cut the old board off and put a 3/4" ultralite mdf board on in its place and turned on the vacuum. Even without glue, I get much more consistent vacuum holding force across the entire board. I don't see nearly as much deflection (still a little bit, which I want to examine further tomorrow before I glue anything). Since the top isn't glued, I couldn't get a good vacuum gauge reading. The needle jumps a lot, from lots of leakage, I assume.

Anyways, just thought I'd update for those who find this thread later and may have similar problems.

I will post pictures of the inside of the box tomorrow and you guys can tell me how bad my design is!

[lambodesigns]

Sounds like you are starting to get it dialed. Be sure and surface the ultralite to take off the outer layer before you glue it down.

[canman77]

Sounds like you are starting to get it dialed. Be sure and surface the ultralite to take off the outer layer before you glue it down.

Slowly but surely, yes. I had read other posts recommending 1/2" mdf, but I have to assume they are using it as a bleeder board over a plenum instead of the box method I have employed. We are looking into upgrading to the CNCRP leg kit which would then allow plumbing to come from under the table. At that point I'd cut the existing spoilboard into a plenum and use a bleeder board, as it would have much more support. That is the biggest challenge right now, getting the necessary support/rigidity to keep the vacuum table surface flat.

I am not sure what happens if I add to much support in regards to the vacuum. Don't know how much vacuum gets sacrificed. I guess in theory since it is all MDF, vacuum wouldn't be horribly impacted because it would just suck through all the supports. I suppose it wouldn't be much different than a plenum grid.

Even without skinning the ultralite MDF, we are getting much better results. My concern with skinning the top and bottom is that it thins the board which might allow for more bowing to occur.