We are developing machine using 11 servo controllers of Allen Bradley K300 .so far 5 drives are blown during mains power on . It works perfectly during operation.

Congrats....
Is that dhoom of 5 drives??
May be have to do some tuning with the PID in current loop section.
How about the high and low current limit setting???
Servo Drive wont blow out easily as its has perfect striping circuit before it malfunction you may just require a fine tuning in ramp and current loop side.

current limit is ok it is 3amp. Rockwell people suggesting to put Semiconductor fuse in all 3 phases.(supply voltage 220v phase to phase).Our management not ready since it is very costly.

Try line reactors which limits the peaks the capacitors inside the drive draw from mains, size and cat. nr. is usually found in the manual.

Apart from any other problem there could be, drives must always be protected with semiconductor fuses which react according to the trip times of semiconductors inside drives which you are intended to protect- If you use an electromagnetic device instead, by the time it trips the drive is already burnt. We have the same problem here in our country. Because of their cost, customers refuse to use them, but the user manuals are clear about stating this spec.

This is one of the lenze brand labeled ACTech PositionServo 940 drives.
The fusing is clearly listed as fast acting semi's, in the manual you can google.
I suspect they will "pop" far too often on you. You may have abnormal di/dt on power
step input. Txfmr used ? what kind...is this 3 wire 3phase or not ? You may have
groung reference problems, and/or high field cable cap is an issue. How is the braking done Dynamic etc... may have a brake issue on starting.... Most likely a Line reactor
"off the shelf for HP" will probably resolve most di/dt issues like this quickly.

Turning mains on is allowing to much disturbance on the power source; any equipment in the on position, draws immediately on mains power, contributing to the overall disturbance. Try to stage the power on process, much like they had to do on Apollo 13 by the way, to minimize the simultaneous disturbance and addition power consumption. You should be able to program it using what ever system you have in place already, sometimes, probably, maybe.

Is there any possibility these blows occur when you are connecting/disconnecting dc power supply cables. Even lifting the ground connection for a moment for unrelated reasons would explain the problem.

If so, I can elaborate if you come back to me.

Upto my knowledge i can suggest you the following point.

1.Use a 3 phase Choke at the incomer side to avoid the spike voltage.
2. Had you used proper Resistor for dynamic Braking as per the duty cycle?
3. Increase the resistance with 2% than the prescribed.
4. Use the exact wattage for the selection of Dynamic Braking Resistor.
5. Never design a circuit without an incomer semiconductor fuse for a drive application.
6. Does the servo motor voltage matches the Controller voltage? If u use a lower rated controller will leads to the problem like the one you are facing.
7.If your incoming supply contains Harmonic then use proper filter to filter it, Since drive are sensitive to harmonics. [I not so confident about this point that will help you, but make a note of it.]
Since servo are sensitive application so check all the possible way to troubleshoot it out.

Kind Request :- People here in Rockwell India are not expert in drive application to give a quick solution. Make note about your problem to Rockwell USA.

Regards,
K.Prabhu

What is the machine you are controlling?? How would you describe the mechanical load and how is the drive configured to manage the load?? If this is not done correctly, many fuses could be blown!. Once I burnt up a braking resistor because I'd underestimated the inertia of a pinch roll+ moving metal-mass used for tail-braking!. It was a wire-rod mill.

AS I have understood the problem, this is happening during power-up and no load.
Line reactors maybe combined with the Apollo trick described above will do the job usually. Another trick taught, way in the past ;-)
Is to turn each phase cable 5 to 6 turns around your palm, so it looks like a coil (Reactor ;-)..)