Hi @Var, these are good ideas. There’s the output impedance for the QA451 that does its measurement with the 4 and 8 ohm loads, and for class D with the big LC output filters, this does make sense. Plus, class D don’t really like driving into open loads because the output LC assumes 4 or 8 ohms
But you are right, maybe the non-QA451 impedance module should focus more on a manual measurement, with the option to run open load, and also plot damping factor.
PS. The math used to compute the Z given two loads is as follows.
If you substitute in your manual measurements, I think the correct answer comes as expected.
@matt- I did the calculations using your equation with the manual measurements and got an impedance of 0.13ohms, which makes sense looking at the graph before it went negative. I guess the question is why the two measurements are not in agreement- it has only happened twice, so most of the time things are fine. I could try measuring from 20hz-1khz every 20hz and see what happens when using the same FFT and sampling ? maybe try doing the measurement at 5w?
If you measured output impedance using the same loads and connections you also use for power output testing, I can see how it could potentially return a negative value.
Amplifier output impedance should be measured at the loudspeaker terminals of the amplifier itself, and the load value reference keyed in should include the resistance of the connecting cable to the load as well, whereas power output testing is connected over the actual load, not at the speaker terminals.
So two distinct connection methods- one Kelvin-like, the other not.
In my case, I use about 3’ of 12gauge wire for all of my connections, but they would start more at the load side of things than at the amp side of things, so there is a very small amount of resistance that is not counted for. I saw something about this relating to the THD improvement that could be had if your connections were at the amp end of things. I wished I had a 4-wire resistance measurement meter to actually measure the cable resistance, but it is not going to be that significant, particularly with what I do. BTW, I measured the Qx8000 today and did not get a negative impedance- it actually was better than the spec!
I remeasured the M501 amp using the same FFT and sample rate as before, but changed the step size to 20hz, starting at 20hz and going to 1100hz and got the same slope and negative value at 20hz, but it did not start going negative until 40hz…?
That sounds good, I will be returning that amp to its owner probably within the week, though. Most likely it will happen at some point in the future.Thanks for the help!
