PolySwitch (and other PPTC) advice needed

[JohnDF]

I am considering using some type of PPTC (specifically the PolySwitch by RayChem) as a "fusing" solution for a board I am designing. I am looking to provide protection on 5 and 12 volt lines.

I'd like some input from others who have used this product before. After reviewing the data sheets, it seems to me that the dependance on ambient temp could be problematic. I am designing for a large operating temp range, and the current to trip the device varies greatly with ambient.

Does anyone have a better explanation than the data sheets as to what exactly Ihold and Itrip currents are? It seems there is a large gap between the two for which operation is undefined.

Thanks for any input,
JohnD

 

[VEBill]

"5 and 12 volt lines" Input or output from your board?

Assuming they are outputs from your board, then you'd need to design the board (its actual failure limits) accordingly.

It might be easier to use a more active current limiting scheme (foldback).

 

[JohnDF]

Ah, pardon my ambiguity. These are the 5 and 12 Volt inputs that provide power to the board. Power is connected through pins on a backplane. I am beginning to wonder if at boundary temps, the PPTCs will not function properly as a fuse (Either not detect the fault or trip without a fault)

 

[VEBill]

If you're protecting inputs, then you need to consider if you really require the PPTC advantage of 'self-resetting'. I mean, if your board fails and draws too much current, it probably needs repair anyway (to fix the root cause) so replacing a good old fashioned fuse at the same time isn't such a big deal.

Most protective devices would be located at or in the PS to protect the PS and the wiring.

If this is a high power system, then you may need to worry about wiring topology. You might not want dasiy-chain high current wiring going along to dozens of mid-power boards.

By way of comparison, have a look at the specs for typical fuses and circuit breakers. Those typically take a verrryy looonnnng time to blow if the overload is 'only' something like 150% of the current rating. If you want them to blow/open 'instantly', then you need to provide something approaching a dead short.

 

[JohnDF]

Ah thanks. Cleared a few things up. This also bringss up another question I had, if you don't mind. In theory, a dead short would produce infinite current. But how much current will actually be produced if a short occurs? Will the PS simply output its max current? Will there be a very large instantenous current? I've been thinking about what actually occurs when a short from Vcc to ground occurs.

 

[VEBill]

Well, Ohm's law still applies, so I=V/R. The question then becomes: how much residual R have you got?

It depends on how the PS is built. High current power supplies would obviously tend to have a higher peak fault current than lower power supplies. For example - a car battery can source 1kA or more, while an alkaline D cell might be good for about 2A (YMMV). Modern power supplies often provide very nice current foldback circuits - these might be set just above the rating.

A more interesting, complicated and important question is when things go wrong, what exactly happens and is it safe?

I've noticed that the household wiring standards (North American electrical codes) are pretty good with respect to making sure that the protective devices (circuit breakers) are in the right place and connected in a manner that provides very good protection. Basically, they make sure that the circuit (wiring and everything downstream, at least as far as the outlet) can handle all the available current.

 

[OperaHouse]

Like nailing jelly to a tree! Actually any thermal device is quite dependent on temperature. The PTTC devices I've used are rated at 25C, but require a temperature of around 100C before they go into protect. I've used the free standing ones and played with them a bit. As they just start to switch, putting your finger on them can reverse the cycle. Use a SMT and that has to make it very complex depending on the ability of the board to sink heat. Given that many board traces will handle 20A, these will easily protect the board from major shorts. I like them, but they are in no way precise. Put fuses under a microscope and PTTC devices look pretty good. In real life it is pretty hard to get currents over 20A when you add up all the wiring resistances.

 

[felixc]

You want to protect your board from backplane voltages that may not be good, or protect the rest of the system in the case that your board fails with a current overload?
Like most said, the polyswitches are slow to respond if the overcurrent that you want to monitor is within the same numbers as the normal board consumption. If the failure modes involve much larger currents, then the polyswitch may be a good device. And yes be very careful with the temperature deratings.
If you use the SMT versions, be careful when you solder them. They really do "fuse" and can get melted down and destroyed by the soldering heat.

 

[OperaHouse]

For those that have never played with these, they require a current to maintain the "protect" state. This is about 3W for the RXE series. As an example, the 3A (up to 50V) version continues to flow about 300ma in a 12V circuit witha a dead short, at 24V it is half that, yada, yada, yada. The current is required to keep heating the element to that 100C.