Bug? Nuclear reactor rejects piped-in coolant cells

I was trying to automate a -SUC reactor by having BuildCraft pipes keep feeding it with a steady supply of fresh coolant cells to replace burned ones. For some reason, it seems only the top row of the reactor's inventory will accept cells. I built a little test setup with a reactor with one external chamber. I got 4 cells along the top only. Any additional cells piped in are spat out of the reactor as though it were one of the items that don't belong in a reactor. Is this a bug with the pipes, the reactor, or just something to prevent automation?

Initially, I had it pumping into the top of the main reactor. In my little test, I tried connecting the pipe to both the top and the side of both the reactor and the side chamber. As expected, the pipes ignored the side chamber, but connected fine to the main reactor. However, no matter which entry point to the reactor the cells came in from, they were tossed out after the top row was filled.
As for each chamber being treated as a separate inventory, I thought the chambers' inventories were organized vertically; three columns for the main reactor, and one for each additional chamber. The pipes were filling only the top space of each column before rejecting more; all four columns, the three main and one extra, had one cell in the top, so it wasn't ignoring the external chamber's part.

Additional clarification: the 'rejection' isn't like the normal problem with BC, where something is full when the item tries to go into it. In that case, the item falls loose when it hits the end of the pipe. What's happening is the cell is accepted into the reactor just fine, but then is immediately spat out of the top, like it would if you tossed coal into it or something of that nature.

The full-chambered reactor didn't work so well, since Buildcraft's pipes don't recognize the added reactor chambers as a pipe target, and leaving one side of the reactor open has the same problem as before. But, thanks to Alblaka's explaination, I see this isn't IC2's problem. I guess I'll just to stick with a 0.5 less efficient better-cooled reactor unless BC's pipes get smarter.