NOISE and 5B's

Noise is usually a difficult problem to overcome. The reason for the
difficulty is that it is so hard to find the source. This technote should
help you to isolate any noise that you may see with the 5B modules.
When you can detect noise in your measurement application that is using 5B
modules you should first try to isolate the problem to the transducer, signal
Conditioner (5B) or measurement system. To isolate the 5B module you want to
disconnect the input from your transducer and apply the appropriate input setup
for each module:

Module 			Input setup 
5B30/31/40/41 		Shorted 
5B32 			Shorted 
5B34 			RTD or 100 ohm resistor 
5B37/47 			Shorted 
5B38 			Excitation is balanced with any 2 equal value resistors greater than 300 ohms with the center of the resistors tied to the (-) input. Then short the (+) and (-) inputs. 
												
These setups should give a steady DC signal on the output. (Either 0
volts or offset.) The next step is to measure the output of the 5B module. And
depending on the bandwidth of the measurement system you can see different
results.

Until recently All 5B modules showed a noise component that is from
the high frequency mod-demod circuitry within the 5B module (from  the
transformer isolation section).  This "noise" is at a frequency  that is the
carrier frequency of the individual 5B module.  It is in the form of a very fast
rise and fall time spike signal. The amplitude of this noise  spike  will
depend  upon  the  bandwidth of the measurement system you use. For example, if
you take a lOOmHz bandwidth scope and look at the output directly of a 5B module,
you  may  see a pk-pk amplitude of 150 mV .. but if you take a scope  with a
l0kHz  bandwidth, and look at the same module you may see a  pk-pk  amplitude
for  this  noise  spike  of only 5 mV.

The reason for this explanation of the "change" in the  apparent
amplitude  of  the  noise spike is to get back to the issue of the bandwidth of
the measurement system. And why this noise will not be a problem in some
applications and why it will be a problem in others.

For example : if the measurement system connected to the output of a 5B
module is a successive approximation type of A/D or an integrating type of A/D,
this noise is not "seen" .. it is much too fast  and  is  never  a  problem.
However, if the measurement system is a sampling A/D  system, this noise spike
can be a problem because that is the type of measurement system which will "see"
fast spikes.

If you are using a sampling A/D system you could try placing by-pass
capacitors on the outputs of the 5B modules, to by-pass this high frequency
noise to ground at the output of each module. Try using 0.1 ufd capacitors.

Because of the increased use of 5B modules in applications utilizing
sampling A/D's, Analog devices has come up with a design change to all 5B
modules to significantly reduce this high frequency spike noise at the output
of each module. The new modules are specified to have a worst case noise of
.6mVpk-pk @ 100KHZ bandwidth and a worst case 14mVpk-pk @ 5MHz bandwidth. These
new specifications for the modules make them the best signal conditioners
available on the market.

Each model has undergone this change at different times. You can check
the 4 digit date codes of your modules to see if the number is less than the
ones listed below. All modules with the date codes below or greater have the
low noise change.The date codes are as follows:

Model 		Date Code 
5B30/31/32 	9316 
5B40/41 		9326 
5B34 		9327 
5B37 		9327 
5B47 		9328 

Note: Please understand that we do not replace any 5B modules that are
before these dates.