Conversation with a new circuit designer

I put voltage/gnd symbols any which way. Since the company requires that schematic symbols match the physical package (except for SSI gates/opamps and BGAs, go figure) the pins come out every which way. ...and so do the voltage symbols.

Everyone has a different way of doing it. I can see the logic of doing it that way, kinda, but if there's a clearly logical way of grouping the pins, that's how I do it.

Where I used to work, the Portland office was big on making schematics easy to read. We had another engineering group out of state which was big on making schematics fast. Manufacturing hated them, and loved us. At one point one of our Portland guys was looking at one of the other guys' schematics. The conversation went something like:

PDX: "It looks like you just dropped parts onto the page and connected them willy-nilly".

other guy: "You mean there's another way?"

That's not an engineering schematic, it's a wiring diagram, and about the level you should stay with.

It's not good to swap C and E on a transistor ;)

Cheers

Smirk >:-} ...Jim Thompson

We try to keep NPN emitters down and PNP emitters up.

That must be awful for signal flow.

If I had a choice, that's the way I'd do it. They allow it (as an option) for SSI stuff and BGAs (won't fit on a D-sized page) but nothing else.

Sounds like ADI's schematics. Yuck! My big bitch is connections by reference, though. What a PITA.

Figures.

You're an idiot, but that's not a surprise, here.

I know, all of your schematics only have a left edge. What an idiot.

It kinda works. Sometimes. ;-)

As do I.

Often it is. I make heavy use of busses/bundles. The tools allow the strands within bundles to be named anything. In an I2S interface, for example, I include all of the data lines and the bit, word, and m-clocks (all with standardized names). They fan straight out of the chip into the bus. On my schematics,

I also try to enter on the left side of the page and exit on the right but it's not always possible to make it look good. Our suppliers think my schematics are pretty easy to read, even with the crap tools we're forced to use. I spend a significant amount of time making them readable.

Others don't do so much or they work for someone who demands even opamps have "physical" symbols. If you want to see schematics with awful signal flows, look at a schematic with a dozen quad opamps on it, with the schematic symbols matching the package outlines. It's a mess but that's what is demanded.

Our stuff generally looks like

which has one transistor sideways.

We create all our own schematic symbols, generally to make the first use flow nicely, but it generally reuses pretty well.

Lately we show all power and ground pins, on the main part or occasionally as another "gate section" off to the side.

I draw schematics by hand, on d-size vellum, and The Brat cads them for me. The finals print nicely on B size.

I *hate* schematics that show, say, a 16 pin SO package as pins 1 through 16 in physical order, which I think you were talking about.

I guess there are at least 4 possible uses for a schematic:

2> Provide a signal flow diagram for an engineer to study. In which case it must look logical.

4> Provide a pretty picture for management. In which case you better follow the management rules.

Something about not being able to please all the people all the time seems to apply here.

On a sunny day (Fri, 03 Jan 2014 21:10:25 -0800) it happened John Larkin wrote in :

No insult intended, but that looks very bad on the 4051. At one point you use numbers, and on the right side the selectors are A0-A2. the numbers could easily be confused with pin numbers, use Y0 Y1 or whatever. I agree with others krw? and would use both pin numbers and real physical layout if possible, PLUS what the signals do inside the square. Like this:

Seconded. Schematics, like source code, are for people to read. Keeping current generally flowing from top to bottom, and signals generally left to right when possible, is a big help.

Cheers

Phil Hobbs

On a sunny day (Sat, 04 Jan 2014 06:38:30 -0500) it happened Phil Hobbs wrote in :

Oh, you make a big mistake. In John's schematics the opamps do not even have pin numbers. As I have worked in both design and service, maintenance too, I dare say that diagram is useless for repair, and documentation like that would not even be accepted by the broadcast organization I worked.

For repair and trouble shooting (and that always needed to happen 'now' and 'fast' as some director would flip out), and as renting a studio and having all artists at the right place and the right time costs more than you can possibly imagine, is often not repeatable either, good schematics were a must. That means, opening a map that has the layouts, pin numbers, symbols, block diagrams, and also training of people by the manufacturers.

Yes, if possible. signals often go both ways though, 'feedback', what not. Upside down text is idiotic and should never be done.

I also object to extremely small pieces of paper that you need to read with at least 10x magnification to find out the ink filled the symbols so you still do not know component values (had one like that last month).

Mine do, because they're often my actual work product, and get chucked over the wall to the client's layout folks. John's schematics go out via his daughter, who sits two cubicles down from his office, and who I imagine chooses which sections of the duals and quads to use. I doubt that their production and test folks ever see the originals.

I generally don't put pin numbers on transistors, except arrays.

If you're building something that takes a lot of maintenance, that's important. Not too many board-level things actually get serviced these days, though, apart from changing caps or obviously burnt components.

Sure, again if maintenance is the driver. Not so in my business.

Yup. I'm in the "4k7" camp for that reason.

Cheers

Phil Hobbs