The Saab Link Forums

The Saab Link Forums (
-   Performance Modifications for the NG900 / Old 9-3 (
-   -   Composites: so easy, anyone can do it wrong. (

jerrit 08-13-2016 12:58 PM

Composites: so easy, anyone can do it wrong.
Near the beginning of my blue car's life, back when it was still going to be a lemons racer for some friends, there was a hood pin "incident" at a test day. The hood flipped up, took a huge dent, and shattered the windshield. Obviously the windshield had to be replaced right away, but the hood was left with the gnarly dent because screw it, lemons car.

Since inheriting the car I'd always meant to replace the dented hood. And fix the sketchy, extremely front-heavy weight distribution. Hey, maybe there's a way to get these two birds stoned at once.

Like a lot of you, I'm crazy about carbon fiber. Nobody makes carbon fiber parts for saabs (except Kevin Yankton, many years ago, and those parts were unobtainium). You can't just go on amazon and prime up a ng900 hood. It's either get someone to build you a one-off at extreme cost, or do it yourself.

I had previously experimented with fiberglass molds and a carbon fiber sunroof delete panel for my c900 7-8 years ago. I didn't have a shop back then, so it all had to be done in the hatch of the car itself. It turned out terrible. It only managed to keep rain out because it was half glue and caulk, and I eventually replaced it with stainless steel in the same build thread where the cage went in. As a learning experience though, I thought I had paid a lot of my dues. This time, things would be different.

What actually happened was a four month odyssey of finding every way to screw up composite fabrication. All I can say in my defense is that this time, NEXT time will be different.

Starting point: junkyard 9-3 hood. Pick n pull had a sale that weekend where everything was half price, so I think I paid $30.

Making a mold sounds straightforward, you just put release agent on your part and slather a bunch of gelcoat, fiberglass, and resin over it.

And end up with junk. Lesson one is that the big sheets of thick chopped strand mat don't like to lie flat. Air gets trapped and the gelcoat and glass don't bond right, so they delaminate and flake apart.

No big deal, all this stuff is sandable so I can just throw more gelcoat on the thin areas and sand it smooth right?

Lesson two is that sanding is an infinity of work. If the part is a little off, yes, it can be patched. If it's majorly screwed up from the start, fixing it is impossible because there is not enough reference for where the surface should be sanded down to. This was a waste of chemicals and time and produced a hurricane of nasty white plastic dust.

Rather than throw it away immediately, I used failed top skin mold #1 to practice trimming to fit, and to test fit the aerocatches that would replace the exposed hood pins. Other than the fact that this is not structurally rigid, it's half believable.

jerrit 08-13-2016 01:11 PM

Moving on to the bottom ribbing mold rather than dwell on the first failure, I was more careful to cut small pieces of cloth and build them up, in an attempt to relieve the wrinkles, and only laid a whole sheet on the very top to back everything up.

This came out pretty good, at first. Yeah there are some cracks but I can make it work.

Here though, I had run out of polyester resin and rather than delay to get more, I used two part epoxy resin I still had a lot of. And I mixed it poorly, and it didn't harden right, and this mold sagged to uselessness. This was really disappointing.

jerrit 08-13-2016 01:15 PM

So back to the top skin mold, attempt 2. I did more reading and learned that for taking panel molds like this you want a flange around the outside. I put this together with foam board, packing tape, hot glue, and plastilina wax-based clay (because I couldn't find any soft wax to seal the gaps).

More carefully applied a new "tooling gelcoat" designed for this and the top layers of cloth.

Flipped over and started on the ribbing, re-using the flanges, all part of this technique as recommended.

Separated the parts. Things look good! Mostly.

jerrit 08-13-2016 01:17 PM

The top skin still had some delamination. In spite of what I learned about sanding, I hold out hope of fixing this.

The ribbing came out great. Only a few tiny spots were wrong, which is almost inevitable.

jerrit 08-13-2016 01:20 PM

I made these captive nut plates out of perforated stainless sheet and m6 rivet nuts.

