Today’s high-powered amplifiers present some challenges on the bench. In part, there’s a lot of power to dissipate. I have 1,000 watts of dummy load capability in four 4-ohm resistors that can be configured in combination to achieve various loads. For these amplifiers, I was able to accommodate both channels driven full into 8-ohm loads (making my office toasty from the space-heater effect), and a single channel into 4 ohms. I did measure into 2-ohm loads, but not at full rated power.
Class D amplifiers bring additional new elements to the table that can affect measurement. There can be a significant amount of energy at the output in terms of out-of-the-audible band noise and artifacts — even inaudible signal components can negatively impact measurement. With instruments like the Prism Sound dScope III and Audio Precision Cascade series used here, the input circuits can be compromised by this hyper audible content. Both companies offer external filter sets that remove the out-of-band signals, but unfortunately I did not have those filters available for these tests (nor the new wideband FFT of AP’s latest test sets). My Cascade is equipped with a single channel of AES17 pre-input 20 kHz filtering, employed for most of the testing here. I measured performance very close to the manufacturers specifications, but would have likely gotten a touch better measurement with external filtering.
The Peavey IPR 1600 at full volume has an internal gain of 35.5 dB, with an input a touch above 0 dBu producing the rated output power. At 8 ohms, both channels driven, the amp is rated at 300 W @ .1 percent THD+N (all THD+N numbers taken with a 20 kHz bandwidth). I cracked the .1 percent mark at 282.5 W and measured .8 percent THD+N @ 300 W out. With a single channel driven into 4 ohms, where the rating is 600 W @ .1 percent THD+N, I measured .1 percent @ 553 W, .5 percent @ 571.7 W. At 600 W, I was into obvious overload.
The Crown XLS 2500 is rated at 440 W/channel @ .5 percent THD+N in the same configuration. I was able to get just over 400 W @ .1 percent THD+N and 414 W @ .5 percent. With one channel driven only, the 775 W spec was easily met: 730 W @ .1 percent THD+N, 784 W @ .5 percent. Amplifier gain at full volume measured 32.5 dB and full output was achieved with an input around 4.5 dBu.
I wasn’t easy on these amps, turning them on and cranking them to full output immediately and so on, and neither showed any sign of folding. A 1W output is typically used for measurement of frequency response. For response and other tests, I decided somewhat arbitrarily to work the amps out a bit, running such plots at a 250W output. I measured close to their specs pretty readily at high level, so I didn’t see the need to go back to genteel levels. Figures 1a and 1b show the frequency response.
Top, Fig. 1a: Crown XLS 2500. Bottom, Fig. 1b: Peavey IPR 1600. Frequency response plotted for both amps at 250W of output at 1 kHz. I left the plots on equal scales including the 50 kHz bandwidth that Peavey boasts, but looking close you’ll see that the Crown is quite flat, within .2 dB across the full audible band. The Peavey response shows an unusual tilt, rising gradually near 1.5 dB from 20 Hz to 20 kHz. The Peavey’s response does extend well past 50 kHz. This represents a design philosophy difference between the two companies in the implementation of Class D design and the permissibility of out-of-band signal elements.
Signal-to-noise ratios (noise relative to rated power) are spec’d at >103 dB for the Crown and >105 dB for the Peavy, A-weighted. With no weighting applied, I measured, relative to the max measured output at the rated distortion numbers, better than -99 dB performances for each amp, 22 Hz-20 kHz bandwidth. FFT plots of the residual noise showed no discernable artifacts, just a gradual rise of about 6 dB from 10 Hz to 20 kHz for the Crown and broad U-shaped plot for the Peavey, falling about 7 dB from 10 Hz to 10 kHz then rising about 7 dB to 20 kHz.
Top, Fig. 2a: Crown XLS 2500. Bottom, Fig. 2b: Peavey IPR 1600. These plots show the THD+N performance versus input level. The 250 W output points fall at 2.7 dBu input on the Crown and at about -.44 dBu in on the Peavey.
FFT plots of the amps’ outputs driven to 250 W out by a 1 kHz tone, on the Crown, shows some harmonic content peaking out of the noise floor (below .02 percent THD+N), dominated by odd harmonic components, with discernible harmonics including a bit of fourth, sixth and eighth, and odds out to 19 kHz. The same plot on the Peavey (at .028 percent THD+N) is dominated by the third and fifth, with significant second and fourth, then odd harmonics peaking out of the noise floor to 21 kHz. Driven to .1 percent THD+N, the harmonics are far more obvious (as would be expected), The visible harmonic peaks on the Crown got back to the noise level at about 15 kHz and the Peavey showed significant spikes well past 30 kHz (these tests used an 80 kHz bandwidth, so apply the earlier stated qualifications). IMD distortion plots showed respectable performance, slightly better with the Crown. With its deliberately wider frequency response, the Peavey does allow the output to contain out-of-band distortion components.
Both amps turn in respectable performance on the bench.