DIY Vacuum Thermo-Former

[SCzEngrgGroup]

One of the things on my ToDo list for some time has been to build vacuum therm-forming machine, so I could produce the vacuum formed 1/8" ABS ABS plastic housings for the ATC in-house, rather than having to drive 40 miles to Silicon Valley and rent a commercial machine. I have finally done that, and here's what it looks like:

First, the framework. The frame is built using a small portion of the several hundred feet of 8020 extrusions and aluminum diamond plate that was given to me by a friend several months back.



The plywood/aluminum vacuum table is at the bottom, supported by the lower square 8020 frame. A second identical frame is at the top, to support the heating elements. Inside the two uprights, you can see the linear guide rails and linear ball-bearing carriages that will allow the frame (now shown) that will support the plastic sheet, and allow it to be moved up, by pneumatic cylinders, to a few inches below the heaters.

The down-facing heater assembly is made of four 24"x5", 1000W carbon-fiber electric heater elements. Each heater is driven by its own PWM for temperature control. The control knobs for these PWMs are on the black control panel at the top of the machine. The wiring is, obviously, temporary, and very much NOT UL-approved.





The vacuum is created by a pair of "vacuum ejectors" which us an air venturi to create a -25 in-Hg vacuum using compressed air. I used an old, failed pressure storage tank from our water well as the vacuum reservoir. The reservoir is connected to the machine through about 7 feet of flexible hose normally used for plumbing a Jacuzzi.



Here is the entire machine. You can see the first formed parts in black near the bottom. This was the very first "pull" done with the macine. The four silver-ish knobs clamp the two halves of the frame that hold the plastic sheet. The whole frame assembly, with the pentagonal black plastic side pieces moves up and down via the pneumatic cylinders.



Operation is simple:

1) Connect air supply to vacuum ejector, to get reservoir evacuated. Full evacuation takes about 15 minutes.
2) Turn on the heaters, and adjust PWMs to obtain an even heat distribution, and heat the entire plastic sheet to just under 350F.
3) Clamp the ABS sheet into the frames, tightening the clamping knobs securely to ensure the sheet does not pull out when stretched over the forms.
4) Once the heaters are up to temperature (just a few minutes), open the air valve to pressurize the pneumatic cylinders, to raise the plastic up under the heaters
5) After two minutes, the plastic should be up to temperature, and sagging 6-10" in the center.
6) Remove the air from the pneumatic cylinders, allowing the clamping frames and now soft plastic to fall back down over the table, and forms.
7) Open the vacuum reservoir to the underside of the plastic, and it is instantly pulled down tightly over the forms.
8) After about 5 seconds, disconnect the vacuum supply.
9) After about 1 minute, the clamping frame can be opened, and the now solid, formed plastic can be removed, and the process completed.

These first parts are, obviously, less than perfect, as the temperature was too low in some places, too high in others, and the vacuum not low enough (due to leaks since corrected), which is part of the reason there are so many "webs". But, once I get all the settings dialed in, the next one should be perfect.



Regards,
Ray L.

Similar Threads:

• Vacuum pump size for vacuum hold down jig?
• Need Help!- Vacuum Fixture with Festo Vacuum Generators ?
• Vacuum table and vacuum pump
• How much vacuum?

[mklotz70]

Wow! Looking forward to seeing video and the finished pieces! Great work Ray!

[LeeWay]

Yep.
Looks very cool.

[SCzEngrgGroup]

Finally got my first good quality parts off the vacuum former today. It took a few tries to get the heater settings right, figure out the right heating time, and how to control the vacuum to get a good result. The ABS needs to be as close to 350F as possible, but NOT over 350F. or the surface starts to bubble. Once heated, and dropped over the forms, it helps to turn the vacuum on slowly, as this seems to prevent the formation of "webs" at the corners of the forms. Once it's mostly pulled down over the form, full vacuum can be applied to really pull it down tight.

I ended up adding a second "vacuum ejector" to get the pull-down time on the big vacuum reservoir down from about 40 minutes to less than 15 minutes. I also improved the sealing, so the vacuum loss with the vacuum valve open to the table is now down to about 1/2"-Hg per second, which is pretty good. With the two vacuum ejectors, and a 10-12 minute heating time, there is more than enough time for the vacuum to recover between runs.

Regards,
Ray L.

