- Published on 05 January 2009
I make no apologies about my enthusiasm for Emotiva. After I reviewed their RSP-1 preamplifier and RPA-1 dual mono power amplifier here about 18 months ago, I replaced a $3000 Plinius integrated amplifier with them. They cost less than $1400 retail for the entire setup, and bested the "high end" Plinius in virtually every area. Emotiva is built on the idea of offering the maximum possible value to home theater and audio enthusiasts. They sell manufacturer direct only, build their products overseas and pass those discounts on to their customers. Their products offer quality and performance you would have to spend at least twice as much, maybe even three or four times as much to equal.
- Design: Single-channel (Monoblock) Power Amplifier, Differential Configuration
- Power Output: 500 Watts RMS into 8 Ohms, 1000 Watts RMS into 4 Ohms
- Gain: 32 dB
- MFR: 20 Hz to 20 kHz ± 0.2 dB, -3 dB at 100 kHz
- THD+N: 0.001% (20 Hz to 20 kHz)
- Input Impedance: 47 kOhm
- Dimensions: 7.75" H x 17" W x 19.5" D
- Weight: 75 Pounds
- MSRP: $999 USA (Check Emotiva Website for Special Sale Price)
The XPA-1 is the "statement" amplifier from Emotiva, and has been hotly anticipated by fans for quite a while now. The specifications are quite impressive: 500W RMS into 8 Ohms, 1000W RMS into 4 Ohms. As we'll see later in the measurements section, these are very conservative ratings. It will drive a 2 Ohm load, but requires a 20A 115V circuit or a 10A 230V circuit to reach its maximum output into that load. If you'd want to try to run a pair of them into 2 Ohms at or near full power, you'd want a dedicated 20A 230V circuit. That's combined with a fully differential design that takes full advantage of the balanced inputs. Other manufacturers typically ask at least $3000-$4000 a piece for amps like this. The MSRP is $999, with an introductory special price of $899 each (January, 2009).
Outwardly, the XPA-1 looks very similar to the XPA-2 stereo amplifier reviewed here at Secrets recently. The XPA-2 is a stereo amp with a single transformer and power supply, with a Class A/B output stage. The RPA-1 amp I reviewed is true dual mono Class H design, with two separate transformers and power supplies driving two amplifier modules in a single chassis. The XPA-1 is Class AB like the XPA-2, and looks similar inside. Rather than the two sides of the amp driving two independent channels, the XPA-1 uses the two "channels" of the amp to drive the hot and cold legs of the balanced input signal. The single ended input is copied, and the copy is inverted to feed the two sides of the amp. This is the ultimate expression of balanced drive, with two completely separate halves of the amp used to amplify the inverted and un-inverted legs of the input signal, and then recombined at the output. This completely cancels out any common mode noise that both halves of the input signal see in the amplifier. Bridging a stereo amplifier, as you can do with the XPA-2, does not offer the noise cancellation offered by a fully differential design.
The XPA-1 chassis is 5RU size, with a black anodized machined aluminum faceplate with clear anodized aluminum trim pieces. A single LED meter, doubling as a status indicator, spans the face. Both the meter and the status LEDs can be independently defeated via rear panel switches. The 1200VA toroidal transformer is mounted vertically in the front of the amp, just behind the front panel, making the amp a bit front heavy. The amp weighs 76 pounds, with a shipping weight of 90 pounds. Large, internally mounted heatsinks run along both sides of the amp inside the enclosure. The center of the amp sports over 130,000 μF of power supply capacitance. Unlike some "statement" amps by other manufacturers that are mostly empty inside, the XPA-1's chassis is absolutely crammed with components. The rear panel offers a mains toggle switch, 12V trigger inputs and pass-through, balanced XLR and RCA single ended input connectors, and a small toggle switch to select between them. An IEC input jack is present for AC input. The binding posts are in the same position as the XPA-2, but instead of serving as left and right, they serve as positive and negative and offer the possibility of bi-wiring. Ominous "Caution: Potentially Lethal Voltages Present During Operation" warnings are silk-screened in red next to each set of binding posts. Delivering its full rated power, over 60V are present at the terminals. Large, sturdy rubber bumpers are mounted next to each set of binding posts to prevent the amp from being pushed back into something to cause a short. This amp could serve as a high fidelity arc welder, so the extra margin of safety is welcome. A comprehensive set of fault protection circuitry helps to keep you from blowing things up with an incorrectly wired XPA-1.
