Very detailed DIY for a Vacuum Bagged Carbon Fiber Car Door

Let me begin this how-to by laying out two things upfront. First and foremost I am not a professional composite person nor am I an engineer. Second, I can’t begin to answer all of the design and engineering questions many of you have already had and will continue to have. This write-up is focused purely on the PROCESS, My process of making molds and parts and NOT how to attach them. I will not address how to handle various load situations either - i.e. will the roof I made work on a 200+ mph Pro 5.0 car. With that disclaimer, let’s jump in.

I will be breaking the process how-to down in multiple steps or phases. These steps are both long and involved. I personally prefer it when I am provided too much information over too little. For those that would prefer a shorter version, sorry. Also, like my one year old son and three year old daughter, I like pictures and as you will soon find out use them extensively.

***Please follow product safety information provided by all material suppliers** Some of this stuff is pretty noxious and I am sure can cause all sorts problems with prolonged exposure. Some of this stuff may be flammable as well and I can guarantee you that fiberglass and carbon fiber bits are not good to breathe so wear a respirator.

Materials:

Below is the stuff I use. I have used more products that are not listed here but have settled in on a these materials. They work well for ME. That doesn’t mean that you can’t accomplish the same thing with other stuff. If you find something that I am not using that works well, let me know as I am always open to try new products and procedures to improve the process.

Tools and Non-consumables:

-Gel Coat Cup Gun

-Paint spray gun with 2mm tip (Obviously you will need a compressor and my 4 hp Craftsman works fine)

-Mold release wedges (I have four which is a minimum for big parts)

-Scale (.1 oz resolution)

-Large table with a stick resistant top. I use a 3/4 in. thick 4x8 sheet of melamine covered fiberboard on top of saw horses. You can get one for $25 at Home Depot or Lowes. This works extremely well as it is cheap, large and very resistant to stick. It is also pretty durable and double sided which is great as the table will take a serious beating.

-Tarp to cover the ground as it will be covered in resin

-Scissors

-Tin Snips

-Circular Material Cutter w/extra blades

-Squeeze dispenser for polyester resin catalyst

-Composite rollers

-Body filler spreaders

-Rivet gun

-Hot melt glue gun

-Vacuum attachment for bagging

-Vacuum pump (lots more on this later)

Consumables:

-Kleen Clay modeling clay

-Fiberglass Chop Strand Mat (CSM) (I use 1.5 oz. I originally bough 75 yards of this which has made a lot of molds. Most of my molds are 3 to 4 layers thick and a yard is 50” wide. So for the door the outer skin mold required 4 yards of CSM).

-Polyester Resin (I use ‘tooling resin’ from US Composites which has worked great for me as it is super tough and is not tacky after cure)

-Polyester Gel Coat (I use black also from US Composites because it is easy to see blemishes and easy to see where the fiberglass has been wetted out)

-Tongue depressors (you will want lots and lots of these)

-Cup gun cups (I use a ton of these too as I use them not only for spraying gel coat but mixing resins as well)

-Sand paper (wet 240 and 400 grit)

-Sanding block

-Masking tape (3/4” and 2”)

-4.5” Grinder with thin cutting disk

-Acetone (cleanup)

-Aluminum flanging materials (I like the 50 foot .016x6” roll I get from McMaster-Car and also use .040 sheet and cut with pneumatic sheers for places that require stiffer flanges)

-Aluminum angle stock (I like to use the 1/16” thick 1x1” and 2x2” 90 deg angles because I can cut it with snips)

-Rivets 1/8” aluminum 1/2” long (I love rivets as you will soon see)

-Hot Melt glue gun sticks (the best of these I have found are the ‘High Strength long cure” variety sold at Home Depot)

-Part-All #2 Release wax

-PVA (poly vinyl alcohol)

-PTFE Release agent

-Nitrile rubber gloves

-Carbon Fiber fabric (I use 5.7 oz 2x2 twill for 95% of my parts. 50” widths are common, 60” are not which stinks as 60” saves lots of material for doors and trunk and is a requirement for a seamless hood on a SN95 Mustang.)

-Nomex Honeycomb Core (1/8” - 1/2”  thick 3.0 lb./sq ft. depending on part - AVT is my source for honeycomb)

-Perforated release film (US Composites)

-Breather Ply (US Composites)

-High elongation bagging film (US Composites)

-Bag sealing tape (US Composites)

-Epoxy Resin (I use the medium cure thin from US Composites for most of my parts and I have been happy with it)

   

Phase 1: Planning  

You will soon learn that everything you do now will affect not only what you are working on but everything that you will make from that point forward. If the original part has a problem that is not addressed, the mold will also have that problem as will the finished part. Planning can prevent these problems from moving forward. Often you will not realize that you are going to have a problem until the lay-up stage and then it is too late and the mold has to be redone. BTW, this phase of the process will get easier with each part you make. There is nothing like learning from your own mistakes. I am the master of this.  

Let’s talk about why a concave part requires a multi-part mold.

These diagrams above are obviously simplistic but illustrate what to watch out for. The first illustration will be fine to use a one part mold; the second is marginal but will probably work ok as a one part mold if the part or mold has any flex to it. The third illustration will definitely require a multi-part mold. Again, this is simplistic and in the real world the part shape will be much more complex. All it takes is one little concave ridge or spot to ruin a mold. The good news is that we have options to deal with these small areas which we will address later. Another piece of good news is that most production parts are stamped which also requires that a part not be concave.  

