Emotiva XPA-1 Gen 2 Monoblock Power Amplifier Review Highlights
Emotiva made their name by producing high power, affordable amplifiers.
Perhaps no amplifier personified that more than the XPA-1. The original XPA-1 offered 500 watts into 8 ohms and 1,000 watts into 4 ohms. The updated model improves on this by delivering 600 watts into 8 ohms and 1,000 watts into 4 ohms. It also offers a Class A mode accessible by a switch on the front panel. When engaged, this provides 60 watts of Class A, but retains the ability to provide up to 600 watts in Class A/B mode.
Emotiva XPA-1 Gen 2 Monoblock Power Amplifier Highlights Summary
- 600 watts into 8 ohms, 1,000 watts into 4 ohms
- Switchable Class A and A/B modes
- Balanced and Unbalanced Connections
- Oddly spaced binding posts
- No handles or rack mounts provided
Introduction to the Emotiva XPA-1 Gen 2 Monoblock Amplifier Review
Emotiva made their name by producing highpower, affordable amplifiers. Perhaps no amplifier personified that more than the original XPA-1.
EMOTIVA XPA-1 GEN 2 MONOBLOCK POWER AMPLIFIER REVIEW SPECIFICATIONS
- Design: Fully Balanced, Fully Discrete, Quad Differential, Ultra-high Current, Short Signal Path, Selectable Class A/AB Monoblock Power Amplifier
- Power Output (0.1% THD): 600 Watts RMS into 8 Ohms, 1,000 Watts RMS into 4 Ohms
- Gain: 29 dB
- Power Supply: 1,200VA toroidal transformer, 120,000 µF Filter Capacitance
- Input Impedance: 33 kOhms (Balanced), 23.5 kOhms (Unbalanced)
- Size: 7.75” H x 17” W x 19” D
- Weight: 73 Pounds
- MSRP: $1,099 USD
- SECRETS Tags: Emotiva, XPA-1, Power Amplifiers, Monoblocks, Gen 2
The original XPA-1offered 500 watts into 8 ohms and 1,000 watts into 4 ohms. The updated model outdoes this by offering 600 watts into 8 ohms and 1,000 watts into 8 ohms. It also has a Class A mode, accessible by a switch on the front panel. When engaged, this provides 60 watts of Class A, but retains the ability to provide up to 600 watts in Class A/B mode.
The XPA-1 Gen 2 is the most powerful Class A/B amplifier that Emotiva makes. We know it has raw power, but does it also have grace and finesse to go along with brute force?
Design and Setup of the Emotiva XPA-1 Gen 2 Monoblock Power Amplifier
The XPA-1 is a balanced, discrete Class A/B monoblock amplifier. It features XLR and RCA inputs on the rear, with a switch to toggle between the two inputs. 12V triggers have a 3.5mm input and output, so you can chain two XPA-1s together if you wish. The front has the switch to move between Class A and Class A/B modes, along with an LED meter showing the relative output level. If you find the LEDs too bright for your room, you can disable them on the rear panel. There are optional rack rails available, though they are not included in the box.
Despite being a monoblock amplifier, there are two pairs of binding posts on the rear of the XPA-1. This allows you to bi-wire speakers with two individual runs of cables if you desire.
I felt that the binding posts are too far apart. Over a foot separates the positive and negative terminals, and most speaker cables will likely not reach them by default. I added a banana extension for my cables to enable them to make it, but you might need to split your speaker cable at the end so that the two legs can reach each set of binding posts.
The biggest internal change to the XPA-1 Gen 2 is the addition of the Class A switch. When you enable this, the first 60 watts from the amplifier are Class A. If you need more power, you can switch it to Class A/B operation. If you have the switch in the Class A/B mode, it operates there all the time. Since both modes have access to all 600 watts, why wouldn’t you leave it in Class A all the time?
The answer is simple: heat. Class A means that the amplifier is always drawing power and Class A amps are only 25% efficient by nature. With 60 watts of Class A power, that means 180 watts of power are dissipated as heat when the amp is idle. Leaving the XPA-1in Class A mode at idle for an hour bears this out, as the amps will become quite warm.
