Fire risk from low power device?

[RogerThat]

Hi,

Does anyone know if there is a lower limit power (watts) which physically can not cause a fire on a PCB?
I'm looking for something in line with "Below 1watt power there is 0.000% risk of fire"

I have a device which runs from 12vdc (consumes 100mA) and is fused at 200mA(PTC internally which trips around 200mA and glass fuse externally ). I want to know the fire risk of 12v @ 200mA if there is a fault and non of the fuses trips.

[Andy Chee]

I'm looking for something in line with "Below 1watt power there is 0.000% risk of fire"

Such a rule of thumb would be too simplistic, because it doesn't take into account the surrounding materials, atmospheric environment, or failure modes of your 12V power supply.

Can you elaborate on the nature of your device?  Given your concern about fire, I'm picturing that it will be used in an oil refinery or similar hazardous environment?

[Haenk]

Running 1W through a 0.25W resistor will likely let it smoke.
Also there are 1W light bulbs, essential a red hot glowing wire, that would match a very thin PCB trace.

So I'd say "yes, 1W is enough to set things on fire".

[Solder_Junkie]

In order to reduce the risk of fire to close to zero, you need to go down the same route as intrinsically safe equipment.

Many years ago I used to service intrinsically safe hand held radios, they had sealed fuse units, tamper proof cases, coated print boards and current limiting resistors in series with any larger electrolytic capacitors.

In addition, you were limited to the number and nature of any repairs. Some had to be destroyed and paperwork supplied to the customer. It was taken very seriously.

SJ

[CaptDon]

A maximum open circuit voltage of 24vdc with a maximum available current of 20 milliamps was chosen for metering loops in the chemical and petro chemical facilities for its supposed inability to spark a fire. It was found however, with excessively long cable runs with an open loop showing 24vdc there were enough Joules of capacitively stored energy to spark a fire if the flammable mixture was exactly right and a spark was generated while shorting the tips of the open circuit. Depending on the design of your power supply it can be 'impedance protected' like motors where the impedance will protect from fire or burnout. Or, some power supply transformers incorporate a 'one-time' thermal fuse deep within the core windings which will trip to an un-repairable open circuit if the transformer overheats. Nothing is 100% fire-safe until you throw a bunch of money at it and test for every possible outcome. You can make a product 'foolproof' but you can never make it 'asshole resistant'!!! The letigious courts prove that every day!!

[coppice]

Its isn't high power which leads to fires. Its high power density. 1W in a tiny spot == fire risk. 1W spread over a large PCB will barely warm it.

[RogerThat]

thanks for the answers.

I've tested two things this afternoon.

1. 2.5W thru a 3W resistor. it stabilises at around 120-135c = not hot enough to cause fire. Especially not if mounted on a PCB board which acts like a big heat sink.
2. 200mA thru my smallest PCB track(0.2mm), not even warm to the touch. @ 3amps you can feel it.

I have electrolytics, they are 105c types and soldered down to a flooded ground plane. So, as with the resistor will not be hot enough even if they failed to exactly the right resistance.

[Solder_Junkie]

Are you looking to minimise fire risk in a domestic situation, or for industrial sites where flammable vapour or dust is present?

For the latter, you need to guard against a tiny strand of wire floating around inside the case and shorting out a circuit with enough energy to cause it to heat up, or cause a spark. Conformal coating helps minimise such risks, but a thick layer of heat resistant silicon is generally safer.

[tszaboo]

Are you looking to minimise fire risk in a domestic situation, or for industrial sites where flammable vapour or dust is present?

For the latter, you need to guard against a tiny strand of wire floating around inside the case and shorting out a circuit with enough energy to cause it to heat up, or cause a spark. Conformal coating helps minimise such risks, but a thick layer of heat resistant silicon is generally safer.

For the latter you need to fulfill the requirements of the ATEX directive in Europe, and have a notified body underwrite your design and production site to be able to make electronics. It takes a lot of time money and effort to make very simple electronics safe, and we shouldn't advise through a forum.

[RogerThat]

To keep on track, it's for domestic use.

Maybe turn the question around a bit. Can anyone make(or think of) a circuit from just passives that can put a PCB(std FR4) on fire with maximum 12v and 200mA (2.4W) at room temperature? Let the games begin.

[Randy222]

To keep on track, it's for domestic use.

Maybe turn the question around a bit. Can anyone make(or think of) a circuit from just passives that can put a PCB(std FR4) on fire with maximum 12v and 200mA (2.4W) at room temperature? Let the games begin.