BTW, when I made my first roof mold I did not think that the concavity of the window ledge flanges relative to each other would be a problem. They were and it cost me a carbon/honeycomb part and a complete mold to learn this lesson.  

This is a good time to address physical mold release. Another problem I ran into with the roof was that the shape did not allow me to insert release wedges up into the main central area of the roof. So there will be situations where even though the part is not concave, you may want to make a multi-part mold. This is exactly the case with the interior door structure for this part. The shape would not allow me to get wedges into to large flat section so I split the mold down the middle of the roof.  

For my door there were several issues that I had to get my arms around up front: How much door did I want to make and if I was going to make an interior structure, how much to make? How was I going to attach the door? Do I want to use the stock latch? How strong does the door need to be in flex and in impact? At what point do I sacrifice the above for weight?  

From here I came away with two options. One, make as light a door as possible - i.e. just an outer skin - and two, make one with a stock matching interior structure but with a smooth interior panel that would create a completely enclosed structure. Ultimately I chose to make the molds such that I would not have to make this decision now. This would require a single mold of the exterior panel and a separate mold of the interior structure. If I wanted a part with an interior structure I would simply lay-up carbon parts in both molds and bond them together (more on this later).  

The first challenge that jumped out at me was the shape of the stock door’s interior structure. As you can see in the picture above, it is quite complex with lots of tight curves and angle changes. I wanted a large flat panel on the inside so I had to create a shelf for a panel to sit on. You can clearly see in the picture how I was trimming the inner structure to accomplish this. I also have to work out how the inner carbon fiber structure was going to bond to the top outer carbon skin. More on this later.

The next item to address was the door handle and lock areas on the outer skin. As the doors I am making are going to be in the 5-6 lbs. weight range I decided to eliminate these parts and to make the door smooth completely smooth. To do this I will filled these areas with clay on the stock part which keeps them from being transferred to the mold. More on this in the next phase.  

After studying the door for literally hours I finally came to the conclusion that the inner structure would require a multi-part mold to guarantee release and that the seam between the molds should bisect the large flat section of the middle of the door. I chose to seam it vertically to make the mold halves square instead of long rectangles. That way the seam would be smaller and the mold halves would have more structural integrity. Also, I could have chosen to seam it along the edge of the inner panel but that would have been much more difficult and as this was not an exterior panel I chose the faster route.  

Phase 2 - Part Preparation  

Now that I had a plan it was time to make the initial mold. The first step in this process is to completely clean the part inside and out. By completely clean I mean 100% free of grit and completely smooth. For this I love to use one of the clay bar cleaning bars. If your stock part has any major defects then this is a good time to address them using common body repair methods - welding, body filler, hammer, etc. Just get the surface perfectly smooth with a minimum of porosity as that will prevent the mold from sticking to the part. Remember any imperfections in the part will be in the mold and any imperfections in the mold will be in the finished part.  

BTW, my original door was in pretty good shape except for an ‘outward ding’ that came from the removal of the side impact door bar. This I chose to address on the mold as it would be an inward depression. More on this later.  

One thing I haven’t discussed yet is flanging. When vacuum bagging it is, in my opinion, highly desirable to have a flanged mold. This does two things, principally it gives you a smooth surface that surrounds the part to enable you to bag up the part (this will make more since later) and it also makes the finished part oversized which makes it easier to release from the mold and allows you to trim the part back to match the original parts dimensions.  

For flanges I shoot for a minimum of 2” and prefer 4”. Much bigger than this and you are wasting fiberglass and gel coat. To make the flange I like to use strips of 6” wide x .016” aluminum and glue them to the underside of the edges of the part. Often this will require trimming and radiusing of the aluminum to insure that the flange is continuous and has good adhesion to the part.  I will then trim the excess flange of to give me my desired 2-4” size. 

Above you will see a great example of the beginning of the flanging process. I held the aluminum under the part and trace the profile with a marker. I then trimmed along the line which resulted in a part that was ready to apply to the inside door flange. BTW, you must make sure the aluminum follows the contour of the part in every plane by pressing the aluminum up to meet the stock part continuously. You absolutely must make sure that there is no floating gap between the flange and the part. If there is then resin will seep into the seam and cause all sorts of problems.    

Click Read More to Continue

Once this piece is trimmed I then glued it down to the flange which in this case was the inside of the door as I am molding the exterior. You can see in the above photo the flange on the opposite end of the door that has been glued down. Notice that I have riveted short pieces of aluminum angles to wedge the flange down. While the thin aluminum flanging was glued down with a continuous stream of hot melt, this glue did not have a ton of adhesion to the stock part or flange and needed the additional support to prevent the flange from peeling off.    

Where the various aluminum flanges overlapped they had to be completely glued together to prevent gel coat from filling the crack. I want the transitions to be smooth as they serve as the vacuum seal later. You might want to also note that in the picture above I had begun to fill/cover the holes in the internal structure. I finished all the filling, trimming and covering of the interior structure before I laid up the exterior mold. Why? Because once the outside mold was laid up I didn’t want it to move, shift, or release from the stock part in any way. It needed to stay locked tight because I was going to lay-up the interior mold directly to it.  

