Hazelrigg Industries VNE

Single-channel PWM Compressor
By Hugh Robjohns

With its unusual method of applying gain reduction, the Hazelrigg VNE is no ordinary compressor.

Hazelrigg Industries may not yet be a familiar name to many readers, but DW Fearn certainly should be and the two brands are inextricably linked — Hazelrigg Industries provide all the manufacturing and distribution for both DW Fearn and Hazelrigg products. Moreover, while Doug Fearn is the President of DW Fearn and the lead designer of that company’s products, he also has a hand in Hazelrigg Industries’ own products, working collaboratively with George and Geoff Hazelrigg to build a range of hardware using his well‑proven circuit designs. For example, Hazelrigg Industries’ VLC (a review of which is in the pipeline) is a microphone preamp which combines elements of DW Fearn’s VT‑1 mic preamp, a DI input based on the VT‑IF, and an inductor EQ derived from the VT‑4/VT‑5 Equalizers.

The subject of this review, though, is Hazelrigg Industries’ newest product, the VNE. It’s a single‑channel compressor that is hand‑built in the USA and claimed to be tonally transparent while applying its dynamic control. This is significant, since many analogue compressors change the tonal balance to some extent, depending on the technology being used and the amount of gain reduction being applied. One reason for the VNE’s neutrality in this respect lies in an unusual approach to achieving gain reduction, which is borrowed from DW Fearn’s VT7 stereo compressor: a technique called pulse‑width modulation, or PWM (see the box for more on this). The unit comes with a most welcome seven‑year warranty (valves excepted), which says a lot about the company’s confidence in their construction and components.

Like the VT7, Hazelrigg’s VNE employs hybrid technology. The input and output circuitry is very traditional, with transformers and valve gain stages, and this bookends a solid‑state gain‑reduction element that’s based on the VT7’s pulse‑width modulation technology. Compressors using the PWM technique are relatively rare — simple FET, optical, and VCA compressors are far more common — but despite the considerable complexity involved in a PWM design it brings significant benefits, including extremely low distortion, the option of particularly fast attack and release times, and reliable tonal accuracy that is unaffected by the amount of gain reduction being applied.

Overview

The VNE has a standard 2U rackmounting chassis, weighs around 5.6kg (12.5lbs) and extends around 255mm (10 inches) behind the rack ears. It has a vintage‑looking, brown‑painted front panel and the styling is austere, with sparse and widely spread controls — an intentionally understated aesthetic that doesn’t proclaim that this device is anywhere near as expensive as it actually is! As someone of the ‘elderly persuasion’, I can certainly appreciate the benefits of not having to squeeze podgy fingers around densely packed controls or squint in low lighting to read the labelling. Though some others might find it a little too minimalist, beauty is in the eye of the beholder and what really matters is what it sounds like!

None of the controls have numerical labelling; Hazelrigg say they want their users to make parameter decisions by listening and not looking. While that is highly commendable it’s also quite nice to know what settings are employed when wanting to compare the sound with other units, or to log settings for future recall. In this case, clock‑face descriptions are the only option.

With only five controls, operating the VNE is very straightforward. The first rotary control, Threshold, determines the amount of compression and actually operates as an input gain control; the compression threshold is fixed internally at around ‑35dBu. Advancing the Threshold control increases the amount of signal entering the compressor above this threshold level and thus introduces more ‘squash’. Unusually, there’s no ratio control — the compressor has a very soft‑knee action, with the ratio increasing gently and progressively as the amount of gain reduction rises. At the highest Threshold setting, an increase of input level from 0 to +15 dBu produced an output level rise of 5dB, indicating an effective ratio of 3:1.

The amount of gain reduction is indicated on a slightly incongruous LED bar‑graph meter. It works perfectly well but in an expensive vintage‑styled unit like this I expected an old‑school moving‑coil meter. Again, there are no markings on the GR meter but the first seven LEDs denote roughly 1dB increments, while the last three have larger steps closer to 2.5dB, to give a display range of 15dB. Interestingly, the DW Fearn VT7 from which the VNE is derived was originally set up to provide a maximum of 30dB of gain reduction, but this was reduced to 15dB in later units to provide greater resolution in the Threshold control when dialling in moderate amounts of compression. It appears that the VNE has been designed with the same intention.

A rotary three‑way switch selects a full bypass (with the input connected directly to the output via sealed relays), normal operation, or operation with a high‑pass filter in the side‑chain. This filter rolls‑off at 6dB/octave below 150Hz to reduce the compressor’s sensitivity to LF energy.