[SCzEngrgGroup]

I'm now busy making the new forms for all the plastics, doing my very first true 3D machining, using Fusion360 Ultimate/HSMWorks. The forms are coming out absolutely gorgeous! The polyurethane "pattern plank" is fabulous stuff to work with - machines quickly and easily (running up to 150 IPM), and the surface is nothing short of amazing - NO visible machining marks whatsoever. The smaller, simpler was first, and part took under an hour. The other took about 1-1/2 hours. Tomorrow I hope to get the two large forms (14.5" x 14.5" x 3") done, and expect them to take perhaps 2-1/2 hours each.





Regards,
Ray L.

[mklotz70]

Very cool!
Any more info on the pattern plank? Seems similar to the Renshape I've played with. Looking for a cheaper alternative to it. A quick google search comes up with PP-1052, but there's lot's of other choices....some affordable, some not even close. lol

[LeeWay]

Yeah. I was interested in that as well. I saw over $800 for a 3" x 16" x 60" piece. I have been looking for a mold for a project I am working on. I could use the right size stainless steel bowl, but finding specifically the shape and size I need has not been possible. I need one about 20" square or round that I could machine on my router and possibly hand finish to get the smoothness I would need. I would be thermoforming 1/8" PET G over it. Not me actually, but my plastics supplier is already setup with the machine and would do short runs for us.
The mold has so far been the stopping point and I am about ready to just use MDF glue ups instead. Maybe plywood.
I estimate it would need to make 20 to 25 parts to start with. We would cut the finished parts in half, so that would render 40 to 50 parts. Plenty for a new product I think.

[SCzEngrgGroup]

The pattern plank is MP-1055 from Axson.com, and cost was $179.99 for a 16" x 60" x 2" plank, and is 48#/cu. ft. density. This is well under half the price I found for anything similar anywhere else. Shipping added another ~$75, and they have a $250 minimum order, which I met by buying two planks, plus a quart of their TCC-230 epoxy and a pint of TCC-102 hardener, as several of my forms have features up to 4" high, so need to be glued up from several pieces of plank.

Forget about MDF for anything other than a handful of prototypes - it simply does not hold up. You'd be lucky to get a half dozen pieces before the form starts falling apart. You'll never get a decent surface finish with plywood, and trying to smooth it out with epoxy, Bondo or some other filler will not end well.

Regards,
Ray L.

[LeeWay]

That makes sense on the form materials, Ray. Thanks for the input there.
This PET G will need to remain clear as well, so that is a consideration. I don't necessarily have to have one large circle since the finish parts will get cut in half. It can be two molds together on a backing plate with a smaller bridge form between the two. This would leave a nice parting line too. That material price is within my R&D budget for this as well. Do you know anywhere I could get a small sample of it? Maybe a 1" x 2" block or something?

[SCzEngrgGroup]

That makes sense on the form materials, Ray. Thanks for the input there.
This PET G will need to remain clear as well, so that is a consideration. I don't necessarily have to have one large circle since the finish parts will get cut in half. It can be two molds together on a backing plate with a smaller bridge form between the two. This would leave a nice parting line too. That material price is within my R&D budget for this as well. Do you know anywhere I could get a small sample of it? Maybe a 1" x 2" block or something?

Lee,

PM me your mailing address, and I'll send you a chunk. It has roughly the consistency/density of cured Bondo, if you're familiar with that. It's quite strong and dense, but machines like butter, leaving a dead-smooth surface. I do my finish passes at 120 IPM with a 0.020" stepover using a 1/4" 4-flute ball endmill, and there isn't even a hint of machining marks when it's done. No need for sanding or anything.

Regards,
Ray L.

[SCzEngrgGroup]

Today I did a part that was just barely within the limits of the Pro. My machine has 15.75" of Y travel, and this part used 15.25" of that. I had to VERY carefully position the blank on the table, and with the table at the programs maximum Y- extent, there was only about 1/8" clearance between the back face of the blank and the front of the column, and about the same up front. But, it came out perfect.

Unfortunately, this part required continuous baby-sitting. Cutting at 135 IPM, it generated a huge volume of chips, which, naturally, piled up quickly on the Y bellows. It didn't take long before the chips got deep enough to risk obstructing Y axis motion near the limits, so I had to constantly clear the chips off the bellows, and away from the cutter.

The blank, mounted on the machine...


Partway through roughing...


Roughing nearly complete...


All done!


Tomorrow is another similar piece, but with a more complex shape, having a "bump-out" on the top to house a couple of connectors.

Regards,
Ray L.

[LeeWay]

Looks great, Ray.
I will be doing mine in my router, so it's not quite as accurate as a mill, but it should do the job.
How did you actually order from Axson?
I didn't see a way to do that.