The picture above shows the beginning of the solution to one of the more difficult challenges of this mold. I wanted the interior structure to be boxed with a smooth interior panel. This required a 90 deg. angle to be tied back into the upper edge of the door. To have something for the 90 deg. to attach to I had to rivet small pieces of angle to the interior of the door. These can be seen clearly in the picture.  

Also notice how I have flanged around the side view mirror support. I debated about cutting this piece off but decided to leave it as it weighs almost nothing and will give me a surface to mount my racing mirrors and my lap timer pickup. Note the various pieces of flange materials that overlap. These are all hot glued to each other and clay will be applied to the seam to prevent gel coat from entering the seam.  

Once the flanging was finished the interior aluminum panel was installed and clay was applied to all of the gaps and holes. This includes the face of the part. A thin, smooth layer of clay must be run around the edge of the stock panel to smooth the transition to the flange and prevent gel coat from seeping into this seam. Think caulking around a window or bathtub and you will have a good idea of the process and the desired result. I also filled in the door handle area and contoured it to the face of the door. This took a lot of clay and was a very time consuming step of the process that consumed several hours of time. It was now time to flip the part over and start the fiberglassing stage.  

Part III - Mold Lay-up  

The first step here was to apply mold release to the perfectly clean and smooth surface. This was the moment of truth so I paid special attention to keep the part clean and free of any lint, grit or grease. Mold release is much more of an art than a science. Unfortunately there is very little published on mold release and what is published is full of holes. I think that the people that do this stuff for a living would just as soon keep it a mystery. I am sharing here what works for ME and do not in any way imply that this is the best, fastest, or only guaranteed way to get the part released.  

Applying Mold Release:  

This is really the most important part of the process. The great news is that everything you do here will be repeated when you lay up your carbon part. I have broken the process down into the three steps that I follow. If you ask 10 people you will get 10 answers on how to do this so keep in mind that this is what I do and you may experiment and find better ways and products to accomplish the same task.  

Step 1: Part-All #2 wax. I apply one coat and remove. Key here is to remove QUICKLY. Once this stuff dries it is really hard to remove. I then give this about 20-30 minutes to fully dry. The wax does a good job in filling a lot of the small textures and scratches. This is the only reason I use it. I should also caution that this wax breaks down at over 120 deg. When it does it is a problem. This was a lesson learned the hard way for me.  

Step 2: PTFE Release Agent. This stuff actually dries kind of sticky. The key to applying it is to keep the coats VERY thin and as even as possible. As shown in the picture above, my preferred method of application is to use a huge cotton ball wetted with a little PTFE. I then sweep it across the surface leaving a thin film. I have found that if I put my head down almost to the surface I can easily see what has been coated. Overlap your stokes as little as possible as wiping over dried or semi-dried PTFE will cause it to smear. This isn’t as much of an issue when coating the black gel coat because it is so easy to see where you have already been but it is a challenge when coating my stock white parts. While this stuff dries to a tack very quickly, I do let it dry for a full 30 minutes prior to spraying PVA.    

Step 3: PVA. I spray this as following the directions that come with the PVA. The first coat is pretty light and will dry in under 10 minutes on a 70 deg. day. The next coat is much thicker. Make 100% sure that the entire piece is 100% covered. You do NOT want to get runs in the PVA and DON’T TOUCH IT after you have sprayed it. PVA creates a thin vinyl coating on the part that is super easy to peal off. If something gets on the PVA that you can’t blow off - leave it and address the imperfection on the mold. Don’t attempt to remove unless you are ready to wash the entire part down and start over.  

Tip: I originally tried to use my HVLP spray gun with 1.3 mm tip. This didn’t work well as the finely atomized PVA dried in the air. A $40 Harbor Freight gun with 2.0 mm tip did the trick and is excellent at spraying PVA.  

Now I was ready to spray the gel coat. For this stage I like to have all of my gel coat stuff ready to go prior to application. By stuff I mean cup gun with plenty of cups, stirs, and gel coat and MEKT (catalyst) ready. I like to get all of the cup gun cups full of gel coat prior the spraying. This makes the application go faster and smoother as you don’t have to keep pouring cups full of gel coat while spraying.  I also like having plenty of nitrile disposable gloves ready.  Above is a picture of my lay-up table with supplies and cup gun. As you can see, this is a messy process.  No matter how much effort you put into being neat and organized, you will make a mess.  

Spraying on the gel coat is pretty darned easy. There are only a few things to watch out for. One, you’ve got a limited amount of time after you mix the catalyst with the gel coat. You don’t want it to setup in the cup gun. I like to mix it and spray it. I never stop spraying until the cup is empty. That pretty much guarantees that it won’t setup. The second thing to watch out for is that you don’t want to release the cup gun trigger without the gun pointing strait up or gel coat will simply run out of the gun on the piece or ground making a pretty big mess. Third, keep the gun at least 12” from the piece and keep it moving. If you concentrate the spray too much the air pressure the blow the gel coat away from the area you are trying to coat. This is particularly the case when you are spraying an edge. If the piece you are spraying has a lot of edges then you will want to let the gel coat setup a little between coats the will prevent the first coat from being blown around by the second. Fourth, spray your part where you will lay on the fiberglass. DON’T attempt to move the part with only gel coat on it. The gel coat doesn’t have enough rigidity to take the bending if you flex a flange. I learned this the hard way. And last, be careful what you get gel coat on. This stuff is nasty and what it gets on, it stays on.  

