In progress
Machines
A pro pneumatic vise for injection molding: Can it be done?
We wanted to make a professional-grade pneumatic toggle vise that leverages mechanical advantage and can be made with easily available materials. These vises have many purposes, but we're going to use ours for injection molding to achieve quick and high clamping pressure. In this research module, we'll walk you through our research process, and show how we tested as many production methods as possible to lower the barrier to entry and still get a decent result. Let's talk theory! One of the biggest problems we need to solve is that injection molding needs LOTS of pressure and a strong clamping action. Since we don't want a huge machine, we're going to use physics to our advantage. Let's look at the following equation, which is the principle of virtual work: Q = 0.5P tgΘ P is the force pushing down on the knee of a linkage Θ is half the angle between the linkages Q is the closing force of the vise When the linkage is straight, Θ is approaching 90 degrees, which makes the tangent approach infinity. This should realistically multiply the force exerted by the pneumatic piston ~5-10 times and let us use pneumatics for pressures where a regular vise fails. To really test this hypothesis out, we are going to have to break our work down into a few phases: Phase 1: Designing and simulating the vise Phase 2: Physically creating the vise Phase 3: Bench testing the vise Phase 4: Using it in real life!compressionother machineinjectionHDPE
More Information
Phase 1: Designing the plates and parts
Created 11 months agoUpdated 3 months ago
We use a completely open stack, so we designed everything with FreeCAD. This design uses the new, experimental "Assembly Bench", so a weekly FreeCAD build is required to open and tinker with it.
We had to take a lot of considerations into account in this phase, like:
* Clamping force requirements
* Durability and material selection
* Precision and alignment
* Ease of operation
* Maintenance and repair
* Safety
* Automation compatibility
* Cost vs. performance
* Ergonomics
Luckily, we didn't have to start from scratch, as there are lots of designs to look at. Here are a few of the big design decisions we made, and why:
1. Orientation: The design allows for usage in both vertical injection machines and traditional, horizontal machines. We made the plates and the spacers for ejector pins are interchangable and can be mounted on both sides. Ultimately, we chose to create our machine to be vertical for space-related constraints in our office.
2. Ease of operation: The additional length of the structural rods on the left side of the vise is to allow mounting moulds of different thickness and for finetuning the vise for squareness and perfect closing strength.
3. Precision, alignment, and durability: We made the plates thick enough that we felt that the vise would be highly durable.
The photos below show the basic mechanics of the vise. A pneumatic cylinder will be mounted above the top platform, with the plunger connected to the lower platform. The photos show the vise when the cylinder is fully extended. When contracted, the cylinder will pull on the lower platform, making the middle plate slide on the four structural / guide rods.
The CAD files are also attached below.
Phase 2: Building the vise
Created 11 months agoUpdated 3 months ago
Once the parts were designed, it was time to build the vise!
I decided to use PA9 (aka AW-7075) 20mm plates because it has an extremely high strength-to-weight ratio comparable to steel, excellent fatigue resistance over many cycles, and excellent machinability. I ordered enough plates for two vises / a couple of mistakes in the process or a redo. I have also added a 400x200mm plate on which I'll be testing various methods of cutting the plates out of a bigger sheet of aluminium.
The cost of the aluminium came to €157 with tax and delivery to Poland.
Now came the time for cutting the aluminium! I mentioned that I wanted to offer as many ways to cut the aluminium as practical, so makers can have options. Here are the methods I'm testing:
- router with a guide bushing
- circular saw with aluminium / multipurpose blade
- drill press with annular cutter adapter
- magdrill with annular cutter
- drill press with a regular twist drill (if I can mount a 20mm one into my machine)
- PrintNC CNC machine
Can a router cut out the plates?
Created 10 months agoUpdated 3 months ago
IMPORTANT NOTE: The steps using the router are optional and could pose a risk of an injury. If you don't have any metalworking experience, it's best to skip this step or get the plates machined at a CNC shop.
Handheld routers are inexpensive woodworker power tools - the one pictured cost €70. They usually come with two or three collets, which accept 6mm, 8mm and optionally 12mm bits. But can they cut through 20mm of aluminium? Let's find out!
I 3D printed a template for the cut, glued it to the aluminium with a strong two-sided mounting tape, and used the guide bushing included with the tool. This let me trace the template with the router and allows for a very simple way to achieve an accurate result.
I used a dedicated aluminium metal bit - a three flute, 6mm DLC coated bit. The exact one I bought is:
https://allegro.pl/oferta/frez-palcowy-hrc-60-dlc-do-metalu-fi-6-mm-75-mm-6x24x6x75mm-15460630163
To prep to cut precisely with the router, two things are needed - a 3D printed template that's glued to the stock, and a "frame" around the template which is the exact thickness of the template itself. I made mine with a small piece of 6mm MDF.
