Final Project:  Modular Vacuum Former

6.24.2015


Lecture Notes: 

Homework:

Resources:

Files:

Acknowledgements:

Many thanks to Andrew Harmon for his advice, troubleshooting, and assistance with this project.  Thanks to Brandi, Angela, and Dennis for their assistance during assembly and testing.

Project:

My final project consists of a proof of concept modular vacuum former system (MVFS) that uses a sealed medium density fiberboard (MDF) structure, programmable infrared heating system, and a sliding frame to heat and rapidly form thin plastic parts.  Each vacuum former module includes check valves and latches on either end so multiple modules can be joined together via their vacuum tables and create a larger forming area.  By joining modules, a single vacuum motor could potentially draw air from multiple tables as opposed to using multiple motors. 


finalproject_72_w

 

Structure:

The structure of the vacuum former was milled from 0.75 inch (19 mm) medium-density fiberboard (MDF) using a Shopbot PRS Alpha 96 3-axis milling machine with a 0.25 inch (6.35 mm) square end mill.  The structure consisted of four principal pieces:  1) vacuum table, 2) bulb/electronics enclosure, 3) frame, and 4) frame rails. 

After assembling the vacuum table's lower enclosure with glue and fasteners, its inner surface (including edges around its valve ports and vacuum motor port) was sealed with wood sealer.  Foam strips and sheets were used as temporary seals between the vacuum table's bed and its lower enclosure, between the vacuum motor's housing and the vacuum table's underside, as well as around the eight check valves.  The vacuum motor was installed in the center of the vacuum table's lower surface by compressing its housing between an MDF plate and razor-cut foam gasket.

finalproject_1_w
 finalproject_4_w
finalproject_8_w
 finalproject_10_w

Draw latches were fastened to the front and rear ends of the vacuum table to join additional vacuum former modules, compress and seal gaskets around the valves, and increase the forming area.  Blocking was glued to the underside of the vacuum table's bed prior to fastening it to the lower enclosure to resist large deflections under reduced internal air pressure. 

finalproject_13_w
 finalproject_18_w
finalproject_19_w
 finalproject_22_w

The vacuum table was elevated on four MDF legs fastened through the its lower surface. 

Check Valves:

Eight reversible check valves were designed to fill the ports on either end of the vacuum table.  The valve bodies were designed in AutoDesk Inventor and constructed of two FDM 3D-printed Acrylonitrile Butadiene Styrene (ABS) valve halves fastened together with screws. 

The valve halves were printed using a MakerBot Replicator 2X with two different filament colors to facilitate identifying the direction of air flow through the valve.  Valve halves were drawn and printed using the nominal port inner diameter dimensions.  Actual valve body outer diameters exhibited a -0.010 to +0.05 inch (-0.25mm to +0.13mm) tolerance and were somewhat oval. 

The valve's sheet metal screws tapped their own threads through undersized 3D printed pilot holes. 

finalproject_26_w
 finalproject_28_w
finalproject_29_w
 finalproject_30_w

The printed valve halves would benefit from additional infill (greater than 10%) to address irregular surfaces particularly evident in the red parts. 

The valves were positioned in their ports with flanges and sealed with foam gaskets (one on either side of the flange).  The natural gum foam selected for the gaskets proved to be unsuitable for laser cutting and disintegrated into ash when heated.  The foam sheets were ultimately cut with a paper cutter into square cross-sectioned strips and adhered to the valve bodies using rubber cement.

Laser cut polystyrene sheet 0.030 inches (0.76 mm) thick was used as the reed within each check valve.    The reed consisted of a U shaped flap terminated with two circular holes to relieve stress concentrations.  The flap rests against the smaller of the two orifices within the assembled valve body to create a seal.  In this case, the valve will permit air flow in the direction of the blue half of the valve pictured above while preventing flow towards the valve's red half.

For a single standalone vacuum former module, the valves would be inserted with their blue half (larger orifice) facing outward to seal the vacuum table.  To expand the vacuum former, the valves would be flipped with their red half (smaller orifice) facing outward and an additional module joined to the vacuum former using the draw latches. 


Heating / Electrical:

I selected common threaded infrared bulbs as the heat source for the vacuum former for ease of installation, ease of modification, widespread availability, affordability, and importantly safety.  The bulbs were threaded into ceramic sockets and supported and distributed above the vacuum table in an MDF enclosure. 


finalproject_33_w
 finalproject_34_w
finalproject_47_w

I wired eight sockets in parallel using terminal blocks and clustered them near the center of the enclosure.  The bulbs were powered through a 3-32VDC / 25A-120VAC solid state relay (SSR) triggered by the vacuum former's control board.  An additional 20A SPST switch was placed in series with the SSR to provide the operator with a manual safety interlock.  Additional socket holes would permit the bulbs to be redistributed in wider patterns in the event modules are joined together. 

The underside of the electrical enclosure was covered in a layer of aluminized synthetic felt to further insulate it from the heat of the bulbs and reflect radiation downwards. 

Laser cut and slotted cast acrylic grates covered the ends of the electronics enclosure while providing ventilation to the SSR mounted to a salvaged aluminum heat sink.   Laser cut holes were also created for the switches and the potentiometer.

