Small scale vacuum resin infusion.

Part one.


26'th July 2006.
This page is not a detailed description of what vacuum infusion (or whatever you want to call it) is. There is a bit of stuff on the web describing this technique but most infusion work on the web is for boat-building.
There is an excellent example here -
also see

But working on the model scale is a little different and that is what this page is about.
To see the sort of things I make see my model paraplane.

I have occasionally done something crudely resembling vacuum infusion by simply adding resin into the bottom of a plastic bag, placing my “job” next to it, then a resin trap made of tissue paper and shoving a vacuum line into the open end and then sealing the bag around it.

This kind of works (sometimes) but there is the same time pressure that you have with other wet layups - you have resin with a limited working time and the clock is ticking. When things go wrong things can get really messy really quickly. The other problem was the infusion was slow and unless I helped it along I could only get around 50mm infused before the resin gelled.

A bit of research (including the links above) pointed to the lack of a resin transport/flow layer as the reason to this slowness.
Today's test was partly see if any material I had around to house would work as a flow medium.
I could have taken a drive to my fibreglass supplier and bought all the proper stuff (and I probably will) but this was only a rough test and the whole set-up was new so I was bound to learn something even if the flow media didn't work.

I tested hessian, chux wipes, paper towel and cotton flannelette. I would also have liked to test – cotton towel, wool, shade cloth, geo-textile and of course real flow media.

The set-up (click to enlarge).

Infusion port (yellow disk on the left).

I turned a disk of plastic on my lathe (LDPE, re-melted milk bottle tops). The disk has a hole bored in it which will take a syringe nozzle with a tight fit. I also machined a hollow around the hole. It might be beneficial to have some radial groves on the underside but this one doesn't. The disk is placed inside the plastic bag, after the vacuum is drawn and everything is sitting in place – a syringe is jammed into the hole. Today the plastic didn't puncture when I did this, so I had to remove it and break the seal with a drill bit. This has to be an air tight seal. The reason I machined a depression is so resin can be poured there as a leak test and leak inhibitor. Having a pool of resin there could also allow the hole to be unplugged under certain circumstances without letting air in - I will explain why later.

I choose this LDPE plastic because resin doesn't bond to it. I'm sure there are many alternatives to what I've done. I would suggest a rubber suction cap or a large pencil eraser could also be easily modified for this purpose.

Underneath the port I placed a strip cut from a scouring pad to give a low resistance path for the incoming resin.

The test laminate.
The laminate was laid up on a plastic cutting board (again LDPE) it consists of three layer of woven glass matt (around 150 GSM). This is covered by peel-ply and then has the flow media candidates on top. Everything (including board) is put inside a clear plastic bag which was sticky-taped shut.

The vacuum port (white disk on the right).
This is quite similar to the infusion port. The airtight seal isn't quite so critical here, apart from that the two ports are interchangeable. I made this one from some scrap silicone (silicone which had gone hard inside it's tube). I cut it into disk and bored a hole through it. The little nozzle on the vacuum line is sold for mini irrigation systems (garden variety). Another piece of pad is placed under this port as well.

The resin trap (at the back).
This is just a jar with a couple of nozzles glued into the lid. It just catches resin before it reaches the pump. To re-use the jar I put a cut-off plastic cup inside so I can get the cured resin out.
The pump is a dynavac single stage vane pump. You need a strong vacuum (don't even think of using a vacuum cleaner) but if the bag is well seal you shouldn't need a big pump. I'd like to try a peristaltic pump some time as the dynavac is very noisy.

This is what happened.

I poured 25ml of resin into the 25ml syringe. The resin infused fairly rapidly. I immediately mixed another batch and only 5ml of the first batch was left when I was ready. The pad rapidly saturated and the resin appeared to be boiling. Resin pooled inside the bag near the port. As soon as the resin front hit the hessian it was clear this was going to win the race. The resin travelled along the hessian about twice as fast as through the glass. The hessian is supplying resin to the surrounding area as well so it would be even faster if the entire surface was covered with it. The chux and paper towel didn't seem to help at all, the chux is only in ahead because it is getting resin from the hessian as well.
The cotton actually retarded the resin flow, the flow through the cotton cloth was clearly lagging behind flow in the glass. The resin didn't reach pad on the vacuum port because the resin started to gel too soon.

Something to ponder.

During infusion the pressure inside the bag near the infusion port is close to ambient. Resin pools here and the laminate is not being compacted as we would like. The pressure in this region only drops after the resin inflow has been stopped and excess resin is removed from the area as it flows toward the exit.
It seems to me that this would be much improved if – after infusion - vacuum is applied to the infusion port to rapidly remove the excess resin from this area.
On the big projects where resin goes in via a tube (of course I could do that too) it would be fairly trivial to arrange the plumbing to allow an easy change over.
In my minimalist system I think I could simply unplug the syringe and plug in a vacuum line.
By the time the resin has completely infused - the flow rate is very slow and as I hinted earlier some extra resin in the dimple around the inlet would prevent air entering. Even if some air entered it shouldn't reach the laminate and it would be immediate sucked back out when vacuum is applied.
I suspect you could drop the vacuum for a few second without wrecking the job. There are other options I won't go into.

