Using the isolation transformer on all the test equipment but not the amp will prevent ground problems. If the test equipment has grounds they should not be tied together and also left floating from earth ground which is connected to the amp chassis ground and also the 4 ohm output tap. On some of my test equipment like the signal generator you can either use it balanced with no ground connection or else tie the negative output to ground with a shorting strap at the output jack set for an unbalanced configuration. However that is not the case with all signal generators where the negative output is sometimes always connected to the chassis. With other pieces like the scope you cannot lift the negative signal input from the chassis ground which is more problematic.
Remember with your amp, the 4 ohm output is always connected to chassis ground. The shell of the RCA input jacks are also connected to chassis ground at some point and therefore are connected to the 4 ohm output. If you connect a signal generator to an input jack the negative side of the generator will be essentially connected to the 4 ohm output. If it cannot be configured in a floating mode and always has the negative side at chassis ground that chassis is connected to the 4 ohm tap. Likewise if the scope cannot be lifted from ground unless it is connected with with a true differential probe, it's chassis will be at chassis ground. If you connect the ground side of the probe to the 0 output of the speaker, it is not at the same potential as the ground of the signal generator. In that case above, the ground side of the input signal generator will be at the same potential as the 4 ohm output and the chassis of the scope at the 0 ohm output.. If the chassis of the signal generator and the scope should touch, you will short out the 0 to 4 ohm set of windings on the output transformer.
It is probably safer with that amp to always connect the negative sides of all test equipment to the 4 ohm tap. That way even with equipment where the test connections cannot be floated, you will not short out the output transformer. In general, it is also not a good measurement practice to leave the individual chassis floating as leakage and induced voltages can affect the measurements. If you connect all the test equipment chassis together and MAKE SURE all the negative connections are either to the chassis on the input side and the 4 ohm tap on the output side you will minimize induced noise problems. What the isolation transformer will do for you is provide some protection if a piece develops a short to the chassis as there is no way for it to get back to earth because of the isolation. The reason I said some protection is that if by some extreme fluke one piece of equipment gets a short from one side of the isolation transformer to chassis and another piece gets a short to the opposite side of the isolation transformer the chassis on the two pieces will differ from each other by the full output voltage of the isolation transformer. If for example the signal generator negative (in an unbalanced configuration) connects to the shell of the RCA jack or ground and the scope negative connection is on the 0 tap of the output transformer, the full 120 volt output of the isolation transformer would travel through the negative test leads from each test piece and the output transformer from the 0 ohm to 4 ohm tap. That is not a good situation, although the chances of that happening are extremely remote.
I still think the best way to do the setup is to not use an isolation transformer and connect all the chassis grounds together including the test equipment. You will then only make negative connections on the input side to chassis ground and the output side to the 4 ohm tap. In that case measurement noise is minimized and there will always be a path for a short to chassis in any piece of equipment to travel safely to ground. However, you MUST, MUST, MUST make sure that the connection connection configuration is ALWAYS maintained as stated above. I really hate working on designs like that amp.
Tom
|