Power to the People
Power amp giant test
Austin Armstrong, BSS, C-Audio, MHA Chameleon, and RCF battle it out
You need more power from your amp than you previously thought. You need an amp that'll stand up to both studio and live use. You need to know what to look for when shopping around for the best buy. Well, your prayers have just been answered. Ben Duncan takes five leading power-merchants for a long, hard drive...
Perhaps I should begin with some practical advice for those of you unwilling (or unable) to shell out £1-2,000 on a new power amp.
Get yourself a heavy-duty rotary switch, so you can switch a stereo amp between nearfield and main monitors. This gives you twice the budget for a single machine. Or go for a PA amp, which you can hire out or use for gigs. Each of the five amps reviewed here boasts that versatility, without compromising their studio performance. So make your considerations practical: can I afford it, have I the room for it, and does it give me enough headroom?
Unglamorous, unsung, and very often unseen, the power amps are perhaps the forgotten music-production tool. In the studio, a beefy power amp can make your whole monitoring system more listenable (not to mention more reliable) because it has sufficient headroom to drive the speakers in question. Live, having a power amp with enough punch for the job can be the difference between a satisfactory performance and a gig that is unexpectedly curtailed halfway through.
So if you think you can cut corners on amplification, think again. Save your pennies, buy the best you can afford - and you may just save the rest of your system...
Austin Armstrong's 409C has found favour with MGM (the movie people), and has also been extensively 'rave-tested'. Interestingly, Austin-Armstrong's Dave Collins also founded C-Audio.
The ST600 from C-Audio is new this year, and aims to provide an 'entry level" amp, for those on tighter budgets.
RCF are an Italian drive-unit and speaker manufacturer whose AS6001 amp is made in Canada. It isn't big in the UK (so far), but has received much acclaim in the US.
The founders of Hill Audio have recently separated to become two separate organisations, Millbank and Malcolm Hill Associates. It's the latter company whose MHA Chameleon power amp is tested here.
BSS Audio's EPC760 is made in St Albans and is the smaller of two models targeted principally at the PA market.
Gain controls are commonly omitted on domestic power amplifiers, but are useful in music production for two reasons.
First, they can protect your monitors when all the mixer controls are wound fully up. Second, most can be used as (preset) stereo balance controls - as budget mixers may omit such a control.
When studio control-room level is set by a stereo rotary pot, and it's anywhere below maximum, L/R balance can swing about wildly. Of the five, only the BSS has ordinary rotary gain controls. The Austin-Armstrong, C-Audio and RCF have click-stop rotaries, but this is a dubious advantage when half the time the exact knob position is delicately balanced at the edge of a 'click'.
The MHA is one of few power amps to have made a radical departure from the norm just described: gain is controlled in discrete steps with a thumbwheel switch - of the kind once used on mixers for subgroup routing. It has a great feel. However, the gain steps (0dB, -2, -4, -6, -10, and so on) may not be quite the ones you need, and as the finest increment is no less than 2dB, there is no chance of using these controls for preset stereo balance trim.
Unusually - and sensibly - the Austin-Armstrong's gain controls are on the rear panel, preventing changes from knocks and unauthorised twiddlings. The BSS also has a (second) set of gain controls at the back, in the form of screwdriver-set presets calibrated against input sensitivity.
All five of these amps are fan-cooled. The RCF makes the quietest début, with a fan that isn't triggered until the amp breaks into a sweat. Since the design is highly efficient (some high-power tests were performed to prove this), the fan is not likely to run unless some seriously loud monitoring into 4ohms goes on for an hour or more. Even then, the fan is not very loud, especially considering it's on the outside.
The Austin-Armstrong, BSS, and MHA all have quiet fans located 'inside the box', which helps keep an even firmer grip on noise. The speed of the BSS amp's fan increases fairly imperceptibly as the amplifier heats up. In the studio, a lightweight cupboard with only basic soundproofing is enough to make these amps acoustically unobtrusive.
C-Audio's ST600 has by far the loudest fan - on full bore, full time. The acoustic nuisance this causes is enough to rule this amp out for serious studio use unless you have access to a properly soundproofed cupboard, or can else install it in an adjacent room or - ignominy! - the corridor outside.
