Kara McCloskey writes: >I am trying to learn how a log amp works in a flow cytmeter. Does anyone >have a good reference for this (more detail than Shapiro's text). More detail on log transformation in general or on the electronic guts of log amps? A discussion of the log scale and dynamic range in somewhat more detail than appears in the book can be found in "A flow cytometer designed for fluorescence calibration", by Shapiro, Perlmutter, and Stein, Cytometry 33:280-287, 1998. For the electronics, the place to start in terms of how log amps work is the book "The Art of Electronics", by Horowitz and Hill, 2nd Edition, Cambridge University Press. Log amps are basically of two types, those built around a single operational amplifier (op amp) and those made of cascaded stages. The best single op amp log amps are made by Optical Electronics, Inc. (OEI), P. O. Box 11140, Tucson, AZ 85734, (520) 889-8811; OEI is a company so staunchly analog in philosophy that it hasn't got a Web site (actually it does, but only for its video processing components, so you'll have to write or phone about log amps). The current leader in the other variety of log amp is Analog Devices; try www.analog.com . While some of the newer log amps are quite good, the basic obstacle to using them in sophisticated flow cytometers is that fluorescence compensation must be done using linear electronics at the input to the log amp; as the number of colors to be compensated increases, the linear circuitry gets more complex, and there is an unavoidable increase in the noise level, which starts to limit the dynamic range; it therefore wouldn't matter if you had a perfect log amp, and most of them are far from perfection. It is now preferable to use high (16 bits or more) resolution analog-to-digital converters to digitize pulse peaks or integrals (as is done in laboratory-built cytometers and in the Beckman Coulter EPICS XL), and even better to use high-speed, high-resolution ADC's and process (baseline restore, threshold, and peak detect and/or integrate) the pulse digitally, as is done in the Luminex 100. -Howard
This archive was generated by hypermail 2b29 : Sat Mar 10 2001 - 19:31:09 EST