GoldenEar Aon 3

ON TEST

GoldenEar Aon 3 Loudspeakers

W

hen Sandy Gross retired from Definitive Technology (DefTech) back in 2010, the company he helped found in 1990, he said: ‘For decades I’ve worked 24/7 running my businesses and I saw the time approaching when I would want to slow down, take a break and contemplate my future plans.’ Just six months later he started yet another loudspeaker company, GoldenEar, whose Aon 3 we’re reviewing here. And when I say ‘yet another company’ it’s because Gross not only founded the company he retired from in 2010, he was also one of the founders of Polk Audio, ‘way back in 1972. (His co-founders were Matthew Polk, George Klopfer and Craig Georgi.) After 40 years in the hi-fi loudspeaker business, why would Gross want to start all over again from scratch? It appears that one of the reasons was that Gross is trying to reduce the cost of high-quality loudspeakers. The company’s website explains it this way: ‘We don’t feel that having less to spend means

40

Australian

that you are looking for lesser quality. We lavish the same kind of care, expertise and experience that is usually reserved for the highest-priced high-end speakers on every GoldenEar speaker that we create.’

The Equipment Surprise, surprise! Whereas almost all most small bookshelf/standmount loudspeakers have two drivers at the most, each GoldenEar Aon 3 has four of them… well, strictly speaking two of these aren’t technically ‘drivers’ as such, because they don’t have voice coils and magnets, and so don’t connect to your amplifier at all—instead they’re just ‘passive’ radiators (also known as ‘PRs’, or drone cones), designed to make use of the energy issuing from the rear of that single front-firing bass/midrange driver… energy that is usually either absorbed inside the cabinet in a sealed box design, or routed through a hole in the cabinet in the case of a bass reflex design. But let’s first take a look at the electricallydriven driver, and this bass/midrange driver

is said to be made exclusively for GoldenEar, which I’d certainly believe, due to the very unusual exposed pole-piece/phase corrector, which is fluted. (GoldenEar calls it an ‘MVPP’ or ‘Multi Vane Phase Plug’ design, and has trademarked the initials.) I have never seen a fluted phase corrector and I’ve never seen one that protrudes so far. The one on the Aon 3’s driver is so long that it very nearly extends beyond the plane of the driver chassis. Indeed if it extended any further, it would be causing a bulge in the ‘sock’ grille that covers the entire loudspeaker (about which more later). The driver itself is very solidly constructed, with the chassis made from a casting, rather than from pressed steel, and it supports a very (very!) large unshielded magnet. The cone appeared to be made from polypropylene. The overall diameter of the driver basket is 155mm, with the driver mounting holes on 163mm centres. The moving diameter is 140mm, but the diameter of the cone itself is 125mm. However, the most important


dimension—the Thiele/Small diameter, which is what loudspeaker designers use to calculate the effective cone area (technically known as ‘Sd’)—is 134mm, putting the Sd of the Aon 3’s front-firing bass/mid driver at 141cm². The two side-mounted passive radiators are identical. Each one is circular, with a flat ‘plate-style’ diaphragm and has a Thiele/ Small diameter of 166mm, for an Sd of 216 cm² per driver. Unlike some manufacturers, whose passive radiators are essentially loudspeakers from which the magnets and voice coils have been removed (or, more accurately, never installed in the first place), GoldenEar has cut costs by using what is literally a ‘plate’ radiator. The radiator is a circular section of MDF, to which is attached a roll surround, and this roll surround is in turn attached to a ring of pressed steel (rather like a giant washer) that’s used to attach the radiator to the cabinet. Using MDF means the cone is quite heavy, but it seems that it was not quite heavy enough to ensure the correct ‘tuning’ for the cabinet because GoldenEar has attached steel discs to the back of the MDF plates in order to further increase the moving mass. I have to say that GoldenEar has done a splendid job on these passive radiators, because to look at them, they look exactly like ordinary bass drivers, however, if you touch one, you can instantly tell that it’s not a bass driver because pushing one side of the diaphragm inwards causes the other side to move outwards, which would not happen with a real bass driver or a PR with a spider. Bass is always a trade-off for any given cone size. You can have loud bass, or you can have deep bass… and you have to pick one,

