Carbon fiber outsourcing lessons

Well where do I start?

We are a small operation and pride ourselves on quality and customer service.  We've made a conscious effort not to expand our core operation for reasons I'm not going to explain right now.  So it is obvious to outsource if we get more orders than we can cope with doesn't it?

We've tried other companies in the UK, we've tried France, Eastern Europe and China.  We've lost shedloads of money shipping moulds backwards and forwards and reject parts.

Lesson 1. 

Start with an precise engineering level quality standard. 

We didn't, we thought some of it was obvious.

Lesson 2.

Don't use places outside the UK unless you are going to put a lot of business their way.  Any price advantage will be lost in poor quality; agency fees and tooling costs.

One consignment from China was 90% reject.  The frustrating thing was that the base product was fine, these were all lacquering problems.

Conclusion.

All our motorsport carbon production has been brought back in house.

All of our phone case production has been brought back in house.

Designing parts, how swearing helps.

I hardly ever swear.  Honest; I just don't.  I don't see the need for it and to my mind the first person to use a profanity in an argument automatically loses even if the other person is in the wrong.  Swearing at a person just hands them the moral high ground and all the power.

But what if you swear at a computer?  Well they never swear back.  Even Siri is very polite (go on, try it).  I've observed that an alchemy of talent and profusive swearing results in stunning designs.  Like this for the iPhone 5, developed in CAD from Apple's drawings. 

 

Our designs start with the end user.  What do you want and need?  We have a cycle of feedback during the design process to try to get it right.

Our rally customers want something that will give them minimum fitting time in the service park and workshop; look stylish and do the job. They are looking for minimum weight, maximum function and good looks. 

Our range of Subaru and Mitsubishi parts were originally designed to fit Group N rally cars but also work for Group R.  Group N/R starts with a road car as it comes off the end of a production line.  As far as bodywork goes the rally teams will strip everything they possibly can out to save weight; have to put some things in for safety,  and add some things needed to get the car to the end of the stage in one piece.  FIA regulations define what they can and can't do. Basically they rip out all of the interior, put in a roll cage and add underfloor and sill guards.  OK, there's a bit more to it than that.  There's the small matter of the electrics, braking and suspension systems, and they may tweak the engine a bit if they have time.

Our carbon parts address the weight issue and our epoxy and glass parts help to address the "getting to the end of the stage in one piece" issue.  So we will make things like mirrors and door cards from carbon fibre and things like underfloor guards from Twintex (superb for impact resistance).

The designs for our rally car parts mostly start from 3D scans of cars.  This involved more scary investment in kit and software.  Do you want me to repeat how I look from the stress of this?  No, I thought not.  If you need to see just look backwards in the blog.

Back in 2007 we got some scarily expensive kit (Faro arm with laser scanner; geomagic and solidworks software, plus fearsome spec computers) our team went out to Japan and scanned the Subaru Imprezza before it was launched.  We worked with a rally team in Japan and STI to prepare parts and they revealed a fully prepared rally car on the same day as the road car was released to the showrooms.  And it had our bits on it!  We did more scanning of the Subaru in 2008 at Vermont Sportscar www.vtcar.com and did a complete scan of the EVO X at Van Heuvel Motorsport www.heuvel-motorsport.com in the Netherlands.

When I first heard about this scanner my imagination worked no further forward than a flatbed scanner making 2D images.  I was so wrong.  It is a fabulous bit of german engineering.

The challenge to me was booking flights for it without paying stupid amounts of excess luggage and the challenge to Michael and Matt was to travel with an extra three boxes weighing 95kilos between them. I'm sure there was some swearing, especially when they ended up having to navigate Tokyo on their own.

The Faro laser scanner creates a cloud point image of the thing being scanned.  Imagine something solid being made out of millions of points in the air.  More software is needed and skill (and swearing) to get a surfaced drawing that can be downloaded to CAD software such as SolidWorks.

So now we have the data in SolidWorks.  We know exactly what the surface is we need to match our parts to, now the design can begin and the swearing goes up a notch.  We work to the four fs of design; function, form and fit.  Yes I know there are only 3 listed, you can work out the other one.

Well obviously the part has to fulfil its function and fit.  Our mirrors would be useless if they didn't fit or the driver couldn't see out of them.  Our designers go that extra mile.  Our rally mirrors are vertically oriented and hug the contour of the car.  This means they still supply the driver with enough visual information but are also less prone to be wiped off in that "on the limit" manoeuvre.  Added to that they look curvy and sleek.