The only other consideration is, “How much gel coat is enough.” My personal take on it is that you can’t have too much but you can definitely have to little. If you have too thin a gel coat layer and have an air pocket behind it (this will happen) then the surface gel coat can crack exposing the hole. No huge deal as you can fill the mold hole later with body filler but you can also prevent it by using plenty of gel coat. Back to how much is enough. For the exterior door mold I used three cup gun cups full (almost full, say 24-28 oz. each). For the roof I used 5. The hood will probably take 6.

Above is the door exterior with three cups full of gel coat on it. After application I immediately start getting my polyester resin and fiberglass chop strand mat (CSM) ready to apply. I also get my composite rollers ready. Once the gel coat dries to the point that I can just leave a fingerprint on it I am ready to lay-up.  

Let me say that the door skin mold was borderline for size of what one person should lay up at 80 deg. Any hotter or bigger and it would have require two people. The problem here is that the resin started curing as I was laying out the fiberglass. This made it difficult to get all of the air bubbles out. With two people I would have been able to move twice as fast and resin cure would not have been an issue.  

Above is a shot of me laying up this part. Notice how the fiberglass that has resin on it appears black. I covered the door with glass to illustrate the difference between the wetted and un-wetted areas. The reason for this is that wetted fiberglass becomes almost completely clear showing through the black gel coat. This makes it very easy to see what has been wetted out and what hasn’t. You can also easily see the bubbles in the resin/glass. These you need to work out as much as possible. Note: Normally I do not cover the entire part with glass. Rather, I work from one end to the other.  

This is a good time to mention that the same time constraints apply to this polyester resin as the gel coat. You can’t keep it in the cup very long as it will setup very fast. Get the resin out of the cup. There is no reason to even try to be neat with this step. Slop it on and roll it out. BTW, don’t attempt to do this without the composite rollers (which I am using in the picture) that all of the composite suppliers sell. These things are indispensable.  

Now for a complicated question, how many layers of glass do I use? I have read that you want to make sure that the mold is three times as thick as the part you want to make. I can’t really buy into this as a cored part changes this equation. I can’t tell you for sure how thick to make your mold but I can say that there are three factors that play into this. Size, shape, and use of the mold. For me, if the mold is big, flat, (think hood or roof) and I want to make a bunch of parts then I make it 4-5 layer. If it is small and contoured then 2-3. For the outer door skin I used three layers and the same for the interior structure. These molds turned out plenty strong and appear to be suitable or production duty.  

8 hours later (depending on temperature) and I was ready to trim the excess glass and resin off the edge of the mold. To do this I first flipped the door over exposing the flanges. I am extra careful with the part as I don’t want it to separate from the mold. I then trimmed the excess fiberglass and resin that extend beyond the flange line with my 4.5” grinder and cutting disk. Cutting into the flange a little isn’t a problem as hopefully I have made it plenty big. The trim line doesn’t have to be perfectly strait or even follow the flange line perfectly. All I am after with trimming is a good clean edge.  

Once the flange was trimmed I flipped the part back over, removed the aluminum flange and all of the riveted supports. If you have difficulty with this you can use a heat gun to re-melt the hot melt glue and it will easily release. After the flanges are off I removed all of the excess glue from the part and went back over the clayed areas to make sure that there were no holes or concave areas. I also smoothed out the clay in the areas that were under the aluminum flanges as removing the flanges will pull up some of the clay. At this point the exterior skin mold was finished and it was time to proceed to the interior molds.  

  For the door I now had to put a separator flange on the part as the internal cavity was to be a two part mold. For this flange I used a 2” angled aluminum piece that ran from the top to the bottom of the inside of the door. Above is a picture of this flange installed (with rivets to the internal panel btw). Note that pieces of aluminum (.040 sheet) were used to seal off the contoured edges of the piece.  

The above picture is important as not only can you see the separator flange but also the exposed flange left from the exterior surface mold. You can also see how I have smoothed and filled all of the seams and holes with clay. This half of the internal structure is now ready to treat with mold release(s) and lay up. You will notice that the other half of the interior is covered as I don’t want any polyester gel coat or tooling resin over there until I get this side finished.  

Above is this section under gel coat prior to lay-up. After the gel coat got tacky (remember fingerprint) I laid up three layers of CSM and polyester tooling resin just as I did on the exterior mold. Now it was time to move to the other half of the interior structure.

Above you will see the first finished half of the internal mold with the separator flange removed. I first trimmed (with scissors) the loose fiberglass off the top of the vertical flange to make a nice neat smooth surface to lay-up the other half to. I then went along the seems at the bottom edge of the separator surface to repair the clay to prevent gel coat seepage under the flange. The same process as above is followed to lay-up the other half of the inside of the door. Continue to take every precaution to not disturb the other molds that are still stuck to the original part. It is imperative that they not move. BTW, you will notice that I used masking tape along some of the edges of this part to seal them. This I regret doing and don’t recommend. Use clay and clay only.  

After the last section had setup I drilled 1/4” diameter locator holes (8) through the vertical separator flange. This absolutely must be done prior to any release of any of the three molds. The bolts are used to join these halves back together for part lay-up.  