One other feature I would like to see on the XPA-1 Gen 2 would be rack handles. While not the heaviest amplifiers I’ve moved around, they are the heaviest without handles.
The Emotiva XPA-1s were set up in my system using a Marantz AV7005 preamp, an Auralic Vega DAC, or an Anthem MRX510 receiver. The Marantz and Anthem connect with RCA, while the Auralic uses XLR connections. The speakers are Revel f208s, which have a nominal impedance of 8 ohms, but dips down below 4 ohms at various points.
The Emotiva XPA-1 Gen 2 Monoblock Power Amplifier In Use
In Class A/B mode, only a few of the LEDs on the front of the XPA-1s lit up, because the speakers ae sensitive. A single LED or two was as often as much activity as I got with some music. Throwing on some Led Zeppelin re-issues changed that, and I was treated to a light show, as the LEDs danced back and forth. Driving the f208s in full range, the XPA-1s had no trouble keeping up with John Bohnam’s insane drumming.
Even more aggressive than Zeppelin is Reckoner from Radiohead. It explodes out of the speakers from the first note. The song is crisp and dynamic, with no strain from the f208s at all. Even pushed to the limits for my volume tolerances, the XPA-1s didn’t go beyond 2/3rd on their LED meters. The XPA-1s have plenty of headroom, even if you have more demanding speakers than the Revels.
Movies also come to life with the XPA-1s. The opening attack scene from Pacific Rim startled me in my chair as the roar of a Kaiju attacking San Francisco filled the room. Another piece of eye and ear candy (with no plot) is Transformers: Dark Side of the Moon. With massive explosions and battles that I pushed through the Emotiva pair, they never hiccuped.
Listening in Class A instead of Class A/B, I admit I did not hear a difference in the sound quality, perhaps because Class A/B mode sounded so good.
However, I did notice two distinct things in Class A mode compared to Class A/B mode. The first is the heat. When both amps were on a rack behind my seating area, I could feel the warmth it gave off after an hour or two. When I moved them to the front of my room to make the A-A/B switch easier to access, I didn’t notice the heat until I got up to turn them off.
The second is a metallic popping sound that the amps would give off in Class A mode. It sounded different than a relay click, and was never audible in Class A/B mode. Since it happened during music playback, where I could hear it, and also after turning them off, I imagine it is a thermal issue. Perhaps the case screws were expanding or contracting and causing the sound.
In both Class A and A/B modes I could also hear a slight hiss from the tweeter. It was very low level, and I had to be close to the speaker, but it was there. It was low enough that I couldn’t hear it in use unless the music was in between tracks.
The Emotiva XPA-1 Gen 2 Monoblock Power Amplifier On The Bench
David A Rich Ph. D. and Chris Heinonen
All AC wattage measurements are in units of Watts average, not RMS.The Emotiva adds a Class A biasing option to remove crossover distortion. The amp transitions to Class B at 60 Watts into 8 Ohms and 30 Watts into 4 Ohms.
The power amp gets hot in Class A mode, reading 110 F on the amp surface and 120 F inside the chassis (using a thermal IR gun) after 20 minutes . In Class A/B mode it only rose by 5-10 degrees.
The XPA-1 design causes the performance in mode Class A/B mode not to be optimized. This is the result of including the Class A biasing option. It is possible to have both Class A and Class A/B optimized, but this requires extra circuitry not in the XPA-1. Details are in the section below called Design Technology in the Emotiva XPA-1.
Power Output of the Emotiva XPA-1 in Class A and Class A/B Modes into an 8 Ohm Load
We tested with the unit into a standard AC line with no VARIAC (a high current rheostat that allows you to keep the AC supply at 120 volts), which may have resulted in higher distortion levels.
Test at 630 Watts into 8 Ohms (Emotiva spec limit)
In Class A/B mode, THD was 0.11 % and the 19kHz / 20kHz IM test result was 0.05 % (average of the IM levels of the first 3 IM products at 1kHz, 18kHz and 19kHz).