The depth of cut needs to be small, around 1 - 2mm at a time, with many passes to get through the 20mm plate.
After repeated cuts with the router, you should end up with a result close to what is being shown in the photos.
The disadvantages of this method are the time it takes to cut each plate, and the chips the router produces - thin, long and they poke into your skin easily.
RESULT: YES, You a router can cut out plates, and it works pretty well. 6/10
Can a circular saw cut out the plates?
Created 10 months agoUpdated 3 months ago
IMPORTANT NOTE: Circular saws are dangerous and can cause injury. If you can, buy the plates already at the specific dimensions from the supplier. This skips cutting them out of a big plate yourself.
A circular saw can be used to quickly cut aluminium into the necessary shape. It takes seconds, but ends up with a rough finish.
I bought a new blade at a local hardware store for my circular saw - 185mm, 60 teeth. It needs to be marked specifically for aluminium, or it needs to have lots of teeth to get through it. The one pictured is marked for wood and was able to cut the plate.
This method shook the whole workshop, was incredibly loud, but it took only half a minute to take a single cut. It requires more preparation than the router method and doesn't guarantee the cuts will be square. A tracksaw would be helpful, alternatively some circullar saws come with a laser attached which projects a predicted line in front of the saw, which helps with cutting precisely.
Hardware for the project - examples
Created 10 months agoPublished 10 months ago
I went through the old hardware we ordered for the project a year ago. These were purchased during the initial design of the plates. Pic contains:
- Hardened shafts: 20mm, 16mm, 10mm in various lengths.
The 20mms are used for the main four structural rods that hold all the plates. The 16mm is meant for the limbs of the knee joint. The 10mm ones are for pivot points on which the knee joint limbs rotate.
- SK10 shaft supports - these hold the 10mm shafts and are bolted with M5 screws to one external and internal plate.
- GN 752-10-M10 joint ends
- GN 705-10-E shaft collars
- square flange LMK20UU bearing
These can be substituted by various other choices as long as the dimensions fit. The vise allows for some leeway in the parts.
There are two major problems that can occur here. Due to the rods being hardened, I will have an extremely bad time in threading them. I'll try anyways. Second, the linear bearings might need to be substituted by a bushing due to the forces in the vise.
3d printing of plates to determine general fit
Created 10 months agoPublished 10 months ago
Since the plates fit on my 3d printer, I made plastic versions of them to avoid costly fitting mistakes on aluminium and taking up pricey CNC time. This will let me experiment with them, and make a working model before recreating everything on the CNC. This might take additional time due to fitting issues with printed parts though.
Assembly of vise with 3d printed plates
Created 10 months agoPublished 10 months ago
I assembled the vise and some of the hardware for it. You can see the vise progress in the photos.
I received the dies to make external threads and tried it on the hardened shafts we got previously. As I had feared in the previous update, the shafts are too hard to make usable threads in them. I'll need to order a different set - either regular steel, or stainless.
Threading of the rods, supports manufacture
Created 10 months agoPublished 10 months ago
The stainless steel rods arrived today. I ordered them cheap, totalling around 50 Euro for the four structural 50cm x 20mm, and one 100cm x 10mm rod. The type of stainless is 1.4301
I made the four short support rods for the lower horizontal plate. These are 10mm x 6cm rods, with both ends threaded about 2cm, leaving a gap of 2cm in between. I then screw in the GN 752-10-M10 steel joint end onto both ends. I also threaded one end of a structural support rod, to see if it is possible to do by hand.
The worst part of making external threads with a die, is starting the thread. The material needs to be prepped by making a less than 45 degree chamfer around the end of the rods. I did this with a bench grinder by hand. You can also put the rod into a drill driver and create the chamfer that way - it'll be more uniform than the "eyeball" method. The threading itself is tiring, especially the M20 thread. Use a lot of thread cutting oil and it'll be easier.
Threading, supports pt2
Created 10 months agoPublished 10 months ago
Threaded all of the M20 rods on one side. The amount I need to thread on the other end will be determined by the size of the entire assembly on the left. We don't want to thread the rod where the linear bearings go. I also made the longer supports. Sadly, I'm missing hardware. Due to a redesign, 4x SK10 blocks needed to be bought, along with two eye nuts, I ordered them last week and am waiting for them to get here.
Up next - manufacture of the plates, more testing of cutting methods. When the necessary items get here, assembly of the 3d printed version.