Now that the molds were cured and locator holes drilled I was able to pop them off the part. This was, as always, THE moment of truth. I always like to start by using a hammer to tap the outside of the mold. I then go around the flanges with the wedges to release them. Once they are released I keep moving them further out into the piece. At some point the mold will literally ‘pop’ off the part. Knock on wood; I have never had a part stick when using PVA. Before I used PVA I did and it wasn’t pretty. From my experience and from everything I have read using PVA is simply the most foolproof product on the market for release. Most people I have spoken with will tell you that a seasoned mold (one that has been used a number of times) will release easier than a new one and may only require wax or PTFE to release. As I have only made a limited number of parts from my molds I have not attempted to eliminate PVA from my process. Given the amount of time and effort involved in making the molds I don’t like the thought of destroying one.  

After the molds were released from the part it was time post finish. I have yet to have a mold come off a part ready to lay-up. I always have a blemish, pit, dimple, high spot, surface irregularity, etc. to fix and this part was no exception. The majority of the work on the door surrounded the surface irregularity at the door handle and lock areas. While these spots were both filled with clay and smoothed prior to molding, my experience is that a smoothed clay surface will always have to be sanded to match the surrounding area.

For sanding I like to wet sand with blocks and sponge pads as shown above. The surface finish left by 400 grit is plenty smooth. There will be no discernable difference between 400 and 2000 grit on the finished part if PVA is used so I find no reason to go to the trouble. This is a good time to address the downside of using PVA. PVA does not go on perfectly smooth or perfectly uniform regardless of how much effort is put into the its application. Therefore these irregularities will transfer to your parts and molds. 

You may remember the outward dimple on the stock part that I mentioned earlier. In the above picture you can easily see it as an inward dimple on the mold. At this point I simply filled this depression with body filler and sand smooth. The difference in surface texture of this small area will be every so slightly noticeable on the finished part but light sanding of the will eliminate most or all of this variance. Note: My experience is that it is easier to address surface depressions than high spots. Therefore I chose to always address depressions. If there is a depression on a part I will fill it prior to making the mold. Also, I should mention that hardened polyester resin is very tough and takes a long time to sand. If you have a high spot of gel coat that you need to sand down it will take 10 times more effort and time than sanding of body filler.

Above you can see the sanded mold with filled dimple. You can also see that I filled a depression on the mirror extension that will not be needed or wanted on the finished part. Notice that the door handle and lock area are no longer noticeable as they have been completely sanded smooth with the surrounding profile. This part was now ready for mold release and carbon/honeycomb/carbon vacuum bag lay-up.  

Part IV: Preparation of Carbon Fiber Lay-up  

Alright, I have completed the long, messy task of making the mold(s) and I was now ready to churn out a 3 lbs. race part. This is the fun part. It is also 1/100th the mess that mold making is. The first step of this process was to organize and pre-cut all of my materials. As I mentioned in the opening of this write-up, I will not go into excess detail on how many layers of material are or how much if any core is required to make a part. I will only address the part I am making.  

For this part I chose to make the outer skin from three layers of 5.7 oz. 60” wide 2x2 twill carbon fiber and a single 1/8” 3.0 sq. foot Nomex honeycomb core. The topology for the outer door skin was 2x1 or two layers on the outside (mold side) with a core and one layer on the inside. Obviously a single layer carbon was required on both sides of the core. The lightest possible cored parts are 1x1 but I wanted more impact protection for the doors. I know, I know, I can here it now, “What is a single layer of carbon going to protect.” The answer is itself. A single layer is pretty darned easy to damage with a slight impact. I can just see the door swinging open into a jack handle and cracking. 2 layers on the outer skin of the doors will definitely make them more resilient. With this said, I have made a bunch of 1x1 parts (my roof, C pillar covers, etc) and they are very structurally strong, just a little fragile.  

Ok, on to pre-cutting. I have found that it is much easier to cut the fabrics, core and vacuum bagging materials on a flat table. This keeps the fabric nice and clean and keeps the edges of the mold from making runs in your carbon fiber. I will say that one of the great frustrations in dealing with carbon is keeping it from snagging on stuff. Think of it as a panty hose from hell. You must be VERY careful at all times when measuring, cutting, moving, and wetting out carbon or you will get a run, pull, or tear. A simple hangnail will pull a carbon piece apart. For this reason and to keep any oils off the fiber, I always wear nitrile/latex gloves when handling the fabric. 

Since I was cutting the fabrics on a table I made a template of the part. For this I used the white breather ply material as shown in the picture above. I just made sure that the template followed the curves of the parts and that it was NOT UNDERSIZED as this was the size that $150 worth of carbon and a $75 core was going to cut to.

After the template was made I then laid it out on the table on top of the carefully unrolled carbon fiber. I then took the 3/4” masking tape and traced the outline of the part. This had to be done very carefully as well as it would pull the carbon weave apart if I attempted to remove or relocate. Once the template outline had been transferred to the carbon fiber I then cut out the template by cutting the tape in half. This left both sides of the cut with have of the width of the tape. Above you can see a picture of the “Peel Ply” being trimmed to fit the taped and trimmed section of carbon fiber. BTW, only the carbon fiber needs to be traced with tape.  

