PCB Trace Width

[PRR]

25mil 0.6mm is the width of US #23 wire. I can't see a capacitance problem here. If you are down in the pico-zone, a condenser mike head, you want wide SPACE more than skinny conductor.

For pedal-size work the current-carrying will be ample for anything you can see naked-eye.

As with much small wiring, you pick the size for mechanical robustness. If you wire a pedal with hair-fine #44 wire it will break on the road, or even under the builder's fingers. I'm cursing a house full of lowest-price #28 telephone wire, and replacing with #23. While the PCB board supports traces, fatter is better support.

My actual trace-width is "Sharpie", because I have not laid-out a board since these online services sprouted, and never enough to invest in photo-tooling.

[GGBB]

Other than making the laying out process more difficult none that I know of.

I've never had any problems at 16mil .016in. Maybe give it a try before you dismiss it for no other reason.

[alexisdroso]

@GGBB Normally, I'm using 0.35mm for audio signals which is around 14mils and 25 mils for power signals. I never go below that, even if the manufacturer sets the bottom limit to 5mils. My point was that there is a huge difference between 14 mils and the suggested 25 mils. I was focused on the amp/voltage relationship and not the mechanichal side of it. As I understand the wider traces are suggested more for safety/mechanical reasons rather than electrical reasons. Lesson learned 

[alexisdroso]

25mil 0.6mm is the width of US #23 wire. I can't see a capacitance problem here. If you are down in the pico-zone, a condenser mike head, you want wide SPACE more than skinny conductor.

For pedal-size work the current-carrying will be ample for anything you can see naked-eye.

As with much small wiring, you pick the size for mechanical robustness. If you wire a pedal with hair-fine #44 wire it will break on the road, or even under the builder's fingers. I'm cursing a house full of lowest-price #28 telephone wire, and replacing with #23. While the PCB board supports traces, fatter is better support.

My actual trace-width is "Sharpie", because I have not laid-out a board since these online services sprouted, and never enough to invest in photo-tooling.

22AWG I believe you meant to say. It never occured to me that a few extra mm would play such a significant role on the robustness of the pcb. Thank you for your very clear explanation.

[Rob Strand]

As I understand the wider traces are suggested more for safety/mechanical reasons rather than electrical reasons. Lesson learned

It's a matter of not creating trouble for yourself down the track.   Thin traces and small clearances do make it easier to layout the board.   However if you always take the easy way out you *might* get unnecessary trouble one day.  For example you will occasionally get PCBs with shorts even when the PCB isn't at the PCB manufacturers limit - mistakes are made.  If you use small clearances there's a risk of shorts for example by solder wisps (hairs) - widely spaced tracks minimize this.

The limit of how small you can go is set by the process tolerances of your manufacturer.
If you have 7 to 8mil clearances some PCB manufactures might charge you more.

If you used 10 mil traces and 10 mil clearances you would be fairly safe provided the boards have solder masks and the prices would still remain at the lowest level.

So somewhere between 10mil and 20mil!

[merlinb]

Urg I do wish people would not say 'mil'. 'Mil' is short-speak for millimetres! No wonder the Mars orbiter broke.
'Thou' means thousandths of a inch. I have spoken.

[Rob Strand]

Urg I do wish people would not say 'mil'. 'Mil' is short-speak for millimetres! No wonder the Mars orbiter broke.
'Thou' means thousandths of a inch. I have spoken.

Yeah it's a pain.  It's from using PCB packages back in the days when everything defaulted to Mil.
The only time I use Mil is with PCBs otherwise I'm hardcore SI units.

[GGBB]

Urg I do wish people would not say 'mil'. 'Mil' is short-speak for millimetres! No wonder the Mars orbiter broke.
'Thou' means thousandths of a inch. I have spoken.

Err - you have misspoken? 

Meaning depends on where you come from (and maybe how old you are). https://en.wikipedia.org/wiki/Thousandth_of_an_inch

Also - we're not speaking here we're writing, so it should be "mm" or "in" really (I scored 50%!).

More:
https://www.dictionary.com/browse/mil
https://www.merriam-webster.com/dictionary/mil
https://www.lexico.com/en/definition/mil

The last one (lexico/oxford) notes that 'mil' for millimeters is informal just as it is when used to mean mother-in-law.

[GGBB]

@GGBB Normally, I'm using 0.35mm for audio signals which is around 14mils and 25 mils for power signals. I never go below that, even if the manufacturer sets the bottom limit to 5mils. My point was that there is a huge difference between 14 mils and the suggested 25 mils. I was focused on the amp/voltage relationship and not the mechanichal side of it. As I understand the wider traces are suggested more for safety/mechanical reasons rather than electrical reasons. Lesson learned 

I failed the math test yesterday (or at least the calculator test) so I was misthinking the RG-suggested .635 mil (by which I mean 635pm -picometers for those not following closely) was .016 mil by which I mean .016in.