because you can’t have both at the same time. If you do want to have your cake and eat it too, the only option is to up your driver area, and your cabinet volume. There’s no shortage of driver area with GoldenEar’s Aon 3 because when you add together the areas of the bass driver and the two passive radiators, you arrive at a total radiating area of 573cm², which means that if this was the Sd of a single driver, it would have a ‘nominal’ diameter of around 330mm! (In fact, technically you can’t ‘sum’ the radiating areas in this way, because the two passive radiators make maximal use of their cone areas at frequencies where the cone makes least effective use of its cone area, but in this case, because of the volume of the cabinet and its tuning, it’s a reasonable approximation.) Although GoldenEar refers to the tweeter as an ‘HVFR’ tweeter, and has trademarked these initials (which stand for High Velocity Folded Ribbon) in the USA, it certainly hasn’t trademarked or patented the design of the tweeter itself, because this type of tweeter hails all the way back to 1972 (coincidentally the same year GoldenEar’s Sandy Gross co-founded Polk) when it was invented by the legendary loudspeaker pioneer Oskar Heil. Heil called his design an ‘air-motion transformer’ because of its operating principle, which is rather like the moving part of a piano accordion, being in essence a pleated membrane that’s driven from its edges. As the ‘pleats’ or ‘folds’ move towards each other, they compress the air between them and move it outwards, to create sound waves. The beauty of Oskar’s system is that because the ‘pleats’ in the membrane are highly efficient (think of the difference between ‘squirting’ an orange pip away from you by squeezing it between your fingers, compared to throwing the pip with your hand!), the motor system has neither to be large nor high-powered. Also, because the pleated membrane is inherently ‘loose’, its resonant frequency is well outside its operating range—something that can’t be said of any dome tweeter I’ve ever heard of. Heil’s ‘air motion transformer’ (sometimes known by its initials… AMT) was licensed by many companies until the patent expired,

ON TEST

after which even more manufacturers adopted the idea and developed their own versions. German company Elac calls its version a ‘JET’ tweeter, while other companies call them AVT (Air Velocity Transformer) tweeters. Adam Audio calls its version ‘ART’ (Accelerated Ribbon Technology). I’ve singled out Adam Audio specially because I think that Adam Audio’s explanation of the benefit of air motion transformer technology is particularly clear, so I’ll reproduce it here: ‘All other loudspeaker drive units—whether they are voice-coil driven, electrostatics, piezos or magnetostatics—act like a piston, moving air in a 1:1 ratio. This is undesirable, as the specific


Brand: GoldenEar Model: Aon 3 Category: Standmount Loudspeakers RRP: $1,595 Warranty: Five Years Distributor: Kedcorp Pty Ltd Address: Unit 8, 509–529 Parramatta Road Leichhardt NSW 2040 (02) 9561 0799 [email protected] www.kedcorp.com.au • Small footprint • Bass depth • High frequencies

• Cabinet shape • Choice of finish • Speaker terminals

LAB REPORT Readers interested in a full technical appraisal of the performance of the GoldenEar Aon 3 Loudspeakers should continue on and read the LABORATORY REPORT published on page 44. Readers should note that the results mentioned in the report, tabulated in performance charts and/or displayed using graphs and/or photographs should be construed as applying only to the specific sample tested.

