Dear Members

I am trying to interface 89C2051 uC with a magnetic pulse generator attached to a DC motor. The problem is that I am unable to count exact pulses. here are few details.

uC is AT89C2051 running at 12Mhz crystal.
Motor is 24VDC,speed 3350 rpm.
Magnetic pulse generator operates at 5VDC and generates 2 pulses per revolution, and has 2 channels.signal rise time and decay time is 85 ns typical and 400 max.

I have used a counter in 16 bit mode to measure pulses, as motor is running on 3350rpm and encoder generates 6700 pulses in one minute(2 pulses per rev). My program starts the motor and count pulses and as counter counts 6700 the motor stops. so in simple words i am trying to run motor for one minute using feedback from encoder.
The problem is that instead of running motor for 1 min the motor runs for only 10 seconds and stops, I have tried different values for counter but no good at all, could anyone please pass any suggestion and share there thoughts on this,

Below is the program.

ORG 00H
LJMP MAIN

ORG 00BH
LJMP T1ISR

MAIN:

MOV P1,#0FFH
MOV TMOD,#05H     ;counter in 16 bit mode

SETB EA
SETB ET0

START:

     JB P1.7,START        ; press button and start motor
     CALL DEBOUNCE
     JNB P1.7,START     
 
MOV TH0,#0E5H             ;loading values for counter, to count 6700 pulses 
MOV TL0,#0D3H

CLR P1.5                  ; start motor
CLR P1.4                  ;LED  
SETB TR0

HERE1:

JNB P1.5,HERE1            ;wait for motor stop and jump back
                          ;to start
 
 CLR TR0
 JMP MAIN

T1ISR:                     ;timer interrupt    

CLR TF0

SETB P1.5                  ;stop motor
SETB P1.4

RETI

DEBOUNCE:

        MOV R6, #09H
        
 LOOP4: MOV R7, #09H
        
 LOOP3: MOV R4, #09H
        
 LOOP2: MOV R5, #09H
        
 LOOP1: DJNZ R5, LOOP1
        
        DJNZ R4, LOOP2
        
        DJNZ R7, LOOP3
        
        DJNZ R6, LOOP4
        RET
END

I have connected the channel A of pulse generator to pin 3.4 of uC which is T0. I am not using second channel of pulse generator at the moment.

My idea is that as it requires 2 machine cycles to recognise 1-0 transition on P3.4 so may be I am missing the pulse, but is there any way of getting around it?

any thoughts would be much appreciated

Regards
James