The m6 studs were used to line things up with the old captive hardware on the metal hood, guaranteeing the new ones were exact copies.

The sanding of the top skin went about as well as you might expect by this point.

jerrit 08-13-2016 01:37 PM

On to the first attempt so far at making a positive, using the ribbing.

Vacuum bagging... more on this.

Ribbing positive attempt 1 came out wrong in several ways.
- Not enough carbon layers, didn't really know how much I should have used.
- Again with the improperly mixed epoxy. This stuff is not forgiving. Use a scale, mix two minutes like your life depends on it.
- Vacuum bagging failed in multiple ways.

The vacuum bagging process could take a thread all its own. You put down the carbon cloth, pour epoxy over it and try to spread it around with a roller or brushes as evenly as possible, put peel ply release film over that, put batting over that to help with air flow, put this bagging film over all of that, use gummy tape to stick it down to the flanges of your mold in an airtight seal, attach a vacuum pump or venturi like mine to a port, and the 20-25 in-Hg of vacuum makes the air pressure evenly squeeze your part. The epoxy should spread out in an even layer and the excess comes through micropores in the peel ply and gets absorbed by the batting.

What actually happens is that you chase phantom air leaks infinitely until you're hallucinating with rage and frustration. I tested my equipment on some little swatches that turned out right.

Attempt 2 of the ribbing positive also failed to draw enough vacuum, but mercifully came out good otherwise. Properly mixed, enough layers to be rigid, etc. Super light. Hilariously light.

There was unfortunately minor release damage to the mold. Localized. Probably fixable.

Not perfect, but this is a success!

jerrit 08-13-2016 01:47 PM

Debugging why the vacuum bagging hadn't worked, I finally concluded that my mold itself was not air tight and no amount of tape at the edges was going to seal it. A bad vacuum bagging results in pools of hard plastic at the low points of your part as it sits in the mold, and lots of flashing. Cosmetically, this is devastating (if you care, I don't really in this case). Functionally, it's weight that doesn't make the part stronger and can mess up fitment. So for next time, I decided to change technique to enveloping the entire part, mold and all, in bagging film like a big empanada. The first opportunity to apply this was to top skin positive attempt 1.

Carefully weigh the cloth.

Some maths to figure out the right amount of epoxy ingredients.


Success! Vacuum bagging achieved!

Except this time, unrelated things went wrong. That carefully measured amount of epoxy? Apply a fudge factor next time. So I had some dry spots instead of pools this time. I had run out of peel ply and switched brands to the place I was buying the tooling gelcoat, and the pore size on these was much bigger so more epoxy immediately ran into the batting instead of spreading out. Maybe they expected the user to be much more careful about spreading the epoxy super evenly first, but on a mold with high and low points this seems impossible. The resin is going to want to run.

Attempting to wet out the dry spots is probably not what an F1 team would do, but no way I was going to throw away something this close to right. Second trip through the vacuum bag and I couldn't get it to seal. Seriously? I must have torn it somewhere when I opened it the first time. I wonder if I can get more durable bagging film designed to be reusable or if this is just a fact of life.

Anyways, the top skin came out frumpy but close enough for the 50-50 rule: 50 yards away at 50 mph, it'll look right.

jerrit 08-13-2016 02:00 PM

The ribbing is what attaches to the hinges on the car, so that part goes on first to make sure everything lines up.

3m automotive panel bonding adhesive to permanently attach the nut plates to the ribbing.

Testing out the aerocatches.

More panel bonding adhesive, which had to be weighted down with whatever I could find.

jerrit 08-13-2016 02:04 PM

Complete hood, painted first with black plastidip. Raw carbon is "thirsty" for paint and I'm hoping the plastidip primary layer will help it look less like a lawn chair.

Gloss navy blue.

Back on the car for stripes.


Rhombus prep.

Silver rhombus, white rhombus, plastidip numbers. I don't know where the nice templates we made for the doors are, so I fudged it using an exacto knife and paperboard. Not going for style points here.