Attack and Release time‑constants are adjusted with another pair of rotary knobs and range from ‘fast’ to ‘slow’, with no specific response times disclosed either on the panel or in the manual. In practice, I found the fastest end of the Attack control to be very quick and it can certainly react to the transients of percussive instruments, but it’s not quite as fast as I’d anticipated — I didn’t run into any transient distortion or modulation problems with things like flutes or bass guitars. The slow end of the Release control seems to take five seconds or more to recover fully, with a dual‑stage release characteristic; a full gain reduction is initially released quickly, but then much slower as it nears the last few decibels. This is similar to the ‘auto‑recovery’ or programme‑related modes found on other manufacturers’ compressors.

The rear panel’s balanced I/O XLRs are joined by a 120/230 V mains inlet and a chassis earthing stud.
The final rotary control provides up to 15dB of make‑up gain, while a single bi‑colour LED mounted above the GR meter provides a rough indication of the signal level at the output XLR. This illuminates green for signals above ‑4dBu and turns red at +11dBu, and the manual recommends adjusting I/O levels so that the unit works mainly in the green zone. I found this quite easy to achieve and despite the simplicity of the single LED it worked well in practice.

A chunky toggle switch, which has an associated power LED, turns the unit on and off, and mains power is connected via the usual rear‑panel IEC inlet. A recessed switch caters for nominally 120 or 230 Volt AC mains supplies, and the unit consumes 25 Watts. The transformer‑balanced audio input and output both use XLR connectors, while a chassis earthing stud is also provided on the rear panel. I was pleased to note that the XLR grounding arrangements are exactly as they should be, with both pin‑1s being connected directly and solidly to the chassis — and nowhere else.

VNE In Use

Using the VNE couldn’t be simpler. Turn up the Threshold control to introduce the required amount of ‘squash’, adjust the Attack and Release controls to achieve the required dynamic reactions, and turn up the Gain control to restore the peak level. Job done!

I used the VNE on a variety of vocals and instruments, some live and others from raw multitrack recordings. Just introducing the VNE into the signal path with the Threshold at minimum adds subtle but recognisable colour, and that character builds progressively with higher signal levels.

As a compressor, the VNE performed very nicely on everything, but I found it particularly well suited to the subtle reining‑in of dynamics on both male and female vocals, as well as guitars and basses, and even percussive acoustic and electric piano parts. With more dynamic sources I often found it necessary to run large amounts of gain reduction, largely because the ratio is relatively gentle over most of the range. But its soft‑knee characteristic avoids it ever sounding aggressive or even obvious, and the coloration from the valves and transformers was always musically complementary and enhancing, even when the levels crept up. True to Hazelrigg’s claims, for any particular working level the tonal character really doesn’t change as compression is applied either — it doesn’t get duller, for example, as many compressors can. The VNE really is quite transparent in that regard.

As I only had a single mono unit I couldn’t apply the VNE to the stereo mix bus, but my impression is that a pair would make a really effective ‘glue’ compressor — although maybe DW Fearn’s VT7 stereo unit would be a more practical (and slightly more cost‑effective) option in that role; the VNE has no facility for linking side‑chains with a second unit.

I found it particularly well suited to the subtle reigning‑in of dynamics on both male and female vocals, as well as guitars and basses.

Verdict

Overall, I really like the VNE. It’s simple to set up but performs extremely well on a wide variety of material delivering genuinely transparent compression while adding some beguiling thermionic magic along the way. Of course, there’s no getting away from the fact that this is a fearsomely expensive single‑channel unit, and while it’s easy to hear what that money buys you, it’s not so obvious to the eye and that may matter to some potential purchasers. Personally, I love its understated looks, it’s simplicity of configuration and, most of all, its delicious sound. More from Hazelrigg please!  

PWM Gain Reduction

These three AP analyser plots shows the gain reduction applied at the different ratio settings, the frequency response and the noise level and spectrum.

The application of pulse‑width modulation as a means of gain reduction in a compressor may not be immediately obvious. Most conventional compressors control the audio signal level via a remote‑controlled attenuator (usually some form of variable resistance), generally configured as a potential divider. For example, an optical compressor uses a light‑dependent resistor to ‘shunt’ part of the audio signal to ground, thereby reducing the signal energy passed to the output. A FET compressor typically uses a special transistor to attenuate the audio signal according to a control voltage applied to its gate terminal, to alter the effective resistance between its source and drain terminals.

However, often the relationship between the control voltage and varying resistance for these kinds of simple approaches isn’t entirely linear, and that results in distortion. More sophisticated attenuation solutions, like Vari‑mu and VCA designs, have been developed to minimise signal distortion but an even more radical approach is to use something like a FET as a very fast on/off switch.

