I'm well on my way with my electronics lab. It's low budget, but functional. I have a nice little assortment of components and can actually go in and build simple circuits with what I have in stock... yay. But, the lab is a LOOOOONG way from complete or well-equipped. I don't have a power supply which makes experimenting really hard. Mostly, I've just used batteries. A Radio Shack 200-in-1 electronics kit was for a long time all I had... I still pull it out once in a while to experiment. Obviously, there are huge limitations to not having an adjustable power supply.

A few days ago, I down loaded Kicad, took a tutorial and began laying out (schematic) an adjustable PSU built around the LM350 and an ATX PSU for the source. I just copied the circuit in the Onsemi datasheet. I built the basic circuit on a breadboard to test, substituting this or that component for what I have in stock. It worked as expected. I then went back to kicad using my actual components and laid the components out in PCB mode, hand routing the traces. I was quite satisfied with the results. I was able to use generic footprints for just about everything.

This past week, I finally received the last of the materials I needed to etch boards, but decided this would be easy enough to do on a protoboard (perf-board) instead. After completing the board, I excitedly plugged it into the computer PSU. Nothing blew. No smoke. Seemed good. I connected the volt meter. 11V. So far so good. I adjusted the variable resistor. Nothing. Voltmeter stayed within a couple tenths of a volt. 20 minutes later I found the problem. From left to right, the printout of my circuit, PCB view, showed (in tiny font) that the pins were number from left to right, 1,3,2. When I wired it, I didn't bother to look at the footprint because I knew how the pins were numbered. But, KiCad apparently did not know. I realize the models can be wrong, but I didn't expect it. Is this a common problem? Luckily, no harm was done this time. I reversed a couple jumpers and now I have an adjustable power supply. But, I'm wondering how often component get fried for similar problems. The LM350 is one of their stock components. Or, is there something I didn't do, that I was supposed to?

 

Always, always, always.... check the footprint matches the physical part. I've even seen transistors go awry and who would expect that?

Well, *I* would now that I've seen it happen, so I always always always check the pin numbers. And when possible I print the gerbers 1 to 1 and put the parts on it to check sizes.

(Did I mention "always?")

 

So, it wasn't something I did. I thought open source was supposed to mean someone in the community would have fixed it before now. What's the easiest way to fix pins that have been mis-numbered?

 

No, it was something that you ought to have done!!!

One of the problems that arises from the fact that a schematic is a representation of the circuit, is just that, It should be designed so that it is easy to read and understand.

It is not meant to represent a circuit layout, and pins are often changed such that their position, makes the schematic easier to read.

Whilst this is most often done where IC's have many pins, and is 'obvious' , it is not always the case that it will be obvious.

The LM350 is invariably shown with the adjust pin in the middle on schematics whereas the pin out shows the adjust pin to be pin1. (shows how easy it is to make mistakes.....)

A lesson everybody learns.....

I don't use Kicad, I use gEDA so I can't help as to exactly how to edit your symbols. but it shouldn't be too difficult.

 

What I said in the OP was:

From left to right, the printout of my circuit, PCB view, showed (in tiny font) that the pins were number from left to right, 1,3,2.

I also said, I had the circuit working on the bread board after drawing the schematic.