Sorry, I didn't mean to sound like a sales brochure but the thought and time that goes into the design stage just amazes me.

So I (or heaven forbid, a customer) might suggest something at the last minute, it seems simple and obvious to me.  What's wrong with that?  I'm being helpful aren't I?  Swearing goes off the scale and there may be some door slamming.

But somehow we end up with things as beautiful as this.

Incidentally these doorcards and our mirrors were fitted to Ken Block's car.  See him playing with James May on Top Gear here. http://www.topgear.com/uk/videos/ken-from-the-block

Making a pattern for a motorsport part

The motorsport parts we make are manufactured using vacuum bagging and heat curing.  So this post is about making a pattern for a motorsport part.

Once we have a design for a motorsport part we send off the CAD files to a CNC machiner to make the pattern from tooling board.  The CNC machine is a big, expensive bit of kit and looks like this:-

The engineering toolmaker that we use is a true craftsman and he is always very busy so we have to build a lead time into everything we do.  We've tried to cut time out by using other toolmakers but we got some really shoddy patterns back that ended up having to be re-cut by the original toolmaker.  So it ended up taking twice as long and being much more expensive.  Lesson learned; when you find someone who does a good job, stick with them.

When the raw pattern comes back to us it has to be finished and sealed. Finishing means a lot of rubbing down with various grades of rubbing down paper. This can involve a lot of swearing as well. I've never got a sensible explanation as to why, I just accept it as a fact of a life.

Even I can help with the sealing bit. We use fine cotton cloth and sealing fluid. I wipe every bit of the pattern with the sealing fluid and it is passed over for more finishing. It then gets passed back to me for more sealing and so on at least five times.

We then end up with perfectly finished patterns. The surface is absolutely smooth and is highly polished with a silky touch. This is one of our patterns.

A pattern may only be used once but the effort put into this stage of the production makes all the difference between something that fits and looks good and something that vaguely does the job.  Your design can be brilliant but if the pattern is rubbish guess what you get?

A 10 second rant

Why oh why don't people pay you on time?  One or two each month is OK but all of them?  We seem to have hit a perfect storm this month.  Then I freak out because I can't decide if they are about to go bust or they are never going to order from us again and they don't care.  Please tell me I'm not being paranoid, please, pretty please.  OK, I don't care, (well I do really).

Thank goodness our chief designer comes up with stuff like this.

For those with an expensive bike and no place to store it.  Put the bike on the living room wall.  Simples.

Acually what is simple is this lesson in how to be two.  Thank you Lily.

Customer service - why we try so hard

I don't understand rubbish customer service.  Why would a company do that to its most valuable asset?  These are the people who are willing to pay you to give them something you want to sell.  These are the people who will tell their friends, family and colleagues ad-infinitum when they're pleased with something they've bought.  These are the people who will buy from you again if you treat them well.

The ones who contact you with a complaint?  These can be the most valuable customers.  Listen to them and you get valuable feedback on how to do things better.  Treat them well and they become your best advocates.

I loathe customer service centres.  I especially loathe the premium number ones that give you multiple menus and then keep you on hold even though "your call is very important to us".  Well sorry, but it was that important they'd employ enough staff to answer in a reasonable time.

I know big companies have to find more and more efficient ways of doing things so they can satisfy market expectations in these hard times but there are limits.  It's so frustrating when you talking to someone who is reading from a script and has to go through pointless steps when you know they can't help and you need to get to the next level quickly.  It's equally frustrating when you hit a brick wall of jobsworth and obdurancy.

Phew, I got a bit carried away there.  And breathe.... 

So we try with all best endeavours to look after our customers.  If they have a complaint we try to resolve it as quickly as we can.  If our customers are happy it can only be good for our business can't it?  We welcome all feedback from our customers as reviews on our website www.m-techdesigns.com for phone cases or www.m-techcomposites.com for rally and other composite parts. Pick up the phone, tell us. We operate on continuous improvement and your feedback helps us to do this.

If you'd like to comment on any companies with particularly good service we'd be glad to give them a mention. 

Carbon Fiber Moulds or why you can't make a silk purse out of a sow's ear

The moulds used to make carbon fiber parts are frighteningly expensive if you want to make quality parts.  I know because I juggle the money.

This is me on a good day, you should have seen me on halloween.