If this switch were placed in the direct signal path there would be no signal attenuation when ‘closed’ but full attenuation when ‘open’. Obviously a compressor needs a bit more subtlety than that, and this is achieved by flipping the switch on and off continually at a very high rate — typically, a few hundred kilohertz — under the control of a pulse signal. (Any switching transients are easy to filter off as they are so far above the highest needed audio frequency.) Since this switching is happening at such a very high frequency, well above our hearing range, we don’t perceive the signal as coming and going at all, but the proportion of time that the switch spends closed or open affects the amount of energy being passed through the switch. That translates directly into a variation of the perceived signal level and so, by altering (or modulating) the proportion of time the controlling pulse signal is in its high or low states (in other words the pulse width), the amount of attenuation can be controlled very precisely.

This technique is known as pulse‑width modulation (PWM) and, despite its aggressive nature, it inherently introduces very little signal distortion and doesn’t change the signal’s frequency response as the attenuation level increases. Since the ‘chopping’ rate is so high it’s also possible to change the amount of attenuation extremely quickly, so the approach is capable of phenomenally fast attack and release times, if required.

Technicalities

Lifting the VNE's lid reveals neat, hand‑wired construction.

Internally, the unit is constructed to a very high standard, with the linear power supply section towards the rear fed from a large toroidal transformer enclosed in a metal sleeve to minimise radiated hum. A metal dividing screen further isolates the audio transformers and electronics, which are mounted on three large vertical circuit boards behind the front panel. The first card carries most of the valve amplifier components, with four 6072A dual‑triode valves protruding horizontally into the rear section of the chassis under metal screening cans. The second card hosts the solid‑state PWM circuitry, while the last serves the gain‑reduction meter. Neat hand‑wiring is evident between the boards, transformers, and I/O connectors.

A compact Jensen input transformer presents a 32kΩ input impedance, expecting a nominal +4dBu input level. The manual claims a maximum input level of +25dBu (specified, unusually, at 20Hz), although my bench measurements using an Audio Precision test set revealed 1 percent THD at +21dBu (measured at 1kHz). This increased to a whopping 10 percent THD at +25dBu, with distortion rising steadily above about +11dBu (where the output level LED turns red). At the nominal operating level of +4dBu, THD measured 0.04 percent, which agrees with the published specs. This progressive increase in THD is nicely aligned to typical signal levels in an analogue environment, so pushing hot signals intentionally becomes increasingly gritty.

The measured signal‑to‑noise ratio of the review unit is 78.3dB (ref +4dBu), matching the specifications, while the highest mains hum component was measured at ‑82dBu (100Hz). The frequency response remained within ±0.5dB between 20Hz and 30kHz, rising to modest transformer‑related peaks at both frequency extremes (reaching around +2dB at 8Hz and +3dB at 70kHz).

Hazelrigg’s valve amplification configuration is more elaborate than I expected, with two valves serving the input side and another two on the output side. Both valve sections are configured in a broadly similar way with a pair of Class‑A gain‑stages followed by an output driver. The input section feeds the solid‑state PWM circuitry, while the output section drives a custom output transformer. This is claimed to be able to deliver up to +22dBu into a normal bridging load impedance of 20kΩ or greater.

Alternatives

Most compressors in this per‑channel price bracket are dual‑channel or stereo designs and include the Cranesong STC‑8 (which also uses PWM technology), the Gyraf Audio Gyratec G22 and SPL Iron Mastering Compressor (both of which are valve designs) and the Vertigo Sound VSC‑3, a discrete VCA design, amongst others.

Pros

  • Transparent PWM compression technology.
  • Superbly musical valve/transformer signal path.
  • Simple but effective user controls.
  • Understated styling.
  • Built to exacting engineering standards.
  • Seven‑year warranty.

Cons

  • No side‑chain linking option.
  • Only offers gentle ratios, with a maximum of around 3:1.

Summary

The VNE borrows from DW Fearn’s VT7 to deliver a single channel of transparent PWM compression with a wonderfully musical‑sounding valve/transformer signal path.

Information

£4395 including VAT.

ASAP Europe +44 (0)20 8672 6618

sales@asapeurope.com

www.asapeurope.com

dwfearn.com/wp/hazelrigg-industries

$4500.

Hazelrigg Industries +1 567 393 3276

support@hazelriggindustries.com

dwfearn.com/wp/hazelrigg-industries

Published March 2022

From the same manufacturer