Re: [kicad-users] Re: Kicad library project

2009-09-14 Thread KeepIt SimpleStupid
Is XML a viable avenue to pursue because of all the trouble Microsoft is 
getting in for using XML as a basis for Word?

Example:
http://www.pcworld.com/article/170148/microsoft_word_lawsuit_xml_explained.html

Terms: Microsoft, XML, Lawsuit

--- On Mon, 9/14/09, Greg Dyess gregory.dy...@yahoo.com wrote:

From: Greg Dyess gregory.dy...@yahoo.com
Subject: Re: [kicad-users] Re: Kicad library project
To: kicad-users@yahoogroups.com
Date: Monday, September 14, 2009, 8:37 AM






 





  There are a number of open-source public-domain XML parsing 
libraries that are cross platform.  I am most familiar with Xerxes.  ACE may 
even have some XML capabilities in addition to its wonderful 
networking, multithreading and job pooling.  Wow, a multithreaded router 
perhaps?
 
Greg





From: Andy Eskelson andyya...@g0poy. co.uk
To: kicad-users@ yahoogroups. com
Sent: Saturday, September 12, 2009 5:01:41 PM
Subject: Re: [kicad-users] Re: Kicad library project

Do bear in mind that KiCad is a multi platform system, so things should
work on both Linux and Windows.

Andy




On Sat, 12 Sep 2009 19:45:02 +0100
Philip Pemberton ygro...@philpem. me.uk wrote:

 waynegramlich wrote:
  I think your idea is a good one.  Not locking it down to a
  particular electronic CAD package is also a pretty good idea.
 
 It's more a convenience thing of anything, that and other people
 might 
 want to use it with e.g. Altium Designer, Mentor or OrCAD. I'd rather 
 like the answer to Can I use it with X? to be Yes for as many 
 values of X as practically possible.
 
  Greg's proposal of using XML is pretty good as well.  At least
  with XML, nobody argues about syntax issues anymore.  In addition,
  this means additional information can be added.  For example,
  some parts can be ordered from multiple vendors with different
  part numbers.  With XML, adding a new tag does not break
  programs that are reading the file.
 
 The only (he says) problem with XML is that as a hierarchical format, 
 it's a bit of a pain to parse. That said, parsing is generally handled 
 with (say) Libxml or ElementTree/ Etree, and ends up being more or less 
 recursive -- if you see a
 Component element, you branch off to 
 HandleComponentElem ent() for a bit, which goes looking for the various 
 tags it needs to build the component.
 
 It is a LOT more verbose than what I proposed. I'd rather have a 
 single-line description to create a component, than 20 or 30 lines of 
 text that all has to be parsed. Though all that verbosity would make it 
 easier to understand the source script, and I guess you could always 
 write a Perl, Bash or Python script to write the component script... :)
 
  The concept of having a package is good one.  Many parts
  come in a multitude of packages.
 
 Well the concept derives from something along these lines:
    - Microchip, Fairchild, National and TI (for example) all make chips 
 in DIL packages.
    - The parameters of a given type of DIL package is
 standardised among 
 these vendors.
    - Why create four footprint libraries, when you can do what you need 
 to do with one?
 
 Admittedly some manufacturers do tweak footprints a little, but for 
 the most part the IC packages are standard. Actually, I seem to recall 
 there being an IPC standard for SMD footprints.. .
 
   It would be nice to
   specify the part once and select the package separately.
   Pin bindings sometimes vary between parts.
 
 This is how EAGLE works (more or less) -- you pick a Part in the library 
 browser, expand it, and then you can see what footprints are available 
 (f.ex. the Microchip PIC18F452 is available in a -I/P (DIL) and a -I/SO 
 (SOIC) package). You accept the selection, then it's added to the PCB.
 
 The catch is that the schematic symbols and PCB
 footprints are part of 
 the same library, and you can't link a schematic symbol and pinout set 
 to a footprint in another library (or at least you couldn't the last 
 time I tried). The Part provides the link between a symbol and
 
 IIRC, In KiCAD symbols and modules are a separate concept, and there 
 is no linking object (the pin numbers come from the Symbol). When you 
 create, say, a PIC18F452-I/ P symbol, you tell KiCAD that MCLR is on pin 
 1, and so forth. So if you wanted to change the pin numbers to create a 
 PIC18F452-I/ SO object, you need to copy the existing symbol. Later on, 
 if a change has to be made to the symbol as a whole, you have to 
 change both of them -- in EAGLE you'd edit the Symbol and both Parts 
 auto-update.
 
