TPS2492 Operation

I am not sure I completely understand the question. The device can operate in an unloaded (open circuit) condition between the OUT pin (11) and GND pin (7).

Actually, I was looking to add the resistor between the FET source and Pin7 of the MSOP pkg TPS2491 (or TPS2492).  The IS examiner can use an arbitrary fault (he is allowed one such fault to analyze what happens in a fault is imposed, such as an IC failing) to short the IC internally. He could therefore short pin 9/10 to pin 7 and completely connect input to output, bypassing the protection function completely. If there were a resistance between pin 7 and the FET source, it wouldn't allow much current to flow. The LT4356 has the same problem and they don't allow any resistance there or the part won't function.

So my question is this: is resistance allowed between pin 7 and the FET source?

Pin 7 provides a reference for the internal zener diode which protects the MOSFET gate. Isolating the MOSFET source pin from pin 7 will degrade the effectiveness of the zener.

Is the described fault injection method only used on a limited case for this test? Then is the circuit re-configured as recommended (no resistance between MOSFET source and pin 7)? The TPS2491 will function with some resistance between the MOSFET source and pin 7 as long as the resistance is limited. How big of a resistor do you want to install?

Is the described fault injection method only used on a limited case for this test?

No, whatever is done remains part of the design as the entire design is analyzed for Intrinsic Safety compliance and is frozen after certification is granted. 

Then is the circuit re-configured as recommended (no resistance between MOSFET source and pin 7)?

How big of a resistor do you want to install?

My purpose for the 2491 is to clamp the output power available in case of a fault in the load. If the load shorts, for example, I need to limit the output current to the preset safe value. If the 2491 is shorted internally by the IS examiners for analysis purposes then a direct connection exists between the input and the output with no current limiting. Therefore the resistor in question would have to be of such a value to limit current to the same level as the 2491 would do if functioning properly. In most cases our input voltage will be 5 to 12VDC with output currents ranging from 50 to 200mA. Taking worse case 12V/0.2A = 60 ohms.

However, this resistance must not affect the clamping time of a load fault transient. The IS agency allows what they call 'let through' energy during such a fault. This is measured by integrating the product of the output voltage and output current (output power) over time to obtain the total 'let through' energy. The limit for this depends on the hazardous gas group being certified for, but in our case it is 120 uJ. So you can see that both current and voltage must be clamped quickly, a few microseconds, to meet this energy limit.

If the concern is a short internal to the TPS249x, another approach might be simply to put a fuse between the source of the FET and Pin 7.  Although very little leakage current flows into this pin in normal operation, a short circuit at the load might try to pull current out of pin 7 unless there is a clamping diode at the output to prevent this. As mentioned above a resistor will impact the accuracy of the Vds detection. Of course if this feature is not being used (i.e. Prog=Vref) then you can put a resistor between the source and pin 7 with minimal impact to functionality.

I hope this helps.

John,

The fuse suggestion is a good one, and would solve any thermal ignition problems (slow) downstream from the converter, but not spark ignition (fast).

The IS agency considers only resistors as a passive solution to spark ignition. For Zone 1 applications we are allowed to use active current limiting, such as the TPS249x if we can demonstrate that the let through energy is low enough, as discussed earlier. For Zone0, no active protection is allowed at all. There are multiple problems associated with obtaining compliance for active current limiters and voltage clamps. Zener diodes are considered fast enough to clamp voltage for spark ignition, but won't limit current. This is a very difficult problem to solve, not so much technically as compliance-wise.  We have yet to find a solution.

If I understand the Prog pin function, defeating it would defeat the current-limit feature of the device, making it useless for my application. Am I missing something?

At this point we are not sure whether we need to go with Zone 1 or 0.

Thanks for your help on this.

Hi Kevin,

The prog pin is used to set the Safe Operating Area (SOA) of the pass FET. By tying this pin to Vref, this function is disabled. However, the current limit to the load is still fully functional. The issue you need to pay attention to is now you will have to size the FET and/or put an RC at the gate to slow the ramp of the gate to help minimize inrush. So, this might be a possible path if you choose to add a resistor. However, with a fuse approach, you will be able to take advantage of the SOA protection feature and have a much better controlled inrush current profile.

Regards,

John

Kevin,
For the voltages and currents described the TPS2420 family may be a better fit.. ( TPS2420/21, 2590). These devices have internal FETS and internal power limiting of the FET.  And they all have UL certification.

Hi Jim,

They prefer the internal FET versions because of the fewer pin connections needed, the less the chance for the IS agency to short around the device.

Thanks,

Kevin