each filter has a Q of 0.7
When you cascade four of these, the Q of the whole circuit is 0.7^4=0.25
This is far too low.
You should alter the Q of each section to deliver a final Q between 0.5 and 0.7
ESP has altered the Q of each section to give a fairly fast turn over.
That's why he warns you not to experiment with changes.
This type of circuit is much easier to design and build if you adopt equal value Sallen & Key filters instead of unity gain filters.
The disadvantage of evsk is that each stage's gain is greater than 1 and the total gain is usually too high. This can easily be solved by using MFB filters for the two highest gain (highest Q) sections. MFB are inverting stages and using two cancels out the inversion.