Using front and back copper for wide power traces... questions

[AmmoMFG]

So I’m trying to make a board that will support Andersen Power Poles, and will draw up to 20-25A, it’s a really simple layout and I’m doing the traces by hand, however I want to use the front and the back to double my copper, without increasing the trace width. However when I lay out the traces, I do the front, then when I start editing the back, the front traces start dropping connections. I’m sure there’s some setting that controls this, I messed around with the “mirror traces” but that doesn’t seem to fix the problem.

Thoughts? Ideas?

[1.21Gigawatts]

On the left hand toolbar, about halfway down, there is an icon with a pair of tracks with an ‘X’ through one of them. Toggle it to enable/disable “Auto delete old track”.

[AmmoMFG]

Excellent, I knew it was something silly I was missing. Thank you much for your help!

Here’s the finished product… dunno if anyone else out there uses Andersen Power Poles for running radio equipment etc. But I designed this as a rig-runner substitute for my group of Amateur Radio friends for christmas. Going to throw it up on OSHPARK as a shared project.

Thanks again for all your help!

[dchisholm]

@AmmoMFG

Thoughts? Ideas?

In a situation like this I prefer to increase the current-handling capability of the traces by turning them into Fill Zones (a.k.a. “copper pours”). By doing so, I can make it possible for the current to distribute itself over the maximum area available with a given placement of components. Of course, before filling the copper zone I establish the basic electrical conductivity by laying down traces with some nominal width, such as 20 mils.

I can also control how the pads are incorporated into the larger current-carrying conductor: Thermal-relief spokes are easier to solder by manual methods, while solid connections are easily handled by wave-soldering processes.

Finally, the large copper areas - with their “industrial looking” closely-spaced edges, squared-off corners, etc - give me a visual reminder that this section of the board performs a special task.

With copper pours you can lay out the filled zone on one board layer then replicate it to the opposite side with just a few clicks.

Dale

[AmmoMFG]

I’ve done copper pours in the past with mixed results. One thing I am really careful to avoid, especially when it comes to a board with a high amp feed, is getting the mounting stand-offs anywhere near the pours. I also really wanted the copper to be in a specific place. This board is set up to do a 2oz pour, with .375" traces double sided. IIRC the calculated minimum was .25" 1-sided for a 5A stable draw, with a pulse of up to 20A (most 100W HF rigs draw about 15-20A while TXing, but I wanted to stay under 20A).

I have some other projects coming up, I’ll give pours another shot.

Also, here’s the BOM:

6x APP Red 25A housings (small)
6x APP Black 25A housings (small)

Mouser Part Numbers:
6x 534-3522-2
6x 576-0ATO020.MXGLO (or similar, 20A automotive fuse)
12x 879-1377G13-BK

Optional: (but recommended)

Diodes:
6x 750-1N5402-G

[davidsrsb]

I corrected the topic title spelling

[dchisholm]

@AmmoMFG

I’ve done copper pours in the past with mixed results . . . [/quote]
What problems have you encountered? A copper pour adds capacitive loading to (physically) adjacent circuitry . . . . which may be either undesirable, or beneficial, depending on the situation. Analyzing that kind of detail is probably outside the scope of this Forum.

I will confess that it took me a few tries to understand KiCAD’s Priority system for copper pours.

. . . . One thing I am really careful to avoid, especially when it comes to a board with a high amp feed, is getting the mounting stand-offs anywhere near the pours . . . .

Yeah, my library of Mounting Holes includes four different styles: some that provide copper-free clearance for a screw head, some with clearance for a standard washer, and some with generous clearance for an over-sized washer (or standoff post). And there are cases where you deliberately place copper under the fastener.

. . . . This board is set up to do a 2oz pour, with .375" traces double sided. IIRC the calculated minimum was .25" 1-sided for a 5A stable draw , . . .

There are several formulas, calculators, and rules-of-thumb for determining trace width. (KiCAD even includes one in the “PCB Calculator” application.) Whether assumed, or explicitly entered into the calculation, the conductor’s temperature rise is the controlling parameter in the calculation. Most of us judge a conductor to be “too hot” well before it reaches a temperature that threatens personal safety or equipment damage so don’t be afraid to add liberal margins to the values you get from a calculator.

This topic is a good example to illustrate the notion that a LOT of engineering comes down to evaluating tradeoffs. Many engineers are surprised that most trace-width calculators allow about an amp of current in a 10-mil trace on 1-ounce copper. Of course, the calculation is done solely from the viewpoint of the trace and board. That 10-mil trace, carrying an amp of current across several inches of board, may have unacceptably high voltage drop, or introduce too much resistance into the power distribution design. As they say, “For every solution, there is a problem - and that helps to keep engineers employed!”.