So did it work? Did it immediately disintegrate on track?

jerrit 08-13-2016 02:07 PM

"Day 2: The Spec Miatas have accepted me as one of their own"

jerrit 08-13-2016 02:39 PM

Here's a few condensed tips for anyone else thinking of doing this.

1. Buy all the fabrics, chemicals, and other supplies you think you might need up front. Getting sticker shock and second guessing how much I'd need was half the reason this took as long as it did. Downtime waiting to buy more or worse, running out halfway through a part and ruining it by trying to improvise.

2. If you're not vacuum bagging, you have to use really small strips of thin cloth to get the bubble-free coverage you need. The ribbing mold that turned out right was layers of really thin, thin, then lastly thick cloths laid in small patches. The thin cloths have their own gotchas; once wetted with chemicals, they wad up easily. Trying to lay down a large sheet of thin cloth over anything but a perfectly flat surface is equally futile for slightly different reasons to the thick cloths.

3. Measure your epoxy with a scale. Mix it 2 minutes. Scrape the sides of the cup. You can't overmix this stuff and if anything about the measure or mix was slightly off, it ruins the whole part in a way that can't be fixed.

4. You can't fix it. You either do it right the first time, or throw it in the garbage and try again. It's not like working with metal. Yeah you can patch little stuff, but that means _really_ little stuff. You can't fix the big stuff. Do not be tempted to sand away your shame. It's a losing battle and will make an unholy mess you'll have to clean up later.

5. Make sure your vacuum bagging is going to work before you mix up the chemicals. Once you've done that, the clock is ticking and you do not have time to find where your air leak is.

There are some great resources on youtube where people walk through the step by step of making parts like mine. I think watching those professionals now and knowing all the screwups they're _not_ doing would be more valuable by far than just seeing them make it look easy.

jerrit 08-13-2016 02:51 PM

Bonus round: I ended up with extra cured carbon from the first failed ribbing positive, felt sick about throwing it away, so used it and some other parts to build an amplifier, which formed the base of a tv/bluetooth/multiple audio signal stereo system I hacked together for the shop.

"100W + 100W ...Ass..." I kid you not, this is what it says.

The little amp board wants 20-something to 32v input. Turn the pots on this variable voltage supply until it says 32.

Worked out the barrel plug for power and a 3.5 mm audio jack.

More of that perforated stainless from the captive nut plates repurposed into a vented lid.

Visible guts.

Tested out.

Tboy 08-14-2016 06:29 PM

This is a cool project, interesting read. It seems like you were pretty ambitious starting with a hood though! Maybe scale it down a bit for the first few parts!??! I realize though that for a race car you don't need to remake interior bits and such (that would be small).

Seems like end the end you had a decent result.

jerrit 08-15-2016 01:16 PM


Originally Posted by Tboy (Post 2150594)
This is a cool project, interesting read. It seems like you were pretty ambitious starting with a hood though! Maybe scale it down a bit for the first few parts!??! I realize though that for a race car you don't need to remake interior bits and such (that would be small).

You're absolutely right that it would have made more sense to start small. Jumping in at the deep end and struggling is part of the fun for me.


Originally Posted by Tboy (Post 2150594)
Seems like end the end you had a decent result.

I don't expect the orders to come pouring in. :)

Yeah, I'm happy with it. It's an ugly hood, but it's MY ugly hood. It opens and closes, doesn't fly off, and it's light, so that ticks all the important boxes.

After a break to forget how much work this was, I'll try making some other parts.

Trionic3000 08-15-2016 03:28 PM

If it makes you feel better, I'm the guy who bought Yankton's mold and it was warped a little--I never made anything with it and basically got ripped off. I left it in the upstairs of my garage when I moved.

