I have started looking at one of my older Power amps, which is rated at 200w/ch into 8ohms. Following the guide in the “Basics- Power Amps”, I have built my 8.1ohm loads such that there is a 0.1ohm sense resistor in the middle of the resistor stack, with banana jacks across it, so that I can measure across it directly when measuring at a higher power level. I decided to do a preliminary test with the amp putting out about 100w. I did one measurement with the QA402 measuring across the 8.1ohm resistor stack and got this:
I then performed the same FR such that I was measuring across the 0.1ohm resistor, with adjustments for the attenuator setting and the dbv offset set appropriately. This is the response I got:
Any suggestions as to what may be causing this difference and how to improve it ? I would think that even with the signal being attenuated so much, there should be plenty of signal for the QA402- they look similar up to about 200Hz or so… Thanks,
Hi @VAR, notice the “range” indicator in the upper left of the last plot you shared. That means that the max input level had to changed during the acquisition (due to overload) and thus the acquisition is invalid. Can you repeat at a higher full scale setting until the “range” indicator doesn’t show up and see what you get?
Thanks once more, Matt. I feel kinda dumb- I did not see RANGE indicator on. I figured that since the relays stopped clicking on & off I was OK. So I repeated the test with these results, which are much better:
This amp was made around 1985, and I had replaced a few of the electrolytics a while back when I had the cover off. When it was new the FR was +/- 0.1dB 20-20kHz, so I can live with +/-0.4dB for the case across the 0.1ohm resistor, but the 1dB drop off at higher freq’s across the 8.1ohm load is a little more disconcerting. I am wondering if changing from 0.1 to .2 ohm would make the reading closer- I doubt I would ever measure more than 400w/8ohms. Maybe the QA451 would be better than my loads?
Hi @VAR, your intuition is right–the fuzziness you are seeing is because the signal has been knocked down so far it’s getting a bit noisy. If you bump up to 0.2 ohms, you’ve doubled the signal you are seeing. But even better would be two 4 ohm loads in series and tap across one of those, or 4 2 ohm load in series, and tap across one of those.
Another thing to experiment with is a longer FFT. When you double the FFT, you double the energy recovered at each frequency. So, you can make the recovered signal twice as “strong” by doubling the sense resistor OR doubling the FFT size. And cutting your sample rate in half, has the same impact as doubling the FFT size. So, lots of knobs to play with! But once you see the fuzz you are seeing in the first plot, be aware the FFT math is being challenged by noise.
Thanks for the suggestions, Matt. I tried again playing around with the FFT size and sampling rate, but could not really get rid of the noise, and I was running the amp at it’s full rated power. One of my loads is made of 4, 100w 2ohm resistors, so I try measuring across that and see if it makes a difference. The other load is 8, 1ohm, 50w resistors, so that change would not be very difficult. In a post a while back you broke down the max QA402 input power levels and a special input mode- I need to review that as well since I would be right at the cusp of it with a direct measurement across 8ohms. Here is the FR plot at full power and a different FFT:
Going from the 0.1ohm to 2ohm sense resistors was simple and did the trick! You are the man!! Here is the plot at just a hair below the 200w rated power (where I started seeing some very small noise):
for the case across the 0.1ohm resistor, but the 1dB drop off at higher freq’s across the 8.1ohm load is a little more disconcerting. I am wondering if changing from 0.1 to .2 ohm would make the reading closer- I doubt I would ever measure more than 400w/8ohms. Maybe the QA451 would be better than my loads?