Every imperfection in your mould is shown up on your parts so they have a perfect surface for a perfect part.  They also have to be able to withstand the production process and be capable of being used time after time without degradation.

Our moulds are lovingly made and live a life of total luxury.  They are pampered and cossetted because we can't make anything of value without them.   

Most of our valuable rally car moulds are made from high temperature tooling material; they are then sealed, polished to perfection and treated with a release agent. The whole process takes several days, lots of craftsman's time and many hours in our industrial oven.  They are a pleasure to touch but make sure any finger marks are polished off. 

These moulds are formed from a pattern which resembles the finished product with added flanges for the moulds.  Oh crikey, patterns; well there's a whole new topic.  "Design to pattern, how much swearing?", coming soon.

Our moulds for iPhone cases are made from highly polished steel.  The steel moulds are cut using CNC (computer numerical control) multi axis milling machines.  We outsource this work to specialists, apart from anything else my blood pressure would not withstand the cost of the CNC machine.

I may update this blog with more pictures but right now I'll leave you with the image of me above.

Why carbon fiber thread counts don't matter but weight does

We already discussed how carbon fiber fabric is woven from tows (or yarns) made up of thousands of filaments to a level of density giving a weight per square meter in the previous post.  Well I don't know about you, but by the end of that my head was spinning with yet more questions and I needed to go and test out my 410 thread count sheets.

The question I most puzzled over was why the yarns were 1K, 3K, 6K and 12K?  Why just those and why not a linear progression by weight?  Well the truth is I don't know and haven't been able to find out and to be quite honest for our purposes I don't need to know.  All I can guess is that somewhere along the line these became de facto standard and various manufacturers in the chain tooled up to make and cope with these.  If anyone knows a better answer I'd be pleased to hear it.

Most of the rolls of fabric we buy in are 1250 cm wide.  Our freezers are this length, our cutting benches are sized accordingly, our cutting templates are all calculated on this size, etc, etc. We would be pretty hacked off if suddenly the whole industry moved to 2000 cm wide fabrics.  Why are they 1250 wide?  Well I don't know, I could make some wild guesses but it doesn't matter, they just are.  See what I mean about raising more questions? 

So back to filament counts.  A 12K tow will be 6 times thicker than a 2K tow but also 6 times heavier.  It will also be stronger but we'll look at that aspect later.  Fabrics made using these thicker tows are useful but we don't usually use them on their own.

Generally the higher the filament count the coarser the material.  It's a bit extreme but try imagining the difference if you stroked a piece of fabric woven from silk embroidery thread versus one woven from silk rope.  The rope cloth will have a bumpier surface, be thicker, heavier and more difficult to tuck into a corner.   OK, you could make the silk rope into a less bumpy, thick and heavy material by weaving it slightly less densely but then it wouldn't be as dense and there may be gaps showing.  If you hung them both on a wall and tried to paint them which one would need most paint? 

It is the weight of the carbon fibre fabric that determines the thickness, (this is a massive generalisation but it will work for now).  A rough approximation is 200gsm = 0.2mm.  So if you want the finished article to be 1mm thick then a 1000gsm fabric should do it, right?  Well no, not quite.  It might work if the finished article was completely flat and the same thickness throughout and you could get hold of a 1000gsm fabric and you didn't mind a bumpy surface. 

Most items made from carbon fiber fabric are made in moulds (2nd wild generalisation but keep working with me) so you need to be able to drape it into the mould, build up layers to the required thickness, with an appropriate surface layer for the finish you want. 

We make door cards for Subaru and Mitsubishi rally cars, FIA regulations say they have to be 1mm thick.  We use two layers of 200gsm 3K 2x2 twill and a layer of 600gsm 12K 2x2 twill.  The 200gsm gives us a fine surface finish and the 600gsm gives the bulk.  This is the most cost effective combination for this application as the 200gsm costs almost as much as the 600gsm and we are saving labour in only having to cut and layup 3 layers rather than the potential 5 layers. 

So those pesky generalisations?  I've not said anything about resins, pre-preg versus dry materials, composite mixes or manufacturing methods yet have I?

So what about thread counts?  Well if you recall from the previous post the thread count is the number of ends (weft) and picks (warp) in a square inch so that gives you the density of the fabric.  We already know that a certain weight is achieved by using an appropriate tow to the right density.  If you know the weight, tow size and weave you know the thickness and the surface appearance so who cares about thread count, it's not as if we are buying cotton sheets after all.
.