something old, something new, something borrowed, something blue

This post is basically a lessons learned post. Thankfully the first thing I did meant all the other lessons learned have ultimately only caused harm to my wallet and my schedule.

After failing to get a seal on my mould, and in discussions with Warren, the best way forward with this tunnel is to take a fibre-glass part from it which won’t need a vacuum. It can just be laid in. This means I can crack out a quick part, ensure excellent surface finish, and then make a mould from that using a full mould making system.

photoThe first thing to do is apply release agent to the mould, and I went for spraying in PVA (to make it all blue).

 

 

 

PVA Spraying Lessons Learned

  • advice – not a lesson because i didn’t make this mistake – use a clean airline, not one that goes through an inline oiler.
  • make sure you attach your airline fully to your spray gun, else it may fire free, hit the deck and spray dust and crap onto your mould
  • you want a light mist each pass – anything heavier and you may get runs
  • when you’re spraying near the bottom, you may blow crap from the floor onto the part
  • because the mould is very smooth and well finished, you don’t need to spray loads of PVA in to go blue, you can do it with a wipe of the cloth. meh – that was plenty of time wasted, but at least I now know how to put it down in such a way that i can take a part off concrete.
  • Spraying a mould this size to get that much blue meant I used nearly a litre of PVA.

Laying in fibreglass lessons learned

  • for three layers of 450gsm chopped strand mat, your remaining last litre of resin isn’t enough. Not nearly, nearly enough.
  • trying to put the glass in in 6 big pieces just means it all falls down
  • don’t try to fit the glass to the mould and flange – let it hang way down each side – that way the weight of the glass stops it falling into the mould.
  • OR, cut yourself lots and lots of 6″ strips of glass and lay them in – it’s not a competition to get the smoothest inside of a part – that comes later with the proper infusion part.
  • it’s handy having a skip outside to throw away the glass that has only some resin on.

Starting again

  • it was sensible to cut my losses when I did
  • I’m really relieved that I used PVA release agent – it’s water soluable
  • I will pull what I can out of the mould, wash the release agent out (from where I manked it up with grit from the garage floor)
  • Blast it with the power hose and start again
  • I’ll wipe in the PVA this time
  • I now have a 25 litre drum of resin and the right amount of catalyst – no chance this will happen to me again.

More on the transmission tunnel mould

Wow – this mould is becoming a saga but it’s worth it to get it right.photoSo, as you can see from the phototwo pictures, the mould is finished and ready to pull a part.

 

 

So, the plan to get a part out of this mould was as follows:

  1. do a trial bagging session to understand how the bag falls in under vacuum and get a good handle on the pleats. write off the bag afterwards.
  2. lay in the first two layers and set them – this gives me a chance to examine the the quality of the finish and lets me decide if I like it before continuing. If I like it, I can put the part back into the mould and infuse the further layers into it.

But … DISASTER.

The mould isn’t remotely air-tight. I think it boils down to a fundamentally bad assumption on my behalf, namely that the wood I ordered would be vac-tight. It’s not. I did a test on a patch of it and couldn’t get it to hold vac at all. And that’s before I go investigating the other bits of it where it mates and the flanges for the three-piece aspect of it.

So, rather than faff about for a week and get nowhere and waste a load of bagging material, I called Warren for suggestions.

His idea was a step beyond my initial one (throw it into the skip) which was to lay a chopped strand mat part in there (no need for a vacuum seal with simple wet lay) and then take a mould from that.

This has many advantages:

  • 3 layers of 450g CSM is cheap
  • I can trial fit the part to the car. It’s not going to be anything like the final part (13mm thick) but will give me ideas
  • it’s a one-off, so best bet is get the part finish as good as possible and then I can use a standard moulding kit to make the final, vac tight mould.

Lessons Learned

  • I tried a hybrid technique of using part making material for mould making, which didn’t work
  • you can use any old cheap bits for pattern making but moulds require precise application of the right materials and chemistry for resin infusion
  • I have a good mould but nothing like strong enough or sealed enough for epoxy resin infusion
  • cut your losses
  • growth through pain. 80% of learning is experiential – I have learned so much getting hands-on like this.

So, I can now see a positive from this, which is I will have a part that can be wrangled into the shape I need and surface finished. Then I can make the right mould.

mould finishing

So, work has been insane and I haven’t had much of a chance to blog about the car.

The mould is finished and the surface prepped. It’s been painted with high-build primer, flatted, painted, flatted and painted again. It would be ready for a part but I’m having a couple of problems.