Right now you are asking, “Why use tape?” The reason I used tape was that the carbon loves to fray. Much like a woven basket, if it starts to unravel it will become a HUGE MESS. The tape ‘helps’ keep this from happening. As you can see in the picture below, it doesn’t totally prevent it. Now that the first layer of carbon fiber was cut I repeated the process for the other two layers and the peel ply (slightly oversized).

Cutting of the core was only slightly more difficult. My preferred method of doing this is to lay the core on the mold (this is why we haven’t applied mold release yet, btw) and mark it. Again I must make sure that the core follows the profile of the part before marking. I often put something on the core to insure that this is the case. For marking I made dots with a black Sharpie marker to show me where to cut. I should mention that the core must stop prior to the edge of the part. I like it to stop no closer than 3/4" from the edge. This give the inner carbon layer a good chance to bond to the outer layers. It also makes it easier when bagging - more on this later. For this outer door skin I have stopped the core well away from the edges as that is where I am going to bond the inner structure.  

After the core was marked I then moved it to the layout table. To cut the core I used a pizza cutter like circular fabric cutter and simply followed the dots. This worked extremely well as the core cut very easily. If I am using a core thicker than 3/16” I bevel the edge by angling the cutter 45 deg. Cutting the core was not difficult to do as the markings were correct.  

Now I apply mold release:

Now that the carbon, core, peel ply, and breather ply were cut, it was time to apply mold release to the mold. For this I followed the exact same process as before. The only addition to this process was that I ran 2” wide blue painters tape completely around the outer edge of the flange as show in the above picture where I am spraying PVA to the mold. I did this so that this area received no mold release. Why? Because this was where the bagging seal tape is going to go. I have made the mistake of trying to get this tape to stick to the release agents and it is a challenge. I am much better off with it sticking to the bare mold.  

Finished Part Lay-up: 

As with the mold lay-up, I like to get all of my supplies out and ready. Above is a picture of them. In addition to the two parts of the epoxy, I have my .1 oz resolution scale, nitrile gloves, cups to mix the epoxy, tongue depressors to stir the epoxy, composite rollers, and spreaders. I also have the bagging sealing tape rolls ready.

Now it was time to start wetting out the part. Once I began this phase I was back into a time sensitive area. The epoxy resin I used starts to cure in the cup in about 20 minutes and on the mold in 3-4 hours. Obviously this is quite a bit longer than the polyester resin used in the mold lay-up. I still proceeded with diligence as the vacuum bagging process requires the resin to flow from the part to the breather ply.  

Above is a picture of the first layer of the carbon on the mold. A couple of notes, here you can see the tape along the trimmed edges of the carbon and the outer edges of the mold. They are still there for a reason. DON’T CUT OR REMOVE THEM YET.  

Once the carbon was in place it was then time to mix and add the epoxy. The epoxy I used requires a 3:1 mix ratio. For accuracy I always recommend using the scale. Most composite shops sell ‘ratio’ pumps. I have several and don’t like them. They are messy and inaccurate. I definitely would not use them without a scale to confirm the ratio. For this part I began by mixing 12 oz. of epoxy. BTW, I usually don’t mix up more than that even if the part needs it as it prevents it from sitting in the cup.  

In the above picture I was adding the epoxy to the mold. I began by using the soft spreader (make sure it is clean and free of anything that could snag the fabric) and working the epoxy around. I have no problem seeing what was wet and what wasn’t. The beauty of the vacuum bagging process is that it will remove almost all of the excess resin from the part. Knowing this I usually will error on the side of too much epoxy as too little will ruin the part.  

I continued spreading the epoxy mixing and adding more as necessary. Once I got the part wetted out I lifted the carbon exposing the interior of the mold. I did this one corner at a time to prevent movement of the fabric. If I ever notice an area that is dry I add a little epoxy. This is usually not needed but this was the time to check.  After this I went over the fabric with the composite roller. This roller does a great job of two things. One, it gets rid of any air bubbles/pockets under the fabric, and two, it fixes cosmetic blemishes to the orientation of the fibers.  

Once this layer is fully wetted with epoxy I preceded with the second layer. As this second layer was not cosmetic I used the least perfect of the three layers I cut. The wetting of this layer followed the exact procedure as above except I did not lift the carbon to check underneath. This layer takes less epoxy to wet out as it will pick up from the layer below. After this layer was completely wet I went over the entire surface again with the composite roller. 

Now it was time to wet out the inner carbon layer. I am sure you are asking, “You skipped the core?” We will get to that in a minute. The inner carbon layer was wetted out first on the layout table as shown in the picture above. The procedure for this was simple; I laid out the fabric and wetted it completely with epoxy. I definitely wanted this layer of fabric to be ‘oversaturated’ with resin.  

Now it was time to lay the core on the outer two layers of carbon. I had to be very careful here as I didn’t want the core to snag the carbon. When the core needed to be moved, I picked it up and laid it back down. I did not attempt to shift it. Also, the core did not stick to the carbon underneath nor did it perfectly follow the contours of the part (and it won’t unless the mold is perfectly flat). I just located it as best I could and followed the rule of getting the core to no closer than 3/4” from the finished edge.  