[SCzEngrgGroup]

Lee,

Ain't that odd? They don't provide ANY obvious way to submit an order. Seems more and more common these days.

Anyway, send an e-mail to axsonmh@axson.com, and you'll get a response from one of their sales people. They're actually extremely easy and pleasant to deal with.

Your sample will go out tomorrow via USPS.

Regards,
Ray L.

[SCzEngrgGroup]

Here is the more complex part I made today - the cover for the top of the carousel, including the little "penthouse" for the cable connectors to the transfer arm drive.

Connector Housing Roughing Done


Main Roughing Underway


Roughing Complete, Finishing Just Started


Finishing Partially Complete


Finishing Complete (Ignore the gouge in the top, caused by a CAM oopsie. It will be fixed with a little epoxy). You can't really tell in the photo, but the top surface is not flat. It's got a 5" diameter flat area in the center, then it slopes down at 10 degrees from there to the edge, so coolant can't collect up there.


I am really thrilled with how well these came out. The machine performed magnificently - never skipped a beat, and the accuracy is phenomenal, even on the most complex surfaces.

Regards,
Ray L.

[mklotz70]

Thanks for the info and the pics of it!!

[LeeWay]

Thanks a bunch, Ray.
Good to know you can repair this stuff fairly easy if they do sustain damage over time.

[AUSTINMACHINING]

Great work Ray,
Thanks for posting.

[SCzEngrgGroup]

The vacuum former is officially "ready for prime time"! Today I made the first "production quality" part. I found it necessary to drill some holes through the form to make it work really well. First, I had to drill a series of small holes all the way through the form, to ensure the vacuum could reach all the features on the top. Then I had to drill a 1/4" hole though the center, which I use after the plastic has partially cooled, to inject high-pressure air to get the form out of the plastic. Both worked beautifully. This first part had a couple of very minor purely cosmetic flaws, but is, nonetheless "good enough". I'm confident the next ones will be perfect.

The dual vacuum ejectors also worked perfectly, and I can now take the vacuum tank from empty to -25 mm-Hg in about the time it takes to heat the plastic. So, re-charging the vacuum reservoir will no longer limit cycle time.

This project is done!

Regards,
Ray L.

[LeeWay]

Excellent news.
Sample arrived in great shape. Cool stuff.
Glad it is rather light weight. I expected it to be a little heavier. This stuff should work great on our router too.
Thanks a bunch.

[SCzEngrgGroup]

Just to finish off this thread.... After far too many hours of testing, tweaking, and improving, I now have the vacuum former working essentially perfectly. In fact, not only far better than I ever dared hope, but actually better than the commercial machine I used to rent at TechShop! The final, really consequential changes were:

1) With the exception of the hose to the vacuum reservoir, all the vacuum plumbing has been re-done with copper pipe, and sweat fittings, to ensure no leaks. The PVC ball valves have been replaced with brass gate valves, as the PVC valves leaks under vacuum.

2) The original perforated wood vacuum table has been replaced with a single sheet of 1/4" aluminum with a single 1/2" vacuum port in the center. The wood table developed leaks that I could neither find nor fix. A piece of 1/4" Wire mesh "hardware cloth" is placed on the table to distribute the vacuum from the port to the entire table. This works MUCH better than the old table, and has a fare lower probability of leaks, due to its simplicity.

3) The original PWM heater controls have been replaced with the ubiquitous "Rex C-100" PID temperature controllers available for under $10 each on E-Bay. Feedback is by Type K thermocouples, mounted about 1/4" above the top of the plastic, under the center of each heater. 1/16" aluminum sheet baffles hanging between the heaters "isolate" the heaters from each other, so the PIDs do not unduly interfere with each other. By FAR the MOST important little "tweak" on the heaters is covering each thermocouple with heat-shrink tubing, as an Infra-Red absorber, just like the plastic to be formed. Without this, temperature control was absolutely abysmal. With the heat shrink, temperature control is near perfect.

I can now make parts effortlessly. It takes about 20 minutes to "charge" the 70 gallon vacuum reservoir to ~24.5 in-Hg (as good as it gets up here at 2700 feet above sea level). Heating the plastic (1/8" ABS) takes about 3-1/2 minutes, and forming takes only 2-3 seconds. During forming, vacuum drops from 24.5" to about 22", which can be re-charged in the time it take to re-load the plastic, and heat it. So, a new part can be done about every 6 minutes - no bad!

Regards,
Ray L.