The BSS has (as ever) the the greatest number of LEDs. Horizontal LED bars indicate each channel's temperature, while vertical bars display peak signal, simultaneously confirming signal presence and warning of clip. One flashing LED per channel indicates a fault, causing 'intelligent' shutdown, the fault being diagnosed according to how and in what order the LEDs flash.
There are soft mute buttons for each channel and mute LEDs, and it's interesting to note that the BSS' output steps up from a low level after the switch-on delay period - less shocking than appearing instantly at full bore after un-muting!
The Austin-Armstrong has signal metering with bar LEDs, like the BSS, and a further LED indicating groundlift. After that, in common with the RCF, MHA, and C-Audio, you're more or less on your own, the one exception being clip indication.
The MHA has an anti-clip facility called 'headlok' which can be manually set, or else engages automatically, to protect your monitors in the event of overdrive. The RCF has a rear-panel button to insert a sub-sonic filter, which is particularly relevant to PA use as protection against dropped and thrown mics. Intriguingly, the RCF also promises 'LF synthesis', shifting any information below 40Hz upwards, where it will be potentially more audible and less harmful All the amps have 4mm binding post outlets. The BSS has two sets, ideal for biwiring, while the Austin-Armstrong has Speakon outlets as well.
Figure 1 shows the Austin-Armstrong's bandwidth - a textbook response if ever there was one. Low-frequency (LF) response is maintained to within 0.5dB of the midband norm, down to 20Hz. Likewise, high-frequency (HF) response extends to just beyond 20kHz before it drops 0.5dB and continues dropping at an accelerating rate. The MHA's response is much the same.
The BSS's extends at least an octave lower, with -0.5dB at about 8Hz suggesting some subtle differences with the lowest notes. It's worth mentioning that phase response is affected at frequencies up to a decade higher by an extension of this kind - so sonic changes to musical components up to 200Hz cannot be ruled out.
At high frequencies, the BSS is also down -0.5dB b 30kHz, but thereafter if falls far more steeply, dropping 'off the bottom' (>-7dB) at 65kHz. Again, this steeper fall will affect phase response some way down into the audio band. However, it is worth noting that the sonic effects of phase shift from steep-cut, low-pass filters are much less serious than many once imagined, and that differences in the amount of phase shift are more audible at low frequencies.
The C-Audio's response is like the BSS at LF, but a HF, it begins rolling off unusually gently and early, at 4kHz. The roll-off is still -0.5dB just above 20kHz, but we might expect some subtle 'dulling' or 'mellowing' seeing as the response in the entire upper half of the audio band is slightly 'tilted'.
The RCF (Figure 2) peaks at both frequency extremes. It is true that the effect of the peak at both frequency extremes will be small (in one sense), because there isn't much information around 30-40Hz where the LF peaks, or at 10-11 kHz where the HF peaks. However, the phase response will be messy, and notice how the frequency response through most of the audio range is tilted up or down, visible even on this diagram's small scale. Such 'micro-tilt-EQ' may be very audible as coloration to some users.
"The Austin-Armstrong's gain controls are on the rear panel, preventing changes from knocks and unauthorised twiddlings"
Spec fiends - and US makers in particular - have long been obsessed with the Damping Factor as a measurement of amplifier virility. High damping was held in be important for a taut or accurate bass. To some extent this is true, but the affects have been overstated, and the consequence has been a drive to achieve high damping at low frequencies - with scant regard paid to any trade-offs that might result.
In turn, few amplifiers have damping factors that are completely constant across the audio band. Most are specified for bass, and may degrade appreciably at mid and high frequencies. Some would argue (correctly) that this shouldn't matter, as HF drivers, such as dome tweeters, don't have much mass or back-EMF, so they don't need as much damping.
Again, though, the point is being missed. Damping factor is really an expression of an amplifier's output impedance. 'High damping factor' means 'very low impedance'.
All amplifier outputs have some finite impedance, albeit a small fraction of an ohm, and from this, the voltage at the amplifier's output will be depressed (albeit only very slightly) by the speaker's impedance. The cables' steady state and dynamic impedance adds another tier of variables.
In practice, speaker impedances vary widely with frequency and even program dynamics, and it follows that as they vary, the amplifier's output signal varies (however slightly) in sympathy. The amount to which this occurs depends on the amplifier's output impedance.
