Organize my previous QCWDRSSTC

[曹靖]

I've been working on some PCB boards lately https://highvoltageforum.net/index.php?topic=2434.msg18070#msg18070 The machines in this post are all hand carved and made of some perforated boards. I have taken some time these days to draw a part using Pcb software. I hope to have the opportunity to resurrect it in the future                Here is a small portion of completed boards. I'm not sure if there are any areas that are not properly done. If there are any, please criticize and correct them                                                                                                                                                                                                 

[Unrealeous]

They look very nice.

[Edit: This is not correct - see below]
My only suggestion is where you have used many tiny via's (aka via-stitching) to allow high current, another option is to increase the size of the via's.  My understanding is that fewer bigger ones work better than many smaller ones.

Another option is you get the boards fabricated, you can then use the 'untented via' covering option, and you can then fill your via holes with solder - which also greatly helps with current between the layers.

Another fabrication option if you have the cash, using 2oz copper outer copper weight over default 1oz likely to help...

[Anders Mikkelsen]

They look very nice.

My only suggestion is where you have used many tiny via's (aka via-stitching) to allow high current, another option is to increase the size of the via's.  My understanding is that fewer bigger ones work better than many smaller ones.

Do you have a source for this? A via cuts a hole in the copper plane, while adding copper plated on the inside of the hole. The lost copper on the surface goes with the square of the diameter of the hole, while the added copper plated on the inside of the hole goes linearly with the diameter since it covers the circumference of the hole. From this it should follow that smaller holes are better.

2 oz copper is definitely a nice upgrade, if needed. Figuring out how much current your planes can handle is not trivial, and I would recommend testing it if you want to know. Having everything connected by short and wide polygons makes a huge difference. My process for making high current boards starts with component placement in order to facilitate short connections. Then adding enough layers to make all the required interconnections, then upsizing the copper thickness if required. In some cases, adding more layers can be cheaper than using thicker copper. 2 oz is supported by most board houses, at least on outer layers, but copper thicknesses beyond this tend to be more exotic and expensive.

[Unrealeous]

Do you have a source for this? A via cuts a hole in the copper plane, while adding copper plated on the inside of the hole. The lost copper on the surface goes with the square of the diameter of the hole, while the added copper plated on the inside of the hole goes linearly with the diameter since it covers the circumference of the hole. From this it should follow that smaller holes are better.

Ah, you are correct.  And now I realise I've had made my via stiching assumption based on a single via and so what I said above is not correct.   I have been doing boards with via-stiching using untented vias filled with solder.

I believe the following to be true:

-An individual small tented via will conduct less than a individual large tented via

-An individual small solder-filled untented via will conduct less than a individual large solder-filled untented via

So bigger is better right?  Not so fast....

-A fixed area with a tight grid of small tented vias will conduct more than using large vias (for the reasons you describe above).

-A fixed area with a tight grid of large untented vias filled with solder will conduct the same as small untented vias filled with solder. 

16 circles of diameter 1 = 4pi units^{^2}
4 circles of diameter  2 = 4pi units^{^2}

[曹靖]

Today caused a phase compensation drive board