EEVblog2: When you REALLY don't want to pay for a Double Sided PCB

[tom66]

I can understand, why the designer goes for a single-layer board. But could anybody explain to me, why did they go for four different models of a 0Ω link?

I have seen this happen - though not quite on such a bad scale as this - when designs are merged and copied across from other projects.  We ended up with a board with 0201, 0402 and 0603 decoupling caps.  All were eventually rationalised to 0201 common parts.  It can also happen when an engineer manually enters in part details rather than using a part management system, or mistakenly selects multiple types of the same value, and then they end up with different parts on the BOM. It should be caught in the BOM review stage, but if they mark them differently "JUMPER 1206", "0 OHM 1206", "0 OHM RES 3216M", "0 OHM JUMPER 5% TOL" etc. etc.  And if there's pressure to get a product out in time for e.g. Xmas sales period, then things get missed like this.

Occasionally we see products like this - not designed by a competent engineer with unexpected priorities or outside influences, but instead, designed by someone who simply doesn't know any better. The designer in this particular case might just be someone who doesn't know what a multi layer PCB even is.

Or, they were told by management it 'had to be single layer' to save costs because that's how they've always done it, and in some cases, that is an appropriate way to control costs. But they didn't appreciate their field enough to go, "hang on, this doesn't make any sense."   Or they didn't feel senior enough to speak up.  Happens a lot!  We practice a policy where I work where every decision by the most senior member of the team can be critiqued, it helps avoids bad decisions like this.

[Kleinstein]

Having the PCB single sided makes sense. What does not make much sense is having 2 seprate PCBs directly side by side and having so many 0 ohm jumpers. Even without alloing for pin swaps and related software adjustments it should not take much more than 1 jumper per LED+switch.
Even if done in steps to replace older versions or include other variants it makes little sense - there is not much difference in designing the slightly larger (if needed at all) to include the odd adapert PCB that is in between.

For the number of jumpers it looks like a true beginner doing a bad job or bored layout person trying to get the most jumpers on a PCB. Handling the jumpers in the PCB software may not be the most efficient way - it may not be high priority in the software.

[rsjsouza]

I can understand, why the designer goes for a single-layer board. But could anybody explain to me, why did they go for four different models of a 0Ω link?

In one of my last designs I had not much board space and needed to use different sizes of zero ohm resistors whenever I needed to jumper over many tracks at once. Having several 0603 (one per jumpered track) was very space consuming, since a 1206 can jump up to four 0.2mm tracks with 0.2mm track-to-track and track-to-pad clearance (I couldn't use thinner tracks on my design).

[Eagcress88]

You can find similar constructions in hifi equipment from early 2000's. I had for example a mini disk player and the complete control board consisting of 3 or 4 large TQFP's, some smaller chips,...all on a single layer board with 0R SMD jumpers on the bottom side and wire links as regular trough hole. The AD/DA board and the remaining control interface was also assembled onto FR-1 board with a mixture of SMD and trough hole components.


Let's say you have a piece of equipment with one or two large boards and e few smaller ones. That's what you typically see in consumer equipment. A small PCB for the power switch, on other one for the incoming mains and fuses, an other one for the front panel controls,...

If you do this with FR4, you need equipment to cut the panel into small pieces unless you used milling to separate the boards. With FR1 you can just snap the boards into pieces by hand and the required cut outs are punched. In TV's and the like you often find pieces of FR1 used as spacers, cable retainers,... all broken of from the same panel.

I can imagine wave soldering being faster than reflow soldering and you don't need to maintain stencils and control a paste printer which will act up sooner or later. But I have no idea about the number of solder defects for reflow versus wave soldering SMD components.

It seems that a single-layer board might be chosen to cut costs or space, but using four different models for 0 ohm communication is a bit surprising. Also, for those who are interested in reliable deposit methods or want to make some money to buy supplies for various research. Pay attention to resources like https://icasinoreviews.co.nz/neosurf-casino/ this guide, which offer great insight into payment or deposit options. I always check payment methods through this resource before purchasing. Also when a connection may need different resistance depending on the design or even the need for flexibility during testing or future modifications.

But maybe the truth is very different. Those unknown to us Japanese CAD systems with libraries with DIL packages which for the PCB as well as the schematic can't do double sided boards yet They are working on that.

I agree that FR1 was super practical for mass production! Snapping boards by hand saved time and cost in consumer electronics back then!

[5U4GB]

Neither sticking to FR-2 boards, nor trying to avoid the cost of multi-layer board explains a design, which requires four different rolls of 0 Ω links in the pick-and-place machine. This would increase the cost.

Well, they went with 0 ohms where possible but had to use a few 000's for higher precision.