Hi every body,

I am making a Liquid flow meter.

I am using a DS12C887 Timekeeping NVRAM with in-built battery and crystal with AT89C55WD,
Very much possible to use EEPROM and RTC on I2C but I have to work with the old hardware :-(.

Problem: -

When the controller is working as a master and I operate the motor the RPM is registered correctly in 12C887, but when I put my AT89C55WD in slave mode ( I am uing RS485 communication (75176 used in half duplex mode) for interfacing it with the PC based s/w through a 232 to 485 converter)
and when I try to operate the motor some times the entire memory is getting Blanked &/or my entire hardware gets reseted with the PC based software. (Like a hardware reset)

I tried earthing the communication cable at both ends i.e at PC end and at my hardware end / at indivisual ends.
I have used EMI filter in my power supply. (SMPS will definately be a better option but I cannt change the current hardware)

There is a supervisory IC by Maxim which is controlling my Reset signal to AT89C55WD (MAX691)

(The three optical sensors give the three in phase outputs. The three sensors have a 0.01uF caps connected to ground for bypassing any noise pulses)

If I simulate the pulses by multivibrators or even by using another controller(Hardware Simulator) then the problem is not faced in communication mode aswell.

But when I start the 230V AC motor (which is normally the case)through a relay the gost comes from somewhere.
A reverse biased diode is connected to the relay.

Please help.

Thanks in advance
Sachin