Lab Report on page 44 avhub.com.au

41

ON TEST


weight of air is much lower than that of the driving mechanics. Speaking in terms of electrical engineering one could say there is a bad match between source and load. The [air transformer] principle achieves a 4:1 velocity transformation between (the) driving diaphragm and the driven air. In other words, the air moves in and out four times faster than the folds are moving. This superior motor system is responsible for the enormous clarity and transient reproduction that is to be heard from the A.R.T. drive units.’ I think that it is worth noting that although GoldenEar uses the word ‘ribbon’ to describe the tweeter, it’s not really what I’d describe as being a ‘ribbon’, and works completely differently to how I’d personally describe a true ‘ribbon’ tweeter as operating… not that this is a bad thing: true ribbon tweeters have myriad shortcomings that work to offset their advantages, which is one reason why you don’t see them used more often. The Aon 3’s crossover network is mounted on a PCB and comprised of two inductors (one air-cored, the other ferritecored), five capacitors (four MKT, one a bipolar electrolytic) and two resistors (one metal film, the other cermet). The speaker terminals on the rear panel of the Aon 3 are gold-plated ‘post and screw’ types on 19mm centres. The terminals are recessed in the cabinet in such a way that I found it difficult to access them to insert and tighten speaker wires, so I’d suggest using banana plugs. Alternatively, because of the 19mm centres, you could use dual banana plugs, which would make things even easier. The peculiar shape of the cabinet (GoldenEar calls it a ‘truncated pyramid’) makes it difficult to imagine the dimensions, so you should certainly check out a pair of Aon 3s in the flesh, but each cabinet stands 360mm high and is 226mm wide at the base (a mere 156mm wide at the top) and 280mm deep at the base. GoldenEar has used the economical ‘sock’ method to finish the speakers, where a tube of stretch cloth is fixed around the base of the speaker then ‘pulled up’ to cover all four sides, then fastened with a drawstring at the top of the cabinet. The drawstring is then covered with a black plastic top plate that ‘plugs’ into the top of the cabinet. (It is plastic, but to call it plastic really does it a disservice, because it’s so beautifully coated in high-gloss black piano gloss that it looks for all the world like it’s made of wood!) It appeared to me that the cabinet itself is made for the most part from 19mm thick MDF with some slightly thinner sections.

In Use and Listening Sessions If I hadn’t heard it with my own ears, I would not have thought it possible. That’s pretty much exactly what I thought about the level

of bass issuing from the Aon 3 speakers. I’ve heard deep bass from small speakers before, but never so deep as from the Aon 3s, and certainly never so loud! Self-evidently this deep bass is made possible by the fact that there’s a bass driver as well as two passive radiators delivering all the bass… and lots of it. As I noted earlier, the diaphragms of the passive radiators are not constrained by a spider and I wondered if the drivers would operate pistonically. Just from eye-balling them, I’d say they do, and logically I suspect that this would have to be the case, since the air pressure inside would act equally over the diaphragm surface, essentially forcing a pistonic diaphragm action. However, although there was plenty of bass, and that bass was certainly extended, the lower bass was not what I’d call ‘tight’. To my ears the low bass

I’ve heard deep bass from small speakers before, but never so deep... and certainly never so loud... from the Aons was rather ‘warmer’ than I hear from my reference speakers and I also fancied there wasn’t quite as much fine detail either… though this only came into play when several low-voiced instruments were playing simultaneously. The midrange sound was beyond reproach. The detailing that’s clearly audible right across the upper bass and midrange means you won’t miss even the tiniest musical nuance, and there’s a ‘snap’ to the sound that gives it a real presence and liveliness, so that it’s not hard to imagine you’re at a live performance. Tonal purity was outstanding… I suspect the inherent distortion of the bass driver in the Aon 3 is very low. I pulled out a recording I’ve been using for years, Sheffield Lab’s incredible capture of Margie Gibson singing ‘The Song has Ended’, which has Lincoln Mayorga tinkling along on a Mason & Hamlin concert grand. The Aon 3s keep Gibson’s voice out front where it was recorded, and the separation between her voice and the accompanying instruments is spectacular. Listen to how the GoldenEars let her voice sustain, even when she’s scatting. Also noteworthy is the way the Aon 3s can deliver Chuck Domanico’s great bass lines… I’ll wager you’ve never before heard double-bass so clearly and so well-balanced against the midrange from such a small loudspeaker. The vocal clarity enabled by the Aon 3s was also a highlight when I was listening to Lover’s Mask (from Michael Ruff’s ‘Speaking in Melodies’, now available from Town Hall Records). The harmonies were incredibly good, with each voice clearly delineated, but

