The call came in on the Tuesday – the client (an electronics design house) needed 6 prototype boards assembled, and they needed them fast. ‘As fast as possible’ was the phrase used.

So we received the gerbers and the BOM, and set to work pricing the job up.  The trickiest part of any rapid build exercise is getting the PCBs fabricated, and this job was no exception as the PCB in question was huge – nearly 700mm long by 400mm wide.  It rapidly became apparent that you can’t make a PCB of over 550mm on a rapid manufacturing basis.  It’s entirely feasible if you’ve got a couple of weeks, but there are size limitations on what you can do in a mere couple of days.  So we went back to the client and suggested that they cut the board into two equal-sized halves, with holes down the sides where the tracks had been cut so we could join them up again with wire.  Plus a few more holes so we could stitch the 2 boards together rigidly.  This was feasible as, although the board was big, it didn’t have any components on the centre line and not too many tracks crossing it either.

This added a few hours to the process while the client made the necessary changes, but by Wednesday afternoon we had received new gerbers, come up with a price, been given the go-ahead, and were busily placing the orders for PCBs and parts.  The PCBs were to be manufactured on a 2-day schedule, which is about the fastest possible using conventional techniques.  We have since found ourselves a UK company that offers a 1-day service, but we have yet to test the efficacy of this.

Two days later, and with another night thrown in for shipping, we had the boards, stencils and all the components and were ready to start assembly.  This was Saturday morning, so we had an entire weekend to get the job done.  The longest task is setting up the pick-and-place machine with the components.  It takes as long to do this for 6 boards as it does for 6000, which is why the price for 6 is a lot higher (per board) than for 6000.  And as we had split the boards into two halves, this job was now effectively two jobs; 6 left-hand boards and 6 right-hand boards.

Once the machines are set up, production speed is impressive.  The boards go through the pasting process inside a minute and are loaded onto the pick-and-place machine which then places the parts at around 1 per second.  It then goes down the oven, which has previously been calibrated for these boards, and emerges at the other end in a finished state. It then goes through inspection, and finally – in this case – gets stitched together with its opposite twin to make the single huge PCB that the design originally called for.

By Sunday afternoon we had finished the work, and as couriers don’t tend to make collections on a Sunday, we volunteered ourselves to drive the boards across to the client on Monday morning.  We can’t do this for all our customers, but this one was within a 90 minute drive so not too painful, and we effectively saved another day there.

The customer was very pleased with the service and the finished product.


Lessons, if any, that we might take away from this:

  1.  When you design your PCB, give some thought to the manufacturing process.  In this case it was size of the board that gave problems, although this is a very rare problem to have. Normally it is something more mundane, such as not observing a minimum track width and gap of 0.15mm, or just specifying the wrong component footprint so the part listed on the BOM won’t physically fit on the board.
  2. If your assembly house is prepared to drop everything and throw themselves wholeheartedly into your job, then a 3 working-day service is possible. Realistically this is only if there is a weekend in the middle of it, otherwise 4 working days is about the best that can be achieved.
  3. More complex boards take longer to build.  This one was very complex, but most prototype small-batch jobs are much simpler, and therefore potentially quicker to produce.

View of the ‘stitching’ between left-hand and right-hand boards