Page 215 - Troubleshooting Analog Circuits

202 F. How to Get the Right Information From a Data Sheet

Even today, an attempt is made to build on the good things WEN TO WRITE DATA SHEETS
learned from the past and add a few improvements when A new product becomes available. The applications engi-
necessary. But, it's important to have real improvements, neers start evaluating their application circuits and the test
not just change for the sake of change. engineers examine their production test equipment.
So, while it's not easy to get the format and everything in it But how can the users evaluate the new device? They have
exactly right to please everybody, new data sheets are con- to have a data sheet-which is still in the process of being
tinually surfacing with new features, applications ideas, written. Every week, as the data sheet writer tries to polish
specifications and aids for the user. And, if the users com- and refine the incipient data sheet. other engineers are re-
plain loudly enough about misleading or inadequate data porting, "These spec limits and conditions have to be re
sheets, they can help lead the way to change data sheets. vised;' and, "Those application circuits don't work like we
That's how many of today's improvements came about- thought they would we'll have one running in a cwple of
through customer demand. days." The marketing people insist that the data sheet must
Who writes data sheets? In some cases, a marketing per- be finalized and frozen right away so that they can start
son does the actual writing and engineers do the checking. printing copies to go out with evaluation samples.
In other companies, the engineer writes, while marketing These trying conditions may explain why data sheets always
people and other engineers check. Sometimes, a commit- seem to have been thrown together under panic conditions
tee seems to be doing the miting. None of these ways is and why they have so many rough spots. Users should be
necessarily wrong. aware of the conflicting requirements: Getting a data sheet
For example. one approach mght be The originat Mi "as completely as possibte" and "as accuratety as possi-
of the product writes the data sheet (inside his head) at the ble" is compromised if one wants to get the data sheet "as
same time the product is designed. The concept here is. if quickly as possible."
one can't find the proper ingredients for a data sheet--good The reader should always question the manufacturer. What
appliitions, convenient features for the user and nicely are the alternatives? By not asking the right question, a mis
tested specifications as the part is being designed-then understanding could arise; getting angry with the manufac-
maybe it's not a very good product until all those ingredients turer is not to anyone's advantage.
are completed. Thus, the collection of raw materials for a Robert Pease has been staff scienbkt at National Semicim-
good data sheet is an integral parl of the design of a prod- ductor Gorp... anta Clam, CalK. for ehven years. He has
uct The actual assembly of these materials is an art which desgtmd numerous op amps, data convertem, wltage reg-
can take place later.
&tors and analog&cuil functbns.