Regardless of passives or not, almost all joules consumed ("power") get turned into heat, for most things.
2.4w in a tight box will cause temp to rise to high number. How high depends on how good/poorly the box can sink out heat. I guess FR4 could "burn" in a box until all the oxygen has been depleted.

 

[Ian.M]

Maybe turn the question around a bit. Can anyone make(or think of) a circuit from just passives that can put a PCB(std FR4) on fire with maximum 12v and 200mA (2.4W) at room temperature? Let the games begin.

I'll go with the nuclear option: an array of 6V solid Tantalum capacitors across the power rails!  Even assuming you limit the charging current to 200mA you are still likely to have a very bad day!

[Berni]

Sure run those 2.4W trough a tiny 0402 resistor and cover the resistor with something flammable. It will be on fire within a few seconds.

As for getting a sustained burn of the FR4 after power is removed. That is rather difficult to achieve even if you have a lot of power available. The FR4 material really does not like to burn. It is mostly made out of glass fibers and those obviously don't burn, the part that does burn is epoxy binders holding it together and those have to be self extinguishing in order to be called FR4. Otherwise you have a fake FR4 clone that is not actually FR4.

Go ahead and try burning some FR4. It won't sustain a flame, but it will smell absolutely god awful and make a crap ton of smoke.

When electronics catch fire it is usually due to the enclosure or wire insulation catching on fire.

[MarkT]

Its not just power you have to worry about, its stored energy too.  A good mitigation is ensure the enclosure will contain any circuit fire, as well as ensuring no fire can start...  A die-cast box is a great enclosure for this and other reasons (cooling, EMI).

[soldar]

A switching wall-wart PSU might supply 10 watt but will not use more than 1 or 2 watts.

Two or three years ago I was awakened in the middle of the night by an explosion in the room and it was one of those SMPS. I think the main filter capacitor had exploded and the entire case had exploded. Quite a loud pop!

[Randy222]

A switching wall-wart PSU might supply 10 watt but will not use more than 1 or 2 watts.

Two or three years ago I was awakened in the middle of the night by an explosion in the room and it was one of those SMPS. I think the main filter capacitor had exploded and the entire case had exploded. Quite a loud pop!

The case of that wart needs to sink out those 1-2watt to keep temp down. Wrap the wart in some styrofoam and it will likely overheat from just that 1-2watt it can't sink out.

[harerod]

Quote from: Berni on Today at 18:18:49

...
Go ahead and try burning some FR4. It won't sustain a flame, but it will smell absolutely god awful and make a crap ton of smoke.
...

Spooky, as if FR4 were made from some flame retardant material.

I am still undecided whether the TO just likes to play with stuff, or if this is some commercial project where management doesn't want to invest in competent help.

[Jwillis]

All fires require Three elements. Heat, an oxidizer (air), and fuel. If you remove any one of those elements you get no fire. The easiest way of making an electronic device fire proof or at least fire resistant is to use flame proof or flame retardant board and ,since the resistor is the most common source of heat, flame proof resistors. Essentially removing the fuel in both cases.
Encapsulation works to prevent electrical equipment that could potentially cause an ignition in an explosive atmosphere.Used a lot in the oil and gas industry. But many resins used for encapsulation are flammable. So choosing the correct type is important.   

[CatalinaWOW]

You are on the right track, but there isn't a simple to apply rule.  The simple rule is to keep all elements of your design below the ignition temperature of any materials that will be in contact with that element.  As pointed out in prior posts some applications may require that to be true through one or more layers of faults.  That may even be true for some domestic applications.

The simple rule is complicated to apply because:

1.  You have to calculate heating which requires both power in (the easy part, current squared times resistance or whatever) and power out (convection, conduction and so on).  And will airflow be the same over life with dust, filters and other factors?

2.  You need to really think through what materials will be present.  Kitchen applications leave the possibility of cooking oils and alcohols.  Garage applications add gasoline and other solvents.  Bathrooms with feminine presence have hairspray and nail polish remover among other things.  What is the ignition temperature after these things have condensed on your circuit cards, or been spilled on the enclosure.

If you are not selling your device you get to decide how thorough an analysis is good enough for you.  I wouldn't worry about it much as long as I wasn't working any components hard and had a good pwb supplier that didn't have any thin or narrow spots on traces, but then I store LiPo batteries and gasoline in my house.  Others have higher standards of safety.

If you are selling your device you need to make it hard to pin the blame for a house fire on you, regardless of  whether your device had anything to do with it.   The charred remains of your device showing significant disregard for industry standards will be very tough on you if there isn't any other obvious source of ignition.