In Class A mode, THD was 0.17 %, and the 19kHz / 20kHz IM test measured 0.05 %
Test at 1,030 Watts into 4 Ohms (Emotiva spec limit).
In Class A/B mode, THD measured 0.1%, and the 19kHz / 20kHz IM measured 0.13 %
In Class A mode the THD measured 0.13 % and the 19kHz / 20kHz IM test measured 0.11 %
Distortion Spectra A/B modes 600 Watts into 8 Ohms
The spectra below is of the amplifier producing 600 Watts (69.2VRMS) in Class A/B mode.
The THD was 0.004%. Note the odd harmonics are significantly larger than the even harmonics as expected for a balanced amplifier The 19kHz / 20kHz IM was 0.0028%
Give the importance of the distortion structure in a balanced amplifier, I have listed the harmonics below. The 3rd harmonic is predominate, with the 2nd the only other significant harmonic.
Below is a plot under the same conditions from 0Hz to 1kHz.
The hum was well suppressed, with the maximum spur at -110dB. This is again expected in a balanced amplifier with continuous current flow in both supply rails and no current flow in ground.
At 600 Watts into 8 Ohms, the 19kHz / 20kHz IM was 0.0028% for the first 3 IM products (the IM products were at 1kHz -115.20dB, at 18kHz -96.50dB and -97.78 at 21kHz).
Signal to Noise (S/N SNR) Measurements in Class A and Class A/B
The measured noise of the XPA-1 was higher than the majority of other high quality amplifiers. The reason for this is explained in the Design technology in the Emotiva XPA-1 section below.
The measured noise at the output was 120 µVRMS with the input grounded, over a bandwidth of 20Hz to 20kHz. Using the measured noise, the SNR can be calculated for different output references values. The full power SNR runs are direct measurements.
78.4dB ref 1VRMS
87.4dB ref 1 Watts into 8 Ohms (2.8VRMS).
115.4 dB ref 630 Watts into 8 Ohms (71.1VRMS)
114.6dB ref 1,030 Watts into 4 Ohms (64.2VRMS)
The amplifier’s heat in Class A results in higher noise because of the nature of electronic circuits and temperature (electronic noise). We measured the noise of the amplifier at its output at 142µV RMS with the input grounded. This is 1.4dB higher, so the SNR is 1.4dB lower.
77.0dB ref 1VRMS
86.0dB ref 1 Watt 8 Ohms (2.8VRMS).
113.8 dB ref 630 Watts into 8 Ohm (71.1RMS)
113.2 dB ref 1,030 Watts into 4 Ohms (64.2VRMS)
The Emotiva XPA-1 has a gain of 30dB (31.55V/V) at 1kHz, although this changes slightly when the amplifier is driven to 1% distortion.
Comparison in Performance Between Class A and Class A/B
Below is a comparion of the spectra of the amplifer in Class A and Class A/B mode outputting at 12.5 Watts into 8 Ohm (10VRMS). The distortion rises at higher power but a problem with the swept data curves prevents showing this here. We also cannot show when the Class A and Class A/B distortion curves merge.
XPA-1 8 Ohm Class A Data 12.5Watts
XPA-1 8 Ohm Class A/B Data 12.5Watts
For Class A mode, THD was 0.009% and the 19kHz / 20kHz IM was 0.0069%
For Class A/B mode, THD was 0.0022%, but the 19kHz / 20kHz IM was similar at 0.0071%.
A comparison of the individual spur values provides insight of the performance in Class A and Class AB. The first line is Class A mode of operation.
Below is the spectrum of each spur.
Strangely, for Class A/B the 2nd harmonic dominates, indicating it must be at a high level single-ended. This was also the case for 5 Watts. This is an indication that the Class A/B is not optimized. Nevertheless, the spur is low at 0.0022% at 10 Watts output.