with all merging together to create spinetingling vocal harmonics. The force and majesty of the percussion is delivered with real impact by the Aon 3s. The clarity of the speakers is such that you can also hear things you weren’t supposed to hear… such as vocal asides about who’s to take the next solo! GoldenEar’s HVFR tweeter is also a stunner. First because it enables a completely seamless transition from the bass driver… try as I might (and I did) I simply could not hear the point where the bass/midrange driver handed over to the tweeter; it was as if the sound were being delivered by a single speaker (eat your heart out you Lowther-lovers!). Second because of the air and spaciousness of the high-frequencies it delivered which, coupled with the effortlessness of the delivery itself, made the high-frequencies issuing from the GoldenEar Aon 3s high-frequencies to swoon for. Third, because unlike most tweeters, which can sound a little ‘light-on’ in the extreme highs, the HVFR just kept powering upwards… and man what a difference this makes to cymbals… and that’s just one example. Fourth, because the imaging and sound-staging I was hearing at such high frequencies was palpable. But… and it’s only a little but… I did find that to obtain this level of imaging and soundstaging I had to position the Aon 3s on stands so the HVFRs were at seated ear level, and I had to ‘toe-in’ the speakers so they aimed at the listening position. The HVFRs most certainly have a defined ‘sweet spot’. Choosing an amplifier to partner the GoldenEar Aon 3s won’t tax your listening skills, because I found they weren’t at all picky about what I used to drive them. I found that 50-watts per channel was plenty powerful enough to get ‘em sounding their best, and even more than enough to overdrive them if I tried… though on the odd occasion I did manage to overdrive them they succumbed quite gracefully, with the sound becoming ‘confused’ rather than obviously distorted.

Conclusion Some people might be put off by the shape of these speakers, and others by the exterior finish (black cloth is not exactly a decorator favourite), but I can’t see anyone objecting to the sound quality of these speakers—particularly that bass!—or to their size. I suspect that when Sandy Gross knew he was going to come out of retirement and start up GoldenEar, he knew that in order to gain some quick publicity and guarantee that retailers (and reviewers) would come on board and promote the new brand, he’d have to come up with a design that would, as they say, ‘knock ‘em in the aisles’… and with the Aon 3, he’s done exactly that. Jutta Dziwnik LAB REPORT ON PAGE 44

42

Australian

LAB REPORT


CONTINUED FROM PAGE 42

Laboratory Test Results Frequency responses don’t come much flatter than the one GoldenEar’s Aon 3 delivered in Newport Test Lab’s acoustic laboratory… you only need to take a look at Graph 1 to see that. From 300Hz upwards, the trace is virtually clamped to that 80dBSPL reference level line, save for a tiny ‘sag’ between about 1.3kHz and 7kHz. As you can see, this trace is effectively a room response, being the averaged result of nine individual sweeps measured at three metres at nine different axes with the central grid point of the nine being on-axis with the tweeter. I am particularly impressed by the high-frequency performance of the Aon 3’s tweeter. Going downwards, below 300Hz there’s a little ‘bump’ that’s partially room related, but it’s no more than +2.5dB. Below 100Hz the response rolls off smoothly but quite slowly, so it’s only 6dB down at 53Hz. The high-frequency response of the Aon 3 is shown in Graph 2, with a lower measuring limit of 400Hz. This response simulates the response that would be obtained in an anechoic chamber, so you’re looking at the true response of the tweeter. The use of a different measurement technique (gated sinus) also allows more precise detailing of the response, whereas the pink noise test signal used in Graph 1 involves some averaging. However, even with this more precise measurement technique, the high-frequency response is superbly flat, essentially 400Hz to 12kHz ±1.5dB. Above 12kHz the response rises a further 2.5dB before rolling off to be at 20kHz the same level it was at 12.5kHz, then rolling off 5dB at 26kHz before picking up and regaining reference level at 35kHz. Across the whole frequency range depicted on this graph (400Hz to 40kHz) the GoldenEar Aon 3’s response varies no more than ±3.7dB. Graph 3 shows the low-frequency performance of the bass driver and passive radiators. The bass driver produces its maximum output at 130Hz and above this is very flat. (There is a slight roll-off, but this is partially a function of the near-field measurement technique used.) Below 130Hz the driver rolls off at 18dB/octave to a minima at around 47Hz. The passive radiators’ output is unusual. They start working quite efficiently at 38Hz then deliver almost exactly this same output level up to 100Hz, after which they roll off at around 6dB per octave to 200Hz. From around 250Hz up to 600Hz the radiators still contribute to the output,