Firstly I have sanded back through to the wood once or twice. Then I put more primer on, and still get it wrong about sanding back. I now understand what Warren means when he says you have to get the orbital in there and dig out the high-spot and then fill it with high build and sand back. Many evenings have gone into this process. I could have farmed it out but bodywork is something I don’t know anything about, so want to learn.

Secondly, this could keep going on forever and I won’t end up with the finish I crave for such a visible part. As such, the right thing to do is flat the mould back again and spray the whole thing with 2k clear-coat. then I’ll have a nice shiny surface for my part to take.

Thirdly, in order to get an ultra-violet proof finish on the part, it will need to be gel-coated, which means spraying the mould surface with GC50, something I’m not equipped to do.

So, I’m off to see Warren again and we’ll finish the mould prep. It’s a fantastic amount of hours to get this far, but the end result should be a showroom quality utterly structural part that weighs a lot less than the transmission tunnel coming out.

I think the layup will involve something like 10 layers of cloth and a 10mm closed-cell foam core with some soric hard points and soric radius support on the inside bends of the part.

I’m really looking forward to this.

New transmission tunnel mould

So, after making the transmission tunnel top panels at great expense, I realised that they weren’t going to work well for me. I’d made and cut them so that the panel edges mated with the curve of the top tunnel chassis rails, as you can see on the left.

 

However, the holes in the chassis rails that you drill through to are far too near the edge of the panel to give proper, structural location. I’m not in this for the bling so whilst I could bolt the panel on, it the holes would be near the edge and probably wear and oval. It would be lighter, but wouldn’t add any structural value. As Warren says: “if you’re going to put anything on there, make it structural, or don’t bother“.

So, my first thought was to weld some brackets inboard to the top tunnel rails so I can drill through much further into the panel and get a secure location. This would also mean having to weld up the holes in the original rails which with TIG isn’t a doddle. It’d be easier with MIG but that would also mean injecting a load of filler in. Either method would require grinding the welds flat again.

Then I thought… why not chop out the top rails and replace them, adding the inboard brackets for a quicker route to location. But at this point it dawned on me that I should be thinking in 3D and I should create a wrap-around tunnel secured to the floor. Then the voice of Warren-Wan-Kenobie whispered in my ear and said take the rails off and make the tunnel totally structural.

So, that’s what I’m going to do. I’ve been working on the mould for a while now and here is the history. When you read this, you should read the titles in a Gordon Ramsey voice:

tunnel 1Clad the tunnel. I feel like I live in “world of clamps”. As it stands, each separate panel is cut in 5.5mm ply and clamped to the rail. This is when you understand why many car manufacturers use soft ally. The intersections of the chassis rails are all in different planes. With Ally, you can just drill it and tap it in place with a soft hammer. With composites there’s absolutely no option like that. The value in the part is it’s strength and rigidity. So, whatever I make must be a good fit from the start. You should be able to see the bolts in the top securing the top panel to the chassis rails to give accurate location for the rest of the panel.

tunnel 2Check the Mould. Here you can see the top of the tunnel with the side panel in place. What is missing from this picture is the subtle compound curves in the large vertical panel to the right. This is also the first opportunity to check if the mould will suffer from mechanical lock. It looks like it will, so a single piece mould will not be an option, and I need to go 3 piece.

 

top shot

top shot

Change Your Plan. So, it was when I was looking at the top panel originally and the compound curves with their implications for mechanical lock that I decided (with Warren’s guidance) that it needed to be a 3 piece split mould. This picture shows a much larger top panel than the first picture because I scrapped the original top panel and made a new one with two integrated flanges, minimum width of +100mm over the panel size. It means the side panels can have a mating flange to come up to this top flange and will split well. No chance of mechanical lock. You can also see one (of many) side reinforcing ribs to hold the wood in the compound curve shapes.

tunnel 3Reinforce like a git. Here you can see the side panels in place, and the mating flange up to the top flange. I’ve also put all the reinforcing ribs in to hold the compound curves. The process is simple if a little time-consuming. Clamps hold the panels to the chassis. Then you get the hot-air gun onto the panel to heat it up quite a bit to try and melt the laminate adhesive. Once done the adhesive should re-set and help hold the shape.

