Cutting foam sheets...with a needle!

That was how David (dkj4linux on the FT Forums) introuced his thread back in January of 2016.  It was an idea he stumbled up on a few years back and posted about on RCGroups but didn't get much interest.  He later posted on the RCPowers forum and got more interest and a later post on RCG got a bit more interest.  Then FT Forum user Tritium contacted him and suggested he share it with our community.  And...well...it's amaing what a good community can do with a good idea!

So how well does it cut?

That's a cut edge from the needle cutter (with a rather old and worn needle no less) also showing a few score cuts that only go through the top layer of paper and halfway through the foam.  Here's a full baby bugatti cut:

And yes...it works just as well on FT's WR foamboard too:

As of today there are over 140 pages in David's original thread and the discussion is ongoing.  It's jam packed with great information not only on the needle cutter but also on the MPCNC (an easy and affordable CNC machine which is "Mostly Printed") CNC in general, laser diodes, drag knifes, and other ideas on cutting foam and other materials used in our hobby.  Despite all of that experimentation the needle cutter still seems to win out as the best way to cut foam on a CNC machine at home.  I highly recommend reading through that thread or at least the first few pages where David shares a lot of background information.  In this article I'm just trying to condense down the very best and most useful knowledge from that thread and sharing my own current needle cutter design.

Brief History of the Needle Cutter

While David gets the credit for bringing this idea to our community he readily admits that the basic idea was not orignally his - that credit goes to Tom McGuire who's foam cutting videos from 8 years ago inspired David to give this method a try:

David's original cutters used brushed motors, he then evolved the design to use a brushless motor - both of these designs used a balanced flywheel with a bearing to convert the rotational motion of the motor to linear motion of the needle.  He then went on to develop a simpler crankshaft based design that eliminated the flywheel.  That was the version I started with.  It worked well for me for about a year - but did have it's issues, I've recently built a new cutter using the flywheel design and now believe that the flywheel is a superior design and is the one I suggest building.  The crankshaft design appear simpler, but can deform at higher RPM's causing needles to go flying and tends to result in more vibrations which will eventually take their toll on the cutter and it's motor.

While most needle cutters like this are being used on CNC machines they can also be built and used as a handheld tool.  Forum user behnt created a "hand held needle saw" that uses the same basic principle.  

Here's a quick overview of my latest needle cutter:

And here it is cutting a full sheet:

It's worth noting that that cut was the very first I tried with the new cutter.   Since the new cutter has proven to be more reliable and capable of higher RPM's than my previous cutter I've been able to increase the speed of cuts on my CNC machine and that same file now takes just 17 minutes to cut instead of the 27 in that video -while delivering a nearly identical quality of cut.

Needle Cutter Basics

The needle cutter is a very flexible design.  All you really need is a way to get a needle moving up and down fast enough that it will create overlapping perforations as it's moved across the foam.  If you additionally have control over the Z axis (up and down) then it's possible to make score cuts and very shallow "marking" cuts that only pierce the top layer of paper leaving the foam intact.  

David's original designs were built from wood.  Other users have made them from metal.  Just for fun David created a super simple cutter just from mousetraps he bought at the dollar store.  The design is really only limited by your imagination.  Some people have considered using a solenoid type of mechanism however it can be difficult to achieve enough stroke on the needle with that kind of design.

The hardest part of building a needle cutter is devising a way to convert the rotational motion of the motor to linear motion of the needle.  The simplest way to achieve this is with a crank or flywheel and a guide - the needle is thin enough that it's able to flex allowing this method to work.  There have been attempts to use more complicated linkages modeled after tattoo guns however much like the solenoid idea it's proven to be troublesome to get it working reliably with the length of stroke needed.

Most of the recent improvements have centered around pre-guiding the needle with "sideboards" (as in David's current design) or bearings (as in my current design) that straighten the motion of the needle before it enters the guide.  

Two other recent improvements pioneered in the FT forum are the addition of an oiler (just a wad of cotton around the needle before it enters the guide) which helps reduce friction and therefore heat.  As well as lengthening the distance between the motor and the guide to reduce the flex on the needle increasing the life of the needle and also reducing friction and heat.

So...Let's build one!

I'm going to outline how to build my version since it's the one I'm most familiar with :)  My cutter is based on a 3D printed design which I mount on my MPCNC machine.   The files for my cutter are posted on Thingiverse here:  https://www.thingiverse.com/thing:2450613  I created this design in Onshape and it's freely available for others to copy and modify for their own needs here:  https://cad.onshape.com/documents/0b010a7a70bd0183d800fa80/w/2427c16d54be94a0c0fb03df/e/6104b1396ce31afa164ac429

In addition to the 3D printed body and flywheel you will need:

Required parts

Small piece of hardwood for insulator
Length of 0.62mm/0.025" music wire (at least 6" long)
5 - 693ZZ - 8mm OD x 3mm ID x 4mm thick bearings.
5 - 2.5mm washers with ID drilled out to 3mm. (3mm washers with an OD that doesn't interfere with the outer race of the bearings would work as well.)
9 - M3 x 10mm machine screws
Assorted M3 screws/washers for counterweight
2 - M3 x 16mm machine screws (for mounting insulator)
2 - M3 nuts (nyloc prefered but regular probably ok.)
4 - M2 x 6mm screws (for attaching the flywheel to the motor)
2212 sized 850kv motor and ESC to control it - needs 2mm holes in face for mounting the flywheel.

