I’ve recently been working on a new product that will be pretty tiny. So tiny in fact there just won’t be any room for screws in the case. So what’s the solution? A case that snaps together is nice, but all those clip features need side pulls in the injection mould tooling which adds a lot of cost. Besides, the clips take a lot of iterations to get right, which means paying for tool adjustments, plus they require a reasonable wall thickness, compromising on size.
How about pillars and holes? Well again you need circular cut-outs on the boards inside and we just don’t have space on this one, plus using glue adds to the assembly cost. Sometimes a clever design means you can just dip the tips of the pillars in solvent weld, which is handy but again this adds to assembly time and requires clamping.
The solution (we think!) is ultrasonic welding. You probably already have a few consumer electronics enclosures that are ultrasonically welded together – your laptop power supply for example. How does ultrasonic welding work? Well it’s pretty simple really – if you rub together anything then you produce heat, right? Well now imagine rubbing stuff together 20 or 30 thousand times a second – things get pretty hot, pretty quick. Ultrasonic welding injects ultrasonic energy into one thing you want to join and where that contacts the other stationary thing you are joining it to, heat is produced. This melts the plastic, creating a weld.
Of course there is a bit more to it than that. That point of contact should be very well defined by creating an energy director feature to focus the heat in just one place. This results in controlled welding and a strong bond. Here is an example of an energy director feature:
However, this post isn’t about energy director design. If you want to find out more, visit the Branson website and read their application notes.
To perform ultrasonic welding you need an ultrasonic welder (surprise!). Well it turns out these are pretty expensive – especially since you really need a welding press to control the force on the joint as well. In fact there are three things to control: ultrasonic power, time the ultrasonics are applied, and the force that is applied. The kind of thing one of those big boring companies would buy is this:
Looks like a serious bit of kit doesn’t it – well it is! These things can easily be 1.5 meters high and generally need an air supply (pressured) for the pneumatics. And the cost? Well for a new one you are looking at something in the region of £10k+, plus the tooling which is extra, of course. You can pick up second hand ones in the region of £5k, but still, it’s a fairly expensive game to be in.
The product we are working on now is a bit of a risk as it is so we don’t want to sink thousands into new machinery right now. Sure, if it works out then maybe we will invest but for the short term we want to at least understand ultrasonic welding and what it can do for us before committing.
The hacker in me stirs.
I figured that really it’s pretty simple – just an ultrasonic transducer and then a mechanically constrained system which allows you to impart a controlled force for a certain time? Having looked at my first port of call, eBay, I realised that there was another type of ultrasonic welder available – a handheld version for heat staking. What’s heat staking? Heat staking is melting a plastic pillar into a mushroom shape to secure something that the pillar goes through. Google it.
Anyway, these ultrasonic heat staking guns are cheaper, a lot cheaper, and they do come up on eBay from time to time. Sure, they don’t have the fancy press part but I reckoned I could knock something together there so I kept vigilant and sure enough before long one did appear on eBay. As luck would have it the guy didn’t even know what it was, or whether it was working. I took a punt and won it for something like £130. Normally they go for more like £500 – £1000 second hand.
Well I was lucky and it did work, with the only downside being that the text was in Italian. As it turned out it was 30kHz which was just perfect for what I wanted (frequency is related to welding depth). However, what I needed now was to mount this in some kind of press system to apply a known force to the parts being welded.
Being a purist, I wanted to do this in the simplest possible way. I was sure that I could cut corners and end up with something much simpler than the fancy pneumatic systems available. Also, as I mentioned, we didn’t have much money to sink into this one, so it had to be cheap.
As luck would have it, a friend of had a Record Power DS19 drill stand for sale. It was partly broken but for this project it was perfect.
I paid my £20 and got thinking. Just mounting the ultrasonic welder in the drill press wouldn’t fix the problem as the force would be relatively uncontrollable, so it needed a second stage linear slide with some kind of force control.
Assembling the press, which was still in its box made me realise how cheap and nasty it was. I mean for the price, it was a bargain, but it’s not exactly Makita standard. A few mm play from side to side just wouldn’t cut it for this project so that had to be fixed right away. After some investigation it turned out that the play was supposed to be taken up by a square section rod inside the main casting which indexed onto the part that was clamped to the upright.
A bit of lateral thinking and I figured I could use some fascinating polymorph I had recently purchased to shim up the space perfectly. The handy thing about polymorph is that it’s a lot like polythene, which is very slippery and makes a reasonable bearing. After heating some up I just pressed it into the indexing slot and pressed the bar in, shimming it up perfectly. Now there is no play at all.
This does sound crazy when I read it back, but it really works well.
So the basic premise was to create a linear sliding part on the already moving part of the drill press. This part would have a spring providing pressure from the main moving part, so that the more the spring was compressed, the more force was imparted on the part being welded. The idea was then to simply use the normal drill press lever to compress the secondary spring by the correct amount (using the existing depth stop) giving the correct force. Don’t worry, you’ll see how it works in a bit (or perhaps you already scrolled down to see, and have already moved onto to another website?).
I started looking for linear bearings on RS for the sliding part, and there was plenty to choose from. However, I remembered that we happened to have an old x/y/z drilling table tucked away in a cupboard that never quite got around to being thrown away. As I suspected, it had most of the parts I needed.
Turns out that there was a decent sliding z head with bronze bearings and steel bars, so that came out and became the sliding part of the press. After removing the drill clamp I mounted the cast head from the drill press in my small milling machine and faced off the front to make it flatter. Some 20mm delrin stock and some milling passes later I had a good solid mounting plate for the reclaimed bearings and bars.
The only problem then was to mount the ultrasonic welder itself (49mm diameter in case you were wondering). This had to be mounted on the hacked sliding Z head. I set to work and 5 minutes later I had knocked up a suitable bracket in CAD:
I converted my small SIEG mill to CNC a long while ago and it made quick work of manufacturing the bracket, again in Delrin. I use Mach 3 for controlling the Mill.
All the running parts were working well now, so I was ready to assembly the lot. Before putting it back together I glued a microswitch to the fixed part of the drill press, which activated when the drill press depth stop reached the bottom.
The handheld welder came apart pretty easily (three grubscrews) and it was a simple matter to add a second short cable coming from the top to use as an auxiliary switch. This would plug into a socket on the main casting which connected to the depth activated microswitch.
All back together now and the principle works well. The force is very repeatable, which was the main goal. I’ll put a small scale on the sliding unit with newton markers on it at some point. If you plug in the auxiliary switch then the ultrasonic welder gets activated when the correct force is met. The welder controller box already has time and power control so that’s all three parameters controllable. Check out the video to see what it looks like:
Total cost was about £170 including the welder, and it took the best part of a day. Here are some pictures of the finished thing (but really you want to see the video).
And does it work? Of course it does! I haven’t made any new horns yet for the welder but I can easily spot weld two pieces of plastic together – see the video. I’m just using an old staking horn at the moment to test it (with the end ground down) which is why it leaves a mark. And no, this isn’t the box we are using it for, it’s just a poor victim that was lying around. The welds are very strong indeed, and that’s without even any attempt at an energy director feature.
Right, where were those mince pies?