I then cut the reinforcing ribs and hot-glue them to the chassis. Hot-glue isn’t anything like strong enough hold these in for good, or up to the stresses pulling full-vac in the infusion process. So, the hot glue only locates the ribs before two or three layers of 300gsm chopped-strand-mat are used to glass them in place using poly resin. Once this lot has set the panel won’t lose it’s shape.

tunnel 4

Get ready for the finish. Here’s the final pair of sides bolted onto the top side. You can just see some of the bolts I used to bolt top to bottom flanges. When taking a part from the mould (unfinished) the flanges will have to be sealed with bathroom sealant.

 

 

tunnel 5Take another look. Here’s the tunnel from another angle. It’s still not right – the long flange pointing vertically up is totally in the wrong place – I should have made this horizontal but couldn’t do so whilst the tunnel is in the car (because there’s a floor in the way). Sorting this and making the mould ready for pulling the tunnel a part is subject to another post – preparing the mould.

Polishing a Mould or Part – Notes

it needs to be revolving polisher

wash mop, and constantly spray the surface.  use a garden pump-sprayer to get a fine mist.

always spray surface and spray cloth. polished needs to be wet to work.

let it use its own weight.  put the polish on the part then the polisher on the dollop. then push polish all over the part. use pump sprayer a lot.

with carbon fibre, always wet sand, then if the water goes grey, you’re through. this is not a good thing.

Making Moulds – how not to do it again

So, in order to prevent a mould distorting under vacuum (especially when you make it from thin absorbent fibre-board) you need to stiffen it, generally by bonding it to a backboard. Previously I bonded the other mould to a backboard with bog. It worked well and made everything solid, but used a load of bog in the process (well over a kilo).

So, next approach was to try bonding with expanding foam, a technique which does work.

What to do – squirt in small amounts at regular intervals so it just makes contact whilst keeping pressure on the top so it doesn’t wobble around

What not to do – squirt a shed load on the backboard, lay the top on and leave it overnight.

photo

This is what it looks like, all distorted and buckled, way beyond the application of bog to fix it. This is what it looks like just before it goes into the skip.

Mould Making, and issues

So, I’m now making the first side panel, which is an intersection of five planes. This is a bit of a step up from my other panels which have been flat.

I have another post to go into the making of the mould but this one is more about the chemistry. As I’ve said to Warren when being frustrated by this before, getting this right is a mixture of talents, namely being a baker, chemist, structural engineer and tailor. He’s got all those skills in spades, and I have few.

So, The mould shape was made from slabs of fibre-board (more on this in another post) clamped to the car and then bonded on the back with glass and poly resin. Then I decided to face the front of it with bog (car body filler) in order to get a reasonable finish.

Now for the trixy bit: epoxy and styrene (the stuff that smells sweet in bog) absolutely don’t mix at all, because the styrene gas breaks down the surface tension in the epoxy and it runs away. Remember this bit for later.

photo copyThe first approach to preparing the surface was to spray tack some release film to it and mould straight off that – it wouldn’t be super shiny and cosmetic but would be very quick to prepare the mould. Spray tack was applied and then release film. Initially it looked lovely and shiny. However, after a few minutes bubbles were appearing under the film, due to styrene gas coming off the part; it was very green at this point. Nadgers.

So, I call Warren and get advice. I don’t have ready access to equipment to spray high-build primer or 2k clear-coat, so the approach I can use is to mix glass microspheres in with the resin to bulk it out and make it easy to sand, then paint it on.

photoand … Disaster

 

 

 

It went off and separated. Three reasons – firstly I didn’t clean the spray-tack off (I assumed epoxy sticks to anything therefore spray-tack can eat my shorts), and secondly didn’t appreciate how sensitive some of this chemistry is. Thirdly it looks like the styrene gas is still bleeding through.

So, how to fix? Go to Warren, sand all the crap off and spray with 2k clear coat.

Job done, and I now have a reasonable mould.

Lessons Learned

  • It takes ages for the poly resin to stop gassing (weeks). Poly is the resin of choice for mould reinforcement though because it’s 20% the price of epoxy
  • it’s a lot of work with an orbital to get that lot off
  • make friends with a local body-shop who can do the spraying as a side-job for beer tokens.
  • It wasn’t necessary to apply the bog to the surface all over, and as thick as I had. All this meant was that I had a mildly uneven, slightly wavy surface. It doesn’t affect the integrity of the part, which fits nicely but it does show up if the sun is bouncing off it and you have a critical eye.

PostScript

I have now pulled a part from this and it fits well. what’s more, it is 546g in weight rather than the original ally panel which is 830g. A saving of 34%. The new part is also way more rigid than the ally. I have a post to follow about the part.