In addition to all the features, the build quality of this amp is exceptional. No whining about the quality of Chinese made audio gear on this amp. It is as well assembled, and as well designed, as pretty much anything I've seen short of a Boulder 2050 (which costs as much as a new BMW, thank you). Board layout, solder joint quality, parts and materials quality needs no apologies. It isn't stuffed with ridiculously priced "audiophile" components, but everything is top quality for the real world.
As far as design and implementation go, I can't find much to fault with the XPA-1. One issue is that the positive and negative amp terminals are 13" apart. I had to send my Wireworld Eclipse speaker cable back to the manufacturer to have them re-terminated to reach the widely separated Emotiva terminals. If you order these amps, be sure your speaker cables can span a 13" terminal separation. Another weird thing was that the LED meters on the two amps did not seem to have the same sensitivity. I was testing a pair of prototype amps, so they could have been adjusted differently. This was a pretty irrelevant criticism though, as the meters are really just for show. Where the rubber meets the road is with the sound.
I listened to the XPA-1s both with my normal Gallo Reference 3.1s, and the fantastic THIEL CS3.7s I just reviewed. They were driven via the balanced inputs from an Emotiva RSP-1 preamplifier. As these were prototype amps, they were already well broken-in, so I could get to listening right away. To be sure I got around any warm-up problems, I left the amps on all the time. Given they do consume 90W each at idle, I would probably use the 12V trigger from the preamp normally, and just turn them on a while before any serious listening. I had no problem running the pair of them, along with all my other audio gear, on a single 15A 115V circuit. My Gallo Reference 3.1s are 87 dB/W/m sensitivity, and the THIEL CS3.7s were 90, so I never approached sustained power delivery anywhere near what the XPA-1 could deliver.
The point of high power amplifiers is not just to drive low sensitivity, low impedance speakers that require the power. Pretty much any dynamic loudspeaker without super high sensitivity can benefit from a powerhouse like the XPA-1. It all comes down to dynamic contrasts and transients. With any sort of transient musical event, like a drum hit or movie explosion, your speakers will demand a tremendous amount of power for a very short period of time. If your amplifier can't deliver, that transient will get smeared in the time domain. This smearing will homogenize the sound and decrease the impact of dynamic contrasts.
The dynamics delivered by the XPA-1 were the first stunning aspect of these amplifiers. After listening to the super high sensitivity Zu Audio Druids, and several systems at the Rocky Mountain Audio Festival that sounded so alive and energetic, I was painfully aware of the main drawback of my system: lack of dynamics. While the Emotiva RPA-1 offered exceptionally smooth and liquid sound, the dynamics were always somewhat limited compared to the most alive sounding systems. The dynamic contrasts present in these great sounding systems lend a presence and reality to the sound that is impossible to fake. One way to deliver these contrasts is with very high sensitivity loudspeakers, like the Zu Druid. Another is with an amplifier with exceptional power and current delivery. With the XPA-1 in control, my Gallos genuinely came alive. What was dynamically flat before was able to compete with the memory of the Zu Druids, but without some of their drawbacks. Dynamic contrast, both macro and micro, were improved across all frequencies. The most dramatic result was in the bass. This might normally be expected, but the Gallos have a separate amp driving the second voice coil of the woofer. Still, the added impact and control of the XPA-1 not only improved the agility and precision of the bass, but also improved the perceived bass smoothness. Maybe it was speaker positioning, but even the measured frequency response in the bass was smoother with the XPA-1s. Driving the THIEL CS3.7s, with their relatively high sensitivity but very low impedance, the XPA-1s were the key to releasing their true greatness. With the RPA-1, the THIELs were a bit flat, both dynamically and spatially. The XPA-1 allowed the full dynamic contrast available to be released, and simultaneously added soundstage width and depth.
