A few weeks ago we revealed 3 silly mechanical design mistakes that you must avoid. Good news – here are 3 more for you! Not everything about mechanical design can possibly taught at university, especially real-world experience. I’ve made plenty of mistakes in my time and here are 3 classic mechanical engineering mistakes you can learn from and avoid with ease.
Design for mis-assembly
If you enjoy stopping production lines in mid-flow then the king of mechanical engineering mistakes is to design something that can be mis-assembled. It is therefore essential to design parts that cannot be assembled incorrectly. DFA (Design For Assembly) is the formal way of educating this point, but fool-proofing your design is actually pretty simple – you just have to be logical in your approach and take extra care to ensure that something cannot be mis-assembled.
When you design an assembly take time to imagine holding the mating parts in your hands. In your mind’s eye visualise the many ways that they could be put together incorrectly, and try to anticipate problems. Think of ways you can deliberately put them together wrongly even if it seems like a really stupid thing to do, and even if it means forcing them together (people will do that).
It’s also worth imagining the worst possible kind of person who is could end up screwing together your parts on the assembly line. Imagine that the person in question is inexperienced, ignorant, definitely hasn’t seen (or wouldn’t understand) your assembly drawing, is not exactly highly paid, is working against the clock, is in a transient job and possibly doesn’t even care if they put it together incorrectly. Don’t get me wrong – in my time I’ve worked with many superb assembly operatives, but not all are like that. You definitely need to cowboy-proof your designs and always assume the worst.
Get a prototype made once you’ve improved your designs as much as you can. Get your parts 3D printed, machined (or whatever method is appropriate) and have a good play. Show them to colleagues and people on the production line if you can. Improve your design as necessary.
Wrong material specification
When you design a part and specify a material for it take care to ensure you choose something with the correct properties for its intended use. In the first instance there are a plethora of mechanical properties to consider (10 examples are here), but bear in mind that not all requirements are related to mechanical engineering. For example, the material may need to be RoHS compliant, or have some type of UL certification.
Materials are often required to have a flame retardancy rating (e.g. UL94-V0 or UL94-HB). If you have an approvals engineer or approvals department at your company you should discuss this with them. Ensure that you specify a UV-stable material for outdoor use to prevent it from fading, changing colour, or degrading in sunlight. Material choice can also be sensitive to geographic location. For instance, some countries frown upon PVC, and you might have to find a suitable alternative if you want your product to be marketable there.
Here are a few very basic things to bear in mind when you select a material to ensure that it will be acceptable:
- Regulatory compliance (e.g. UL, RoHS)
- UV stability if it will be exposed to direct sunlight
- Operational temperature range
- Flame retardancy rating
There are a tonne of other things to bear in mind. I suggest that you make your own personalised checklist and refer to it every time you specify a material for a part. When putting that list together talk to people, particularly approvals engineers, to ensure you don’t miss out anything important.
A beautifully detailed finished design that doesn’t work
It’s best not to assume that something will work because it looks OK. Always make a Proof of Concept (PoC) model (or Proof of Principle (PoP) model as it’s sometimes called) to guarantee that it will definitely be OK. A design can usually be broken down into very simple functional elements and it’s easy to test these for reliability by designing extremely simple test rigs. For example, the one pictured on the right is to test the reliability of an o-ring seal when the mating bore is machined to maximum manufacturing tolerance.
We recently published an article how 114 people died because of a simple joint design error. Even if the designer, for some reason didn’t bother to do the calculations, imagine how quickly and easily a PoC model of that joint could have been knocked up in the workshop and tested. An afternoon’s work could have saved the lives of all those people.
We hope you’ve enjoyed this article and that it’ll help you avoid some avoidable mechanical engineering mistakes! Do you have any silly mechanical engineering mistakes you want to admit to? Tell us in the comments below.
6 thoughts on “3 More Really Obvious Mechanical Engineering Mistakes You Must Avoid”
From own experience. You have finished your project. Lay out it from twoo-three weeks. Then look at it carefully again. You can see new mistakes and probable some silly a your design.
Good article
Yes, I got caught out badly with not proving a concept with a physical prototype once. The seductiveness of CAD models and FEA results!
Not using GD&T to dimension and tolerance parts and assemblies. Traditional coordinate dimensioning and tolerancing does not fully define design intent and the resulting ambiguities are a significant cause of parts not fitting and functioning correctly.
Well, personally, I don't make any such mistakes… in the alternate universe that my ego wishes to live in…The fact is that I've made some horrible clangers in my time, but they're so indelibly seared into the DNR (Do Not Repeat) component of my brain that I honestly can't dredge one up right now – but I can happily tell you about a massive whopper made by someone else – one of my suppliers (a Die Casting company), makes Alternator housings for the automotive industry, mostly for Japanese models.One day, they get an order for 10,000 housings, so they pull the bolster and inserts from storage and happily start casting – 1 month later they realise that they've been casting one half of the housing for manufacturer X and the other for manufacturer Y… oops…So to amplify Gareth's comments about a DFA being required, if you're going to automate an assembly process, for heavens sake make sure that there is a deliberate incompatibility between supposedly mate-able parts, such as a die casting bolster…
Hi, I'm a mechanical design engineer from Russia. I work for a company that is supported by the ministery of defense. We have a short-list of materials and regulation, helping us to avoid problems while our devices fly on the board of aircraft. And we have also test operations. Of course, we have to develop our assemblies that can resist temperatures from -60 up to 80. Our production lines consist of people described in this article and locate in the other city. So we have a lot of troubles.