|
Group: Forum Members
Posts: 2.5K,
Visits: 28K
|
Hey all, I have been doing some tests the last few days to compare various infusion core materials. My primary purpose for these tests was to determine the resin uptake of each core, as well as performance as an internal flow media. My findings were interesting, so I thought I would post them up for others to see. The tests were basic. I used 4 types of core that I frequently use in infusion; Soric, 3D PET, and 2 different types of Amorim Corecork: NL-10 and NL-20. Unfortunately I only had 3mm thick Soric/3D PET, and 2mm thick of each type of Corecork. I would have liked to have tested equal sized cores, but it doesn't matter for the purpose of these tests. So, I cut each core into equal sized pieces of 180x175mm (would have done exact squares, but I cut the PET core first and just cut it between the cells). I then weighed each core:  The PET core came in at 11 grams.  The Soric was the lightest dry core at 7 grams.  The NL-20 was the heaviest dry core at 14 grams.  And the NL-10 came in close to the PET core at 10 grams. I then cut 2 plies of 200gsm carbon for each core, weighing in at 13 grams:  So I could record how long each piece took to infuse, I placed the resin inlet directly on the part (as equally spaced as I could estimate) without any flow media, and bagged each part separately.  One thing to note here is the print of the Soric and the PET core compared to the Corecork. Now the fun could commence! Believing that the Soric would have the highest resin uptake, I used the manufacturer's data to calculate how much resin the Soric panel would need to infuse. For an area of 0.0306sqm, the carbon plies should take 8 grams, while the Soric should need 50 grams. I allowed for a bit of wastage in the pot and resin lines, and mixed 100 grams of Araldite LY 3600. I also added some translucent red pigment to give a visible indication of the resin in the core when I dissect the panels later.  So, Soric first.  Note the very even flow path. Soric certainly works very well as an internal flow media, and the piece was completely infused after 3 minutes 55 seconds. Moving onto the PET core, again mixing 100 grams of resin:  Again the PET core works well as an internal flow media with an even flow path, and slightly more interesting to watch! A touch slower than the Soric, this piece was completely infused after 4 minutes and 5 seconds. Then the NL-20:  This core is completely useless as an internal flow media. This photo was taken at 4 minutes 5 seconds to show how much faster the PET core and Soric facilitate flow. Also note the poor flow path throughout the part, with the fastest channels being around the outer edges, no doubt due to the slight bridging from the 90 degree edge. This piece was completely infused after 20 minutes. Pretty safe to conclude that this core doesn't operate well as a flow media, and should be used with mesh to assist flow. And finally, the NL-10:  This core worked much better as an internal flow media. Whilst not quite as good as the Soric or PET core, it was certainly much faster than the NL-20 and maintained an even flow path. Interestingly, the first half of the sheet infused in 3 minutes which is when this photo was taken. However the infusion was only fully completed after 10 minutes. I would suggest that this core may be useful as an internal flow media across smaller parts, or with flow mesh assisting on larger parts if pot-life is going to become an issue. With the infusions done, this photo of the remaining resin from each pot gives a pretty good visual representation of the resin uptake of each core:  From left to right, Soric > 3D PET core > NL-10 > NL-20. I allowed the sheets to cure for 24 hours before demoulding. I then filed each sheet down to 175x165mm to get rid of excess resin and wastage on the edges. For those following the maths, this brought the total area down to 0.028875sqm, a 5.6% reduction from the original size. To allow for this, I subtracted 5.6% of the weight from my original dry measurements, so the carbon plies became 11.55g, the Soric became 6.6g, the PET core 10.38g, the NL-20 13.2g and the NL-10 became 9.44g. I then weighed each sheet in the above order; Soric > PET core > NL-20 > NL-10.   Noting again the print through from the 3D PET core.   