Prototype assembly pt1
Created 10 months agoUpdated 10 months ago
The necessary hardware arrived which lets me continue on with the build. Previously I was missing two of the eye nuts and eight SK10 supports. I managed to assemble the entire knee mechanism. I made a mistake when buying M20 nuts for the project, and bought some with a coarser thread, I'll need to exchange them for the correct ones. The pneumatic SC80 x 100mm piston is ordered and should be here on Tuesday, which would let me see how the machine works when completely assembled.
The method of manufacturing the supports for the mechanism in the middle of the vise makes it hard to align the eye nuts and SK10 blocks on the two middle platforms. Luckily, unthreading the eye nuts enables some leeway with the length of the arms. Still, percussive maintenance required.
What's left to do is mount the piston, make the plates . The only questionmark is the piston alignment working or not.
The blue mark in the 4th pic is the max stroke of the vise, which is ~9cm.
Prototype assembly pt 1.5
Created 10 months agoPublished 10 months ago
I found an old pneumatic cylinder that I then mounted on the prototype. It is of wrong dimensions both in stroke and width of the piston (targetting SC80x100), but it makes it easier to visualize and test for alignment. It's a bit crooked, but should work properly.
Prototype assembly finished and working action of the vise
Created 9 months agoPublished 9 months ago
The correct pneumatic piston is here - SC 80 x 100 according to some specific iso forms that I'll probably include in the "How to?" later. The vise moving jaw isn't backing up as much as it can on the prototype - I didn't screw in the piston to the lower plate far enough, should still get a lot more travel in the final version.
Production of plates in aluminium ongoing
Created 9 months agoUpdated 9 months ago
I'm CNC milling the plates in aluminium. Had some trial and error with tolerances between the threaded rods or the bearings, and the holes in aluminium. I've marked the threaded hole locations with a spot drill and will be drilling them out on the drill press, then tapping. Would've had the moving plate too, but sadly lost the cut / zero to Windows logging me out because of "inactivity" and stopped the CNC.
Production of plates in aluminium ongoing
Created 9 months agoPublished 9 months ago
Continuing on with the production. Need a better way to hold down the stock when doing the contour pass. I broke some bits because I was getting distracted and made amateurish errors. I need to order some more material for the horizontal plates. I have one plate left, which will be used for the mounting plate of the mold. Todo: drill out all the marks, and tap them. I can start assembly then. I also ordered a small piston for the ejection pins.
Main plates done!
Created 9 months agoPublished 9 months ago
I've finished the three main plates - including drilling and threading the M5 holes. You can see the ejector pins' pneumatic cylider that will push on the ejector pin plate. It's a small SC50x50. The idea is to have the plate screwed to the back of the mold half with long screws and springs on them. When the pneumatic cylinder is extending, it'll push on the plate, the return action will be the springs. This lets you use less length of the ejection pins for smaller parts and not have the piston be in a half-actuated position.
I have also started cutting out the spacer that the moving mold half will be screwed into. I'm having workholding issues with my contour passes on stock that's almost to size. I finished it by hand with the hand router method. It's ugly, but it'll work. Luckily, the next couple of parts will be cut out of a larger pieces of aluminium, where I can clamp them without interfering with the contour pass.
Files updated, STEP format available
Created 9 months agoPublished 9 months ago
I've redone the CAD design to better fit the current state of the vise. I added counterbored holes for the screw capheads, fixed a bugged reference in the SK10 supports. Also exported a .step file. The files can be found in update #2.
Piston plates machined
Created 9 months agoPublished 9 months ago
I machined the upper and lower piston plates. Looks like the tap fits the lower plate fine and will make a thread. Next up - drilling and tapping. After that, I should be able to assemble the mechanism on the left side of the vise in metal. Right side is missing the spacers for the ejection pins. I think I'll also add an additional spacer for the pneumatic piston for the ejectors, which will maximize the available travel of the ejector plate.
I'll need to order more aluminium. I'll try to make the spacers out of what I have left, but it won't be enough. I also decided to go for more spacers, three instead of two, which should get me more ejectors travel and max out the 50mm piston - assuming 5cm of travel and 1cm thickness of plate.
Assembly of the metal version of the vise, other minor info
Created 8 months agoPublished 8 months ago
Having finished machining all the plates, I started moving the parts from the plastic prototype into the final aluminium vise. I had some small fitment issues here and there - stemming from the CNC not producing accurate parts and my drill press being the cheapest thing you can buy. The reverse of the throughholes I drilled were oblong for instance.
This lead to the most significant problem, holes being in wrong places for the main SC80 cylinder. They wouldn't match at all.
