ideas for Voltage surge on MAINS protection ??!!

[Electro nS]

hello everyone

At my house when voltage is switched from generator to mains (220v ac) in the ups/aps, sometimes the computer restarts and states overvoltage has occured , but most of the time nothing happens .

to solve this problem i am proposing 2 solutions :
1-design emi/surge filter using iductors/capacitors/and varistors.

2-using rectifying diodes and convert to Dc and filter by electrolytic capacitors . (i think computer PSU accepts DC 200v although not mentioned, right ?)

NOTE : WE HAVE NO GROUND ! (earth) in the building , so only Line and neutral

please state your opinions

 

[D.A.(Tony)Stewart]

Power Supplies for PC's are cost reduced so much that there is no excess storage energy to sustain a missing power cycle or two, so DC drops like a brick.I dont know the dead time on your UPS AC power transfer switch.A synchronous SSR transfer switch may help reduce the glitch by switching at a zero current crossing of the generator, but not cheap for commercial unit.Best advice would come from the UPS supplier since it is their design. But if the dropout on transfer is too long for your power supply, you could get a much bigger power rating PC supply to gain some storage time in the X rated caps. But this is only a guess and depends on dropout time or glitch that may be coupling thru the PS in transient suppression.THe probability of switching during the peak voltage may only be 10% ~ 20% so it may only happen 10 ~ 20% of the time, so it can be diffiuclt to diagnose without a scope on DC good signal and AC switch.

 

[Electro nS]

Power Supplies for PC's are cost reduced so much that there is no excess storage energy to sustain a missing power cycle or two, so DC drops like a brick.I dont know the dead time on your UPS AC power transfer switch.A synchronous SSR transfer switch may help reduce the glitch by switching at a zero current crossing of the generator, but not cheap for commercial unit.Best advice would come from the UPS supplier since it is their design. But if the dropout on transfer is too long for your power supply, you could get a much bigger power rating PC supply to gain some storage time in the X rated caps. But this is only a guess and depends on dropout time or glitch that may be coupling thru the PS in transient suppression.THe probability of switching during the peak voltage may only be 10% ~ 20% so it may only happen 10 ~ 20% of the time, so it can be diffiuclt to diagnose without a scope on DC good signal and AC switch.

thank u for the input.

i fully understand what you have said , but the problem is no the turn off of the computer due to the loss of electricity in the switching inside the UPS . what happened is that the contactor (house relay) switched very fast and the ups didnot notice a blackout and continue to supply normally which is good , but what happened is that the contactor switching caused arcing and high voltage spike or surge which was feed directly into computer power supply . which was outside the safe input region.
now i have designed a circuit that could deal with voltage spikes can u check out and give me feedback;

 

[chuckey]

Have you tried snubbing the coil of the change over contactor?
Frank

 

[D.A.(Tony)Stewart]

you have the right idea to filter incoming RF transient spikes. Caps must be rated for X or Y class voltage , plastic film, X-cap (=Line-line), Y-cap(= line-gnd)

Household power meter internal electrodes only clamp to 6kV slowly with ionization, so it is possible to exceed 1kV rating on 1st stage Cap. before MOV reacts. (depends on dv/dt)

MOV must sustain current from flyback of switched current, which is short, but adds stress to MOV unless choke inserted before, which may be my suggestion.

SO Caps ought to be rated for 3kV across line input before filter, additional snubber is useful.

Line filters are usually design for both ingress and egress, but in this case you are only concerned about ingress pulses.

 

[Electro nS]

you have the right idea to filter incoming RF transient spikes. Caps must be rated for X or Y class voltage , plastic film, X-cap (=Line-line), Y-cap(= line-gnd)

Household power meter internal electrodes only clamp to 6kV slowly with ionization, so it is possible to exceed 1kV rating on 1st stage Cap. before MOV reacts. (depends on dv/dt)

MOV must sustain current from flyback of switched current, which is short, but adds stress to MOV unless choke inserted before, which may be my suggestion.

SO Caps ought to be rated for 3kV across line input before filter, additional snubber is useful.

Line filters are usually design for both ingress and egress, but in this case you are only concerned about ingress pulses.

thank u very much , but as i said in the first post in the NOTE. i have no earthing so how does the above post diagram or schematic become ??
check this link i found online : https://www.diyaudio.com/forums/blo...6-wide-band-rf-filtering-mains-mainsfiltb.jpg
2- x-y caps are really expensive and hard to find here. but i have 1uF polyester 1000v , could they work even if not perfect for the job ? and if capacitors should be rated for 3kv , can i put 3 in series ?

 

[D.A.(Tony)Stewart]

It may be worth your while to install an earth ground wire. ( water supply plumbing )

The common mode chokes work better when there is a common mode capacitance to earth to create a high series and low shunt impedance ratio.


Polyester caps work well. X & Y class is for safety, which seems to be ignored without a ground. Transformers have a earth ground on neutral but long inductance and neutral can rise up to 10% of line voltage at full current.