Very low leakage semiconductor switch. 3

Your approach seems counter-intuitive to me.

Assuming that your "test" premise is true, then such conditions are likely to be short-lived, as the small components will eventually fail into an open-circuit condition, whereas replacing them with high-power capacity components means that your so-called "test" conditions will prevail for a longer period of time and potentially cause more damage, because the devices that you will insert WILL continue to supply current into a potentially damaged load and subsequently damage the battery as well.

Additionally, by altering the off-the-shelf battery, you potentially incur additional handling damage as well potentially NOT providing the same level of functionality that the existing circuitry provides, thus lowering the overall safety of the system.

And, even if a particular part exceeds some surface temperature, so what? If it's not immediately accessible to the user or maintainer, what is the issue? Usually, a user or maintainer will need to disable the system to service any failure condition, so the issue really should be whether the part in question is still at an unacceptable condition by the time they get access to the part.

Furthermore, if the issue is simply a burn hazard, wouldn't it be easier to simply pot or enclose the devices in question?

TTFN

Irstuff,
I have to disagree with you to some extent. A semiconductor device can just as easily fail short-circuit, in which case the power dissipation can be low enough that it can continue to pass the fault current. A melted junction is not guaranteed to fuse and should never be considered to do so for safety purposes.

Thanks IRstuff.
My thoughts exactly when I got involved in this project.
In the normal world everything you say would be true BUT in the Intrinsic Safety world there is much more to consider.
The end product will be used in an atmosphere which could be explosive. The test gas mixture they use has a low ignition temperature.
The obvious thing to do is put the whole thing in an explosion proof container, but for reasons which I am not at liberty to disclose, this is not possible.
The next thing to do is use an epoxy potting compound, but this introduces problems we would like to avoid.
The circuitry we have in prototype form fulfils all the criteria required except for the battery protection circuit.
It is fused using a Baseefa approved 1 Amp fuse.(i.e. wont ignite an explosive mixture when it blows.)
The test parameters assume the fuse will carry 1.7Amps for a sufficiently long time before it ruptures.
The test requires that you introduce two worst case faults anywhere in the circuit and the surface temperature of any component shall not rise above the specified temperature for the protection level you require. The only components which are deemed unlikely to fail are specific classes of resistor.
At 1.7 Amps the battery should be very happy so overcurrent discharge is not really a problem. The resistors we have used are happy at worst case fault currents and run cool enough.
The semiconductors we use in the high current path are beefy enough to run cool if worst case faults occur.
The problem we have is the battery manufacturer insists that a cut-off switch should be included in the circuit to isolate the battery if the voltage drops below a critical level. This needs to be a semiconductor device if possible because mechanical contacts are bulky and need to be properly enclosed.
To prevent deep discharge this switch (they claim)should have a leakage current of less than 1 microamp.
Their circuit uses a miniscule S08 package which achieves this. BUT the test method assumes that this device could fail in such a way that it could be equivalent to 4 ohms(the worst case for max power disipation at 1.7Amps) !
Under those conditions it fries for long enough to ignite the test gases.
I suppose one could argue that this is highly unlikely, but when lives could be at stake we have to stick to the rules.
Which is why I need a bigger device with very low leakage if it exists.
I live close to a little village called Abbeystead where a group of visitors died in an underground reservoir because one of them carried the remains of a stamped on fag under the sole of his shoe into an atmosphere laced with methane.
They didn't know it was laced with methane, he didn't know the fag had stuck to his shoe. A highly unlikely scenario that did happen.
So I have to take the test house experts at there word when they say "if we can make it happen then the chances are it WILL happen"

Any of these?


TTFN

Thanks again IRstuff for the link.

The leakage figures are fine BUT I don't think the devices will pass the 700mA required by the circuit when switched ON.

Back to my hunt!!

If someone makes a device with those leakage currents but with a 0.5 ohm ON resistance under normal operating conditions then we might be in business. BUT it must also be capable of dissipating 4 watts in fault mode at a maximum case surface temperature of 150C when mounted on a suitable heat sink. This more or less implies that it needs to be a device with a reasonable bonding pad area to channel the heat into the sink.

A reed relay is hermetically sealed and will not cause a problem in an explosive atmosphere. However, the reed switch will not switch more than about 60W. However, any relay can carry more current than it can switch. Switch the current off using the semiconductor, complete with gross leakage. Then isolate the battery with the reed relay.