I had to drill a 10mm hole for each bolt attaching the cylinder to the plates. I used a few drills, from 4.5mm increasing by ~30% each till I got to 10mm. I think this made the holes "walk" from the already incorrect spot drill locations made on the CNC.
I went for the lazy solution, mounted the plate in a machinist's vise and using the router, I widened them a bit in the directions where they were overlapping the thread on the pneumatic cylinder. It removed just enough of the material to make it all fit.
I also took lots of photos for documentation during assembly for a how-to article I'll post later.
My feeds and speeds are:
12000 RPM - I have a 4 pole 400Hz spindle which generates higher torque in lower RPMs. 12k is as high as it goes.
Surface speed: 226 m/min
Cutting feedrate: 1200 mm/min
This gives me a feed per tooth of 0.03
6mm 3 flute DLC coated cheap flat end mill
For the upper piston plate. this leads to ~15mins of machine time, excluding the manual toolchanges.
I machined the plates using a sacrificial MDF piece 8mm thick, and went 2mm further than necessary to make sure the material is cut out. For smaller plates I used small square tabs which I could break by hitting the piece with a hammer. In hindsight, I should've bought larger pieces of aluminium, out of which the machine would cut out each larger plate.
The cost increase would be minimal, but I would avoid a lot of issues with clamping the stock to the table.
For workholding the large plates, I ended up using automotive doublesided tape, but not before testing around 10 - 15 different random brands / carpet tapes. In hindsight, painter's tape + CA glue with an accelerant should've been my go-to, I haven't tested this method yet though.
For every part other than the three large plates, I cut the pieces out of the 400x200 plate I ordered, so I had plenty of room for clamping there.
There is a lot of tapping and drilling required in this project - according to the Bill of Materials, around 90 holes. Having a pneumatic / electric tapping machine would speed up the manufacture a lot.
There is still a little bit of work to be done on the rear external plate - I need to make the spacers for the ejection pins travel. I've decided on 3x 20mm thickness, should be more than enough for my needs.
Vise completely assembled, pneumatics explained
Created 8 months agoPublished 8 months ago
I've finished cutting out the upper and lower spacers and mounted them to the back end of the vise. This lets me assemble everything into a functional product.
The vise needs to be setup for each mold thickness. The way it works is you adjust the external plate on the opposite of the toggle mechanism, to bring it closer or further away. There's a "leveling" process for the external plate - you need to keep it as parallel as possible to the moving internal plate.
You achieve that by disconnecting the pneumatics from the air supply. You then move the vise into the closed position - by pushing on both the lower piston plate and the moving plate. Put the molds in between the spacers mounted on the external plate, and the moving plate. Push down on the lower piston plate and try to wiggle the molds You can then feel where the molds are loose and where they are clamped already, and which nuts to adjust in or out to spread the clamping force.
After tuning the vise to a mold and securing the external plate nuts, screw in the mold into the vise. Test the vise and see if it's toggling properly. If the mechanism (and the shorter links) isn't locked straight, move the external plate out. If the links go "too far" and don't stay straight, instead bending in the other direction, bring the external plate closer.
The setup for pneumatics is really simple. Start out with a reductor with a water filter. This will let you change the force the piston is exerting onto the lower piston plate by lowering the pressure. Then run a line into the middle hole of the 5/2 pneumatic lever valve, on the side with three holes. This is your input.
Run two lines from the other side of the valve, to the upper and lower input of the piston. Depending on the position of the lever, this might extend the piston when you provide air to it, so be extremely careful when plugging it in. If you want to switch this behavior around, switch around the pneumatic outputs going into the piston.
The speed control valves let you increase the speed of either closing or opening the vise. You achieve this by turning the knobs on the piston inputs.
Proof of concept - holesaw for structural rod holes
Created 8 months agoUpdated 8 months ago
I tested a holesaw for making the large diameter holes for the structural rods. They work great!
Pictured is a 20mm holesaw for the two endplates. The finish they leave is great. I had to drill a 6mm pilot hole where I wanted the 20mm hole to be. It leads the spot drill on the holesaw. The deeper I went, the worse the drill press performed - with the motor stalling if using too much pressure. It might either be a wrong speed on the drill press, or simply a weak motor. I ended up turning the piece around and drilling from both sides.
The hole fits the threads on the structural rod well, althought I woud recommend getting a saw a milimiter bigger than the actual diameter of the hole. This maximises the tolerance and lets you be less precise with replicating the holes on all plates, with the end result still working as it should.
How-to detailed instructions for making the project
Created 8 months agoPublished 8 months ago
I've created a how-to for this project, you can see it here:
https://community.preciousplastic.com/how-to/pneumatic-vise-for-desktop-injection-molding