Location of hot-swap controllers - backplane or peripheral card?

Hi Rob,

Below are the reasons why we generally see customers put hot-swap controllers on the peripheral card rather than on the backplane.

• If the hot-swap is on the backplane and the the back plane is live (i.e. hotswap is ON), once you insert a peripheral card which has huge capacitance at its input there will be current limited charging of capacitors. If the capacitors are not able to charge (with current limit) within the timer duration, the hot-swap turns OFF. The timer duration cannot be increased beyond a certain value due to limitation of SOA margin of the FET's. In these cases, we generally opt for dv/dt based startup which is possible only when the hot-swap is turned ON along with the total capacitance at its output.
• Every peripheral card have different current ratings and different values of capacitance at their input.  The design of hot-swap is dependent on these two variables. As there is no one design fits all kind of hot-swap design, it is better to have a hot-swap on each of the peripheral card designed for its specifications.
• As you mentioned, backplanes are generally passive in design

Thanks for the quick and helpful reply. I understand your points.

In our design we can use a separate controller to ensure the hot-swap is off when a card is not inserted (by detecting the card using an I/O signal). We can then turn the power on only when the full card capacitance already present. Furthermore, the peripheral cards have essentially the same power requirements in our case.
Our backplane already has active components on it, so it's not an issue to put hot-swap on there too.

I therefore think it will work for us to put the hot-swap on the backplane, even though it's unusual. The advantage of being able to current limit prior to the peripheral board is useful for us: mainly because if there's a short on the peripheral board, we can prevent over-loading the power connector.