27'th July 2006.
Today I bought some proper flow material. It is useless as a flow material for my stuff. It is way too thick and drapes about as well as a piece of toast. When you cut it - the spiky bits can puncture the thin bags I'm using.

All images link to larger versions.

Today's test laminate is carbon fibre, four layers of 135 GSM cloth (25/75 weave). The flow material is shade cloth. The enkafusion flow material is under the two ports. Note the the infusion port is still sealed in the above photo. Inserting the syringe punctured the plastic and the port worked perfectly.

This is what happened.

I mixed up a bigger batch of resin than yesterday (60ml). It infused at alarming speed and by the time I hurriedly mixed a second batch the resin had already reached the vacuum side of the laminate. The resin appeared to be boiling (and it probably was).

The mid line connecting the ports was fairly bubble free but further out towards the edges seemed to be mostly vapour. I don't know what was happening in the carbon. I made a little video clip which shows the bubbling accompanied by the sweet sound of a dynavac (1.3 meg).

It only took a few more minutes for bubbles to appear in the vacuum line. Here is a longer clip (2.6 meg).

After a while there didn't seem to be much point in continuing to let resin into the system. I tried reducing the vacuum a little by uncapped the second vacuum line but it didn't make much difference. When the flow media is as open as this - my previous comment about dual vacuum lines doesn't apply but I wanted to try it anyway. The flow of resin into the infusion port had slowed quite a lot. I pooled extra resin in around the infusion point and unplugged the syringe. I delayed a little to obverse the flow and a little air did get sucked in. I plugged in the second vacuum line and the boiling increased and resin immediately began to bubble through the line. More vapour appeared in the shade cloth even along the mid line. The resin in the resin trap was also boiling (supporting the theory that these weren't just air bubbles)

And another clip (2.5meg)

Being impatient I shone a desk lamp onto the job to warm it and speed up the curing. Eventually the bubbles in the lines slowly down as the resin thickened and the left line more so than the right. The reason was probably that there was an air leak on the right side due to a puncture from the enkafusion.
When both lines stopped (sealed off by resin) I disconnect the pump and a while later heated the job to 70 deg in my oven to cure it.

Once cured, the carbon was very difficult to separate from the other layers. In particular the enkafusion becomes rather tough to move once it is full of cured resin. This could be a real problem on a complex shape.

The laminate had some blemishes on the underside (the cutting board side). This was probably due to my hurried mixing on the second batch of resin. The top side looks as good as any I've ever made.

More to ponder.

Today's results are looking promising.

It is pretty obvious resin will boil if you pull a hard enough vacuum. I don't remember anyone mentioning this before or flagging it as a problem. On a large job like a boat - the pump has a lot more work to do and it may not be able to maintain a high enough vacuum for noticeable boiling. On this small scale my pump can pull a hard vacuum in a few seconds and maintain it even when the resin boils. Despite all this the resin seemed to cure OK.

The initial infusion was too fast for comfort. If I use these materials again I will add a valve or clamp to reduce the vacuum in the early stages. I think I will still go to full vacuum in the later stages.

I don't really think I will use this enkafusion/shade cloth combination for the model work expect maybe for flat laminates. At the moment I think hessian is the top candidate because it drapes better. I think it will also be easier to separate it from the laminate. Strips of shade cloth or even enkafusion will be used but not over the entire surface. A better flow material may still be found and maybe a suitable release film.

I used a ridiculous amount of resin today. I mixed 120 ml to make a 23 gram laminate. Around 20 ml was unused and almost 40ml ended up in the resin trap. These aren't totally wasted because I can use the hard pieces for machining. I think the amount of resin wasted can be significantly reduced.

28'th July 2006.
Today's test is again 4 layers of CF. The flow media is hessian and strips of shade cloth were used at the ports.

Sixty ml of resin was carefully mixed. The vacuum line was clamped to reduce the initial panic phase but was not needed. The initial infusion was rapid then progressively slowing down. There was a bit of boiling but very little compared to yesterday. I estimate it took under five minutes to reach the stage in the first photo.

It took another seven minutes to get this far. Notice how well defined and straight the resin boundary is.

It took almost forty minutes to reach the shade cloth at the vacuum end. Five ml of resin still remained unused. Note the resin at the edge is a little more advanced than further in, this is because there is no CF at the edge.
I swapped the resin supply for a vacuum line and resin immediately went up the line to the resin trap. Initially there was a lot of resin around the infusion site but it was eventually removed. At the other end resin continued to creep towards the port and all the hessian was eventually filled with resin.

The result.
The laminate was the driest looking laminate I've ever made without it falling apart. There is virtually no resin in the spaces in the weave and there are pinholes right through were the weave patterns had line up. This laminate would not be water proof. However it appears to be well bounded there are no signs of de-lamination. It held together even when severely flex. Again it was difficult to remove it from the other layers but not as difficult as yesterday. The piece is of similar size to yesterday's sample but three grams lighter. I'm guessing a ten percent weight reduction which probably means twenty percent less resin in the fibre. I suspect the capillary action in the hessian has something to do with it. The looser weave of the shade cloth I used as the flow media yesterday might not compact the CF as well as the hessian does. These laminates are about 0.8 mm thick.

Continued on part two.

Cheers Eddie.M.

Up one level and my home page is here.