Tolcap enables you to predict the manufacturing process capability of the parts you design. Out
of tolerance parts are a big problem, adding significant costs that eat away at the profitability
of your product and swallowing scarce engineering resources. This video outlines why you need Tolcap.
1. Why use Tolcap
- Why on earth would we want to subscribe to a website to help engineers put tolerances on drawings?
- That's what they do, isn't it?
- So what is Tolcap?
- What does Tolcap actually do for you – why should you use it?
- What is the value of Tolcap to you and your organisation?
Let me try to answer those questions.
What Tolcap enables you to do is to predict the manufacturing process capability of the parts you design:
Will the dimensions of the part be within tolerance, or will there be an unacceptable number of parts out of tolerance?
Out of tolerance parts are a big problem, adding significant
costs that eat away at the profitability of your product and swallowing scarce engineering
resources as you struggle to put things right rather than getting on with the next project.
2. What is Design?
Well, what is design?
When I've trained teams in ‘Design for Manufacture and Assembly’
that has been my introductory question.
We think about the verb ‘to design‘ and maybe get to the idea that we are transforming a required function into a form.
And maybe that gets us to the noun ‘a design’ - or even ‘the design’ as we say: a list of parts
each defined on a drawing.
But what is this design?
Not just a picture of what we built
and tested and seemed to be acceptable – no, a design is an instruction and a contract.
Each drawing of each part in the design instructs the supplier what is to be produced.
Then, if the part conforms to the drawing, the supplier will have fulfilled his contract – no
one further up the supply chain right through to the customer will be able to say “It
doesn't work” or “It doesn't fit’.
Once you accept this definition of the design, the rather
daunting consequence is that it is down to the designer to ensure that the contract is
feasible, and the instruction is capable.
Much effort goes into analysing designs, even
more into testing prototypes, to ensure the drawing captures every constraint necessary to
ensure that the function is delivered. That is a large part of most designers' working day
and, hard work though it is, the task is pretty much satisfactorily achieved. But often the
process capability (or otherwise) of the design tolerances emerges only after the start of
full production, when problems in assembly or of dimensions out of tolerance then have to
be addressed. And oh, of course, the design still has to manufacturable at the lowest
3. The Cost of (Poor) Quality
I said that poor tolerancing is a big problem. A study carried
out when Tolcap was being developed showed that poor tolerancing really is a big
problem. We could readily account for quality costs of 10-20% of sales – a figure of the
same order as profit to sales in some organisations. It's not just factory scrap and rework
costs: it's those design changes and associated warranty and recall costs that really hurt,
not to mention your loss of reputation.
When we look at the reasons for these in production
changes we note that they come firstly from the customer – who is of course
always right - but next come changes caused by poor design for manufacture or excessive
product variability, and these significantly outweigh ‘engineering’ failures – so at least
all that testing helps a lot. And perhaps if we looked in more detail at the customer
driven changes, some of those must come from variability in parts that interact with ours.
So achieving process capability is important.
4. Measuring Manufacturing Process Capability involves maths
I'm hoping you know the basic statistics of mean and standard deviation.
If not, if we were to measure a particular dimension on a number of parts:
The mean would be the average of the measurements.
The standard deviation is a measure of the variability of the
measurements: take each one, subtract the mean, square the result, add up all the
squared errors and divide by the number of measurements and then take the square root.
The universal symbol for standard deviation is the Greek letter sigma, σ, and we always call
Process Capability is measured as a ratio relating the specified tolerance for a
dimension to the standard deviation, or sigma, of the manufactured parts. The usual
measures are Cp and Cpk.
Cp is calculated as the upper tolerance limit minus the lower
limit divided by six sigma. Cp assumes that the mean or average size of the dimension is
equal to the nominal on the drawing
- whereas Cpk takes into consideration that the mean
value may not be at the centre of the tolerance band. To calculate Cpk, work out the
lesser of the upper limit minus the mean (or the mean minus the lower limit) and divide
that by three sigma. Note that Cpk will be less than Cp unless the mean is exactly centred
in the tolerance band.
When we talk about process capability we mean in the long term and
we recognise the mean of the distribution will move, and Tolcap predicts Cpk.
5. [Design for Six Sigma]
I always think that the quality improvement people who promoted
the ‘Six Sigma’ methodology stole the name and then rather reluctantly discussed
‘Design for Six Sigma’ as a bit of an afterthought. In fact ‘Design for Six Sigma’ came first!
I found it in a paper from the early days of semiconductor development. A batch of
prototype new transistors would be characterised – that is the parameters of interest
measured for all the samples - so a data sheet specifying the device performance could be
issued. Setting the data sheet limits plus and minus six sigma from the measured sample
mean would allow 1.5 sigma either way for the mean to vary (there's that probably
confusing ‘process shift’ that came up if you had six sigma training!), and the remaining
4.5 sigma gives a projected Cpk of 1.5.
Tolcap allows you to design for six sigma, though –
and as a one-time master black belt I probably shouldn't say this - we suggest that a
Tolcap predicted Cpk of 1.33 is quite satisfactory in practice. But to design our part
dimensions for six sigma and to ensure a capable design, we do need a way of predicting
process capability, Cpk.
6. Predicting Process Capability.
So how many sigma or what process capability will we get in production?
Some dimensions will be easy, some may be impossible and redesign will be
necessary – face this before production starts. This sounds like bad news for the design
team and the organisation: making this part of design will cost us time and money. Well
that's true, and these costs are visible in the time sheets and accounts. But remember
those quality costs of failures and rejects, rework and redesign hidden away in the
‘overheads’. Less visible they may be, but someone has to pay somewhere and it will be
a lot more expensive in the not so long run.
Some dimensions may be marginal – and need
to be identified as ‘Special Characteristics’. Some customers require you to identify
special characteristics on the drawing.
Which ones are ‘special’?
Surely all tolerances need to be met or you could open them out?
Special Characteristics are exactly those
marginal cases where you cannot just assume adequate process capability. You need to
identify these and then have a rational engineering discussion with your supplier.
7. The Key Question to Suppliers.
Predicting process capability enables you to ask the right
question when reviewing your design with a supplier.
The wrong question is: “Can you make this part (capably)?”
the answer is likely to be ... “Yes!”.
If the design is good, then all is well.
If the design is a problem, it will be a problem to other suppliers… and they
won't be able to do any better.
If the design is really bad, you might just be told
“NO” ... but don't count on it! It takes a very confident supplier to say this to you.
question is: Analysis shows that the predicted process capability for this dimension is
(0.95), and we need 1.33 (or start from 1.5 for a six sigma design) ... so what evidence do
you have, or what tests can you do, to demonstrate the process capability you will really
What this does is transform a possibly confrontational meeting implying that the
supplier is not clever enough to make your part into a cooperative engineering analysis of
whether a particular measure – a Cpk - is achievable and what data can be analysed or
generated to demonstrate that.
Sometimes suppliers will be able to achieve the marginal,
perhaps by special consideration of the tooling, perhaps by allowing for control charts.
Sometimes suppliers can achieve the apparently impossible – but usually only because they
have put a lot of effort, sweat and tears into the problem in the past – but all engineers
understand the need for proof with data.
8. Predicting Process Capability.
So how can we predict process capability from the early
design concept stages?
If we can predict process capability we can foresee
potentially costly design problems, discuss them constructively with our supplier, and make
changes if we have to in order to release a design to the supplier which he can meet
capably and fulfil the design contract.
Have a look at Tolcap, maybe analysing a part that
you know has given you problems in the past, and see if Tolcap shows up the issue. It's
quite intuitive, but the next video tells you how to get started.