On the exterior, are you using a UV resistant gel coat first? If you lay down a nice thick gel coat first with a spray gun, let it tack up, and then lay in your fabric on top of that you'll have a real nice shiny piece when you pop it out, provided that your mold is good and and it's all waxed up and you have good release agent and none of the other 100 things that can go wrong don't happen lol.

turbojohnny 08-16-2016 03:29 AM

That Hood really is ugly, but I applaud your efforts! It's always encouraging to see someone take on a creative project like this. and it's even better to see all the pictures of your progress.

keep up the good work, I'm sure you'll come up with a good looking finished project soon enough!

srp 08-16-2016 06:11 AM

Very cool.

Maybe I missed the part, but did you give any details as to the performance impact you noticed after swapping?

jerrit 08-16-2016 05:11 PM


Originally Posted by srp (Post 2150666)
Very cool.

Maybe I missed the part, but did you give any details as to the performance impact you noticed after swapping?

I didn't test the hood in isolation, sorry. I wasn't very scientific about this. I did rear axle bushings and a thicker rear sway and rear spring perches at the same time as the hood, and last time the track was cool and rainy versus this time was dry/sunny/96F. So even though subjectively the car feels much better to drive now and objectively lap times are ~5 seconds lower than last time I had the car out, which is huge, that's mostly because of weather and secondly because of suspension. I can only guess at how many tenths of a second the weight was worth.

I could cut the aerocatch holes in the metal hood I used to make the mold, take both with me next time, and do a back to back comparison. Or a lot easier and less destructive to that part, I could add the weight difference as ballast somewhere over the front axles to approximate the metal hood and try that.

sha-d 08-19-2016 08:00 PM

1 Attachment(s)
Good effort!

Making a mould is pretty simple, and anyone can make it, it's only time consuming.

Using the right cloth and laminating process makes all the difference.

I use Frewax as a releasing agent, apply a few layers of tooling gel coat, which gives the mould a solid surface, not allowing cracks etc. Let sit for 12 hours, tacky or not, as long as it isn't wet, start by applying 3 layers of 1/4oz chopmat. I do this as it's easy to work around the bends/creases/corners/edges.

Use a bubble roller after each layer is wet, roll the cloth, forcing all air pockets out (possibly why you had leaks too). This also helps by marrying the upper layer to the other end under.

Move onto laying a heavier gauge cloth, 3/4 oz chopmat. Usually you'd like a 3/8" mould wall thickness.

Moving onto making a part.

Cut the cloth to size.

I'd spray an in mould clear, I use duratec clear. Let it tack up, start laying your cloth, be nice to it, no kinks!

You've got the rest covered from peelply to bagging. Advise would be to put a full vacuum perimeter around the part with peelply strips acting as toes from the part to vacuum. Its also best to have some sort of flow media placed ontop of the peel ply. This helps move the resin equally onto the part. Also decreases thr chances of having dry fiber.

There are many ways you can place the feed line, feed run lines, and vacuum lines on a part. Each will give a different response.

Have good vacuum, 27 to 30 inHg, start the infusion. I'd mix one half kilo at a time. If you think you mixed enough, mix a little more until the resin/hardener is uniform with no resin toes (resin stringies).

I've been thinking of making a carbon hood, but the shop is very busy. Group buy is a possibility with enough interest. We can make one offs too. Top notch quality.

Check us out at

Take advantage of the Canadian Dollar

jerrit 09-13-2016 02:42 PM

I was actually in Toronto for a few days right after you sent this. Not close enough to stop in and chat sadly, but a lot closer than usual. I've been to see some professional composite operations before, like Mosler Automotive (sadly no longer in operation). Always fascinating.

Ballpark estimate for a custom carbon hood like mine? I wouldn't be surprised if it was less than I spent doing it myself badly, given how much I had to do twice.

All times are GMT -8. The time now is 01:42 PM.

Powered by vBulletin® Copyright ©2000 - 2020, vBulletin Solutions, Inc.
Search Engine Friendly URLs by vBSEO 3.5.2
vBulletin Security provided by vBSecurity v2.2.2 (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
All content is copyright The Saab Link and it's original authors.

Garage Plus vBulletin Plugins by Drive Thru Online, Inc.