At 600 Watts output, the Class A/B spectra above the 2nd harmonic is actually lower in level. The 3rd harmonic dominates the 600 Watt measurement at 88dB down.
For Class A, at 12.5 Watts output, the 3rd harmonic dominates as expected for a Class A amplifier. The 3rd is the only harmonic that can be said to be significant.
The 19kHz/20kHz IM test at 12.5 Watts into 8 Ohms does not show significant differences between Class A and Class A/B.
For 5 Watts in Class A operation, THD was 0.0005%, and the 19k/20K IM was 0.0032%
In Class A/B operation, THD was 0.002% and the 19k/20K IM was 0.0037%
The 1kHz spread between Class A and Class A/B is now 1 to 4, but these are low in level.
4 Ohm Performance at Lower Output Levels
In Class A mode, THD was low at 0.0009% at 5 Watts. It increased to 0.002% at 35 Watts.
In Class A/B, THD was at 0.0045% from 5 Watts to 60 Watts.
Problems with the swept measurements prevent filling the data at higher power levels.
The spectra for the amplifier driving 25 Watts into 4 ohms A/B mode is shown below.
The 2nd harmonic is the dominant distortion at 0.004%. The 19k/20kHz IM was 0.009% for the first 3 IM products.
Design technology in the Emotiva XPA-1
David A. Rich Ph. D.
Emotiva supplied a complete schematic. I thank them for the openness they provided on the design. This section assumes the reader is familiar with the introductory presentation on balanced amplifiers: The Fully Balanced Power Amplifier – Advantages and Design Challenges
Please refer to this article directly for a discussion on the sub-optimal Class A/B bias circuit mentioned in the measurement section above.
The XPA-1 topology is based on a single voltage gain stage. Almost all power amplifiers have two stages of voltage gain. A single stage of voltage gain is easier to stabilize. That is important in a fully balanced amplifier with 3 feedback loops.
With only one voltage gain stage, the open loop gain of the amplifier is attenuated, and this reduces the return loops gain of the feedback.
With less loop gain, the amplifier is less able to reduce distortion by the use of negative feedback. Emotiva uses other design approaches to improve the open loop distortion of the XPA-1 amplifier. This includes the fully balanced amplifier design to cancel even ordered distortion as discussed in the introductory piece. A fully complementary design is also used as is the Class A output stage up to 60 Watts (30 Watts into 4 Ohms).
The circuit topology of the Emotiva XPA-1 is close to Figure 2 of the document below which is a differential opamp.
Karki, J. “Fully-Differential Amplifiers.” Texas Instruments, Application Report SLOA054D, 202, p. 4
Copyright restriction prevents me from showing the schematic here.
The gain of a voltage gain stage can be increased by increasing the stage’s output resistance. One way to do this is to use a technique call the folded cascode. Explaining this technique is beyond the scope of this review. The folded cascode can be seen in Figure 2 of the reference noted above. Since only one voltage gain stage is used in the XPA-1, techniques to raise its gain are mandatory.
In the Emotiva XPA-1, the voltage amplifier is on regulated ± 64V rails to isolate the stage from the power supplies noise on the unity gain current amplifier sourcing and sinking current to the pair of output stages. The designer must take great care to sequence the turn-on times of the multiple power supplies to prevent fault conditions.
The output stages are on unregulated ± 52V rails. This allows each single ended stage to swing approximately 35VRMS. In balanced mode, that is 70VRMS. Into 8 Ohms, this is 612 Watts average. In practice the amp produced 71VRMS before starting to clip.
In the Emotiva XPA-1 every PNP transistor is matched by an NPN transistor for the entire circuit, so the voltage gain stage is also complementary to reduce distortion. To accomplish this, one additional differential pair with PNP input devices is added to the schematic in Figure 2 of the paper by Karki.
The common mode feedback circuit shown in Figure 2 of the Karki paper is not the same as the Emotiva XPA-1. We will discuss the circuit Emotiva uses in the next section.
