Injection / Vacuum Moulding in Polyurethane

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

Ken I

Project of the Month Winner!!!
Project of the Month Winner
Oct 31, 2010
Reaction score
Cape Town, South Africa
Vacuum / Injection moulding of polyurethane parts – home workshop style.

Polyurethane comes in grades from marshmallow to wood in hardness and is useful stuff for moulding all sorts of parts.

I have been meaning to do a post on this for some time but nothing came up – but I had to make some seals for one of my industrial customers, so I’ll show you the process.

Some bits that I have made :-


Polyurethane can be gravity cast into open moulds (degas mixture in a vacuum pot before pouring) or injection moulded into a closed cavity mould.

I am going to demonstrate a closed cavity mould.

Firstly the usual safety precautions – polyuerethanes contain unreacted isocyanates – prolonged skin contact should be avoided. Observe safety instructions supplied with product.

Cured and uncured polyurethane can be removed with Methelene Chloride – volatile but not flammable – decomposition products if burnt can be nasty. Minor / brief accidental skin contact O.K. but longer will produce burns.

Despite the warnings it is not too hazardous – just don’t get the stuff in your eyes – methelene chloride in your eyes is excruciating.

The process consists of using two “pressure vessels”, an upper vessel to hold the resin for degassing prior to injecting it into the mould held in the lower vessel.

My upper vessel is a discarded industrial water filter and my lower vessel is a discarded truck oil filter (pretty low tech). The lower vessel only needs to handle vacuum (in this case I just drop the lid on – it does have hold downs but these are never required), the upper vessel needs to handle vacuum and pressure – in this case it is “O” ring sealed and has a locking collar – You don’t need to use the collar under vacuum but you’d damn well better remember to fit it before applying pressure or you are going to end up with one helluva mess (guess how I know this ?).

For a vacuum, I originally used the suction side of my compressor but this was not very effective – only getting down to about 0.9 of full vacuum. I then used a discarded fridge compressor (a big one) – that got down to about 0.95 and was a big improvement but eventually I scored a 5 lpm Edwards vacuum pump which can pull a pretty mean vacuum.

You need vacuum to degas the resins, mould etc – followed by pressure (from the compressor) to push the resin into the mould cavity.
When degassing the resin any aeration bubbles from mixing as well as dissolved gasses will expand dramatically rising rapidly to the surface of the mix during degassing.

Note: During degassing the mixture will froth up to about 3 times the depth of the original mixture – so be sure you have enough headroom in your upper vessel.

This frothing can be diminished by cycling the vacuum – releasing it at intervals – before reaching full vacuum – however polyurethane resins go off fairly quickly and you don’t have time to muck about.

Any residual bubbles will be squashed by the application of pressure (example a 5mm diameter bubble at 1mBar will become a 0.3mm bubble at 5 Bar. First prize is to eliminate bubbles – best done by a good vacuum – you may note in this application the mould has riser sprues to allow any residual bubbles to migrate up into – you can see the vestiges of bubbles in the top of two of the riser sprues later.

Part being made this post :-


This is the finished part, riser sprues & flashing removed. It’s a custom seal for Nitrogen pressurizing shockabsorbers (I designed the filling machine).

Being made in this mould :-


The mould is screwed on to a spigot (1/8” BSP fitting) through the lid of the lower vacuum vessel – this in turn is connected to the upper vessel via 6mm bore Festo PU pipe and standard Festo fittings.


The mould, upper vessel and connecting pipe are coated with release agent – I use a brush for the mould surfaces and slosh some around in the upper vessel and drain it out via the pipe.

Most release agents are silicone based – this can create havoc with spray painting (fish eyes) so keep well away from spray painting areas or make sure you get a non-silicone release agent.

I then connect my vacuum pump to both vessels for about 10 minutes – this pulls a vacuum in both vessels and the mould, this degasses any residual solvents in the release agent.


The pipe between the vessels is then pinched closed with vice grips and the vacuum closed off at the lower vessel (no point in losing the vacuum you have already built up).

The photo above shows the rig after initial vacuuming with the pipe and vaccum line pinched closed. The additional set of grips is to close the pipe to the lower vessel (the leg of the T immediately above the grips) which I had not done yet but must still do so.

The upper vessel is suspended on a cable from my roof.

Release the vacuum in the upper vessel and open it.

