3 phase over and under voltage controller

Hi guyz...........im currently working on a project in which i have to control a 3 phase pump that if the voltage of any phase goes below or over a thredshold the pump is turned off. For that im first stepping down the 3 phase via transformer to 12Vrms then passing each phase via a bridge rectifire to generate D.C voltage which should give 17 volts peak after capacitor filter and then ill give the D.C to a potentiometer and scale it to a level and give it to the A/D channel of microcontroller. But the problem is when i simulated a part of circuit on proteus i got something strange the two phases are giving proper voltage 17 volts D.C but 1 phase gives 19 volts D.C. sometimes more than 20 also. Whats the problem with that please tell me or im using the wr ong approach from top to bottom. The picture of proteus is attached for help.

Only problem I see is the rise fall time using no load and 80 uF.
Perhaps you need to define the duration for detection before Disable and recovery.

.-.-.-.-.-.-.-.- Disable .-.-.-.-.-.-.-.- Recovery
Vmin , T1
Vmax, T2

may i sugest that you OR all these dc to a single point and use it for detection +/- one detector

You've got one capacitor the wrong way up.

You need a load resistor in parallel with each capacitor, even if it is only for the simulator.

You need to make sure that the 3 phase supply is referenced to ground. On a real circuit, the lower side of each transformer primary should be connected to the neutral wire and not earth.

Which output is giving the incorrect voltage?

For better accuracy you might RC (or active) low-pass filter the output to get the average voltage value rather than the peak. That will be less susceptible to spikes, harmonics, and noise on the power lines.

Unless there is a direct reason, I'd separate the AC system GND from the measurement system GND. There is a tolerance value in a transformer's secondary rating, but you can compensate for that with your pot in each channel.

If you upgrade to bridge rectifiers, you can shorten the timing of both detection windows.

ak

analog1 said:

may i suggest that you OR all these dc to a single point and use it for detection +/- one detector

Click to expand...

That would work only for the overvoltage condition.

ak