SparkFun Forums 

Where electronics enthusiasts find answers.

Have questions about a SparkFun product or board? This is the place to be.
#144797
No, the MDF vacuum table that's in there now is a single piece of 3/4" and is cross-drilled, 5x in the X, 8x in the Y, not completely hogged out on the inside...which again, would lead me away from the flexing theory. I mean, really?, how can a 3/4" chunk of MDF, that's only cross-drilled, with plenty of internal support structure flex like that? Weirder stuff has happened...but really? And I'm using a 10 gallon shop vac for vacuum. If I was using an industrial strength vacuum setup and drawing the pressure down to .00001 psia, well, maybe I could see something like that happening. But, a shop vac? I haven't put a gauge on it, but I can't imagine I'm pulling more than a handful of in.hg. of vacuum...maybe 15in.hg at the most? Who knows...Maybe I'll make me a manometer while I'm at it... :)

As far as the bed bending...
When the bed is being milled, I used a 3/4" flat bottomed router bit, multiple overlapping passes, so the same area gets milled at least twice, if not three times, and the bed is NOT under a vacuum.
When I'm etching PCBs, the bed is under a vacuum, the outer edges of the blank PCB get cut deeper than the center-ish of the PCB.
Very odd indeed since I'm using the same process as far as setup and movement (although different bits obviously) to do the flatness-milling and the PCB etching.
I'm going to get a bunch of dowels tomorrow, plug the vacuum holes, and see if I can verify the bending under vacuum.
#144799
Not 13 minutes later and I think I solved my own problem..........(sigh......)........

The MDF cracked a couple days ago along the long axis, where I have the fittings installed in the sides, so if you had to visualize it, the 3/4" MDF kinda split into 2 pieces of 3/8" MDF where the fittings go into the air channels. I put screws thru the MDF to reseal those cracks, but I just had the thought that it's entirely possible that the crack goes all the way into the middle of the board.
Mill it flat without vacuum, no problem, crack stays open, even if only .0001"
Turn on the vacuum, the crack sucks itself shut, gain .0001" of clearance in the middle, but not the outside edges...

A big "Derrrrrrrrrr" going on at this end of the keyboard...
#144802
Another thought/question...

When I get around to routing out the new MDF boards tomorrow, on the flat part of the board where I'll be placing the blank PCBs, ya think I'd be better off with just holes (and a whole lot of them) or a serial of interconnected air channels (just deep enough to let air thru) with holes?

I can see +/- either way...
-Only holes would mean I could use various sizes of blank PCB on the table and cover up the exposed holes (i.e. leaks) with other material (i.e. tape)
-Interconnected air channels with holes might mean a more uniform vacuum across the entire surface, but would limit me to a single sized blank (not a problem at the moment)
#144807
skimask wrote:Another thought/question...

When I get around to routing out the new MDF boards tomorrow, on the flat part of the board where I'll be placing the blank PCBs, ya think I'd be better off with just holes (and a whole lot of them) or a serial of interconnected air channels (just deep enough to let air thru) with holes?

I can see +/- either way...
-Only holes would mean I could use various sizes of blank PCB on the table and cover up the exposed holes (i.e. leaks) with other material (i.e. tape)
-Interconnected air channels with holes might mean a more uniform vacuum across the entire surface, but would limit me to a single sized blank (not a problem at the moment)
I don't really see why you need the holes interconnected on top they already are connected underneath/inside the board.
#144837
You don't actually need holes going to the top side of the MDF at all. You can mill a grid of channels in one sheet of MDF, leaving support structures between the channels. Paint lightly with wood glue to seal it, then let dry. Next, take a sheet of 1/4" MDF and mill both sides just a tiny bit, enough to skim off the hard polished layer and get it level. Glue this to the first sheet of MDF (the plenum) and let dry. Seal around the edges with wood glue, too.

Once you skim off the surface of MDF, it's porous enough to pull vacuum evenly though the whole surface. And if you have an area that's not covered by PCB, the air doesn't flow through fast enough to seriously decrease the pull elsewhere.

This arrangement is used by thousands of people for ShopBot vacuum hold downs, to firmly hold a 4x8 sheet of rough plywood.
#144843
Too late :) I've already got the 2 pieces of MDF cut, channeled, drilled, milled flat, and so on. Just gotta get it all mounted and try it out today/tomorrow.

