Fabula 



investigators using an empirical concentration correction to the t^ formula, in 

 which [rj] is replaced by (v- Vo)^^'Vq, so that 



c'NAkgT 



The concentration dependence is not very large for c' [r;] « 1, as in this work, 

 and its effect on Eq. (7) is reduced by the factor v- ^^^. 



We can make only a crude prediction of t^ with Eqs. (9) and (10), because 

 of the unknown molecular weight distributions of the industrial Polyox polymers 

 used. Thus for P301B with Iv] ~ 20 dl/g and fri ==4x10^, we obtain t^ « 0.002 

 sec from Eq. (9). With this value and v = 0.01 cm^/sec, Eq. (8) gives a required 

 e of about 10"* cm^/sec-^. This value is extremely high compared with values 

 in the present grid-turbulence flow. With m/Uq^ ~ 12/34 sec Vcm^, the required 

 value of eMAJg^ is about 3, while from data in Ref. 1 the extrapolated value for 

 x/M = 5 is about 0.01. Five mesh widths is as close to the grid as one might 

 hope to find some degree of turbulence homogeneity. Thus the dissipation rates 

 in the grid-turbulence experiment are too low by a factor of 300 to expect any 

 effect on the turbulence spectrum. If we assume that the ratio of times in Eq. 

 (8) needs to be only about 1/3 instead of 1, the dissipation rates are still too low 

 by a factor of about 30. Thus no viscoelastic effect upon the turbulence spec- 

 trum in this work is expected. 



The need for matching the viscoelastic relaxation time to the minimum tur- 

 bulence characteristic time became clear early in this work. It was thought 

 that polymer concentration could be increased to produce the desired magnitude 

 of t_^. The task of solution preparation became formidable when the master - 

 solution technique was introduced in an attempt to solve the raggedness prob- 

 lem. Nevertheless, much higher concentrations than 140 ppm could be used in 

 the towing tank. However, the present results indicate that this is not worth- 

 while until the signal raggedness can be eliminated or circumvented. 



The raggedness problem also prevents a useful increase in (e/-u)^'^ which 

 might more nearly satisfy Eq. (8). This is seen when possible changes in the 

 grid-experiment parameters are considered. While energy decay rates can be 

 increased by going to higher grid solidity, problems of flow instability (17) limit 

 the increase in solidity. Thus eU/V^ versus x M for the present grid can be as- 

 sumed to apply. As discussed in Ref. 1, there is little dependence on u^u u. 

 Thus at a given x/M, e can be increased in proportion to the increase in Ug^/M 

 which is consistent with other requirements, namely avoidance of raggedness, 

 adequate sensor frequency response and spatial resolution, and a suitable tur- 

 bulence Reynolds number. The raggedness problem imposes the most severe 

 restriction, since the time from grid passage for a given x M is proportional to 

 MAJq. If the decay of raggedness is a process of disentanglement or relaxation 

 of clusters formed or changed in character by the bars of the grid, as discussed 

 next, then if Ug/M is increased, it may be presumed that even greater solution 

 aging will be required before signal raggedness disappears at a given x m. 

 Since major solution degradation was suggested by the loss of specific viscosity 

 when signal raggedness disappeared at x m = 10.7 in the present experiment, 



68 



