….. beavering away.
OK, so we’ve been a bit quiet of late, but that doesn’t mean we haven’t had our heads down and been busy beavering away.
One of the development jobs we have on our “To Do” list – and this is probably the biggest job we need to successfully complete and has been going on for a while already, is to upgrade the “front end” of our datalogging system – the part that handles all the various sensor inputs and takes their outputs, samples them, and sends them off to the brain to be packaged up, tagged with meta data, and then sent to storage.
Meta data is useful information required to understand anything and in this case it would be sensor type, the time the sample was taken, what the sampling settings were and so on.
So one of our team members is currently writing the software that runs the data sampling process, while another is designing the prototype of the data sampling hardware, while another is checking that design concept (always have a second pair of eyes look at everything – another of the Boss’ mantras) and actually building part of it to enable on-the-desktop testing.
Its a real team effort, and it can be quite an effort.
And here is how part of it is done.
Most printed circuit boards (PCBs) these days have more than one layer to them, with each layer doing different things like routing data or power to/from individual components.
These layers need to be generally isolated from each other, except in very specific places that act as the route between the layers.
We take for granted these days the complexity of the PCBs that lie at the heart our mobile phones, and do we ever wonder how they are made or how difficult they are to make?
We have a facility within our lab to make PCBs which we use to make prototypes of our designs.
Only when these designs are proven and very thoroughly tested do we send them off to a PCB manufacturer to make en masse, otherwise we have to wait between 8-12 weeks just to have a single prototype PCB made, which really slows down development progress – especially when we are up against hard deadlines of research vessels leaving ports.
Time and the tide of research wait for no-one.
So here is how we do it – and we are making what would be classed as a quite simply PCB these days as it has very few layers.
The first job is to mill (or cut) the separate layer template of each layer into its very thin matrix.
This includes the perimeter of the PCB itself. So already you can see what shape and size this one is going to be.
After all the layers are milled they are assembled into a sandwich, which then gets pressed for a period of time to make sure all the parts are in contact, and to ensure that the sandwich is flat.
After pressing, holes for the through-PCB components, and to allow for screws to pass to attach it to the datalogger frame without shorting the power or data lines, are drilled.
The board is then dunked in a tank for a while and copper etched, the copper “sticking” out of the milled out template.
Then all the intra-component copper “rails” are revealed by milling – and of course this has to be done to both sides of this PCB as it has two “layers” – front and back.
A solder mask is then applied because we only want the solder to “stick” to the parts we want to solder components to.
And at this point the PCB is only hanging on to the rest of the sandwich by a few remaining “threads”.
And a quick tidy up – and the PCB is done – here is the top ….. and ……
….. here is the bottom.
And you can start to get an idea just how small some of the components are that are going to get soldered to this board.
Then the long road to assembly starts, and it is very fiddly, requires a steady hand, and certainly good eyesight.
The main component on this board is less than 2 cm x 2 cm in size (less than 1″ x 1″ in old money), and can you see how many individual pins (or legs) it has around each of its four sides. Nightmare!
And the other side – the bottom.
And after a final check for stray bits of solder, and a scrub down in the sink, a thorough dry out and a buff up – time to power it up and see if it works.
Or in the simplest case – do the lights at least come on!
Remember this is just a prototype – and so it can look a bit weird with stuff hanging off everywhere at this point in time – and this board is currently plugged into our “datalogger in a box” test rig which makes it look even weirder.
The board on the front is the very high precision clock from our standard datalogger, so this prototype board can at least know when it is, even if it doesn’t know where it is or what its doing yet!
The lights are on, so that is a good starting point.
Well done team!
As you know our lab is in the northeast of England, which has a heritage of heavy industry and coal mining, whose many sites are in various stages of reclamation.
Not far from us is Consett which used to have a large steel works, but is now more famous for a particular brand of crisps.
But you should always remember your past, even if that remembering is done in weird and wacky ways.
Is this when industry becomes science, and science becomes art …. ?
Clearly the artist has a thing about astronomy!
But is well worth a visit if you are passing.