During operation, the IR bulbs and vacuum motor were powered from separate 20A/120VAC circuits. 

Control / Programming:

The MVF's control board used an ATMel ATTiny 44 micro-controller.  Besides signals from the fabISP, micro-controller inputs included:  a momentary "on" switch, a momentary "off" switch, and a 5KOhm potentiometer.  Heating of the plastic was controlled by regulating the time interval the bulbs were fully energized.  By adjusting the potentiometer, the operator could heat the plastic for longer or shorter intervals. 

Timing was regulated by an external 20MHz resonator.  Outputs were limited to a single voltage to trigger the solid state relay. 

The board was designed in Eagle and milled using the Fab Modules.  Solder pads were used for many of the input and output signals rather than headers and sockets so that dissimilar wire gauges could be used.  Board power was supplied from a fixed 5V AC/DC power supply. 

finalproject_36_w
 finalproject_37_w
finalproject_63_w
 finalproject_64_w

The micro-controller was programmed through the fabISP using an Arduino sketch.  The sketch was triggered by the momentary "on" switch and then read the potentiometer's output.  An if-for statement then incrementally added an integer of 1 from zero to the digital value equivalent to the potentiometer's output and delayed the sketch one instance for each addition an interval equal to a user-defined scale factor.  The sketch had the effect of turning on the bulb array for a time interval equal to the product of the potentiometer's output and a scale factor in units of milliseconds. 

The sketch and board were successfully tested with an LED prior to installing the board with the SSR and IR bulbs.  The 1KOhm pull-up resistor on the board's relay output was found to limit the current below that which was necessary to trigger the SSR.  Substituting the 1KOhm resistor with a 100 Ohm corrected this issue. 

A logarithmic audio potentiometer was mistakenly used instead of a linear output potentiometer which increased the sensitivity of the time delay to the potentiometer's rotation.  Substituting a linear output potentiometer would correct this issue.

The control board was externally mounted on the electronics enclosure for the purposes of testing but would be mounted internally otherwise.


Frame:

A rigid, vertically-adjustable, single piece MDF frame was used to suspend the plastic sheets above the vacuum table and beneath the IR bulbs while heating.  Double-sided foam tape was adhered to the underside of the frame to hold the plastic sheets to be formed.  The tape provided a simple, low profile, and reusable means to hold the plastic sheets and would be periodically replaced as the adhesive became fouled with dust. 

finalproject_38_w
 finalproject_39_w
finalproject_65_w
 finalproject_40_w


The frame was guided by means of two MDF slotted panels on either side of the vacuum table that also served to support the electronics/bulb enclosure.  Bolts fastened to the frame rode against the edges of the panels and in the center slot to restrict the frame's motion as well as hold a length of bungee cord.  Two additional bolts within the panels' center slots served as a hook for the bungee cord and a mechanical stop to limit the upward travel of the frame.  Decreasing the spacing between the frame's outermost bolts and the panel edges and using four rather than two bolts per end would further constrain the frame's motion and reduce the extent of its roll.  The bungee cord ends were formed into loops using vinyl electrical tape and terminated with heat-shrink tubing. 

The inner surface of the frame was covered in a layer of synthetic felt identical to that used on the underside of the electronics enclosure and secured using brads.

The current frame design is suitable for only one vacuum table.  Either a graduated set of rigid frames or a segmented frame would be necessary to hold larger sheets in the event more than one vacuum table were joined together.


Testing:

Incremental empirical tests/inspections were performed during the development of the vacuum former module.

These tests included:

Vacuum Forming:

Of a handful of plastics (ABS, polystyrene, lexan, acrylic, PVC) and plastic thicknesses investigated for vacuum forming with IR bulb heat sources, black ABS plastic 0.060 inches (1.52 mm) thick was the most responsive and was selected for additional tests. 

MVF operation uses the following steps:


The series of images below show the MVF used to form 0.060 inch (1.52 mm) ABS around objects ranging from 0.335 inches (8.5 mm) to 1.670 inches (42.4 mm) thick.  Also pictured are infrared camera images captured during these tests that show the temperature distribution on the plastic's surface.  The camera images illustrate the spotted pattern of hot and cold regions which was an expected consequence and a design compromise when using round IR bulbs rather than linear bulbs or linear heating elements.  Localized plastic temperatures approached 330F (166C) in some cases.   

finalproject_53_w
 finalproject_67_w
finalproject_55_w
 finalproject_68_w

Increasing the distance between the sheet and the bulbs and heating the sheet for several minutes helped to soften the plastic more evenly. 

Vacuum formed part definition was improved by locally post-heating areas around sharp corners with a hot air gun while still under vacuum (see Fab Academy logos below).  The 42mm thick guitar body shell shown below was not post heated and exhibits a draped appearance with less definition to its lower edge. 

The vacuum motor uses its own exhaust to cool itself and is therefore susceptible to overheating if used extensively to pull a vacuum on a covered vacuum table (see IR image below).  It was discovered that reversing one of the check valves to bleed air around the vacuum formed plastic sheet, cooled the motor somewhat without significantly affecting the vacuum table's performance.  Intermittent vacuum motor use would be recommended. 

 
finalproject_56_w
 finalproject_60_w
finalproject_49_w
 finalproject_69_w