Breaking News
Comments
Newest First | Oldest First | Threaded View
<<   <   Page 5 / 6   >   >>
betajet
User Rank
CEO
Re: Paper models?
betajet   6/3/2014 3:30:35 PM
NO RATINGS
doug also wrote: When building a micro processor symbol, it usually ends up being either one big box, or multiple boxes. One part smears accross 4-5 pages.

Whenever possible, I like to have schematic symbols of components with lots of pins to match the actual pinout, so that it's easy to correlate the schematic and the PCB.  This is quite doable up to a 208-pin PQFP on a B-size sheet.  It's harder with BGAs -- with those I'll put the power and ground pins in a separate block and if I'm feeling ambitious I'll make a rectagular symbol that matches the internal pad sequence, which you can get from the chip's BSDL file.

Regarding the alphabet soup of multi-use pins, there's nothing to stop you from taking a generic component for the part and editing the pin names to be just the setting(s) you actually use.  I do this with CPLDs and FPGAs, since names like "I/O32" are pretty useless.

Creating new symbols and customizing existing ones does take time up front, but will save lots of time in the long run when it's time to debug the design when the boards come back 6-12 weeks later.

JMO/YMMV

betajet
User Rank
CEO
Re: Paper models?
betajet   6/3/2014 3:16:32 PM
NO RATINGS
doug wrote: I've been wondering, now that most of us have a computer to work on, why do EE types cling to "sheets".

It depends on what I'm doing.  If I'm probing a board, I like to have B-size schematics so that I can see "the big picture" all at once.  Otherwise it's like those AAA custom travel guides that show you each segment of the route on a separate page.  Some people like to navigate that way -- I like the "chart room" where Captain Nemo unrolls a chart to get the full picture of where one is and where one is going.  Scanning around a schematic on a computer screen is a PITA when it's already covered with terminals and debug consoles.

OTOH, if I'm considering a logic design inside a chip, I'm not going to be able to probe it so it makes more sense for each piece to be documented on separate pages.  But that's because a complex chip is more like a data book of smaller chips that are connected together at a higher level, with a block diagram that shows the big picture.

I once heard a great talk on map making.  A completely accurate map that shows everything is often useless.  Each map has to be created for a particular audience, and it's drawn to be most useful to that audience with some features drawn inaccurately so they don't confuse the important features.  They call this "cartographer's license".

DU00000001
User Rank
CEO
Re: Paper models?
DU00000001   6/3/2014 2:46:18 PM
NO RATINGS
"If the symbols and text can not be searched, I want to run screaming out of the room."


To be honest, Im rating non-searchable PDFs as one of 2 things:

a) Declaration of war (we once had a supplier delivering the
    schematics of an ECU (3 pages only readable if printed
    A3 or beyond) as a non-searchable PDF.
    This WAS the start of a war.

b) Demonstration of imbecility of the creator

tb100
User Rank
CEO
why create a schematic
tb100   6/3/2014 2:23:11 PM
One of my pet peeves is when engineers draw their schematics by putting parts on the sheet then rely on net names to create the connections. Hey, it is easier to draw than lining things up and putting in all those annoying nets, and as long as it connects things up to create a correct netlist then you are doing it right. Right? 

No. If you do that you might as well just avoid the schematic altogether and just create the nelist from scratch. (In fact there are vendors that have products that will create a netlist from a spreadsheet, which actually works really well for things like backplanes).

But a schematic has another purpose besides creating a netlist for layout: it is documentation.. Therefore, it should show connectivity and function. Ideally it should almost look like a block diagram.

I draw my own symbols so that processor memory interfaces line up to DIMM interfaces so the nets flow straight across. I modularlize, so functions are grouped together.

Layout engineers, board fabrication and assembly engineers, and debug technicians will all use your schematic. Will they be able to figure out how the circuit works from the schematic, how signals flow? Is the schematic kept up-to-date? In other words are the parts shown on the schematic the actual ones on the finished board? 

A well drawn schematic makes the entire process of building and testing the board go well. A poorly drawn schematic will slow things down in ways you might not even realize.

betajet
User Rank
CEO
Re: Paper models?
betajet   6/3/2014 1:24:31 PM
NO RATINGS
Max wrote: There is [a happy medium] -- she's always telling jokes...

You should strike her some time :-)

[from the Old Jokes' Home]

Max The Magnificent
User Rank
Blogger
Re: Paper models?
Max The Magnificent   6/3/2014 1:06:35 PM
NO RATINGS
@dougwithau: My thinking is there has to be a happy medium...

There is -- she's always telling jokes -- she has an office just around the corner from here LOL

dougwithau
User Rank
Manager
Re: Paper models?
dougwithau   6/3/2014 12:45:57 PM
NO RATINGS
We look at a lot of schematics from many of our customers. The only feature that matters is the PDF export. If the symbols and text can not be searched, I want to run screaming out of the room.

