Need good resource for PCB traces

[aalbri]

Hey guys.

What are your favorite resources for determining PCB trace width, distances, pad sizes, etc? I am new to designing PCBs and want to learn the 'standards' or at least the right way to do things.

[kizzap]

Proper way:



-kizzap

[Falcon69]

I like using this guys site.

http://www.mcuexamples.com/PCB-Trace-width-calculator.php

[Neilm]

I use a programme called MiscEL, available at http://www.miscel.dk/MiscEl/miscel.html

It not only has PCB trace calculators but many other useful bits (simple circuits, heat calculators, serial string decoder and much more)

[free_electron]

IPC2221. that is the gold standard

[John steave]

I like the pcb trace site for better understand and performance. So far as i know it's good for all other pcb manufacturing goods. http://actpcb.com/Printed-Circuit-Board-PCB-Manufacturing-Capabilities
I’m not recommending but perhaps you could check out this area. 

[WarSim]

IPC2221. that is the gold standard

Many countries issue their own design standards.  The IPC series is used in North America.  Only one thing to add.  There are several additional IPC standards others than IPC2221.  IPC2221 is the general standard and enough for most designers, just keep in mind if you do something special check for additional standards.  For example medical has additional requirements.

[free_electron]

IPC is used all over the world. They do all the testing and write the guidelines.
There may be localized criteria for flammability and electrical safety and other things.
But for the base electrical and manufacturability and assembly of the design the IPC rules are the de facto standard.

there are many factors to take into account.
- is the trace on an outer layer or inner layer.
- is this board going to be used at high altitude ( lower ambient pressure )

[ignator]

Trace width calculators work when the current has an absolute maximum known level.  If the trace is used for power, you need to consider fault current under short circuit faults.
Fast blow (mechanical thermal) circuit breakers, at 1000% over current take nearly 1 second to trip. So your PWB traces need to support this fault current for the CB trip duration without releasing their magic smoke.   If your board is multilayer with a ground plane, the FR4 material (or G10 etc.) will become carbon, which is no longer an insulator, and my experience has been the CB does not trip, and the board glows orange where the copper trace left connected is now burning, and the trace burns like a fire cracker fuse until white ash is all that is left.  If this fuse runs across other power traces in the multilayer board, they start burning also.  And the CBs never trip.
Also consider traces for charged large micro farad capacitors, as these will source welding currents.

[Frank3point14]

Well there is a question that is...HOW many layers????Is the trace is on outer layer....

By the you can visit here and check that if its HELP!!!
http://actpcb.com/ 

[moemoe]

http://frontdoor.biz/PCBportal/HowTo2152.pdf is a good resource to start, they also have a nice excel sheet linked in on the last side.

One problem with this document is that it's only one-way, so getting the needed width for a given current is no problem.

I just tried to reverse the direction, give maximum current for given deltaT and I, but it's not as easy as I hoped at first.

[digsys]

Formulas are nice, but if you REALLY need verification of the batch you just received - an old trick -
Either on a spare part of the PCB, or outside edge, make up 50/100thou or whatever track lengths, and place large pads at each end.
I run some with resist and without. Then on delivery, run 1-10A etc and get an actual value. It confirms they were made correctly.

[codeboy2k]

Formulas are nice, but if you REALLY need verification of the batch you just received - an old trick -
Either on a spare part of the PCB, or outside edge, make up 50/100thou or whatever track lengths, and place large pads at each end.
I run some with resist and without. Then on delivery, run 1-10A etc and get an actual value. It confirms they were made correctly.

Good test.  I'm going to add that to my batch of tricks.