 I think I've explained that quite confusingly, but it's not an easy 
 concept to explain anyway...
 even with a whiteboard to hand... :)
 
  More control over drill selection is a good idea.  Some vendors
  use metric, others use imperial.  Some have a set of free
  drills and charge extra for other sizes.  Just being able to
  specify the minimum and maximum hole size for for each component
  lead would be a huge advance.
 
 Well 

Re: [kicad-users] Alternative footprint editor: proof-of-concept prototype

2009-08-14 Thread KeepIt SimpleStupid


--- On Fri, 8/14/09, Werner Almesberger wer...@almesberger.net wrote:

From: Werner Almesberger wer...@almesberger.net
Subject: [kicad-users] Alternative footprint editor: proof-of-concept prototype
To: kicad-users@yahoogroups.com
Date: Friday, August 14, 2009, 5:56 PM






 





  For a long time, I haven't been too happy with KiCad's module 
editor.

While it eventually does get the job done, I feel it relies overly on

manual work. With computers, we should be able to do better :-)



In the context of the gta02-core project, which depends heavily on

peer review, two more problems became apparent:



- the module editor doesn't capture the construction steps that led

  to the positions of things



- it's hard to review the results of one's work



So I hacked a little footprint editor called fped that tries to get

this right. Its purpose is to explore the parametric design approach

and also to help us make the footprints for gta02-core.



The README with build instructions, etc., is here:

http://svn.openmoko .org/trunk/ eda/fped/ README



You can design a footprint in two ways: either by writing a definition

in the human-readable language fped processes, or by drawing things

via the GUI. You can also combine the two, but be warned that fped

strips all comments when saving a design.



A fully developed example made through the GUI is here:

http://svn.openmoko .org/trunk/ eda/fped/ sc89.fpd



And this is a partial QFN that serves as my playground for

experimenting with the language:

http://svn.openmoko .org/trunk/ eda/fped/ qfn.fpd



Fped writes the KiCad module format and it also has some rudimentary

Postscript output.



Fped is still under construction and has a few severe limitations:



- you can't define holes, not even mechanical ones



- pads always affect component, solder mask, and solder paste layer.

  There's no way to specify a solder paste that doesn't cover the

  entire pad.



- it's easy to generate footprints with multiple identical pads and

  fped won't even notice



And there's a lot of minor issues as well. Of course it's also still

fairly crash-happy and the code organization sucks.



- Werner


 

  




 

















  

Re: [kicad-users]Re: How to make little holes in the pads centers printing from Gerbview?

2008-02-22 Thread KeepIt SimpleStupid
That's an excellent idea and so is the pilot drill
size.  I leared that from a machinist although he
would use something like 1/3.


--- Alain M. [EMAIL PROTECTED] wrote:

 
 Vladimir Kalyaev escreveu:
  If I'm not mistaken Print prints out FULL-SIZED
 holes, not SMALL 
  centering hole!
  Other explanations: for .6-8mm drill open area
 should be 0.2-0.3mm.
  For 0.8mm thru-hole element a drill open area
 produced by Print 
  command equals to the drill diameter. I think it's
 impossible to use 
  0.8mm open area for centering a 0.8mm drill!
 I did some tests (some years ago) and I finaly
 settled that holes *half 
 the real size* are ideal. That is because dril
 centers are bigger for 
 bigger drills, if you have a .8mm center for a 3.2mm
 hole it is just 
 useless. What I used to do was to define all my pads
 with half the real 
 diameter and fix that in the docs acompanying the
 drill file...
 
 Alain
  In OrCAD there is an option keep drill holes
 open which produces a 
  same-sized holes for each thru-hole element.
  I can write a script for overruding every
 thru-hole padshape to force 
  a Print command work correctly. Is it an only
 way?
 Just the same, a script could be made to fix that
 and generate the print 
 file. On option could interest some people...
 
 Alain
 
 



  

Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs


[kicad-users] Feature? Invisible wire segments

2007-12-17 Thread KeepIt SimpleStupid

While servicing a piece of equipment they used a
interesting construct which I will call an Invisible
wire segment .