Now I transferred the pre-wetted carbon layer from the table to the mold. Special care was taken here and for this operation I required an extra set of hands. The fabric needed to be laid out as evenly as possible and centered all the while not disturbing the location of the core. This might not sound like a big deal but it was the most challenging part of this lay-up.  

Given the fact that the core did not follow the form of the part, the third carbon layer also did not follow the contour of the part. The key here was to make sure that the carbon was not stretched or folded. As you can see in the above picture, there was plenty of excess material around the edges to be taken up by the vacuum.  

Once this layer was applied it was time to trim the excess carbon. This is pretty simple as the edge of the blue tape is the line I use to cut to. I just make sure I have good sharp scissors and take my time as I did not want to disturb the fabric and core. Note, trimming the excess carbon also trims of any remnant masking tape that was left on the edges of the carbon. Once the excess carbon was trimmed I then removed the blue tape from the flange.  

You will notice that you now have a perfectly clean (no epoxy, no mold release) flange. The first several parts I made I did not use tape and would have to clean the flange edges. This added a lot of time and effort to the process. Definitely use tape. It not only leaves a clean flange but it also ‘smoothes’ the rough trimmed fiberglass mold edges which reduces the chance of snags on the carbon and bagging film.

I now made a final pass around the part making sure that all of the carbon is nice and smooth, particularly around the flanges. Once that was completed it was time to apply the peel ply as shown above.  

There are a couple different options to transfer resin off the part. The first is perforated release film. This is simply a sheet of polyethylene with tiny little holes spaced at about 1/4” from each other across the entire surface. This I have used for all of my parts up until this one. It works pretty well and is very easy to remove from the cured part. For this part I used nylon peel ply. This is a simply tight nylon fabric that has been treated to prevent it from sticking to the part. Unlike the perforated release film, this fabric does not have holes rather the resin will flow through the entire surface. Also, the peel ply is advantageous in that it leaves a fabric imprint on the carbon which is great for adhesion. This is important for this part as I am going to epoxy the inner door structure to the inside of this panel.  

Neither the perforated release film or nylon peel ply are that flexible. Therefore it is important that these layers drape across the part and do not ‘span’ the contours. If they span a section of the lay-up they may pull the carbon and core when the vacuum is applied.  

I now applied the breather ply. The breather ply is used to absorb the excess resin that is pulled through the peel ply/perforated release film under vacuum. The only trick here is to make sure that the breather does not touch the carbon fiber. If it does it will seriously stick to it. Therefore I want the peel ply to be slightly larger 1/4” than the carbon fabric to prevent this.   Also, I want to determine the location of the vacuum bagging attachment fitting at this time. The fitting is approximately 2” in diameter and must be placed directly on top of the breather ply. Also, I greatly prefer to have this fitting off of the finished section of the part. The flange is excellent for this but you can also use a fold in the vacuum bag to serve the same purpose. I will usually use scrap strips to make vacuum channels for larger parts. This prevents the possibility of resin blocking areas of vacuum. You can see how this was laid out in the picture above. I also ran this extra strip down the part and to the vacuum fitting location.  

Once the breather ply is on it is time to put down the bagging adhesive. BTW, this is some unbelievably sticky stuff. To install it I start in a corner and proceed around the perimeter of the mold. I don’t peel off the outer covering of the tape until the bag is installed. When I come to a corner I simply twist the tape around it as shown in the picture above. The outer cover of the tape will break but the tape will be a continuous run which will GREATLY reduce the possibility of leaks. After I have laid down the tape I go back over it with a lot of pressure to insure a seal.  

I now lay out the bagging film. As with the release film, there are a number of options for bagging film. I LOVE the green high elongation bagging film and highly recommend it. It will stretch a ton and is very easy to work with.  

I start by rough cutting a piece and laying it on the part. BTW, this film does not have to be trimmed to a perfect size. It just needs to be bigger than the part. Once the film is cut and placed it is now time to stick it to the bagging adhesive tape. I like to start in the location that I am to place the vacuum fitting. In this case it is right behind the mirror extension on the flange.  

Now for a tricky part. The key to getting a great vacuum seal is getting the bag smoothly adhered to the tape. This took me a few tries to get down and you may want to practice on a workbench or table. This is also a process that is dramatically easier with two people as I really need three hands. One hand will be necessary to pull the cover off of the tape, another is needed to lightly press the bag on to the tape, and a third is needed to pull the bag tight to prevent wrinkles. I can’t stress enough how important it is that this is perfectly smooth with no even tiny wrinkles. In the picture above of the process you will notice that there is about to be a small wrinkle in front of my finger. Here I pulled the bag film tight to eliminate it as it would be a source of a leak otherwise.  

I am much better off with one big fold vs. a bunch of small ones. As you can see in the picture above I ended up with a big one. To remedy this I stretched and folded open the seam and pressed in a section of tape.  

I now have 95% of the bag taped down except for the area right next to the vacuum fitting. The vacuum fitting is two parts, one that goes under the bag and another that attaches to it on the outside. Therefore I made a small incision in the bag, inserted the bottom section of the fitting into the bag, and connected the two parts. It is better to make a small cut (1/4”) and stretch the opening than one that is too big and will leak. I also make doubly sure that the hole is in the exact place that I want the fitting.  

