39 

 BIOLOGICAL AMPLIFIERS 



In this chapter we consider types of amphfier for bioelectric potentials. 

 Amplifiers for transducer outputs are dealt with in the appropriate section 

 of Part III. 



The amount of gain which can usefully be employed in a recording channel 

 is, as was pointed out in Part I, determined by the noise — including drift — 

 generated by the preparation and the first amplifying stage. Suppose one is 

 looking for spike potentials in a noisy trace on a cathode ray tube. Then the 

 appearance of a spike of ampHtude, say, 100 times the R.M.S. noise voltage 

 could be taken to represent a physiological event with a probability amounting 

 to almost complete certainty. If a spike of duration 1 msec and of amplitude 

 10 times the R.M.S. noise appears in a trace 100 msec in duration, the 

 probability that it represents an action potential rather than a particularly 

 large noise deflection still exceeds 0-99. However, the possibility of spikes 

 smaller than this being noise artefacts has to be taken into account. 



In practical terms, an a.c. coupled amplifier with a pass-band extending 

 from 50 to 10,000 cycles might exhibit a noise level of 2 /aV R.M.S. with the 

 input terminals short-circuited. With the input terminals fed from a glass 

 microelectrode the noise will be very much greater, but in the case of a low 

 resistance preparation and electrode system the noise level should still not 

 be much above 2 //V R.M.S. A reasonable procedure would be to arrange 

 that, at full gain, a 20 //V event at the input, i.e. one which represents a 

 signal with moderate certainty, produces half of full-scale deflection on the 

 cathode ray tube. That is, it deflects the spot, from the tube centre, half of 

 the way to the top or bottom of the screen. A typical 6 in. cathode ray 

 tube with an EHT of 2 kV has a deflection sensitivity of 0-5 mm/V, so that 

 the voltage required to produce 1| in. deflection is about 75. The maximum 

 total amplifier gain needed is then 75 V/20 //V = 3-75 X 10^, which may be 

 obtained with 4 stages. 



It is frequently not possible to use as much as this. In an amplifier 

 habitually to be used with microelectrodes the maximum gain could probably 

 be reduced by a factor of 10. Again, in RC coupled amplifiers of long time 

 constant (e.g. 4 couplings of 4 seconds time constant each) the pass-band 

 would extend down to a frequency range where flicker and carbon resistor 

 noise would become important. The stabiUty of the base line would also be 

 affected by displacements attributable to small mechanical movements within 

 the valves*. In this case, too, the maximum usable gain is probably nearer 

 400,000. This can be obtained with 3 stages, and easily with 4. 



An additional difficulty with high gain amplifiers of long time constant is 

 'blocking'. This refers to a lengthy paralysis of the amplifier following the 



* Not so much microphony, as movements to relieve stresses caused by the valve warming 

 up. 



616 



