Kuba Fuel PSU Adapter - back in stock

micrex22

New member
Jan 8, 2020
7
0
1
"The adapter has exposed traces (for heat dissipation reasons) - be sure not to have any exposed wires touch the boards."

I'm surprised nobody brought up the fact that you have live exposed **power*** traces haha. Surely a better adapter could have been built with a wiring harness instead...
 

Elf

Storybook
Feb 4, 2019
273
63
28
It is true though that PCB traces do heat up proportional to the current through them, and that high power traces are often left without solder mask to be covered in solder to enhance their current capacity.

There is even a nice calculator for such things in this toolkit (which I use regularly): http://saturnpcb.com/pcb_toolkit/ :)

The gist of the calculator is that the current capacity of a trace basically just comes down to how hot you want it to get and what its width (or in this case, cross sectional area) is. The more, the better, of course.

That said, if there is a danger of shorting, provisions should be made accordingly.
 

kubatyszko

New member
Jan 29, 2020
4
6
3
All,

Heat dissipation is a valid concern, as Elf says, and it's always a compromise when designing a PCB.
The PCB has VERY limited size due to small space available between CPU board RAM slots and the metal case.
Add the VERY high currents required by the Fuel, and you end up with traces that need to be really wide.
Simple calculation on a "standard" 2oz pcb at 40 amps, means traces that are INCH wide - good luck fitting that on a roughly 4x4 PCB.
AND that is for EXTERNAL trace IN THE AIR - fully exposed, with 10 degrees C temperature rise (this means that the traces are "allowed" to heat up by that much).
Trace covered up by a mask basically becomes internal, and would take us to 2.5 inches width for a SINGLE 40A voltage rail.
Hence, exposing traces allows me to reduce the width requirements by nearly 2/3, and using thicker copper (but the cost GOES UP a lot) allows even further reductions at the compromise of assembly and part selection (copper etches in trapezoidal fashion, thick copper means that the pad distance requirements are higher and higher, going from 2oz to 6oz I had to use bigger IC chip already to allow for it).

To manufacture board that allows so much current and is able to keep the traces masked, I'd have to go to 12oz or thicker copper or more (6oz vs12oz is a 3x in bare PCB cost). That would drive the cost of the product to probably ~$200 a piece, and the goal was always to keep these adapters affordable yet useful. If you'd like to place an order for boards that are specifically "short-proof", I'd happily make some as a special exception but it is going to get expensive!.

In the 2011-2012 I made a few harness-based prototypes and there's nothing wrong with that, other than the fact it's very time consuming.

The reasons current adapters are PCB based are:
*) Assembly time
*) Repeatability of the process (much less stress for me when I have order for 25 boards)
*) Additional features such as LED's, FAN signal generation, KICKSTART button etc.

As for danger of shorting, I do not feel there is much risk, the Board is vertically mounted at nearly the top of case, and with nearly 100 adapters sold, not even once I heard of any issues of shorting. You have much higher chance of striking a sensitive component on the motherboard or PIMM...

Like massiverobot said, the adapters are in stock to whoever wants to buy them.
I have 3 pieces ready for shipping, and assembling about 40 (forty) more.

Thanks for supporting my effort.
 

About us

  • Silicon Graphics User Group (SGUG) is a community for users, developers, and admirers of Silicon Graphics (SGI) products. We aim to be a friendly hobbyist community for discussing all aspects of SGIs, including use, software development, the IRIX Operating System, and troubleshooting, as well as facilitating hardware exchange.

User Menu