Das war night nur nicht richtig, es war nicht einmal falsch!
... OH thats GOLD;
I havent heard that since I was at high school where our Chemistry teacher (
a very austere German) had it printed above the blackboard in German with an English transliteration underneath. Thanks for reminding me
Hi Guys;
While it is admirable to try and measure the effects of vibration born interference, the results of one single test cannot be considered conclusive with out fully understanding and considering all other possible inputs.
Looking at the data set, there is apparent interaction between the channels. However, if we draw an averaging line through the graph, (
ignoring the initial noise spikes caused by the unshielded cable) the apparent interaction is already 96 dB down at its worst case. Allot of amplifiers already struggle to achieve this.
BTW there is no crosstalk between the power amp channels; they are 5 totally separate amps in the same case with separate transformers, power supplies and grounds
Before accepting the dataset we need to confirm that there is no cross talk between the channels. I dont quite believe that you can achieve complete isolation in the real world. Even with separate power-supplies for each channel stray capacitances, induction between leads or even leakage between apparently isolated power-supplies can happen, it as also a possibility that the crosstalk may be coming from the pre-amp.
I would use a fully shielded cable to feed the amplifier used to generate the test signal then try to eliminate any noise, also try to get a dummy load for the signal amplifier, short the input for the amplifier under test and disconnect all other equipment (
including speakers) from any other unused channels on the amplifier. Run the test again and measure the inherent crosstalk within the test scenario, then record the results.
Once this is done you can be a little more certain of just how much crosstalk there is in the system. Only then can you replace the dummy load with a speaker, re-run the test then subtract the known crosstalk from your new test results. You can now be a little more certain that what you are measuring is actual interference caused by vibration/micro-phonic activity with the amplifier.
You could also re-run the test with, then without a speaker attached to the test channel to see if the passive speaker is behaving like a microphone and feeding a signal back into the amplifier channel under test.
Accuracy and consistency in the method is just as important as knowing how to read the results and understanding what mechanism have been eliminated or allowed for, all this is necessary for interpreting the results correctly. Anybody can make a graph, its knowing how to read and interpret the data contained in the graph that makes all the difference.
Well done for having a crack at it though...
LPSpinner.