Why are the ISS' Sequential Shunt Units not on the inside?

Recently Tim Peake and Tim Kopra carried out an EVA to replace a failed Sequential Shunt Unit, some form of electrical regulator for the solar panels, on the outside of the ISS.

Why aren't these electrical boxes on the inside of the ISS, which would make replacing them much easier?

In short: electrical and fire hazard. From Options Studied for Managing Space Station Solar Array Electrical Hazards for Sequential Shunt Unit Replacement (PDF), explaining a bit what SSU are:

The U.S. solar array strings on the International Space Station are connected to a sequential shunt unit (SSU). The job of the SSU is to shunt, or short, the excess current from the solar array, such that just enough current is provided downstream to maintain the 160-V bus voltage while meeting the power load demand and recharging the batteries.

If SSU units malfunction, as the sunlight hits the station's solar panels, it could present substantial risk to anyone or anything near them:

The open-circuit voltage of cold solar-array strings can exceed 320 V, and warm solar-array strings could feed a short circuit with a total current level exceeding 240 A.

That's the reason why Tim Peake and Tim Kopra had to carry out replacement of a faulty SSU when the station was in Earth's shadow and there was no current going through the units, and this is why they had to work fast:

Normalized short-circuit current, $I_{sc}$, versus orbit solar $\beta$ angle (angle between the orbit plane and the Earth-Sun line). Velocity vector for the International Space Station (ISS) is aligned with the ISS +x-axis in this figure.

Should such units be installed on the inside of the station, with oxygen in the breathable atmosphere, on top of being an electrical hazard due to high current and voltage, it could also present fire hazard if any of them shorted. And as you can imagine, pumping up to 76,800 W through them, they would also get pretty hot, so there is also a good reason to move them far along the station's main truss and away from the habitable and experiment area, and that's why Tim Peake and Tim Kopra took so much time to translate during their EVA to get to the unit.

• To add to this, the SSU is the first of many power conditioning units between the raw power collected by the solar arrays and the conditioned DC power delivered to ISS systems. It makes sense to keep as much of this conditioning equipment close to the source as possible. Jul 21, 2016 at 21:23