But now, i understand because the HCxxx buffers works in AT89sxx programmers. The IC´s datasheet show that buffer reponse time is 9ns!! What more fast than 9 ns?!

I do not understand. What I said was not for you to find something faster than a HC chip, I said not to use the CD4xxx which is slower. HOWEVER, if I recall, rhe authorized design uses 74VHC which IS faster. Many have stated that HC is 'fast enough' I have no idea, but when VHC is the 'official', HC may or may not work I would not even try CD4xxx

CD4010 tPHL Propagation Delay @ 5V: max 70ns
74HC14 tPHL Propagation delay @4.5V max 25 ns
74VHC244 tPHL Propagation delay @5V max 6.5 ns

PS: when you gaze at a CMOS datasheet like http://www.fairchildsemi.com/ds/74/74VHC244.pdf you will see how much even small stray capacitances will affect the propdelay of a CMOS chip which explain why so many "flying lead" 'designs' fall flat.

Erik

let me repeat: when copying 'programmer' circuits make NO substitutions, if it states a red resistor, get a reed one