Overvoltage protection

Mark Ebert wrote: > I'm looking for a method of protecting a circuit from overvoltage > (perhaps by turning off a FET in the presence of higher voltages). > The input voltage is normally +5v, but if plugged into the wrong > system it approaches 30 volts which destroys the downstream > circuitry. I'm trying to find a way to isolate the overvoltage > without shorting the input (so a simple fuse or TVS is out), but > have been unsuccessful so far. My original plan was to use a > comparator to switch a MOSFET off when a pre-determined threshold > is exceeded, but I'm unable to come up with a proper driver for > the FET gate. > Any suggestions would be greatly appreciated! >

A problem with turning OFF devices is it assumes they still work. Series fuse, 5 watt 6.2 volt Zener to ground. Fuse is a 'Dang' indicator Use a self healing poly fuse if you don't want to replace the fuse every time.

I don't get how a fuse shorts the input.

GG

There are things called LDO's. If you can't find one suitable then make one. A PMOS type set to regulate at say 5V +15% will be fully driven to RDS,ON with 5V input/output, but clamp at 5.75V output. There will be thermal protection considerations if the loading is heavy. Or maybe you want to just shut the thing down if 30V is applied, just extend the LDO idea except now it is the input and not output that is sensed.

Please define "a circuit". Is the input a logic signal to the circuit, or does it power it? How much voltage does the circuit actually need to function?

John

John Larkin wrote in news: snipped-for-privacy@4ax.com:

overvoltage

voltages).

circuit,

need

Actually, the "circuit" that is driven by the 5 volt supply is just a series of backlight lamps for a control panel. The normal voltage range is 0-5 volts (dimming), but if the panel is plugged into the wrong product (we have field problems with this), it burns out the bulbs. We have used fuses to protect this in the past, but because of the location of the PCB its a pain to replace so I want to improve on it to provide some sort of overvoltage sense, then open the output in the presence of this. The current draw is 2.5 amps max. and the inout voltage can reach 32 volts, otherwise I would use a commercial overvoltage device, such as a On Semi NCP346 driving a FET.

Polyfuses are near useless in overvoltage situations. When the voltage goes too high and the crowbar fires, a polyfuse will usually be destroyed, unless something else protects it..

Check out the "Max. current" ratings of Polyfuses.

Sounds like you could use a polyfuse followed by a 5-volt power zener or transzorb.

Or just a 5-volt LDO regulator, although the 2.5 amps makes it tricky. If you could split the load into sections, 3 maybe, you could use three small ldo's. Below 5 volts, output would track input pretty well, but above 5v in they'd regulate and eventually thermal limit.

John

In article , Mark Ebert wrote: [...]

How about this:

TIP-36 e c --+-------+------ ----------- To bulbs ! ! \\ / / ! ----- \\ ! ! / 0.5W \\ ! ! 2.2K / 1W \\ zener/-/ \\ 1K / 50R 4.3V ^ ! \\ 2W ! ! 22K ! +-----------/\\/\\-+ ! ! ! ! ! ! ! ! !/ c ! --+------! 2N2222 ! !/ !\\ (2 EA) +---! ! ! !\\ GND / ! \\ GND / 1K ! GND

Maybe a LMV431, BJT, MOSFET + 5R would do the trick.

Best regards, Spehro Pefhany

--- This is easier:

PCH Vin>--+------+--------+---S D----->OUT | | | G | [10K] [10K] | | | | | [10k]--+-----+ | +--| /U1A LM393 | | | [LM385-2.5] | | GND>--+------+-+------------------->GND | +--------+ | | +--|+\\ | | | >--+ +--|-/ U1B

You don't need the gate resistor, and just set the pot for whatever turn-off threshold you want.

Here's the LTSPICE circuit file:

Version 4 SHEET 1 880 680 WIRE -256 272 -256 -16 WIRE -256 432 -256 352 WIRE -256 464 -256 432 WIRE -144 272 -144 160 WIRE -144 432 -256 432 WIRE -144 432 -144 352 WIRE -16 -16 -256 -16 WIRE -16 16 -16 -16 WIRE -16 192 -16 96 WIRE -16 272 -16 192 WIRE -16 432 -144 432 WIRE -16 432 -16 352 WIRE 144 160 -144 160 WIRE 144 192 -16 192 WIRE 176 -16 -16 -16 WIRE 176 144 176 -16 WIRE 176 432 -16 432 WIRE 176 432 176 208 WIRE 256 -16 176 -16 WIRE 256 16 256 -16 WIRE 256 112 256 96 WIRE 256 176 208 176 WIRE 256 176 256 112 WIRE 416 112 256 112 WIRE 464 -16 256 -16 WIRE 464 96 464 -16 WIRE 464 272 464 192 WIRE 464 432 176 432 WIRE 464 432 464 352 FLAG -256 464 0 SYMBOL Comparators\\\\LT1017 176 176 R0 SYMATTR InstName U1 SYMBOL res -32 0 R0 SYMATTR InstName R1 SYMATTR Value 10k SYMBOL res -32 256 R0 SYMATTR InstName R2 SYMATTR Value 8200 SYMBOL voltage -144 256 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 2.5 SYMBOL res 240 0 R0 SYMATTR InstName R3 SYMATTR Value 10k SYMBOL voltage -256 256 R0 WINDOW 3 24 104 Invisible 0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V2 SYMATTR Value PULSE(0 30 0 1) SYMBOL pmos 416 192 M180 SYMATTR InstName M1 SYMATTR Value Si9407AEY SYMBOL res 448 256 R0 SYMATTR InstName R4 SYMATTR Value 2 TEXT -290 506 Left 0 !.tran 2

I simulated with an LT1017 because I don't have a model for an LM393, but I can't imagine why the LM339 wouldn't work.

-- John Fields Professional Circuit Designer

..........This might help....

Regards

AJ