Top ten (10) stupid pcb layout mistakes

On Tue, 31 May 2016 05:58:46 -0500, "Tim Williams" Gave us:

No need to see one. It is flawed logic to think otherwise. For proximity violation reasons alone.

Reply to
DecadentLinuxUserNumeroUno
Loading thread data ...

I use multiple sub-sections for gate-like things (and BGAs). The supply sections then go in the corner of the page. I almost always show opamps '-' up. The only exception I can think of at the moment is if it's part of a differential output amplifier.

If you use the same symbols for all of your opamps, it shouldn't be too hard to make sure they're right. Are there any opamps with strange pinouts? All of the SO (8 or 14) parts I've seen are the same. I'm pretty sure the TSSOP we're switching to is the same pinout (same chip - different package).

Reply to
krw

We have separate "power gates" for complex parts, but not for opamps.

The problem wasn't that the decal was wrong, but that people automatically connected +15 to the top pin and -15 to the bottom pin

*after* flipping the part! That error is in plain sight on the schematic.

Some opamps do have strange pinouts, like THS4303, HFA1130, LT1124, TCA0372, a few others.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

Some want us to use single, pinout looking, symbols for quad opamps. Yuck! I refuse. The other guys have been stealing my schematic pages so they can get away with readable schematics, too.

Do you do the same for logic gates? 74xx? 74xx1Gxx?

I've seen that, too (probably have done it). Having a separate power "gate" fixes that problem.

OK, I haven't run into those.

Reply to
krw

The SOT23-5 and SC70 ones from Microchip are infuriating. Maybe they roll a dice or flip a coin each time they do the bonding diagram. Either that or some marketing guy who is sure he can stop people switching to a competitor by finding the one pinout that nobody else uses.

Chris

Reply to
Chris Jones

Funny. I use them (MCP6291/4, mostly) and hadn't noticed that. I guess I've never bothered to look at the competition. Microchip has given us a price that the others can't even come close to. Nice parts, too.

Reply to
krw

I have been unable to justify moving from one Microchip opamp to another, because whilst they offer the same footprint in SOT23-5 (as well as different ones...), the footprints in SC-70 don't include any that are the same between the two opamps, and I had used SC-70 so I would need to re-do the PCB.

Reply to
Chris Jones

Huh, I think most (of my) opamps have separate power gates.... It started with power pads, and now makes a cleaner schematic, is layout "better" if I force the power back in there?

George H.

Reply to
George Herold

I just used a LTC part (LT6238). It's effectively like a standard 14-pin quad opamp but they added 2 pins on the right. It is a TSSOP16 where pin 8 and 9 are unused. That means the two opamps on the upper row have different pin numbers.

--
Reinhardt
Reply to
Reinhardt Behm

The opamps I'm using aren't available in SC70s, or I'd probably have found your problems, too. ;-) I like SC70s. I'm always in a space crunch.

Reply to
krw

You'd just have to re-calculate. The step is a problem but I never had a kilovolt application that went into the gigahertzes. If there ever is then another trick can help, stepping or "ratcheting" the layer down under that trace and widening it accordingly.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

DIP-looking schematic symbols are popular in some quarters, mostly among amateurs. Yuck! is the proper professional term.

We draw logic pretty much like opamps:

formatting link

formatting link

PADS supports hidden Vcc and GND pins, which are a bad idea. They call them "signal pins" for some odd reason.

You can still connect that wrong! Or better yet, forget to invoke that gate section.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

They insist on physical-looking symbols for QFNs and QFP, too. Fortunately they will allow BGAs to be broken up logically. Otherwise they wouldn't fit on a page.

So you have the power pins on one subsection of a mult-gate part? I don't like that because it restricts the layout guy. It's also uglier than hiding "power gates" (and any decoupling) off in the corners.

Agreed. Connection by name (or reference) is a dumb idea. It should be banished from all tools.

I don't know how you get it wrong. There is no need to flip it over. DRC should check for missing gates. I wouldn't even mind if it forced a connection on all pins (open pins would have a specific open designation). Checking is good.

Reply to
krw

The layout person can move any gate section anywhere she likes. My only restriction is that any gate chip must have power and ground.

"Uglier" is a matter of opinion. I like clean corners!

One can always connect the wrong supplies, or forget to invoke that gate section.

PADS doesn't complain about unused gate sections. Our checklist includes stuff like that. We rarely make mistakes like that... any more.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.