Mix 2 part resins (accelerator if required) and pigment to suit – mix it in a disposable paper cup using a tongue depressor as a stirrer. (all this stuff generally available from the resin supplier) – mix enough resin for the component and the residue in the upper vessel (I generally add about 90g which is a lot of waste for a 30g part – but what the heck).

You may have noticed the mould has an “out” hole – which is plugged – sometimes I stack moulds and do two or three parts at a time.

Weigh the residue after your first cast and reduce the amount if you can.

Beware of short changing yourself – if you come up short, when you are using the compressed air to push the resin into the mould and you push air into the mould, the show’s over and you can start again.

If you do screw up, never make the mistake of opening a “wet” mould – you will face an unholy mess – wait three or four hours for the gunk to cure before attempting to strip it down.

Note: Unlike epoxies & polyesters, Polyurethane resins are NOT tolerant to mixing ratio errors – so use a digital scale and be precise.


Mix thoroughly – adding pigment helps in that when you reach colour consistency then it is properly mixed – same caution applies – polyurethanes are intolerant of mixing errors.

Next – pour mixture into upper vessel.


Note top end above my left hand penduluming on its cable (blurr).

Reattatch and apply vacuum to degas.

Reclose the vacuum line and remove from upper vessel to let the atmosphere in slowly – if you just simply pull the pipe off the rush of air into the vessel may stir up the mix and entrain air once again.

Remove the clamp from the pipe between the upper and lower vessels. Atmospheric pressure is now “pushing” the resin into the vacuumed mould.

Attach a closed compressed air line to the upper vessel and apply (as much pressure as your compressor will produce) pressure – again apply it slowly to avoid aeration of the mixture and potentially causing the mixture to “rathole” air down into the mould.

Keep an eye on the consistency of the residue in the paper cup – if it starts to thicken – you are running out of time and might need to get a move on.

Other than that pace yourself – it works best if done slowly – but you can’t take too long.


Rig with pressure to upper vessel, intermediate pipe open and vacuum to lower vessel (pipe previously connected to upper vessel was pinched off and removed before opening the vessel)

Pinch off the pressure line (so you can reclaim your airline) to leave pressure on the upper vessel. Pinch off the vacuum line to maintain the vacuum in the lower vessel – turn off the vacuum pump & disconnect (it has happened to me once that the vacuum pulled the compressor fluid out of the {stopped} pump and into the lower vessel – so now I disconnect as part of the routine.)

Wait – generally 3-4 hours – so I normally leave it overnight.


Here’s the mould removed – you can see the weeping – it does not seal entirely (this is not injection mould standards here – flycut aluminium).

Also showing the sprues and scrap material from the upper vessel and connecting tube.


Here’s the part removed from the mould “warts and all” – there are two very tiny bubbles in the tips of the two tallest risers – but note the bubbles in the leader scrap – I came perilously close to blowing air into the mould – probably because I got all out of sync taking photos – screwed up my usual routine – I let the air in way too fast and it “ratholed” – sometimes you get lucky, sometimes you don’t.
That said a checklist might not be a bad idea.

Some Additional Tips & Tricks

The resins can go off really fast in hot weather – if this becomes a problem, chill pre-weighed quantities in a fridge before mixing.

If your mould leaks slightly, try pre-heating the mould to 50-60 C° - the resin trying to sneak through the leak cures rapidly to seal the mould.

In very cold weather is it advisable to pre-heat the mould to 50-60 C° anyway.

If your mould leaks like a sieve, in addition to heating, remove the pressure and vacuum as soon as you are confident that the mould is filled (not best practice but it works – kinda halfway between injection and gravity casting).

Mould leaks can also be cured by cleaning one surface of the leaky joint and adding epoxy resin with release agent on the mating face, assemble the mould – gap will be exactly filled by the epoxy.

Gravity cast parts can be cast in silicone rubber moulds thrown from sample parts.

Closed moulds can also be cast off sample parts using hard polyurethanes – obviously cast in two go’s – set part in modeling clay to split line – cast one side – remove clay and repeat for the other side – drill and tap clearance etc for securing bolts, sprues etc. This will only work for parts with a suitable draw angle – otherwise use silicone moulds – see next.

Injection moulding is also possible with silicone moulds cast about a part in a suitable vessel vis :-


This is a louver door segment on an Ami E-120 jukebox (1953) that I restored. It’s the biggest part I have made – its made of rigid polyurethane.