Even then it is only a text search, no context.

Netlist import and export or a standard is what really matters, I have that as the number one feature.

My thinking is there has to be a happy medium between drawing little boxes and ligning up wires, and the blinking cursor in a blank text editor when you start working with an HDL.

antedeluvian
User Rank
Blogger
Re: Paper models?
antedeluvian   6/3/2014 11:49:39 AM
NO RATINGS
dougwithau

When building a micro processor symbol, it usually ends up being either one big box, or multiple boxes. One part smears accross 4-5 pages. 
Even worse, if you look at the symbols for the newer ARM processors, each pin can have 4-5 functions. The pin names are a paragraph all by themselves. "USBB2_ULPITLL_DAT2/USBB2_ULPIPHY_DAT2/SDMMC3_DAT2/GPIO_163/HSI2_ACFLAG/DISPC2_DATA18/USBB2_MM_TXSE0/SAFE_MODE" Yikes!

Wouldn't it be nice if you could (right?) click on the pin of the symbol and the select the chosen function from a drop down menu?

If there was a better desgin capture tool would you use it?

As long as it tied into the netlist format used in the layout. Or am I stating the obvious.

I've thought about making a web based attempt at something better, but would anyone change just based on "easier to use"? Is the schematic to ingrained in us old timers?

Part of me says that it makes sense to have one big schematic, but another part says that partioning is good for conceptualization. I also seem to have trouble moving around from end to end on a large schematic even now on a screen- it may be necessary to have the ability to have multiple views of the same design on different screens.

When I go to site and try to debug on a small laptop screen- I am not sure.

And as for resistance to change- you betcha!

dougwithau
User Rank
Manager
Paper models?
dougwithau   6/3/2014 11:39:21 AM
NO RATINGS
I've been wondering, now that most of us have a computer to work on, why do EE types cling to "sheets". The whole model of schematics is based on the number of pieces of paper. Some vendors sell based on "number of sheets" to differentiate between the lite and pro versions.

When building a micro processor symbol, it usually ends up being either one big box, or multiple boxes. One part smears accross 4-5 pages. 
Even worse, if you look at the symbols for the newer ARM processors, each pin can have 4-5 functions. The pin names are a paragraph all by themselves. "USBB2_ULPITLL_DAT2/USBB2_ULPIPHY_DAT2/SDMMC3_DAT2/GPIO_163/HSI2_ACFLAG/DISPC2_DATA18/USBB2_MM_TXSE0/SAFE_MODE" Yikes!

If there was a better desgin capture tool would you use it?

I've thought about making a web based attempt at something better, but would anyone change just based on "easier to use"? Is the schematic to ingrained in us old timers?

Duane Benson
User Rank
Blogger
Your own PCB symbols
Duane Benson   6/3/2014 11:31:48 AM
NO RATINGS
Here at Screaming Circuits, we see quite a large number of different board designs come through. One of the more common problems we see is having a footprint that doesn't quite match up with the actual part.

It's real easy to pick a symbol that's close, but ends up being a QFN form-factor when a QFP part will be used, or someone did a poor job or creating it and the rotation is off, or a few  pins are swapped. There are some diodes with different size pads. Some people will take a standard land pattern, assuming it's close enough. That can lead to the part moving, or uneven melting.

A lot of issues can show up due to footprint problems. If in doubt, I create a new footprint. Even if there's little doubt, it's best to do a thorough check before using the symbol/footprint.

<<   <   Page 5 / 6   >   >>


Flash Poll
Top Comments of the Week
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Life
Frankenstein's Fix, Teardowns, Sideshows, Design Contests, Reader Content & More
Max Maxfield

Feast Your Orbs on My Jiggly Exercise Machine
Max Maxfield
52 comments
Last weekend, I was chatting with my mother on the phone. She's all excited that I'm coming over to visit for a week in November. "I'll be seeing you in only seven weeks," she trilled ...

Glen Chenier

Missing Datasheet Details Can Cause Problems
Glen Chenier
3 comments
It is often said that "the devil is in the details." All too often those details are hidden deep within a datasheet, where you can easily overlook them. When a datasheet reference circuit ...

David Blaza

RadioShack: The End Is Nigh!
David Blaza
123 comments
I'm feeling a little nostalgic today as I read about what looks like the imminent demise of RadioShack, at least as we currently know it. An old ubiquitous cartoon image popped into my ...

Larry Desjardin

Engineers Should Study Finance: 5 Reasons Why
Larry Desjardin
47 comments
I'm a big proponent of engineers learning financial basics. Why? Because engineers are making decisions all the time, in multiple ways. Having a good financial understanding guides these ...