With a new job, comes new methods, and of course frustration over the unwillingness to change.

I've been working for my current employer for just over 2 months now. The thing that bothers me most, is the work holding. There is one part in particular that I think the holding method is the problem for most discrepancies - probably about 8-10% of parts are discrepant. Most (80% or so) of their fixtures are made primarily of plywood. Base, and walls are usually wood, then a fiberglass 'cavity' for the part to sit in. Most, if not all of these fixtures use vacuum to hold the part. I checked one of them in Z axis with and without vacuum, and saw almost .040 difference. Kind of makes me wonder if it was programmed with or without vacuum. And yes, alot of the parts are programmed at the machine. Going a bit off subject, I have to say that for the types of parts we do, the Thermwood machine/controller is excellent for programming at the machine (even 5 axis).

I was wondering (or maybe just venting) if anyone has done any comprehensive data collection/testing on expansion/contraction of plywood with multiple variables. For example, the shop is not air conditioned (chair) and I have seen humidity from 20% on up to 90ish% - and of course temp varies from 75-100F recently. Does anyone have any idea on how this affects the wood?

Back to the part that is kicking my ass. I have to use a .050 thick, 1" diameter diamond coated wheel saw to cut a groove around an oval shaped cutout. Tolerances are +/-.030. If the slot is too low, it cuts through the part, thus either scrapping it, or requiring rework (gotta love composites - everything is fixable to a point). If cut too high, then it can somewhat be reworked by hand (fill in the slot with resin, and use a piece of sand paper folded in half to redo the slot. If the groove was 'flat', it wouldn't be as hard to trouble shoot or find a band-aid fix for it. Unfortunately, it is dish shaped. And the problem area hardly seems to be in the same area. Could the plywood tooling be the major contributor to the problem?

Of course, my first thought is to rebuild the fixture using aluminum. But that is usually countered with one of two things......too expensive or.....why change we've always done it this way. ARGH!

Forgot to say that we do at least 1 run of 12 parts a week. Sometimes up to 3 runs. And it's definately not isolated to just one run.

So are you trying to trim formed composite components? If so, does the trimming have to be done on the cnc? Or can they be done as a 2nd. Operation using a dedicated fixture?
I'm sure you’re dealing with movement with the plywood. How much? Depends on a number of things, the number of plys, the species of wood itself, and how the tool is constructed, etc.
I've been machining tools for a local composite shop. The majorities are MDF. But we are now using high-density foam as well. We've found each material has issues unique to each.

I doubt it is the wood specifically. It sounds more like nesting error. If the dish shaped part is shifting on the plug it sets on or in, it will show up in the depth and location of the groove in relation to everything else.

You mention this groove runs around a cutout. Is this molded in or is it a trimmed cutout? My thinking is that you need to take advantage of whatever is consistent in the parts "as molded" features or request an indexing boss(or 2) be added to the mold in an area that gets trimmed away. As long as those indexing bosses are used on your fixture, all should have better nesting repeatability. Multiple cavity mold tools the parts are formed on can also add to the problem if the molds are not dimensionally identical.

Selecting a better sequence of the trim operations, like cutting the groove prior to trimming the cutout may keep the part a bit more stable and less risk of a vacuum leak contributing to the part shifting on subsequent cuts from either vibration or cutting forces.

I did a lot of this in vacuum formed plastics. We used MDF or Benolex and Bondo or Urethane filler as a plug made from the run of parts in production. The big headache I had to deal with was inconsistent shrinkage or warping of the parts after forming. Getting the part to set on the fixture and stay in tolerance could be a pain from one run to the next. If the part is going to dance around, you must find a way to make it submissive. That can be done as stated with features in the mold or secondary operations in the trimming. I would have the trimming people drill all holes as a first sequence and use pins through the drill bushings to then stabilize the part prior to any routing cutout operations. If no holes, then I would put scribe lines on the molds non-critical area and indexing reference lines on the fixtures to assist with orientation on the fixture.

Just don't expect to get any credit for resolving quality problems no one else has ever made an issue of. Crude tooling of this type does not mean that will not do the job. It often means the person using it must exhibit more skill in making it work for the job at hand. A good toolmaker should not put that part of his job off on someone else!

DC

Sounds like my last employer. Trimming parts with a cnc mill, made in epoxy and aluminum moulds, in plywood or fenolic fixtures. I also tried to make improvements to them and was generally rebuffed( if it ain't broke don't fix it). When I was put to work trimming I also had those issues, I was the mold tool maker and they still didn't listen. Do the best you can with the tools at hand I was told. Man I sympathize with you. Been there, done that, moved on. :eek:

Alot of questions to answer. So I'll address each....

CNC Pro - So are you trying to trim formed composite components? - Yes.

If so, does the trimming have to be done on the cnc? Or can they be done as a 2nd. Operation using a dedicated fixture? - It doesn't have to be done on the CNC, but then what's the point of having the cnc? :)

I'm sure you’re dealing with movement with the plywood. - Me too, but I'm not sure that is the only problem now that I've revisted the headache (started the same parts again today).

One of Many - It sounds more like nesting error. - This is the direction I'm leaning towards more. I'm sure I'm getting some movement in the wood, but I think the real problem is the nesting.

