METHODS OF MEASURING BLOOD FLOW 1317 



so that, on the plateaus, the field strength (B = B,„„ r ) 

 is constant and, because (dB dt = o), no transformer 

 emf is induced. During these periods, the system be- 

 haves like a d-c flowmeter while the advantages of a 

 carrier-frequency a-c procedure are achieved by the 

 fact that B = B max in each first half cycle and B = 

 — B,„„j in each second half cycle. Therefore two flow- 

 signals of opposite polarity are delivered in each cycle 

 which are picked up by simple metal electrodes, 

 amplified by an a-c amplifier, and then converted to 

 congruent polarity by the discriminating demodula- 

 tor. It is obvious that spurious input signals which do 

 not change their polarity synchronously with B, 

 such as ECG, can be cancelled out if the carrier fre- 

 quencv is high enough. During the short time in- 

 tervals of magnet-field reversal, dB dt is very large so 

 that high spikes of transformer emf may be picked up 

 by the electrode circuit. This emf can be reduced by 

 the aforementioned split-lead method. However, the 

 preferred way to eliminate the transformer emf con- 

 sists in blocking the amplifier during the field re- 

 versals. Satisfactory separation of the flow signal from 

 unwanted a-c voltages is therefore, at least in theory, 

 quite possible. The electric circuitry is more complex 

 than for sine-wave flowmeters. Figure 29, shows the 

 principle of operation. .-1 is the idealized rectangular 

 time course of the magnetic field strength. B shows 

 the assumed course of blood flow as well as the input 

 signal which consists of the amplitude-modulated 

 flow signal and the transformer spikes. The blanking 

 periods C indicate the time intervals during which the 

 amplifier is blocked. Periodic gaps in signal voltage 

 due to the blanking operation are filled out by pro- 

 longing the duration of the voltage level reached im- 

 mediately before the beginning of each blanking 

 period ("filler voltage"). The demodulated voltage 

 shown in D has a steplike appearance which is, for 

 clarity, exaggerated in figure 2g and will be smoothed 

 by filtering. For the block diagram of the circuitry 

 see figure 30. The low-pass inverse feedback between 

 output of spike-blanking circuit and input of pre- 

 amplifier raises the lower frequency limit for sup- 

 pressing the ECG. 



The method is used for flow recording on intact 

 vessels in human surgery as well as in acute and 

 chronic-implantation experiments on animals (see 

 fig. 31). Flow measurements in extracorporeal de- 

 vices have been described (126). 



The heat produced by the magnet-coil current is 

 relatively great, since the electric power used to ob- 

 tain the rectangular wave current is greater than that 



Magnetic Field 

 240 n^per sec 



Flow Generoted Voltoge 



Zero Flow Voltage 



Transformer Effect 



Blanking Periods H H H 1-1 H n M H H H M H 



Rectified Minus Voltage 



Filler Voltage 



fig 29. Principle of operation of the 240 cps square-wave 

 flowmeter of Denison and Spencer. For description see text. 

 [From Spencer & Denison (126).] 



needed for an equally effective sine-wave current. In 

 addition, a higher ampere-turn value is needed to 

 saturate the core material and make the magnetic- 

 field plateaus flat. Even so, it is difficult to make these 

 plateaus perfectly flat so that a small transformer emf 

 may be still effective during the sampling periods. A 

 minor difficulty is the influence of the blood 

 hematocrit on the flowmeter's sensitivity, the nature of 

 which is not understood. 



A square-wave flowmeter similar to the 240-cps 

 model of Denison and Spencer was also described by 

 Ferguson & Wells (34) while Abel (1) developed a 

 400-cps chopper-operated square-wave meter. Shirer 

 et al. (123) built a 480-cps square-wave device pos- 

 sessing a frequency response up to 1 50 cps. The reader 

 mav refer to their thorough considerations of problems 

 of carrier frequency, filtering, gating, and noise. 

 Besides rectangular or trapezoidal wave shape, Spen- 

 cer and Denison suggests a sawtooth-like time course 

 of the magnetic field (126). 



At the present state of flow recording technique, the 

 electromagnetic method is a superior procedure. Any 

 decision whether the preference should be given to the 

 sine-wave, the square-wave or another wave-shape 

 type must await further development. 