Optional - strip of 3 12v LED's.

Flywheel assembly

The holes in the flywheel are sized so M3 screws can self-tap. You can use a tap to start the threads but I suggest NOT tapping the full depth with a tap so the screws fit tighter.

The holes around the outside of the flywheel are so you can balance it before fitting the bearing/needle if necessary.

I use the inner most of the 3 possible locations to mount my bearing - on my flywheel this is 7mm from the center so gives 14mm of stroke on the needle (dkj4linux's original design was 6mm from the center for 12mm of stroke - but I needed to move mine out so the flywheel could fit closer to the motor.)

Behind the bearing you'll want a washer that fits over an M3 screw but which does not rub against the outer race. The M3 washers I was able to find locally were not suitable for this as their OD was too large. So I opted to drill out an M2.5 washer so it fit over a M3 screw.

dkj4linux suggests grinding a groove into the outer race of the bearing to help retain the needle - I tried this but damaged my bearing in the process. I'm currently using an ungrooved bearing and it's working fine for me I just made sure the coils on my needle were undersized so it fits VERY snugly on the bearing. A 623vv grooved bearing may be another good option but it's 10mm OD instead of 8 so a little bigger. Moebeast is using a 10mm bearing on his and has also created a printable needle retainer that he's testing: https://www.thingiverse.com/thing:2429886

Once the bearing/washer/needle are mounted the flywheel should be placed on a shaft and balanced. I used a shorter 8mm M3 screw with 2 M3 washers in the outside hole which seemed to balance out perfectly against my needle.

The flywheel is then mounted to the motor with 4 2M screws these can be hard to find...I used mounting screws that came with a 1806 brushless motor that I had worn out. Another option may be to disassemble an older notebook as M2 screws are sometimes used for assembly on notebook computers.

Needle Details

The needle is made from a length of 0.25"/0.62mm music wire (sometimes called piano wire). It's best to start with a piece that's longer than you'll need since it makes forming the attaching coil easier.

When bending the wire it's best to use round nosed jewelers pliers to avoid putting any stress risers into the wire. I formed mine by selecting a drill around 4mm (roughly half the size of the bearing) and just grabbing it about 6" from either end and twisting it around the drill shaft twice then trimming it off. I then used the round nosed pliers to slightly tighten the coils and to put a small bend where the needle comes off the bearing so the wire leaves the bearing centered instead of off to one side. This turned out to make a BIG difference in how smoothly the cutter runs - but if not done carefully can create a stress in the wire making this the most likely place the needle will break.

Needle in motion showing symmetrical deflection after putting final bend in just below flywheel bearing

The needle is then installed into the cutter and trimmed to length before sharpening. I like to sneak up on the length since I usually loose a bit of length while sharpening. For cutting foamboard you only need about 6.5-8mm of needle protruding from the guide. The less that protrudes the more accurate your cutter will be. The needle should retract fully into the guide on the upstroke.

I sharpen the needle by removing it from the cutter and rotating it in my fingers while holding it against a cutoff disc in a dremel at a sharp angle. It doesn't take much and even a fairly dull needle cuts well.

Needle guide

The two most popular needle guides are either a MIG welder tip or an sports ball inflation needle. In my experience the sports ball inflation needle runs cooler - but wears out quicker and is more likely to deflect in use and cost accuracy.

(Note: You may notice that in this photo my guide is slightly crooked because I was rushing in assembly.  The cutter will work like this but the guide gets a bit warmer than it does when it's properly aligned with the motion of the needle)

Because the guide can get warm from the friction of the needle running inside of it it's best to use a heat resistant insulator to mount the guide to the 3D printed cutter body. I originally used soft pine but had my guide come loose mid-cut and ruin a sheet of foam so have since switched to hard oak. I cut the insulator from a piece of 1/4" x 1" wood which I hold up to the cutter then trace around the cutter body so I can cut it to final shape on a bandsaw (or just sand it down to match.) The guide hole should be drilled to a size that allows your guide to self-tap (or you can tap it if you know what size tap matches your guide.) The guide should be centered under the needle. The two retaining screws get their holes drilled using the cutter body as a guide to position them.

Guide Bearings.

These are optional but highly recommended. They help straighten the motion of the needle before it enters the guide greatly reducing the friction in the guide itself. They are mounted with M3 x 10mm screws self-tapped into the cutter body with M3 washers behind them. Drilled out M2.5 washers would probably be better but I haven't had any issues with M3 washers on these.