44

Australian

dBSPL

100

Newport Test Labs

95 90 85 80 75 70 65 60 55 50 45 40

20 Hz

50

100

200

500

1K

2K

5K

10K

20K

Graph 1. Averaged frequency response using pink noise test stimulus with capture unsmoothed (red trace) Trace is the averaged result of nine individual frequency sweeps measured at three metres, with the central grid point on-axis with the tweeter. [GoldenEar Aon Three Loudspeaker]

dBSPL

100

Newport Test Labs

95 90 85 80 75 70 65 60 55 50 45 40 400 Hz 500

1K

2K

5K

10K

20K

40K

Graph 2. High-frequency response, expanded view. Test stimulus gated sine. Microphone placed at three metres on-axis with dome tweeter. Lower measurement limit 400Hz. [GoldenEar Aon 3]

100

though at a fairly low level (around 20dB below the level of the bass/ midrange), and in a relatively nonlinear manner. The GoldenEar Aon 3’s impedance is quite benign (as in ‘very easy to drive’), as you can see in Graph 4, with the impedance dipping beneath 4Ω only around 200Hz and below 14Hz and the phase angle never going beyond ±42°. I would, however, have liked to see the impedance rising above 20kHz, rather than falling as it does. It’s not a problem, just a better design approach. Although GoldenEar’s specification says the Aon 3’s impedance is ‘compatible with eight ohms’—and it most certainly is compatible with 8Ω—it’s technically a 4Ω design. The impedance curve shows the classic ‘double hump’ of a bass reflex design (which is what the Aon 3 is, with the passive radiators just substituting as ports), with the saddle

Frequency response was 60Hz to 40kHz ±3dB, which is an excellent result!

dBSPL

Newport Test Labs

95 90 85 80 75 70 65 60 55 50 45 40

20 Hz

30

40

50

60

70

80

90 100

200

300

400

500

600

700

800 900 1K

Graph 3. Low frequency response of side-firing passive radiator (red trace) and woofer. Nearfield acquisition. Radiator/woofer levels not compensated for differences in radiating areas. [GE Aon3]

Ohm

30

Deg

Newport Test Labs

180 150 120

20

90 60 30 10

0

9 8

-30

7 -60 6 -90

5

-120 4 -150 3

10 Hz

20

50

100

200

500

1K

2K

5K

10K

20K

-180 40K

Graph 4. Impedance modulus of left (red trace) and right (yellow trace) speakers plus phase (blue trace). [GoldenEar Aon Three Loudspeakers]

100

dBSPL

Newport Test Labs

95 90 85 80 75 70 65 60 55 50 45 40

20 Hz

50

100

200

500

1K

2K

5K

10K

20K

Graph 5. Composite response plot. Red trace is output of bass radiator. Light blue trace is anechoic response of bass driver. Pink trace is gated (simulated anechoic) response above 400Hz. Black trace is averaged in-room pink noise response (from Graph 1). [GoldenEar Aon 3]

40K

between the two resonant peaks (at 32Hz and 87Hz) at just under 50Hz, making GoldenEar’s low-frequency response claim of 38Hz for the Aon 3 look rather optimistic. GoldenEar specs the sensitivity of the Aon 3 design at 90dBSPL, but doesn’t mention under what test conditions it obtained this figure. Under its usual test conditions (which make it a lot more difficult for loudspeakers to achieve ‘higher’ figures than manufacturers like), Newport Test Labs measured the sensitivity of the Aon 3s as 87dBSPL at one metre, for a 2.83Veq. input. This is 3dB lower than GoldenEar’s specification, but it’s still a good result for a small two-way loudspeaker design, and means you won’t need a lot of amplifier power to get the action happening. The final graph we’ve published is a composite graph that marries all the different methods used to test the Aon 3 on one graph to give an overall picture of how things ‘fit’ together. What it does show is that overall, Newport Test Labs measured the frequency response of the GoldenEar Aon 3 as being 60Hz to 40kHz ±3dB, which is an excellent result. Steve Holding