This is the mould :-


An MDF cavity with draw angle (liberally laquered to seal), steel plates, “O” rings etc. Silcone was first cast around sample part in the cavity. Silicone cut to split line with a scalpel (plan ahead) and subsequent parts cast in rigid clear (pigmented to white translucent) polyurethane.

There are about 20 M6 readybolt tiebars holding this together.

Rigid polyurethanes can also be mixed with Aluminium powder 1:2 - when cured it can be machined, drilled. Tapped etc. it even takes a polish and looks like aluminium – I’ll do a post on this some other time as this is waaaayyyy too long already.



Well-Known Member
Mar 18, 2008
Reaction score

Thanks for the tutorial on Injection / Vacuum Moulding in Polyurethane. I am sure that a few here will put this to good use. I am going to put this in my "how to" file for future use. :bow: :bow:

Cheers :)




Well-Known Member
Oct 28, 2009
Reaction score
THANKS, now more projects to clutter up the shop ;D ;D


New Member
Mar 17, 2012
Reaction score

Holy moley! That was an awesome tutorial. :bow:

Questions for you...

Q: What is the primary advantage of injection casting with your method versus the typical method of pouring the casting material into a mold and removing the bubbles with a pressure pot?

Q: Do I understand you said I can use your injection method with a rubber mold so long as the rubber mold is in a container box for support? I do not currently have the ability to create metal molds.


Ken I

Project of the Month Winner!!!
Project of the Month Winner
Oct 31, 2010
Reaction score
Cape Town, South Africa
The primary advantages are :-

1) You don't need to worry about "blind" voids - cavities which refuse to fill under gravity casting - you normally get around the problem with orientation but via vaccum / pressure it makes little to no difference. As you said you can still place your gravity poured mould into the vacuum chamber to accomplish the same thing - so no biggie.

2) For reasons unclear to me there is practically no shinkage by this method so you can make accurate parts (which is important on some of my applications). Also with gravity you have an uncontrolled surface which must be oversize and trimmed or you live with whatever comes out and its miniscus.
The parts I make do not have any unfinished sides.
Again some of this is moot if you can drop a flat plate onto your gravity cast part to finally close off to finish the unfinished side.

3) Far less trouble with bubbles, particularly where surface tension in detailed or thin sections gives rise to poor reproduction or bubbles "clinging" to those areas.
Again the vaccum expands all bubbles but at some point whatever bubbles are remaining when you let the atmosphere in will be significantly squashed - by applying 7 atmospheres this "squashing" is seven times better.
There are always going to be some bubbles and with polyurethanes going off as quickly as they do you are limited to how long you can apply the vacuum (diminishing returns and all....).
Again if your "pressure pot" can do both the vaccum and the pressure and you can close off your "unfinished" side, you have essentially accomplished the same thing via a different route.

As to your next question - Yes you can make non-metal pressurised moulds - the Louvre in my tutorial was originally built completely from MDF including the side plates but they started to deform with use so I replaced them with 10mm metal slabs (flame cut, dressed and hand drilled) - but the rest of the "cavity box" is still MDF.

I have a number of silicone moulds cast inside MDF boxes - in smaller cases I clamp them closed with G clamps.

Also this louvre box is too big to fit into my vacuum pot.

I suppose you could also use a vacuum bag ?

So you can make your mould in much the same way you might make your gravity moulds - the only difference you do it in a closed box (ie a box you can open and close) to handle the pressure.

A word of caution - make sure that the box can handle the pressure you don't want it blowing up on you (the material is too viscous to actually blow up but there is a loud crack of the box splitting followed by goo all over the place).

In my louvre example the exposed area is about 40 sq" @ 100 psi = 4000 lbsf - near enough two tonnes trying to push the plates apart - allowing for the pressure up to the "O" rings nearly trebbles this to guestimate 6 tonnes - plates held together by 24 M6 threaded rod tiebars or about ¼ Tonne each which are good for about 3/4 tonne in commercial or 1.5 tonne in high tensile - plenty safety factor.

But you can also see why the original 16mm MDF side plates started to deform after a handfull of parts.

Hope this enlightens.

P.S. I still do ordinary gravity casting where is is appropriate, if there's no advantage to the application then simpler is better.