Hello girls and boys ! Today I’m starting a long-term project : the design of an open source flying probe tester. Why ? What is it ? Do you need one ? The answers to those questions lie in this introductory article, along with the (not so) complicated etymology of its name, Apertuisane. Next time, I’ll start looking at the challenges involved in this project and share my ideas of how best to address them without breaking the bank.
As of this second, all you need to know is that even if you think you don’t need a flying probe tester, you still have good reasons to follow this project. And it’s not just for your love and admiration of Nefastor. You’re obviously smart enough to realize that I’ll be applying various technologies I write about, and also some I haven’t yet. Plus, it’s like Game of Thrones : with me at the helm, you never know which component is going to be killed next, and there may be tasteful nudity.
OK maybe not that last part. If you’re nice. We’ll see.
Apertuisane is a what ?
A flying probe tester (FPT for short) is a robot that jabs probes on a circuit board to take measurements. It does what the average electrical engineer does when looking for shorts, opens, or measuring things. If that engineer had eight arms and could place eight probes at the same time, dozens of times per second, with sub-millimeter accuracy, like this :
Looking a bit deeper, it’s a CNC machine tool which shares a lot of specs and parts with 3D printers. Does that pique your interest ? Indeed, the commonality extends very far, which is one of the reasons I’ve decided to tackle this project.
FPT’s have two major uses :
- You can feed them naked boards to detect shorts and opens. You’d want to do that if your boards have four or more layers and visual inspection is therefore impossible. Better be sure your copper is in the right places before you waste time and money soldering components on it, right ?
- After you’ve assembled your boards, you may also want to test them again. Solder can make bridges, heat can delaminate copper traces and derate components, and even if none of that happens, you may simply want to check that your boards work, by hooking up a power supply, multi-meters, etc…
There’s a simpler way to do all that : it’s called a Bed Of Nails. Just build a grid with as many probes as you want, press it onto your board, and you’re connected. It’s even faster. Just one problem : you’ll need to make a dedicated probe head for every revision of every board design you need to test. An FPT is flexible, so it’s perfect if you’re making small batches of many different boards.
Apertuisane… is that French ?
Kinda. Well be thankful I’m not German or maybe this project would be called LuftSondenPrüfer 9000.
The name is both a pun and an homage to the open-source project that inspired me to start this project : Apertus°. If you don’t know the name but love digital movie-making, you should definitely go take a look. After I’m done with you, if you don’t mind. Consider it dessert.
Apertus° makes video cameras, hence the name, which refers to the hole through which you can form a picture onto film (as in a camera obscura). And a pertuisane is a medieval spear that is both very long and very pointy. Pertuis, as it happens, is an very old French word for “hole”. Where did you think the word “aperture” comes from ? Cave Johnson ?
Take its name to mean that Apertuisane will poke circuit boards with sharp sticks to find any weaknesses they might have. You could also say that Apertuisane will poke holes into electrical problems. More importantly, it sounds a lot like Apertus°, so that’s nice.
You know what’s even nicer ? That orange “degree” symbol in the logo. To quote the Apertus° website :
The “o” stands for “open”, “openness”, “open source” and is placed where a “TM” symbol (indicating patents, trademarks, protection) would normally reside.
Who is it really for ?
Just like the song, it’s just me, myself and I : the robot army I’m going to conquer the world with is not going to suffer the crappy quality control we’ve witnessed from the Trade Federation and Skynet. Nah, just kidding ! It’s for the people, of course ! Besides, if I had a robot army, I’d make them test each other, duh !
FPT’s cost a fair bit of cash, and as such they are beyond the grasp of a growing segment of electronics manufacturers : the crowdfunding community.
The Kickstarters of this world are full of niche electronics projects which, when successful, lead to the production of anywhere from a hundred to several thousand boards. This is a dark area : enough boards you wouldn’t want to test them all by hand, but also far too few to justify buying an FPT.
If your project happens to fall into that category, the logical solution is to outsource your testing. But that isn’t especially cheap either. Or practical.
That’s where a small and cheap FPT will come in handy.
It would also be a nice tool to find in any hackerspace. If it is easy to setup, perhaps as easy as feeding it your CAD files on an SD card, then putting every board you make into the FPT down at your local hackerspace could become a time-saving reflex for almost any hobbyist. That’s my vision, at any rate.
Also, it will look nice in your hollowed-out volcano hideout. Think what Goldfinger could have done to James Bond if he had had an FPT instead of a laser…
Where is this going to happen ?
Apertuisane, or rather, Apertuisane°, is an open-source project. This means everything about it will be published online.
However, it’s going to be a joyous mix of specification documents, CAD drawing, circuit board designs and two or three flavors of software. And it all needs to be under version control because a/ this isn’t the Dark Ages anymore and b/ while I’m starting on my own, anyone of you lot could end up contributing.
For now, here’s where I’ll put everything :
- Apertuisane GitHub repository : any file that can be handled by git without it getting in the way of efficiency, will be stored in that repo. That means any engineering document, source file, text file, etc… and I must warn you, unlike my website it’ll contain no attempt at humor.
- OnShape cloud : OnShape is a free web-based mechanical CAD system with capabilities on the level of SolidWorks. OnShape documents are public by default, and under version control. You can learn more about it right here.
- Nefastor Online : since this is my site, what better place to post my progress and any news about this project ? Also, if you really must bother me, this contact form is the channel I’d like you to use.
It all started from this page on the Apertus° website.
The concept is simple : to leverage components that have been made really cheap from the economies of scale born out of the rise of DIY 3D printers.
So which parts exactly can be taken advantage of ? Initial analysis suggests it’s pretty much everything but the hot end, extruder and hot bed. This doesn’t mean we could just take out those elements and jam an oscilloscope probe in their place, though : unlike 3D printing, probing a board requires at least two probes, which need to placed at different points on the board.
There are many different ways to do that, search for “flying probe tester” on YouTube and be amazed. But we’re on a budget, here. It means we need to be smart.
The initial idea is to use a two-axis probe carrier :
You can see a green disk on this render. It carries two probes mounted on a linear axis. This axis, as you no doubt have guessed, controls the distance between the probes. Combined with the disk’s rotation and the X and Y axes that carry this contraption, you get the ability to place two probes at two different points on the same BUT (Board Under Test). It’s mathematical : you need to place two probes on two dimensions each, so you’ll need four axes. Good thing they’re practically giving stepper motors these days.
It’s a good start. Especially if you place a second set of probes on the other side of the circuit board.
And now, off to work !
So there you have it : a need, an idea, a statement of purpose and a place to keep track of all that. All that’s left is to roll up my sleeves and be about it.
If you’ve ever attempted to design and build your own CNC machine tool, you’re probably thinking what I’m thinking right now. That is, no plan survives contact with the battlefield. I’m pretty much certain that by the time you see an actual Apertuisane° at your local hackerspace, it’ll only bear a passing ressemblance to this design. Call me crazy, but this uncertainty, this surprise, is something I’m always excited about when I start a project.
And truth be told, there is definitely some uncertainty here. While I have designed various machine tools and robots over the years, there’s something I’ve never done : designing a robotic probe. So that’s where I’ll start : tip of the spear !
In conclusion, the next article in this series will look at my first experiment in computerized probing : making cheap linear actuators out of whatever is handy. And once I’m reassured I can do that, we’ll talk high-level requirements and specifications.
See you then, mortals !