Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
08/29/02 17:36
Read: times


 
#28187 - RE: 12mHz and 11.095MHz crystals
erik malund wrote:
-------------------------------
The experience says that a little error in the baud rate is acceptable

I am not arguing that the baud rate has to be be perfect to "work", what I am saying is : "select your crystal for the UART, not the uC". Why use the wrong part when the right part is readily available. If you go from 11.0592 to 12 to squeeze a little more performance out, you are already too close to the edge and will fall off.

Errors and variations will occur, why add to the load by buying anything but the right part. Is this about "let us see what we can get by with" or is about "what is the right thing to do"

Do ir right, it is more fun,

Erik

------
While I almost always agree with Erik, and value his experience and opinions, this is an exception. Choosing a crystal is a design consideration, and not a matter of "right" and "wrong." Sometimes a project spec says that the board has to last nn hours on a 9V battery. So reducing the clock speed drops the power requirements resulting in a satisfied customer. Some of the designs I have worked on don't even have (or don't use) a UART.

Designing a power supply and reset circuit without a supervisor chip means that it is more likely to fail. This is also a design consideration. In almost all cases, it is also a bad idea. The only advantage is cost. Which is also a design consideration. Ask the guys in marketing if saving a small chunck of change per unit is OK, if it means that the product will not come to life under certain conditions. If they say yes, ignore them and then go tell the guys in the service/support department to go have a chat with them.

This can NOT be equated to designing within a spec. If a design calls for a 120ns SRAM chip, using 70 or 85 ns chips is OK, as long as all the timing specs are satisfied. Choosing a crystal with a known error of 2% in a design where the spec calls for an error of less than 10% does not mean that it will fail.

Read the data sheet, understand the specs (and the trade-offs), then have fun.

Now climbing into flame-proof suit...



List of 17 messages in thread
TopicAuthorDate
12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
RE: 12mHz and 11.095MHz crystals            01/01/70 00:00      
DS89C420            01/01/70 00:00      

Back to Subject List