The upshot of all this is as follows. First, the frequency response of amplifiers with low damping factors - alias a relatively high output impedance - varies more strongly when the amp is connected to different speakers. The sound will also vary accordingly.
Secondly, amplifiers with high damping/low impedance at some frequencies, commonly degrade to higher impendance at other frequencies, leading to more variable sonics in the frequency band where impedance is higher. This means that an amp's high-frequency performance may be its most unpredictable tonal aspect, as you try different speakers.
Finally, low damping/higher impedance magnifies any sonic aberrations in proportion to the speaker's own impedance behaviour: the most predictable match is where both speaker and amp have the most constant impedance across the frequency band, even if it isn't the lowest.
Figures 3-7 show pairs of highly accurate, magnified frequency response plots. By measuring the tiny difference between each channel's unloaded and loaded output voltages, output impedance can be calculated (Table 2).
The differences are in any case directly visible, as the size of the gap between the each pair of plots, except that the scales in Figures 5 and 7 have been changed to fit in the whole picture.
Looking at Table 2, note the MHA and RCF have considerable inter-channel variation - something which, subjectively, can be expected to affect stereo imaging. One of the RCF's channels has the lowest output impedance. The BSS is almost as good and both of its channels are almost identical. The C-Audio 600 has the highest impedance, but again, at least the channels are consistent.
When considering sound quality, however, it's more important to note the frequencies where the midband output impedance changes. For example, in Figure 3, the gap closes up in the low bass on the left, signifying increased damping and hence less interaction with any speaker in the low bass. By contrast, in Figure 4, the distance between the plots increases slightly in this region, signifying less damping and increased sonic dependence on the loudspeaker's impedance qualities.
So, on this basis, how do the various amplifiers in our test score?
The Austin-Armstrong's impedance increases from low bass upwards, but is still not excessive at 20kHz. The BSS's is constant except at extreme high and low frequencies; the C-Audio's is steady to 3kHz, then degrades; and the RCF's is similarly regular to 4kHz, then increases wildly.
The MHA Chameleon is the best all-round performer, with relatively consistent impedance, even at extreme frequencies.
All the amps have balanced, XLR inputs. The Austin-Armstrong, MHA, and RCF have balanced jack inlets as well, but these would be less reliable in a fixed setup.
Common Mode Rejection (CMR) measures how well the inputs reject unwanted garbage. The BSS and Austin-Armstrong had the best CMR: better than -90dB at the principal 50Hz hum frequency, and still -40dB at ultrasonic frequencies.
The BSS had by far the lowest CMR in places, but the channels were less well matched. The C-Audio had a more average CMR of just better than -50dB. At least both channels were almost identical, and the same -50dB holds out to ultrasonic frequencies.
The worst CMR was still -40dB, no worse than the others. The MHA had divergent CMRs of -43 and -63dB for each channel. Again, at least both were almost maintained up to 200kHz.
The RCF managed a steady and close-matched -57dB in the majority of the audio band, degrading from 10kHz upwards, to a minimum of -40dB a couple of octaves above audio.
In any event, all five amps will provide satisfactory rejection with correct wiring, shielding, and grounding techniques. But the BSS and Austin-Armstrong would provide by far the most-potent hum rejection - useful in extreme cases, such as when you have to route the input cables alongside high current mains cabling.
Next, each amplifier's noise was plotted against frequency. The Austin-Armstrong and BSS had the lowest LF (hum or low buzz) noise. The C-Audio had the highest 100Hz 'sawtooth' ripple (an edgy, buzzy kind of hum). One channel of the MHA Chameleon had LF noise as low as the Austin-Armstrong, but the other channel had the highest 50Hz pickup (a rounded 'hummm'). This channel also had the highest midband noise. The RCF also had rather high 50Hz hum on one channel, but it was otherwise quieter than the MHA and C-Audio. The unweighted, averaged figures rank almost the same, as shown in Table 3.
Going on these figures, all five are perfectly acceptable, with the possible exception of the MHA's 'bad' channel. But bear in mind that buzzes and hums can affect subjective sound quality by subliminally 'warming' or otherwise tainting what we hear.
Depending on your speakers' bass response, overall sensitivity, and whether they're nearfield monitors or not, some of the amps' hums 'n' buzzes may be more of a nuisance than others.
