High Voltage switch off system

[casemod]

Hi all,

I have a backup system that operates at 450VDC.

During Battery/Backup operation, the accessories operate from the main battery pack, however, for efficiency reasons I need to switch over to the rectified DC BUS when AC is available.

The Battery Backup has a DC voltage of 450V. The Rectified DC Bus is just rectified 240VAC (0-340V @ 100Hz).

What would be the simplest solution to cut the battery power and force the devices to draw power from the rectified AC line instead? I Though perhaps a couple of IGBT's, but I need to avoid the current to flow from the battery to the rectified input.

Any ideas welcome

- - - Updated - - -

Right, seems I cant edit the first topic.

Just to simplify what I need is a switch over system that feeds the load either from the 450V or from the rectified AC supply.

 

[Orson Cart]

A relay?

regards Orson.

 

[casemod]

A relay?

regards Orson.

I was looking for some solid state device to avoid arcing.

The point is to make the system as simple, and arc free, as possible.

 

[chuckey]

If you include a blocking diode in the output of the DC PSU with a reverse rating greater then the difference between the battery and DC voltages (450 - 340 = 110V, so 200V PIV), then the battery can never feed current into the Dc supply.
To switch off the battery you need a transistor in series with it with a voltage rating as above, or greater. You do not say the current capacity has to be.
Frank

 

[D.A.(Tony)Stewart]

Common cathode diode OR bridge a.k.a. "High man wins"

Since AC source already has a diode bridge , you can use wired OR logic with a reverse blocking diode for Vbat rated for I and V

 

[casemod]

Common cathode diode OR bridge a.k.a. "High man wins"

Since AC source already has a diode bridge , you can use wired OR logic with a reverse blocking diode for Vbat rated for I and V

The attached circuit represents my circuit. For simplicity i only drawn one of the PFC converters.
The top diode avoids current from flowing back to the PFC regulator. The load is PFC corrected, so I can supply it from the bridge rectifier.

What I am trying is to use for the switch over circuit is something on a single package, if possible, as the currents are about 50Amps.

Most 'brick' type IGBT's have a reverse freewheeling diode, which would complicate stuff (A pre-charge circuit would be required), as the current would flow to the uncharged DC-Link capacitors if the battery was off and AC was plugged. One option would be to replace D2 with a thyristor to switch off the AC input, but then I would end up requiring two devices...

Suggestions for simplification welcome.

 

[Orson Cart]

A bistable relay with snubbers fits the bill. In the commercial world we have used 2 x IGBT (600v 100A), emitter to emitter with isolated gate drive and mosfets too (source to source, 500V 80A) this gives an AC switch that is OFF when off (the gates are connected), you need to control the turn off time (~ 10-20uS) to avoid over volts at turn off and/or use snubbers/zeners...

 

[casemod]

A bistable relay with snubbers fits the bill. In the commercial world we have used 2 x IGBT (600v 100A), emitter to emitter with isolated gate drive and mosfets too (source to source, 500V 80A) this gives an AC switch that is OFF when off (the gates are connected), you need to control the turn off time (~ 10-20uS) to avoid over volts at turn off and/or use snubbers/zeners...

Have you got a part number for those IGBT's?

How would you calculate the snubbers for a relay application?

 

[casemod]

Ive done a few simulations. Pretty neat on a semiconductor, but useless for a relay operating on DC unless one would use a large R/C to divert the load long enough and slowly enough for the contacts to fully open.

I believe a better solution would be a PTC across the relay contacts. It should hold the load long enough for the contacts to open before the resistance starts to increase and the load current starts to fall, and even when it does it would do so linearly, controlling di/dt which in turn would avoid a large dv/dt.

Only drawback is that the circuit could not be reengaged before the PTC cools down, say 1 minute delay.

 

[Orson Cart]

A suitable IGBT is: IXXK100N60C3H1 (TO264, i.e. large TO247) two of these in inverse series (emitter to emitter, with gates linked) there are a number of others from other makers out there - easy google search.
To calc a snubber for a relay, for 600Vpk and 50A initial, choose 12 ohms & 100nF cap 630VDc rated and 4 x 120V zener in series with 4 x 120V zener the other way to limit volts on the cap.