Done tests now:
- First I left it idle, un-muted (no input) at 30V overnight. No problem.
- My voltage adjustable PS can deliver 6A / 32V
- So I obviously could not test more than one channel at high power
- With a 4 ohm load the 6A would be reached at around +/-24V minus voltage drops.
- So I hooked up a 4.7ohm, 50W resistor. and power up. It reached a frying 200 deg C in about 10 minutes so I had to hook up my big 4 ohm 600W load. I will post pictures later today.
- The PS / amp combination typically wobbled a bit so the highest nominal voltage I could use was 28.5V. Above that the amp would trip for over-voltage.
- I used a pink noise generator as signal source and tested with up to 4V Vpp (peak-peak) signal however, the current protection of the amp or PS (TBD) kicked in at those levels, making it run intermittently. If I limited the output to a bit above +/-22V, (probably limited by the power supply over-current setpoint) it was all running smoothly. This would be equivalent to around 60WRMS
- My TrueRMS meter is limited to 2A so I had to rely on eyesite and estimate an average of around 2-2.5A average, with peaks up to 6A (power supply current limit LED blinking)
- The setup has been running for a couple of hours like this now with no problems.
- Summary so far: 30V at idle no problem
- 28.5V average at high power is also OK
- Over-voltage kicks in at around 30.5V and restarts again at around 25V
- I have not managed to do damage despite moving up to the voltage limit under high output power a large number of times
The over-voltage limit is not a true protection. It shuts of the amp at too high voltage but does NOT protect from excessive over-voltage. It simply shuts down the amp and this has two advantages:
- You get a warning that something is wrong
- The shutdown stops any power supply instability that may have caused the over-voltage (peaks).
1) Power supply = 24Vdc, input signal sinusoidal 0.41Vrms, 1kHz. Load 4 ohm resistive. The average current was about 2.7Arms
Here we see a nice clean output signal, at around 17.5Vrms. There is a small distortion at 2kHz but I am not sure if it's in the input signal or not.
2) Same as above but now 0.47Vrms input. Signal is now clipping which is clearly visible by all the additional spikes. If we had a normal scope type graph, we would see that the signal is no more nicely sinusoidal but the tops are flat because the output can not be any higher than the supply voltage. The level where the output clips is SUPPLY VOLTAGE dependent. The average current was a bit above 3Arms in this case.
3) Pink noise signal, 28.5V supply, 4 ohm load. The input signal was turned up until the amp over-current protection turns it of in the peaks. The gap you see is about 0.55 seconds long. The peaks are nicely at around +/-23V peak to peak, indicating that the signal is clipping more or less all the time. The average current was somewhere between 1.5Amp.
(pink noise has an average frequency distribution that is similar to the average frequency distribution of music)
With a 4 ohm load and 24V supply, I could not reach the over-current limit, even with considerable clipping. This is the way we want it. The amp should not reach the over-current limit even under high power. It should only happen for example if outputs are shorted.
A PS on measurements: amps with bridged outputs can not be measured ground referenced. The measurements must be floating, using a floating, non-grounded instrument or an isolation transformer or other differential, non-ground referenced measurement. Most measurements shown here where done with an instrument with fully differential inputs.
More AMP9-BASIC measurement results.
First a case with 24V supply, 1kHz signal, 4 ohm load and varying the output level. We can see a sweet spot at around 10Vrms output and when output reaches 11Vrms, the output starts clipping and the THD goes up.
This is at about 30Wrms and the supply current is then about 2.8A per channel. At 1% THD, the current is around 3.4A and output around 13.7Vrms or 50Wrms
Same but 12V supply, 4 ohm load:
And here 12Vdc supply case with 2 ohm load:
Finally a measurement at 28Vdc supply, 13Vrms into 8 ohm, equivalent to 20W output, while varying the frequency. Over a wide range the THD+N is in the 0.01% region, 0.007% at the best, 0.07% at worst.