# Thumbs rule in electric design

Discussion in 'General Electronics Chat' started by nepdeep, Dec 16, 2012.

1. ### nepdeep Thread Starter Active Member

Sep 14, 2011
140
0
Hi,
When designing an electric circuit there are a lot of things not calcuated but assumed as the thumbs rule, eg. 1nH per mm of wire...or 10 times the value of resistor on the compensating circuit...or sometimes 4.7kohm as pullup or pulldown....etc....
specially from the seniors...with a lot of experiences....can we get get some popular design patterns or ....thumbs rules...

2. ### WBahn Moderator

Mar 31, 2012
23,568
7,214
Let's see, we used to make metal traces 1μm wide per mA of current.

If that seems a bit arcane, then you might want to tighten up your question a bit.

nepdeep likes this.
3. ### Wendy Moderator

Mar 24, 2008
21,428
2,952
One of the most basic is KISS, Keep It Simple Stupid. Many people tend to overcomplicate things in an effort to cover everything.

I think it was Forrest Mimms III that said, "Make it as simple as you can, but no more."

nepdeep likes this.
4. ### Jaguarjoe Active Member

Apr 7, 2010
767
89
A resistor's physical wattage should be 2x or even 3x the calculated wattage.

If a component is too hot to touch, then it is too hot.

Vbe is ~0.65 v

nepdeep likes this.
5. ### Blofeld Active Member

Feb 21, 2010
82
18
KISS is the best of them all. I have always associated the quote with Einstein, but perhaps it is even older.

Here is a collection of 100 rules of thumb:

In addition to John Willison who is named as the author of some of them, some others are from Bob Pease book "Troubleshooting Analog Circuits".

But I guess the really tricky (and most important) part is to know when to break the rules.

strantor, #12 and nepdeep like this.
6. ### bountyhunter Well-Known Member

Sep 7, 2009
2,498
510
On PCB's trace width, I use the motto:

The copper is free

In other words, wider is always better.

For resistors: the power dissipation rating should always be at least 2X maximum power dissipated by the resistor.

For semis: average juntion temp never more than 110C worst case.

Capacitors: derate voltage rating by at least 25%.

Breakdown voltage on diodes and transistors: at least 50% headroom.

nepdeep likes this.
7. ### #12 Expert

Nov 30, 2010
18,076
9,676
Some of mine include:

Never stand behind an electricians elbow.
Never bet your life that the power is off.
(subset) Never bet your life that the disconnect (circuit breaker, etc.) actually disconnected the power.
Never hand a tool to the guy on the ladder, hold it near him and wait for him to (forcibly) take it away from you.
The guy carrying the heaviest object has the right of way.
Everyone that is not doing the dirty part of the job is there to serve the guy that IS doing the dirty work.
You don't have to do it my way, but if you don't, you will find out for yourself why my way works better.
Never allow any kind of lubricant to be on or near the setscrew.
Be sure you know what the consequences are before you ignore the warning label.

nepdeep and gerty like this.
8. ### WBahn Moderator

Mar 31, 2012
23,568
7,214
Never turn on the persistent mode heater on a superconducting magnet unless the power leads are connected.

Always assume that the power leads are disconnected when turning on the persent mode heater on a superconducting magnet.

The first will prevent you from lighting yourself up in a rather spectacular way and the second will prevent you from being what get's lit up in a rather spectacular way when you are unsuccessful at ensuring the first.

9. ### strantor AAC Fanatic!

Oct 3, 2010
4,988
3,069
When drawing your PCB in CAD, make sure that top layer pads for trough-hole components are actually accessible when the part is on the board, before planning the entire PCB around the preconceived notion that they are.

nepdeep and #12 like this.
10. ### WBahn Moderator

Mar 31, 2012
23,568
7,214
Ooohhh. That one sounds like the voice of expeerience.

Here's one from my past. Before you mirror part of a PCB layout (a common thing to do in an IC layout), don't forget that some parts, like IC packages, don't mirror well at assembly time. I got to "fix" that one by dead bugging a bunch of 8-pin DIP ICs (optocouplers, if I recall) and bending the legs down the other way. The customer was impressed that I found a fix and didn't have to spin the board, even after I pointed out to him that he should be even more annoyed that I had to find a fix in the first place.