Secondary Injection Tester Going into Overload

https://electrical-engineering-portal.com/secondary-injection-tests

One possibility is that there is a source of power on the circuit being tested that is confusing the instrument, however it cannot be seen from here.

1. It may be an elementary point - you have opened feeder A links - but have you short-circuited feeder A CT ?? If not, there would be arcing which could apply spikes to your test set & cause overload symptoms.
2. How do you know your test set is overloaded?
3. Have you measured currents to /from test set/ relay/remaining CT with a meter?

Unfortunately, I cannot download the diagram, although other diagram from site do work OK.

If as you write, the CT's secondaries are wired in parallel, then the relay has only one current input per phase. The current A minus current B function is inherent in the CT connection according to the polarity of connections[reversing the leads of one CT would give sum of currents, not difference] .

To test the relay, you isolate its current input from the CTs and inject to the relay CT input: with < and > required operating current difference between CT secondary currents.

But the actual result of two currents....

I = A sinwt - Bsinw[t + T]

depends on the phase/time difference - assuming same frequency [you can draw a vector diagram to get the result I]. If you inject with a test set, it may come from an LV supply of different phase to the cables your CTs enclose or even have its own generator of different frequency to CT currents.

Note that, theoretically, the relay and its connecting cables to the point where CTs are paralleled has negligible impedance/resistance compared to the CTs & their cables or at least makes the CT loads less than rated burdens.