A second great improvement with the XPA-1 was the retrieval of detail from recordings. Certainly helped by the dynamic contrasts available, the true magic might be from the fully differential design. The noise floor was stunningly low. Tape hiss became clear on virtually every analog recording. Not obtrusive, but clearly audible and identifiable. The blackness of the background clearly revealed loads of detail that was barely audible or inaudible before. Examples included the piano pedal noise and breathing on Tori Amos' Icicle from the album Little Earthquakes. While aware these sounds were there with the RPA-1, I was only able to clearly hear their detail and texture with very good headphones with lots of isolation. With the XPA-1s, they were all there as clear as day. Bass agility and clarity were on display with Slide Hampton's album "Dedicated to Diz." On the track "Blue and Boogie", George Mraz's acoustic bass solo, the detail revealed in every note was fantastic. This solo was not close miked, and has realistic dynamics, meaning it is way less loud than the rest of the band. An amp needs to be able to pull out lots of subtle detail in the face of much louder competing sounds to get this solo right. With the XPA-1, the result was the best I've heard in my room.
The detail presented by the XPA-1 also improved every aspect of soundstaging and imaging. Spatial cues were all much clearer, improving soundstage width, but especially depth and room ambiance information. Images remained razor sharp as with the RPA-1, but the added detail made those images sound more real, since subtle changes in spatial location and image size were more clearly conveyed.
Compared to the RPA-1, the XPA-1 did have a brighter presentation. The XPA-1 is very detailed and has lots of high frequency extension. This extension could be unkind to poor recordings, or components that sounded hard on top. The Oppo 983, which is brighter sounding than my Bel Canto DAC-1.1, was not the best choice for bright, hard recordings. They sounded a bit better with the Bel Canto. The RPA-1's sweet and smooth treble did not pull out as much detail, but was kinder to bad recordings. Overall the RPA-1 was a bit more "tube like," while the XPA-1 reminded me of high dollar solid-state amps from the likes of Krell, Mark Levinson and Jeff Rowland.
The high gain of the XPA-1, when combined with the relatively high sensitivity THIEL CS3.7s, made the volume knob of the preamp very sensitive, and useful over only the first ¼ of its travel. The high sensitivity also revealed some high frequency hiss when the amps were on but the preamp off. This was inaudible with the Gallos, but loud enough with the THIELs that I left the preamp on all the time too. If any components in your system are noisy, the high gain of the XPA-1s might reveal a little more of that than you want if combined with high sensitivity loudspeakers.
For me, those small caveats were irrelevant. The XPA-1 brought the dynamics to my system that I was craving after hearing several great systems at the RMAF. Now, I have that sound, and without sacrificing any other area of performance. The bass power and extension of my system, and most importantly the smoothness of the frequency response, were greatly improved. On top of all that, retrieval of detail also saw a significant improvement. And that's over an amp, the RPA-1, which I absolutely love. The RPA-1 will be going up for sale, to be replaced by this pair of XPA-1s.
On the Bench
I measured the XPA-1 using SpectraPlus FFT analysis software with a Roland Edirol UA-101 24 bit/192 kHz sound card. I measured the amp into an 8 Ohm, 300W load resistor. Unfortunately, while trying to measure THD vs. frequency, I managed to set the generator level of the sound card too high, and blew out the input on the sound card. I was unable to make any measurements into 4 ohms. In addition, I could not measure output power as the power resistors I have are only rated to 300 W. Emotiva provided me with THD versus power output plots measured with an Audio Precision system, which I also include here.