Distortion was plotted against frequency at high and moderate levels, and against level at a mid frequency, and a whole article could be given over to dissecting the entrails. Alas, as far as distortion is concerned, even such sophisticated test results as these don't correlate very well with what's heard.
For example, the Austin-Armstrong's distortion was lowest in the low bass at 30Hz, increasing past 0.1% at 1kHz to almost 1% at 20kHz. As the HF rise is caused by a 'crossover spike', you might imagine the unit would be unlistenable, but this was not the case.
The BSS had high distortion at high drive levels, falling off after it heats up, or at lower levels. C-Audio's ST600 (Figure 8) had mostly low distortion, not rising above 0.1% below 15kHz at any level below clip, but it rose to above 1% below 40Hz. The Chameleon and RCF behaved similarly overall.
"The BSS' output steps up from a low level after the switch-on delay period - less shocking than appearing instantly at full bore"
Before listening, all the amps were warmed up by leaving them switched on for an hour, without signal. Other than using minimum, quality source equipment and known, quality-recorded music, listening to amplifiers requires great care. If one unit is slightly more sensitive, the increased level can make the sound more attractive, and change many other characteristics, too. This can be overcome to some extent by listening over a range of levels. Interestingly, all the amps required slightly different stereo balance settings.
Each amp was listened to, then used as a reference for the next. The test was then repeated in a different order, and without recalling the past results, until all the permutations were worked through 'blind'. Listening hardware included an Audio Synthesis D/A converter for the clearest CD replay, and Tannoy DMT-II monitors.
As the amplifiers' different output impedances are (for once) documented, it shouldn't be necessary to mention that sonic results will depend on the speakers used - but to a considerably lesser extent with the BSS and Chameleon than with the others.
The Austin-Armstrong had deep imaging with notably fine layering and HF 'zing', and also bass low enough to go 'thud'. As mentioned earlier, the sample unit had a spikey residue on both channels. This may explain a mid harshness and a nasal quality.
The BSS showed perhaps the greatest and finest detail, and fine imaging. It revealed plenty of sonic texture but there was some brittleness among all this.
The C-Audio ST600 sample developed RF oscillation (detected with a 'scope) on one channel, after managing the bench tests.
To be fair, its Zobel network is not designed for sine-wave sweep tests, and would probably not burn out in normal use. Thereafter it could only be tested in mono, where it too exhibited percussion 'zing' absent in others, along with above-average detailing.
MHA's Chameleon had a mid hardness and lacked HF zing, but it also had the deepest bass - though seeming a little less well-damped than others. Dynamics were uniquely 'punchy', while imaging was strong but somehow less deep, as if the stereo picture had been 'telescoped'. Before listening to the MHA began, a hum was noted and moving the ground switch successfully counteracted this.
The RCF was rather hard or harsh, or else emphasised high-midrange sounds that were. It too lacked percussive zing, but on the credit side, its bass went deeper than most of the others. The optional LF synthesis' switch position was tried but had no discernible effect on bass, perhaps indicating that there wasn't any programme information in such a low frequency band. Overall imaging and layering were above average, but the image seemed shallow. Textures in the upper mid and high frequencies were better developed than others, but inferior or lacking in the bass/low-mid region - compared to the Austin-Armstrong, say.
"The RCF makes the quietest début, with a fan that isn't triggered until the amp breaks into a sweat"
These are all serious power amps with serious capabilities. As well as the studio listening tests discussed above, all five amplifiers had previously been rigorously tested for touring PA use - leaving us in no doubt that all five can drive high levels in a studio all day into 8ohm or even 4ohm loads, without faltering.
The five were also 'abuse'-tested. If you were stupid enough to short the speaker outputs with loud programme present, you'd have to replace some solderable fuse links inside the Austin-Armstrong. And you just might risk damaging C-Audio's ST600 if it was already hot and then shorted.
Otherwise, all five amplifiers are certified rugged, subject to proper ventilation and racking - which, of course, is where you come in.
Price: £999 inc VAT
More from: Harman Audio, (Contact Details)
Price: £1,524 inc VAT
More from: Malcolm Hill Associates, (Contact Details)
Price: £1,108 inc VAT
More from: Stable House, (Contact Details)
Price: £1,757 inc VAT
More from: RCF UK. (Contact Details)
Price: £1,880 inc VAT
More from: BSS, (Contact Details)
Gear in this article:
Review by Ben Duncan
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!