Depending on the dimensions of your motor and how your needle mounts to the flywheel bearing you may need to adjust the washers behind these or the flywheel bearing to make sure the needle runs as close to centered as possible through the bearings. The printed bearing standoffs should allow bearings to be mounted with no washers if the need arises.

LEDs / Wiring

The LED's were simply stuck to the bottom and then wired to the power input connections on the ESC for the brushless motor.  A few drops of hot glue keep the wires out of the way.

I'm using a 25a ESC but this is overkill, the motor draws very little current so even a 6a ESC is more than enough for this application.

Motor Considerations

My original cutter used a higher kv motor (actually the motor and ESC from a FT "B Pack") which I had to run off 7.4v to keep it at the ideal RPM range. For this cutter I switched to a slower 850kv motor so I can run it directly off 12v (the same as the LED's require and what the rest of my MPCNC runs on) This cutter body is really designed around the motor. The motor came from a bin of parts I have but is most likely this one:http://store.jdrones.com/AC2830_358_Motors_p/ac2830358.htm However similar motors are available from a number of other sources. While jdrones refers to it as a 2830 motor (because they measure the bell) most other manufacturers/retailers classify this as a 2212 motor (based on the size of the stator) There is a model of the motor included in the Onshape document so you can confirm dimensions.

Tuning the Cutter

I find it very helpful to have an optical tach for dialing in the speed of the motor on the cutter, it can be adjusted "by ear" but until you know what the right speed sounds like that can be difficult to gauge.  The flywheel style cutters are more tolerant of higher RPM's.  I've run mine over 10k RPM for 30 minutes with no problems.  The crankshaft style cutters however can deform at higher RPM's causing them to change their cutting depth and possible throw a needle off the crank.  I could not run my crankshaft cutter above 7,000 without major issues so the tach was critical with it.   You may notice in some photos that there is a small piece of reflective tape on the side of my motors for the tach to register off of.  Since my current motor is chromed I initially tried just putting black tape over most of it but the chrome was not reflective enough to reliably trigger the tach.  

A minimal RPM of about 6,000 works well with cutting speeds of up to about 15mm/second.  I'm currently running around 8k RPM and cutting at 20mm/second and am very happy with the results.  In my video above I was only cutting at 9mm/sec.  They key is to have the needle moving fast enough that it overlaps between punctures creating a clean cut.  You can work out the math of the RPM and needle thickness vs. feed rate...but a little bit of experimentation is easier for me.  A higher RPM on the needle does not tend to improve the cut quality.  In fact it can degrade it by causing more heat which could melt the foam instead of cutting it.  At high RPM's some people have even reported melting their foam into the spoil board.  Speaking of which.

Spoil Board

The spoil board goes under your work and gives the needle something to extend into after it penetrates the workpiece.  The spoil board is slowly consumed as you use the machine but will last a long time.  Mine is a year old and after cutting nearly two cases of foam board it's still totally usable despite being covered in marks from previous cuts.

I use common 1/2" pink insulating foam from my local hardware store.  This is cheap and readily available.  I simply attach it to my worktable with carpet tape and then use T pins in the corners of my DTFB to hold it down.  It's best to mount your workpiece so it curves down in the center and the edges curve up.  This way the pins in the corner will act to flatten it.  If you mount it with the curve up in the middle it can be difficult to keep the center down resulting in overly keep cuts in the center.  

Some people have built and use vacuum tables to hold their workpiece, I started on one but have yet to complete it.  And honestly the sound of a shopvac running in my little shop is kind of obnoxious.  The current needle cutter is quiet enough I can run it without hearing protection and is much quieter than my vacuum.  However a vacuum table does solve the problem of taking the curve out of the workpiece and eliminates the need for needles.  

Preparing files for cutting

There are multiple ways to prepare a file for cutting - this is simply the method I prefer for FT plans:

Essentially the basic idea is:

1. Open a 20x30 document in inkscape
2. Import a page from the PDF
3. Use the group/ungroup tools to separate out the parts
4. Arrange them on the 20x30 document
5. Save as SVG and DXF
6. Open in estlcam
7. Define your cut paths (6mm deep for full cuts, 2.5mm for score cuts and 0.75mm for marking cuts)

I hope this has been helpful!  I highly recommend the MPCNC and needle cutter as a low cost but high quality way to cut foam on a machine.  The MPCNC can be built for a few hundred dollars and as well as cutting DTFB can also be used to machine plastic, wood and some light duty aluminum work.   You can also attach a low cost (~$100) 2.8w laser and cut balsa up to 1/4" thick as well as thin plywood.  You can even get a cheap ($12) drag knife and use it to cut vinyl to make your own decals.  It's a very flexible machine that for less than the cost of a commercial laser cutter can do a large number of tasks for the hobbyist!

Please ask below if you have any questions - or better yet jump into the thread in the forums and join us in continuing to develop this tool!