THD was measured at 1 kHz and 10 kHz at 20V p-p output level into 8 ohms. Measured THD is well under 0.1% at both frequencies. The THD+N measurements are much worse, but this is an issue with my sound interface, and do not reflect reality. The 60 Hz signal is a ground loop caused by an interaction with the amp and the sound card. This noise is completely gone when disconnected from the UA-101 at the input, as measured with a Tektronix oscilloscope.
Emotiva XPA-1 THD spectrum of a 1 kHz sine wave at 20V p-p output amplitude.
Intermodulation products were over 60 dB down from the fundamental tones at 1 kHz and 500 Hz, which is very good performance. Here is the Emotiva XPA-1 intermodulation spectrum of a 500 Hz and 1 kHz sine wave each at 20V p-p output amplitude.
Frequency response was flat as a board from 20 Hz to 48 kHz, the limit of the SpectraPlus software. The slight linear rise towards higher frequency is the response of the sound card input, not the amplifier itself. The same slope is present running the UA-101 output directly into the input. Shown below is the Emotiva XPA-1 frequency response at -30 dBV RMS level into an 8 Ohm load.
THD at 1.535V RMS input voltage (maximum rated output power level) stays at around 0.003% until about 1 kHz, and then rises gently to about 0.02% at 20 kHz. At 0.1V input voltage level (about 5V output voltage), THD is flat as a pancake at around 0.002% over the full frequency range. Results were slightly worse (a bit less than a factor of 2) into 4 Ohms (not shown). Here is the THD+N vs. Frequency graph for the Emotiva XPA-1 at 500W output power into an 8 Ohm load. (Measurement provided by Emotiva.)
THD+N vs. Frequency for the Emotiva XPA-1 at 5W output power into an 8 Ohm load. (Measurement provided by Emotiva.)
Power output into 8 Ohms and 4 Ohms is incredibly impressive. At the rated powers of 500W and 1000W, the THD levels are a very low 0.003% and 0.05%. At the 1% THD spec for output power, the XPA-1 puts out 680W into 8 ohms (37% higher than spec) and 1100W into 4 Ohms (10% over spec). Below is the THD+N vs. output power level for the Emotiva XPA-1 into an 8 Ohm load. (Measurement provided by Emotiva.)
"THD+N vs. Output Power for the Emotiva XPA-1 into a 4 Ohm load. (Measurement provided by Emotiva.)
Overall, the bench results for the XPA-1 are excellent, and prove you are getting even more than advertised. Quite amazing, given the low cost of these amps!
Addendum from the Editor 4/1/09:
I (JEJ) was so impressed with Chris' review of the XPA-1, I asked Emotiva to send me one so I could run some of my own tests. I borrowed a different power amplifier from a friend that cost more than 6 times the price of the XPA-1, which I will call Amplifier "X" here. I ran the same tests on both amplifiers. Here are the results.
With the XPA-1 and a 1 kHz sine wave at 40 volts output into 8 ohms, THD+N was 0.017%. For Amplifier "X", it was 0.023%.
IMD for the XPA-1 was 0.007%, while for Amplifier "X", it was 0.013%.
The frequency response also differed. The XPA-1 was flat to 50 kHz, and Amplifier "X" started rolling off above 20 kHz.
Now, I have to say that I have listened to Amplifier "X" many times over at my friend's house when we get together, and I consider it a fine amp. However, these tests show that, at least within the 200 watt tests (40 volts into 8 ohms) that I ran for this comparison, the XPA-1 comes out ahead. It is a surprisingly good amplifier for its price. A set of three XPA-1's across the front, with their smaller amplifiers for the sides and rear, would be a dynamite home theater setup. JEJ
A pair of XPA-1s costs $1998. They're currently on sale for $1798 per pair. To equal or exceed their specifications and performance, you will need to spend at least three times this amount (Parasound JC 1). In the value for money department, they are absolutely unmatched. They are some of the best amplifiers I have ever heard, period, and they will have to be be pried out of my cold, dead fingers.