The current gain stage connected to the speaker is an emitter follower triple. Six transistors in parallel are used in the final stage of the triple. Since this is a balanced and complementary map, 24 transistors are connected to the speaker.
With the single voltage gain stage, it is important that the input of the current gain amplifier has a high resistance at its input. The gain of the voltage gain stage is directly dependent on the loads it sees.
Common Mode Feedback
The fully complementary topology allows Emotiva to make the common mode feedback very simple in the XPA-1. Emotiva observes that as both speaker terminals move away from ground, producing a common mode error, each of the complementary differential pair tails also moves
The amplifier senses the movement of the differential pair tails. If the speaker terminals drift in the positive direction followed by the differential pair tails, more current is forced into the NPN-based differential pair, pulling the speaker terminal back down
If the speaker terminals drift in the negative direction, more current is forced into the PNP-based differential pair, pulling it back up.
The common mode feedback loop degenerates to being just resistive in the actual design, instead of the standard common mode feedback circuit that uses an additional internal op-amp (Figure 2 of the Karki paper).
How Emotiva designed such a simple circuit that works so well is kept proprietary. The very simple common mode feedback circuit makes stabilization of the common mode feedback easier.
Compressive Fault Protection
The Emotiva XPA-1 has a relay in the speaker line to prevent any fault condition from damaging the amp or the speaker. If the fault condition is a shorted transistor, the relay will prevent the short form being seen by the speaker. The unit has a special chip common to modern amplifiers (uPC1237) that monitors for excess current draw of the output transistors (perhaps an external short circuit or a defective transistor which will require service) and DC offset at the speaker terminal output. Disconnection of the power supply on over-temperature is done with a different circuit.
The IC (Integrated Circuit) times when the relay closes at power-up and opens at power-down. It will keep the relay closed until it is sure a fault condition exists, and the amp will not open the relay when the amp is driving a difficult speaker load.
The fault condition status is coordinated by a micro-controller. Some faults are displayed on the front panel.
The unit has a standby power mode also controlled with relays.
SNR levels lower than expected
Amplifier noise was discussed in: The Fully Balanced Power Amplifier – Advantages and Design Challenges
The measured noise at the output was 120 µVRMS over a bandwidth of 20kHz with the input grounded. The Emotiva XPA-1 has a gain of 31.5V, so the input referred noise is 3.8µVRMS.
This is an inverting topology, which can have the noise floor dominated by the input resistors thermal noise.
The balanced input has a 20kOhm resistance. At 800 F, the thermal noise is 2.6µVRMS. I found this value using an on-line thermal calculator.
The excess noise beyond the thermal component is mostly the enhancement from base current noise, although the amplifier has other components that produce noise and could add to the total.
When the amplifier is switched to class A the SNR was found to decline by 1.4dB. The internal temperature increased by more than 400 F and resultant increase in thermal noise accounts for some of the decline in SNR.
The SNR can be improved if the 20kOhm input impedance is dropped to 200Ohms. The input would then be driven by a buffer that would present a light load to the preamplifier. Indeed, the buffer is in place for the single-ended to balanced converter, although the op-amps in place cannot drive 200Ohms. The circuit would be discrete or an op-amp – discrete composite. The buffers must be single ended for reasons outlined in the introductory article.
Conclusions about the Emotiva XPA-1 Gen 2 Monoblock Power Amplifier
Emotiva made their name with big, beefy amps, and the XPA-1 Gen 2 continues that tradition. Building upon the original with a Class A mode, it has lots of power in reserve. When it comes to pure wattage value, the XPA-1 is a tough beast to top. Running a full range tower speaker, the Revel f208, I never came close to maxing out the XPA-1s’ reserves. Even those with large planar speakers should not worry about running out of juice here.
The only ergonmic faults I have with the XPA-1 are the placement of the speaker binding posts so far apart, and the lack of handles to move them (they are really heavy).
With performance I find little to disagree with on the XPA-1 Gen 2. If you have power hungry speakers, or you just want to be certain you have plenty of headroom when you play orchestral pieces with pipe organ, this power amplifier will do the job.