Hi-Amps Dummy Load - my take

[erupter]

Hi people.
I have some current sensors laying around and I figure I better test them.
But I don't have a decent power source.
So I found out EEV's Dummy Load but it doesn't cut it in terms of power: way too small.
My sensors are for 30~50 amps, I gotta test them with tens of amps.

So I took pen and paper and started designing my own dummy load.
It makes use of one of the same sensors it ought to test, you can find them at this link here.

These essentially output a voltage proportional to current with vcc/2 being zero amps.

The circuit is powered by AC/DC, I figure +/-12 V is sufficient here and gives any opamp enough room.
The idea is that on one side of the main opamp you select the current with a trimmer. The ratio is in the sensor datasheet but is inessential, just some math.
The other side, the feedback loop, comes directly from the sensor.
The opamp drives an array of mosfets mounted on heatsinks.

What worried me is the transients when attaching or removing the supply under test.
So I devised what you see in the scheme.
First of all components are not what is shown on the scheme: I didn't take the time to search for real components, I just used the schematic symbols.
So the 7805 is not a linear regulator, rather depicts the current sensor.
The arrow Vcc is the circuit power supply, while the ball Vcc is the connection to the external supply to be tested.

There is a shallow low-pass filter on the opamp output to the mosfets, to dampen spikes. But it's shallow anyway: leaves plenty of bandwidth (I figure something in the KHz range or maybe even 100Hz).
The same opamp output is clamped to ground by an always-powered humble transistor T1. This is my security.
But the base of this transistor is clamped to ground too by T2 when at the external input is applied power.
There is a very low low-pass on this line as I want current to ramp up slowly when connecting.

But I just figured that a simple bjt may fail if the external power doesn't have high enough a voltage to polarize it.
So since the current sensor can be powered as high as 5V, I think the logic will run at 5V (apart from the opamps that can get 12V).
So... with anything higher than 5V I'm sure the BJT will be polarized.

Suggestions? Comments? Corrections?