Is this molded in or is it a trimmed cutout? - It is a molded oval shaped pocket, where the floor is trimmed out on the cnc. Make sence?

request an indexing boss(or 2) be added to the mold in an area that gets trimmed away - Interesting idea that I will definately suggest.

Multiple cavity mold tools - Not an issue. Only 1 mold.

Selecting a better sequence of the trim operations - Again, a good idea that will be looked into.

I would have the trimming people - that would be me. :violin: Trimmed and drilled on the CNC.

It often means the person using it must exhibit more skill.... - Where would I find some of that? ;)

SPEEDRE - don't suppose your last employer was in Wichita?

I found out something interesting a couple weeks ago that I forgot about. And I"m sure this is probably part of the problem. But it seems most of these fixtures I've been using to trim/drill the part were made using a part. In my opinion (not worth much since I only have about 2 1/2 months experience with composites), wouldn't it be better to use the mold to make any type of fixture?

I'm going to take my digital camera with me tomorrow, and see if they will let me get a couple photos for clerification.

Alot of questions to answer. So I'll address each....

CNC Pro - So are you trying to trim formed composite components? - Yes.


One of Many - It sounds more like nesting error. - This is the direction I'm leaning towards more. I'm sure I'm getting some movement in the wood, but I think the real problem is the nesting.

I found out something interesting a couple weeks ago that I forgot about. And I"m sure this is probably part of the problem. But it seems most of these fixtures I've been using to trim/drill the part were made using a part. In my opinion (not worth much since I only have about 2 1/2 months experience with composites), wouldn't it be better to use the mold to make any type of fixture?

Is this a full mold cavity or a single sided lay up? If all surfaces contact the mold, there should be less chance of thickness variations. Better surfaces help register the part with less error.

I would take the parts made off from the mold and let them set for a few days. These continue to shrink some, but should be reasonably consistant as long as nothing changed too much in the molding process, not demolded too early, then post cured, which could lead to distortion etc by the time secondary operations are under way.

When I would make a plug I would wax the surfaces of the part with parting compound to get the part off the plug. Have a crude core structure ready to create the plug on, made of MDF with a few known true locating surfaces to control the parts alignment to a datum plane on the print(usually the table surface). Mix up some bondo and glob it into corners of the part that would provide the best register and on the MDF core, and place the part on the core before the bondo cures. Making sure the part actually rests on the cores original locating surfaces. I always kept the plugs surface as minimal as possible. More surface=more friction which makes it harder to get the parts to nest on the plug consistantly.





Is this molded in or is it a trimmed cutout? - It is a molded oval shaped pocket, where the floor is trimmed out on the cnc. Make sence?

I would suspect this groove to be an optional size knockout for the oval? Possibly a seal groove? It may not even be that critical other than keeping it intact until that option is required for the application. +/-.03 may be their ideal for location, but it sounds like they may rarely get that quality in a run. It would be interesting to know what the reject rate was before you came on board. Not that you intend to apply rocket science practices to say out-house venting shrouds for instance.


DC

One of Many. That is exactly the same way I used when making a trimm fixture.The only difference being I would use a two part eurethane tooling compound and pour a duplicate of the part and relieve the surfaces to get a possitive nest then registered the part by a drilled hole or feature. Esentially using the mold for the trim fixture. In this way I gradually changed the way trimm fixtures were made at that company for any future fixture fabrication. I eased in a far better and less expensive final part without anyone being the wiser. Eventually Destaco clamps stratigically spaced found there way onto the fixtures making trimming a no brainer.

I used the urethane on vacuum fixtures periodically. It being flexible compared to tuf-fil(bondo), I prefered the bondo. If you know all the parts raw material ever used on the mold will shrink consistant, creating a plug from the mold would be adequate providing the shrinkage of the plug materials is relative. Creating a plug from the parts should cancel out one step in shrinkage since the parts will be the one to fit the plug anyways.

Thermal formed raw materials as in ABS, POLY, PVC's etc. have different shrink factors. These are extruded sheet, but when capped with color and/or texture, the rollers put some stretch to the material as it cools. That stretch or tension can vary from run to run(or during the run) and it shows up after parts are formed. For this reason we would have to relieve the fixture on one run, then use duct tape shimming layers on other runs to get tight or loose parts to register repeatably for trimmed in tolerance. Not much could be done on our end, but to futz with it to keep product flowing out the door.

I'm not sure how much of that would apply to composites, but there should still be ways to compensate for variables out of your immediate control.

One other thing to make note of. If the part were shaped like a bowl or symetrical such that the parts fit the fixture in more than one position with no indexing reference on the mold. If the mold were machined, the parts should have a consistant surface. If the mold surface is made from a casting or by hand, the parts will technically only fit the mold in one position. That goes double for the fixture the parts will be trimmed on, if any hope of repeatable precision is expected from the parts.

DC

I can't find any reference to the size of the machine so the following comment may be irrelevant.If you have a tool length sensor on the machine,do a tool length run immediately prior to running the sequence.I have found on a larger Thermwood that the thermal expansion can make a difference of around half your tolerance going from totally cold to working temperature.I would not expect the temperature change in the workshop to make much difference to the dimensions of the plywood but the humidity change might have a very small effect.