Browse over 10,000 Electronics Projects

How Frequency Counter Works? And Build a Nanocounter using an FPGA, STM32 and a Bluetooth Android App

How Frequency Counter Works? And Build a Nanocounter using an FPGA, STM32 and a Bluetooth Androi ...

Manufacturing the board

Elecrow is my normal go-to provider for two layer manufacturing because their prices are so good. Four layer is more pricey though and anything above 5x5cm sees a large rise in the manufacturing cost. This board is a 10x10cm layout and at the time of writing PCBWay was the best price available.

I’d never used PCBWay before and a quick google around revealed that the online reviews cannot be trusted because PCBWay will pay reviewers for positive reviews posted on internet forums and websites. I really wish they’d stop this because if the service is good then the reviews will come naturally. I’m going to relay my honest experience here with the hope that you can trust me not to make stuff up just to get a few dollars off my next order.

Anyway, I picked the yellow soldermask because this PCB will be exposed and I’d like it to look more distinctive than the standard green. Because I’m going to be doing controlled impedance traces I needed to know the PCB dielectric constant and the thickness of the PCB copper and FR4 substrates. My emails to PCBWay were answered within 24 hours by Lynne and I got the answers I needed. The dielectric constant is 4.29 and the upper and lower layers are 18µm copper (note that since this was written they’ve now changed to 35µm).

I selected the China post option which means the standard 2-4 weeks went by before the boards arrived.


The boards look great to me. What’s really important is that they’ve managed to get complete and unbroken soldermask slivers between the tiny pads of the QFN.

The QFN is by far the most annoying and difficult part to reflow and frankly you need all the help you can get. Having soldermask there to prevent bridges will play a big part in my chances of a successful reflow.



Advertisement1


If you look at the upper and lower right corners of the QFN you’ll see that I’ve used side-exit traces from the pads. Be careful when you do this because QFNs with a ground pad have that pad held in place by very small metal spurs that are exposed from the package at the end of the row of pins. If you don’t have solder mask covering your traces then there’s a small but real risk that you could accidentally ground a trace.

The drill positioning for the vias also looks great for a prototype service and the soldermask apertures appear to be well positioned with good tolerance. The silkscreen has the familiar dot-matrix low resolution style that you get with prototype services.

I’ve heard that the prototype services use different board houses for four layer manufacturing and that may account for the better quality results and higher price to go with it.

Here’s a shot of the back and front layer stack tell tales. I can see that they’ve got the stack correctly ordered so my controlled impedance traces will be correctly referenced to the ground plane.

Here’s my attempt at a spark gap. That’s an 8 mil gap between the two pointy ends. Do they work? I don’t know for sure but they’re free so I may as well include them. If you’ve never seen one before and don’t know what they do then there’s an entertaining video over at the eevblog that’ll tell you all that you need to know and plenty that you didn’t expect!

 

NEXT: 

Building the board

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

 


Top