Once the vacuum fitting was installed and put together I finished taping the bag down. Now I can turn on the pump. For my pump it takes a minute before the excess air is out of the part. At first I always wonder if it is working. Then, all of the sudden, the bag will contract to the part. If it doesn’t, I check the pump. If I am sure that it is working then I have some major leaks. To find them I call on a good set of ears and some chunks of the bagging tape. I simply go around the parameter listening for leaks. When I find them, I stuff in some tape and squeeze it around. Then I go looking for more. This is where having a quiet pump is very nice to have as it is easy to hear air moving into the bag.  

My DIY pump will pull 28.5” of vacuum which is awesome and higher than many $250 pumps will pull and it is 1/1000th as loud. This is made even better by the fact that I have $20 in it. If I am not pulling 28.5” on a part then I need to go looking for leaks. I personally regard 20” as a minimum for acceptable vacuum bagging carbon parts. Less than that and you will need a new pump or you will need to do a better job sealing your bag. BTW, A part the size of a hood is what I consider the maximum for my pump.  

12 hours and a six pack later…  

It was now time to release the part. First I turned off (unplug) the pump and removed the vacuum fitting from the bag. I then pulled the tape (not real easy to do btw) off the part with the bag attached. Then I peeled off the peel ply as shown above - also not easy to do. BTW, with the peel ply I pull back parallel to the part instead of up.  

Once this was off it was time to pop the part from the mold. As with popping the mold from the original part, I began by working the wedges all the way around the flange. I always have to be careful not to get in a hurry even though I am dying to see the finished part. In a way it is kind of like opening a Christmas present. With this part I then ran the wedges into the short end of the mold and it popped almost instantly. With other parts this can be a long process that requires sticking wooden yard sticks way up into the mold. Hopefully yours will pop out as easy as mine have. All hail PVA.  

Above is the finished outer skin after I washed off the thin layer of PVA. BTW the PVA will stick to the part and not the mold. This makes moving on to the next copy of the part easy but requires washing of the finished part. Fortunately PVA removes easily with water. The hotter the water the better but I will usually just hose my parts off and then wipe down with a soft sponge.

Here is the inner panel. The nylon peel ply left a very nice textured finish to the interior of the part. It is now ready to bond to the inner door structure.  

At this point I exposed the part to an extended post-cure. I did this by laying the part out in the sun for a day. 12 hours at 160 deg. flat did it. Post curing makes the finished part 10-15% stronger than not post-curing. 

The stock gutted door skin (no glass, no mirror, no interior panel, no lock, no handle, no speaker, no switches, no side impact door beam, no interior skin, etc.) weighed 30.5 lbs. The stock door with all of the stock stuff in them probably weighed 100 lbs. each. This carbon, honeycomb, carbon reproduction is 3 lbs. 5.6 oz. Not bad as it should be right at 3.0 lbs. when trimmed.  

To trim this part I simply used a Dremel tool with a fiberglass reinforced cutoff disk. Carbon fiber this thin cuts easily, almost too easily. The flange that I put on the mold leaves a nice line where to cut. This part of the process was not hard but I had to be careful, steady, take my time and definitely wear eye protection and a respirator as it made a TON of dust.    

Appendix:  

DIY Kentucky Redneck Vacuum Pump!  

Many people have asked about my vacuum pump and here it is in all of its DIY glory. Keep the redneck jokes to yourself. BTW, full credit for this idea goes to my good friend Gene Young so point the redneck jokes at him.  
This pump started life as a drinking fountain refrigerator compressor. I have converted it from compressing Freon to vacuuming air. This setup is so simple it is comical to think of spending $350 for a rattle trap to do the same thing.  
Here is what you do:
1. Go to a local refrigerator repair guy and buy a used compressor. I got mine for free.
2. Bolt it to a piece of wood.
3. Wire it up to plug strait into the wall - screw an on/off switch
4. On the ‘suck’ line, attach with a worm clamp a 1 foot section of 1/4” section of vacuum line
5. At the end of that worm clamp on a $3.00 fuel filter (needs to have 1/4” barbs) to prevent epoxy from getting into the pump.
6. Then run another foot of vacuum line to a shutoff valve (optional)
7. ‘T’ into the line a vacuum gauge as show. I got mine from McMaster-Carr for $8.
8. To it attach a 15 foot run of vacuum line. This will give you a ton of placement flexibility.
9. To the pump outlet, attach a cheap filter as shown to prevent oil from spraying out of the pump.
 
Every once and a while you will want to lube the pump with refrigerant lube that you can get at AutoZone. Also, I run a fan over my pump to keep it from overheating. This probably isn’t a potential problem but they do it on refrigerators so I do too. Eventually I am going to add a fan to the mounting board. This is why it is oversized.  
Parts list:
1 Small refrigerator pump
1 12 ft. electrical cord
1 1’x2’ piece of wood
4 small lag bolts to bolt the pump to the wood
1 cheap fuel filter with 1/4” barbs
1 cheap filter to keep the pump from making a mess
1 1/4” NPT female to female shut off valve
1 1/4" NPT male, female, female ‘T’ fitting
1 vacuum only 2” gauge w/ 1/4" NPT male fitting
2 1/4" NPT male to male barb fittings
1 1/4" NPT female to male barb fittings
1 Six pack of your favorite beer to enjoy while watching your $20 pump for 12 strait hours while waiting for your part to cure.