Best 1968

Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

NUCLEAR INSTRUMENTS AND METHODS 64(1968) 171-172; 0 NORTH-HOLLAND PUBLISHING co.

SQUARE ROOT DISPLAY WITH A SMALL COMPUTER


G. C. BEST

Electronics and Applied Physics Division, UKAEA Research Group, Atomic Energy Research Establishment, HarweN

Received 9 May 1968

A simple method for square-root display is suggested in which the data is collected directly in square-root form, avoiding the com-
plicated linear to square-root conversion.

Statistical data from nuclear counting experiments is one large. If this is acceptable a square root method of
usually presented by means of a linear or logarithmic display becomes both simple and attractive.
CRT display. It has been pointed out’) that both these Consider a single channel in a multichannel spec-
methods are unsatisfactory, since linear display gives trum. At a given time its content is N2, and a square
insufficient emphasis to low numbers, while logarithmic root display would show N. The next deflection,
display gives them too much emphasis. Square-root representing N+ 1, is made when the count reaches
display is preferable since it takes the statistical signifi- (N+ 1)‘. This will happen after Q more counts where:
cance of the data into proper account.
Q =(N+1)2-n;2
In current practice both data accumulation and
= (2N + 1).
display are frequently achieved by use of a small com-
puter. The necessity for frequent repetition of the se- Suppose now that two locations are provided for
quence makes it essential that the display program be each channel. One holds the current square-root value,
as simple as possible. Ideally, it should consist of the other a tally number. Initially the square-root is 0,
sequentially reading numbers from memory without the tally set to - 1, [i.e., -(2N+ l), where N = 01.
program manipulation. For this reason the square-root Counting now proceeds as follows. When a count is
method has been little used, since the function is received in the channel the tally is incremented. Unless
awkward to calculate with a restricted instruction code. it reaches zero no more is done. If the tally does become
The difficulty can be overcome if information for zero, the square-root location is incremented and the
each channel is held as two small numbers rather than tally reset to the new value of -(2N+ 1). This is
extremely simple since N is the content of the square-
SOUARE ROOT FOR THIS CHANNEL IN SA root location at this time. The algorithm for the
TALLY FOR TMIS COUNT OBSERVED IN operation is given as fig. 1, and a counting sequence
CHANNEL IN TA CHANNEL A
I
shown in table 1.

TABLE 1
Normal (add 1 per count).
(all numbers decimal)

Count Tally [displayed]

Initial -1 0
1 0 1
-3
2 -2 1
3 -1
4 0 2
-5
5 -4 2
AND ENTER IN TA 6 -3 2
7 -2 2
8 -1 2
9 0 3

Fig. 1. Square-root
CA3
EXIT

counting algorithm.
-1
etc.

171
172 G. C. BEST

TABLET
Increased resolution (add 4 per count).
(all numbers decimal)
NO
Count Tally [displayed]
Initial -1 0.0
1 +3 0.5
+3+(-3)= 0 1.0
o+(-5)- -5
2 -1
3 +3 1.5
+3+(-7)= -4
4 0 2.0
o+(-9)= -9
5 -5
6 -1
I f3 2.5
+3+(-ll)= -8
8 -4
9 0 3.0

YES

A possible disadvantage of the system arises if the


1’
count in the most active channel is not particularly
IEXlTl large. For example a maximum count of 214 gives a
Fig. 2. Increased accuracy algorithm. square-root range of 27. If the display system has 10 bit
resolution it will be inefficiently used. This difficulty can
No loss of counting accuracy is involved in this be overcome by a slight extension to the algorithm
method. The total count P at any time is given by: whereby a number greater than unity is added to the
P = (S+ 1)” + T, tally for each count observed. Suppose M is chosen,
where when M= m2. After P counts have occurred the
S = current square-root value, number recorded will be P-M. Suppose that P =p2,
T = current tally value (negative). then P*M=p2-m2, and the square-root observation
However, both S and T are small numbers compared displayed will be p *m.
with P. This is a great advantage both in display and The observed value has thus been increased by a
data collection with a small computer, particularly factor m. In the binary counting operation it will be
where the memory word size is restricted. For example, convenient to make m = 2”, i.e., M = 22”. This has the
in the 12 bit word PDP-8 computer more than one effect of increasing the square-root accuracy by one
location is necessary for P, whereas both S and T can binary place for each step in X. The algorithm for this
easily be limited to a single word each. The display extension of the method is given in fig. 2, and a
sequence then becomes a direct read-out of consecutive counting sequence in table 2.
locations. For a typical 10 bit resolution display system
restricting Sto 10 bits allows a total count of S2 = (210)2 Reference
which is more than sufficient for most experiments. l) I. N. Hooton, Nucl. lnstr. and Me:h. 56 (1967) 277.

You might also like