If there happened to be a connection on say the top
left-hand corner of the schematicPCB to the say,
connected to a component in the lower left hand corner
of the schematic they would draw a line segment with
an arrow at both locations pointing at the other
location.

All that was required for the user was to align a
ruler to find the other corresponding mark.  This
reduced clutter in the schematic.

This could by menu choice be shown as a dotted line
between the arrows pointing toward each other.

It might be worth considering, might not.



  

Looking for last minute shopping deals?  
Find them fast with Yahoo! Search.  
http://tools.search.yahoo.com/newsearch/category.php?category=shopping


Re: [kicad-users] Re: Comparison

2007-11-16 Thread KeepIt SimpleStupid

--- Dan [EMAIL PROTECTED] wrote:

 First, I have a question: what degree is your son
 working on?  I have
 a BSEE degree from a top state university, and I
 never did any PCB
 design while in school.  This is generally not
 covered in BS-level
 undergrad classes, and probably not in grad-level
 classes either. 
 It's considered something you pick up on the job, if
 you need it.

Agreed.  I learned PC Layout at Hewlett-Packard, now
Agilent in a Boy Scout Explorer's post in 1972.  I
taped a few major electroncs hobbyest projects and the
Leach Amp (80's).  The only layout software I've used
extensively so far was DOS based, EasyTRax.

The larger the library, the easier and faster the
design.  It's almost like what should probably happen
is that the component industry should get together and
define a universal importable footprint/3D model etc.


  

Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs


[kicad-users] While I was reading the docs, I thought of a wish list item

2007-10-22 Thread KeepIt SimpleStupid
Documentation for any project seems to be a necessary
evil.  I'm not sure how how this suggestion might be
implemented, but here goes.

Create a mechanism to link the datasheet of a
component to the component.

Each component has a few optional fields.

1) URL of manufacturer's data sheet
2) Local directory to retreive data sheet.

Whether or not to 1) Use manufacturer's URL, 2) Local
directory and/or whether or not to place a copy of the
datasheet in the project directory.

__
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 


Re: [kicad-users] Re: Mils

2007-10-13 Thread KeepIt SimpleStupid
Let's face it, humans like things easy.  It's much
easier to say 5 mills, than 5 thousanths of an
inch, although saying 250 thousanths seems to be
more of the norm when machining in english units. 
Thickness, such as a plastic bag seems to be expressed
in Mills.  Confusion, of course.

When connecting thermocouples, RED is negative.  Go
figure.

Even the metric system has it's pecularities, the
Angstrom and the micron.

My boss once said, he likes numbers to be between 1
and 10.  In other words the exponent expressed in
words such as nano, micro, pico.

I prefer metric too, but when the parts and the tools
are English, do as the Romans do and use the unit of
measure that the parts are in.






  

Shape Yahoo! in your own image.  Join our Network Research Panel today!   
http://surveylink.yahoo.com/gmrs/yahoo_panel_invite.asp?a=7 




Re: [kicad-users] Kicad project for Low TIM (Leach Amp)?

2007-10-13 Thread KeepIt SimpleStupid
It's a nice amp. You'll like it.   I taped it in 1980
or so.  Just like everyone else, I made subtle
changes.
A couple notable ones: metal film resistors where
possible.  10 turn pot for bias adjustment.  I added
audio ramping and thump supression as well.

The power supply makes a big difference in the sound. 
I used a 20 AMP constant voltage transformer and the
AMP sounded great and the hum was horrible.  I then
changed to a somewhat under rated quad custom made
torroidal transformer and the AMP was quieter, but not
as bassey.

I accidently mad a mirror image of the board and
because the design is so symmetric, I was able to use
it with a few external changes and switching the NPN
and PNP transistor positions.

Go for it.


--- j3r3m3l33 [EMAIL PROTECTED] wrote:

 Has anyone KiCAD'ed this amp:
 
 http://users.ece.gatech.edu/~mleach/lowtim/
 (http://www.pavouk.org/hw/leachamp/en_index.html)
 
 The second link's author said he did it in Eagle.
 
 By all accounts is a nice unit when built. I've
 asked
 the author where to get manufactured PCBs but no
 reply
 yet. So failing that, unless someone here has
 already
 entered the design into Kicad, then I might do it
 and
 post the result.
 
 
 
 
 



  

Catch up on fall's hot new shows on Yahoo! TV. Watch previews, get listings, 
and more!
http://tv.yahoo.com/collections/3658