So, the obvious conclusions looking purely at the weight is that the Corecork produces much lighter parts than the Soric, and moderately lighter parts than the 3D PET core. The similarity in weight between the NL-10 and the NL-20 is interesting given their densities, which gives you an insight into their respective resin uptake. So, my calculations for the resin uptake of each core went like this: The carbon plies were a known weight at 11.55g, and required 7.7g of resin. Therefore the total weight of the carbon in each sheet can be assumed to be 19.25g. Subtracting that weight and the DRY weight of each core from the total weight of each sheet leaves you with the weight of the resin in each core, giving me the following values: Soric = 49.15g resin 3D PET = 26.37g resin NL-20 = 12.55g resin NL-10 = 17.31g resin To express that in a term that can be compared to the manufacturer's data, I divided each of those values by 0.028875 to give me the resin uptake per square meter, and then divided that figure by the thickness of each respective core to indicate the resin consumption per square meter per millimetre thickness, giving me the following: Soric = 567g/m2/mm 3D PET = 304g/m2/mm NL-20 = 217g/m2/mm NL-10 = 299g/m2/mm For what it's worth, these are the figures quoted by each manufacturer (excluding the PET core which has no published data, hence my tests): Soric: 550g/m2/mm NL-20 = 170g/m2/mm NL-10 = 270g/m2/mm I believe the variance between my measurements and the manufacturer's measurements is likely down to the fibre volume fraction. When I infuse, I normally let the resin reach the outlet before clamping the vacuum and allowing the resin feed to stay open for 10 seconds. On small panels such as these, this is likely to increase the resin content of the part quite substantially, creating a margin of error. Also, it may have been a mistake placing the resin inlet port on top of each panel, as this of course meant that there was no compaction on the part in the grooves of the connector, again allowing a higher resin content. Despite that, I believe my results to provide a representation of the resin requirements of each of these cores, and hence I draw the following conclusions. In terms of all-round performance, I believe the 3D PET core to be the best. It works very well as an internal flow media, and doesn't consume a lot of resin. The biggest drawback from these tests appears to be the higher print-through from the core cells onto the surface of the part. In terms of out-and-out weight performance, the Corecork is the obvious leader at nearly 20% lighter panels than the 3D PET core, and 40% lighter panels than the Soric. There is also very little print through onto the surface of the panel, with just a slight random texture becoming visible. The difference between the two types of cork is negligible, with the lower density of the NL-10 being negated by its higher resin uptake. The obvious downside of the cork is its performance as a flow media, and in most cases you would probably want to assist flow with mesh. An unfortunate circumstance if you are interested in reducing resin wastage. Anyway, I hope this information serves useful to someone!
|
|
Group: Forum Members
Posts: 2.5K,
Visits: 28K
|
Out of interest, does anyone know if the figures I have measured increase linearly with thickness? So say the 3D PET core for example, if you were to use a 15mm thick core on a 1sqm 2-skin panel; if these figures were linear then you would need 4.5kg of resin just for the core? I just feel like that isn't correct. I feel another test coming on...
|
|
Group: Forum Members
Posts: 292,
Visits: 4.5K
|
Interesting test and thanks for sharing, I wonder how the core materials would compare regarding strength. Chris Carbon fibre works LTD
|
|
Group: Forum Members
Posts: 2.5K,
Visits: 28K
|
Yeh I'd like to test for structural performance, but given the Soric and 3D PET cores are 3mm and the Corecork pieces are 2mm, the results wouldn't correlate. I've attached the Corecork TDS if you're interested, they have done a strength comparison in there with some unspecified synthetic cores. Next time I get some Corecork I will pick up some 3mm and do a direct comparison.
|
|
Group: Administrators
Posts: 2.5K,
Visits: 8.5K
|
Very good test. The main way I could see an improvement would be as you say, different thickness to confirm a linear resin uptake as thickness increases (which is what logic would dictate). Also a control sample with no core would be useful as then you could calculate your fibre/resin fraction and from your figures exactly subtract a more accurate amount of resin used in the cloth leaving more accurate consumption figures for the cores alone.
Warren Penalver
Easy Composites / Carbon Mods - Technical Support Assistant
|