But, now that you've mentioned about milling the "skin" off the MDF, actually pulling air THRU the MDF like I've read about makes much more sense.
I'll remember that for the next try, if needed.
#144907
New "vacuum table" done...
2 pieces of 3/4" MDF...
Bottom piece hard mounted to the moving table, has 4 holes for fittings drilled thru the back (rather than the sides where they interfere with the sides of the machine itself! der...), 1/2" channels spaced 1/2" apart in a grid, .1" deep, with a sort of ramp cut deeper leading to the holes where the air fittings come thru.
Top piece just has holes drilled to match where the channels are in the bottom piece, spaced 1" apart.
Tried it out, but was able to pull the PCBs off, not very easily, but was able to do it nonetheless. Gonna try drill more air holes, twice as many, 1/2" apart and see what happens.
And, most importantly, after milling the surface of the top piece of MDF, I didn't see any deformation under vacuum. Bonus points...
#144943
Added more air holes and "holding power" has increased a fair amount. Can't pull the blank PCBs off unless I grab them at the corners and sliding them around takes a fair amount of effort.
BUT...still got issues with the Z axis, depth of cut, etc. Can't get that right. Surface was milled flat, but not while under vacuum. I don't see any deforming when under vacuum vs. no vacuum, using a straight edge with a flashlight and not seeing light thru it, but who knows, maybe it is actually moving a bit.
AND...not only that, but I was getting ready to re-mill the surface while under vacuum and the spindle motor burned out! Took off the covers for the brushes, and, well, what brushes? :D
I'm having fun...I think...
#144948
Can't help you with the motor but for the Z problem. How did you glue the two pieces of MDF to together? If you only put glue around the edges two things would happen. One there is a thickness of glue and two the fibers would swell a bit. The problem I see is that milling under vacuum would only work if you plugged up the holes or maybe milled it before you drilled the holes. Right now if you mill it under pressure, the surface won't be pulled down and your channels will be clogged up.
#144995
MDF is screwed together in multiple places, no glue, and I didn't reef on the screws and cause any undue stress on the MDF, so I think I'm good there.

And you're right about milling while under vacuum/plugging holes/etc. Thought about that one (sigh)...
I'm gonna get a handful of dowels/toothpicks/etc and blocks the holes as best I can to get the best seal on all the holes I can before re-milling the surface. Then maybe, just maybe, when I'm done I'll get lucky and be able to blow those dowels out with pressure. Might work, might not. Worst case (well, one worst case), I have to re-drill those dowels/toothpicks out.

The spindle motor - turned the brushes around, and it works again. Lost about 1,000 rpm, only (only?) spins ~23K rpm now where it used to do a bit over 25K. And LOTS of arcing :) Good thing I've got a decent vacuum setup for the dust being generated. I'm hoping it'll hold out long enough for another surface milling run tomorrow.
#144996
On another front, a friend of mine brought over a laser printer and a laminator and we tried that method out with some glossy photo paper and all the rest of the stuff that goes along with the toner transfer method. Worked pretty good except the etchant (ferric chloride) is really old (not sure if that had an impact), the paper itself is a few years old and has been sitting on a shelf (fairly sure that probably had some sort of impact), our test design was mainly 8 mil traces (a handful of missing traces and undercutting), with a bunch of tight clearances (lots of shorts between traces, but I'm sure we also didn't etch it long enough), and just not very well setup overall...50 degree etchant, not agitating the board/fluid the whole time, and so on.
Even with all of those issues, the results were fairly impressive.
Assuming we fattened up the traces a bit, opened up the clearances a bit, got fresh photo glossy paper, new etchant, put the whole thing on a warmer of some sort, kinda swirled the stuff around a bit, while keeping an eye on it rather than concentrating on consuming various beverages during the whole process, it's got the possibility of turning out to be a good way of doing things.

And if the CNC machine fails profusely at cutting decent PCBs, well, at least it'll work for drilling holes/vias/etc.
#145007
Multiwire boards used thin wire embedded in resin to make the interconnections between pads rather than etched copper. It was high density, but had the annoying habit of breaking if the board ever got flexed. It disappeared around the time multi-layer PCBs became common.