Hi Bill,
The history of these charts is pretty funny, maybe sordid is the wrong
word. It's nice having a 'safety margin' but 4X is way overboard,
especially if people are going to go to 2 oz, 3 oz or 4 oz for
innerlayers. Time to get a little more accurate in our design work. 1955
was a LONG TIME AGO, before transistors and ICs. Like you have said,
these curves won't lead to any failures, the reverse is true... the charts
have more than enough margin to prevent someone from having a heating
issue if they use them correctly within their given ranges... and they
have a long track record of successes... however they do restrict the
designer a great deal and the board is most likely over designed when they
are through... like you say many more times than need be...
Like Mike has on his history papers on his web site," NBS Report No. 4283
dealt with tests conducted on the current-carrying capacity of etched
copper conductors, resistance measurements on samples having a variety of
protective coatings, and dielectric properties of a number of metal-clad
laminates at various temperatures.
A chart showing the relationship between temperature rise, conductor
width, cross-sectional area, and copper thickness was developed as a
design aid for determining conductor sizes in printed circuit application
utilizing etched conductors and is shown in Figure 1. This design chart
was assembled for use primarily with phenolic (XXXP) and Epoxy-glass
materials of 1/32 (0.031), 1/16 (0.0625") and 1/8 (0.125") inch thickness
and copper thickness of 0.00135 inch (1 ounce) and 0.0027 inch (2 ounce).
In 1955 they were not thinking about multi-layer circuit boards, therefore
all of the testing was performed on external traces only. Presently, all
that is known about the internal conductor charts is that they are based
on half the current from the external chart data collected in 1955...The
external charts from 1956 would have been more useful if they had been
created as multiple charts. There should have been 10 charts for each of
the test samples. Instead, they are all jumbled together and the results
are skewed. This is significant when trying to design with currents that
go above several Amps. The internal chart is even harder to work with when
you don't know that it doesn't directly represent test data!
It is important to design with information that is relevant to the
technology that is being used. It is also important to take into account
the environment that the electronics will be operated, such as in Space
(vacuum). Using conservative charts and keeping track of the (energy
losses) power dissipated by the traces, vias, thermals and planes is where
the industry is going. Optimization of a design must start with
information that is relevant to the technology being used and has a basis
of understanding. It is also important to follow the rules that are taught
in regard to parallel traces.
Parallel traces are not only those that are side by side, but also from
layer to layer. The parallel trace rule:
This is a case where the power dissipation and power density of the trace
can be used. The new charts in the "Thermal Man Calculator" provide trace
heating characterized in just the board material and we make note of the
power dissipation and power density for the current and temperature rise.
When we have copper planes and various mounting configurations this brings
the temperature down, although the power dissipation stays the same. In
any design you don't want to lose sight of the fact that the traces
dissipate power and they contribute to the over-all heating of the circuit
board."
As you say, "To say they are 4X is really a technical generalization...
it's not a linear relationship across the chart, and it varies depending
upon materials and conditions and location and cross sectional area of the
heated tracks...also the presence of planes in the board stack up can
greatly affect the curves... What Mike is creating is really a much better
tool than the charts could ever address...
Thanks for the comments.
Happy Holden
Westwood Associates
"Brooks,Bill" <[log in to unmask]>
Sent by: DesignerCouncil <[log in to unmask]>
04/06/2005 02:59 PM
Please respond to
"(Designers Council Forum)" <[log in to unmask]>; Please respond to
"Brooks,Bill" <[log in to unmask]>
To
[log in to unmask]
cc
Subject
Re: [DC] PCB TECHNOLOGY (2 oz copper)
Happy,
I wouldn't call the charts 'messed up'...
They have de-rating factors built into them and they were based upon
Phenolic material in the beginning and have been used for many, many
years...
Mike Jouppi is helping to make more accurate thermal modeling available to
Designers and engineers who need more accurate charts to work from... but
these charts work, and work well... because they have a lot of margin
built
into them... personally I don't like designing without some safety
margin...
it's like insurance.
If you are designing against the cutting edge then Mike has a program that
will let you calculate the thermal performance of a given trace based upon
all of the variables... http://www.thermalman.com/3.shtml
Best regards,
Bill Brooks - KG6VVP
PCB Design Engineer , C.I.D.+, C.I.I.
Tel: (760)597-1500 Ext 3772 Fax: (760)597-1510
e-mail:[log in to unmask]
http://www.dtwc.com
http://pcbwizards.com
---------------------------------------------------------------------------------
DesignerCouncil Mail List provided as a free service by IPC using LISTSERV 1.8d
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF DesignerCouncil.
To temporarily stop/(restart) delivery of DesignerCouncil send: SET DesignerCouncil NOMAIL/(MAIL)
Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
---------------------------------------------------------------------------------
|