[quote] . . . . Optional: (but recommended)
Diodes:6x 750-1N5402-G

The seat of my britches says this part is marginal for an application that could pull 10 - 20 amps. (The 1N540x series is rated for three amps, as I recall.)

I think the diodes serve two purposes in this design. They swallow any inductive kick-back that may happen when equipment is turned off, or switched between transmit and receive. Since these transients are quite short, the current spike is probably within the part’s transient rating.

The second reason I see for the diodes is to add a bit of idiot-resistance to the power distribution system. If the power supply is accidentally connected with the polarity reversed, the diode will conduct and blow the fuse before expensive equipment is damaged. That could mean the diode must conduct 10 - 20 amps (enough to blow the fuse) for a second or more. That’s asking a lot from a diode rated for 3 amps - the diode may destroy itself fast enough to protect any fuse you or I can afford to buy!

I think several companies - possibly Diodes Inc, and Vishay - offer big, slow, lossy, and cheap rectifiers in leaded “DO-xxx style” packages, rated for 10 - 15 amps. Diodes packaged in variants of TO-220 packages go to 20 - 30 amps or so, but require a re-work of your layout.

(I have often used similar circuits in equipment where a user might connect a battery backwards.

Observation 1. The circuit is idiot-resistant but not fool-proof: we fools are just too ingenious! Did you know that fresh AA alkaline batteries can source enough current to melt the solder joints securing the polarity-protection diode? Then the diode falls off the board . . . meanwhile, the user notices the equipment doesn’t work . . . replaces the fuse . . . and inserts the batteries - backwards - AGAIN!

Observation 2. When the development team is rushing to meet some arbitrary, and unrealistic, deadline somebody invariably connects the power supply backwards. In the weekly staff meeting you refer to the incident as an “unscheduled development test” - all second-level supervisors will think you’re doing a good job dealing with previously overlooked problem areas. The guy who connected the power supply backwards is obligated to provide the other team members with a serving of their favorite malt beverage, to ensure that they will not reveal the true story to anybody in management. )

Dale

[AmmoMFG]

Dchisholm,

The specific issues I’ve encountered are more or less as you state… there seem to be some ambiguity in terms of the pour area (I’m still a relatively new user, and haven’t had a lot of need for cutting really fat traces on the board.) With this project specifically, since it’s all through-hole, I wanted to use a top/bottom trace configuration just for simplicity’s sake, and then all my power lines are where I know they are.

At the moment, I’m kinda using my own mechanical board outlines I created in solidworks. I’ve open-sourced them as DXF files to make them easy to import… https://github.com/AmmoMFG/PCB-Board-Outlines

Typically all of the board holes are copper-free. I’m slowly re-doing all of my board layouts as kicad parts, that have both edge-cuts layers, as well as electrical and copper layers on them. However, this is a very new endeavor. For the moment, I’m more focused on cleaning up my footprints libraries I’m slowly creating as I need new bits…

https://github.com/AmmoMFG/KICAD_Parts (I always appreciate feedback)

You are correct on the diodes… they were absolutely a compromise in that they have a 3-4A max under a constant draw condition, however they have a max 40A, it’s entirely possible it will smoke-check the diode in the process of blowing the fuse, however that’s going to be a rather unusual case as the APP’s are biased to prevent this, but I added them more as an “Oh S—” feature, with the intention of causing the fuse to blow. 20A is maximal fuse rating, I would hope people would be more conservative with their fuse choices, and not run 20A everywhere, including over thin 24ga power wires for recharging an HT or some other use where the wire-gauge at that amperage would be more descriptive of an exploding bridgewire rather than a useful electrical device.

I’ve not sure I’ve ever put batteries in backwards in a way that mattered… I remember I accidentally smokechecked a brand new 10M mobile, it did indeed blow the circuit protection diode before the fuse went, however that was enough to get me to stop and take a hard look at wtf just happened. Major consideration here was keeping the costs low, the .30c for that diode makes sure it’s included in the package, where a $1-2 diode might be excluded. This was kind of the problem I ran into. IIRC the rig-runners don’t even use a reverse bias diode, so I consider this a slight win.

I also spec’ed a fast-blow type fuse. As this was totally among the considerations, but alas compromise rears it’s ugly head.

Generally speaking when it comes to stuff I do professionally, write down everything ahead of time, and then while you’re doing it you read it again. I’ve seen too many operations blow things up and hurt people when they try to hurry up. This is a major reason why I’m trying to push more into using standard connectors for everything. This was a major problem for me when I first got into the e-cad work (normally do mech) a #10-32 machine screw comes in some variations, but other than that it’s always the same. Eagle ended up being quite frustrating in terms of picking parts to use. KICAD, in it’s wisdom divorced itself from this way of doing things, and speeds up my development cycles quite a bit. Still learning though, but that’s what life is all about.