(is that better Merlin? )

But now I realize it's .025in and yeah - I agree that would make layouts more challenging. I also though you were planning to use .254mm which is .010in (which I still think is smaller than you really need to go). .014in is a fine choice IMO.

[mimmotronics]

It is tempting to go for a shorter width to save on space, but if you have the real estate you might as well use it!

[PRR]

> No wonder the Mars orbiter broke.

The land deed for my other house traces back to metes expressed in chains and links.

Thank goodness not in rods! Traditionally the length of a rod was different in every town.

I swear the studs in that house are spaced 1-1/3 cubits.

In the US, mil=0.001" persisted until even more recently than routine use of cubits, rods, or links. And yes electronics is very wedded to customary UK/US units!!

And we've never said "mil" for mm, cuz we never mm; or when we must we say "MM".

I grant that in an international forum, such casual use of local dialect could confuse.

[rockola]

And we've never said "mil" for mm, cuz we never mm; or when we must we say "MM".

Which would mean megamolar, quite a high concentration indeed.

[LiLFX]

PCB fabricator checking in....

Someone commented that mil is short for millimeter. In the PCB world (and mechanical engineering for that matter) that is false. Mil is one thousandth of an inch. Fractional millimeters are discussed in terms of microns. Anyway...

If you really want to figure out the exact minimum trace width you can get away with then download the Saturn PCB Tool Kit. However, keep in mind that the cheap Chinese board shops typically only advertise that they can do a 6mil (.15mm) trace on 1oz copper. The reality is that you likely shouldn't need to go that fine anyway. A 10mil (.254mm) trace on 1oz copper would be perfectly fine in a pedal. You might be tempted to start packing things tighter when using thinner traces, but you need to tread lightly. Imagine an electromagnetic field radiating off of every trace. Will that field hit a neighboring trace before it hits a power or ground plane? If so, you have a potential crosstalk/noise issue. Every single time you route a trace you need to think about the field created by that trace. The fields spread out within the dielectric space of the board until they find a return path. That return path better be power or ground otherwise you could have a problem. In high gain circuits the problem usually manifests as oscillation. In digital circuits, you might have a clock show up in the audio.

[marcelomd]

As an engineer, I'd like to take a moment to express the frustration (directed not at at any of you, but society in general) I feel whenever I see an invalid symbol for a unit.

'mts' for meters in road signs and 'lt'  for liter in the super market come to mind. Uppercase letters when the symbol is lowercase too.

'MM', for millions, is used a lot in financial reports. MM is a million times a million... you'd think they'd care more about accuracy with money (Also, megadollars and gigadollars sound cooler).

Things like this are everywhere. I'm sure lay people don't care (why would they?) Or notice. I'm also sure I am constantly butchering other's jargons all the time.

I also want to thank my physics teacher in engineering school for drilling the importance of correct use of units into our skulls.

Sorry for the rant. =*

[Rob Strand]

Someone commented that mil is short for millimeter. In the PCB world (and mechanical engineering for that matter) that is false. Mil is one thousandth of an inch. Fractional millimeters are discussed in terms of microns.

For written units that's true.  In au we are essentially metric (but were imperial until the 70's).  The spoken word "mil" would mean millimeters and the spoken word 'thou' for  one thousandth of an inch.   That's true even for older people (say 60's to 70's).   However in the PCB industry the spoken word 'mil' would be a little ambiguous unless the conversation (or paper work under discussion) pinned it down.   I have seen plenty of PCB spec'd with Mil dimensions and mm holes!  Often you know the aspect you are talking about (holes vs tracks) and from the size 0.7 would be 0.7mm but 50 would be 50 Mil.  Crazy shit from a scientist or engineers point of view.   Mechanical drawings here are pretty much mm. 

[PRR]

> 'MM', for millions, is used a lot in financial reports. MM is a million times a million...

Roman Numerals.... M is 1,000.

Joseph The Carpenter paid his taxes in Roman Numerals.(*)

However MM in Roman Numerals is clearly 2,000. This system did not use a concept of multiplication.

The Arabic notation, with implicit multiplication, became a fad. So in respect of two powerful number notations, "MM" could be read as "1,000,000". Not good usage; but if everybody in one field is using MM then I guess they know what they want it to mean.

(*) While official counts in the Roman Empire used Roman Numerals, some say that mechanics of the day who actually *figured* (not just counted) tended to know the Babylonian or Egyptian numbers, where figuring is easier (not easy).

[Rob Strand]

Not good usage; but if everybody in one field is using MM then I guess they know what they want it to mean.

There was an era where uuF (mu-mu F) was used for picofarads.   It very often got misinterpreted as uF on schematics and the circuits never worked.  Of course experienced people could see the error.