I just found an excellent diagram illustrating log/linear
conversion on the URL
http://nucleus.immunol.washington.edu/Research_facilities/Apps/logscale.html
IMHO, the diagram (logscale.gif) is much more persuasive and easier to
understand than many pages of plain text I have read before on the
same topic (actually, not so many).
I spent some time with a ruler and a calculator and came to conclusion
that
this diagram also helps to understand how log amplifiers transform data.
For example, when a log amplifier receives an "X" mV pulse from a
preamplifier, it then
produces the new pulse ("Y"), which may be calculated according to the
formula:
Y= lgX x 256 x 9.8 mV
or
Y= lgX x 64 x 39 mV
or
Y= 2.5 lgX.
For example, if there is a 300 mV pulse in, then there's a 6193 mV
out (2.5 x lg300 = 6193); if 30 mV in, then 3692 mV out; 3 mV in/1192
mV out and so on.
Could anybody tell me whether my calculations are correct?
I'd be very pleased to hear your opinion, because I had to mobilize all
my residual memories about school arythmetics doing the calculation.
Does anybody have some other *.gif files illustrating how pulses are
processed in
the electronics? Maybe, someone have a diagram like this?
----------------------------------------------------------------
Preamplifier ------> Log amplifier -------> ADC
analog signal, analog signal, digital scale,
mV mV channels
----------------------------------------------------------------
And, finally, could anybody explain me what the following paragraph
means?
"This assumes that the log amplification is perfect (i.e., each of the
four decades is accurate) which we know is likely a fiction. The first
and last decades can be very inaccurate"?
Thanking you in advance,
Vadim Chromiak