Designed to get the best from passive ribbon and moving-coil mics, this preamp offers 80dB of gain and switchable input impedance.
The humble ribbon mic has recently enjoyed a remarkable and well-deserved resurgence. A key reason for its lack of mainstream popularity for so long was its low output level, and consequent requirement for high-gain and ultra-low-noise preamplification (an expensive luxury throughout the last century). Modern magnetic materials enable far stronger output signals from contemporary passive ribbon mics, and many modern designs also incorporate active impedance converters, to match the sensitivity and convenience of typical capacitor mics. But vintage ribbon mics still require high-gain, low-noise preamps, and a handful of manufacturers produce devices optimised for this niche market. Such products offer higher-than-usual gain (80dB or more), high input impedances (often 18kΩ or greater) and no phantom power (removing the risk of damage to the ribbon element). One such manufacturer is Integer Audio. Based in Rochester, New York, they were founded in 2015 by David Cicero, an electronics engineer and quality audio aficionado with over 30 years’ experience, and they currently offer a single product, the RMP2.
Essentially, the RMP2 is a simple, two-channel mic preamp, but it’s clear that it’s been developed with sonic accuracy and engineering excellence in mind. It’s equally clear that it is built by hand to a very high standard. It’s housed in a smart 1U rackmount chassis that extends around 20cm behind the rack ears, and weighs 2.25kg (about 5lbs). The front-panel controls are limited to a rotary gain switch and a three-position input-impedance toggle switch for each channel (a fixed 10kΩ impedance version will also be available by the time you read this), and a mains on/off rocker switch with blue power-on LED. That’s it. There’s no metering, no high-pass filter and no polarity reversal.
There’s no phantom power either, of course, as the RMP2 was designed for use with passive ribbon mics, but it’s worth noting that it can be used with any moving-coil (dynamic) mic, any valve mic with its own mains supply, or any battery-powered electret mic. In short, any mic which doesn’t require phantom power.
The rear panel is just as simple. Neutrik XLRs cater for the inputs and outputs, and an IEC inlet with integral fuse holder accepts AC mains. The review unit accepted only 120V mains supplies, but when I enquired about a model suitable for 230/240 V AC regions, Cicero told me he’s happy to construct 230V AC models — each unit is built to order anyway.
A double-sided printed circuit board extends over most of the chassis floor area, and incorporates a large ground plane. The electronics are constructed using high-quality through-hole components, and the regulated linear power supply is a fully discrete design, generating individually fused ±18V rails, with power-MOSFETs employed as series regulators, rather than the far more commonly used (and potentially noisy) three-terminal LM-series regulators.
Unusually, the audio signal path is completely balanced throughout, and there are no transformers or capacitors in the signal path. The DC-coupled front end is built around a trio of THAT320 matched-transistor arrays, providing a very low-noise input configuration, constructed using five paralleled PNP transistors for each leg of the balanced input. Two dual op-amps are also employed in each channel’s circuitry: a Texas LM4562 serves as a unity gain buffer and output driver, while a Burr Brown OPA2134 is employed as a dual-servo to control DC offsets and common-mode errors. The gain is adjusted with a 24-position Elma A4 rotary switch, carrying a neat resistor chain to vary the amount of negative feedback around the gain stage. This ensures precise gain-matching between channels, and accurate gain increments.
I was pleased to see Bourns EMI T-filtering components mounted directly on all four of the input and output XLRs, and the IEC mains inlet is also an EMI-filtered type. Also, pin 1 of each XLR is connected directly to the chassis and nowhere else, while the ground references from each preamp and the PSU are all brought together at a bolt on the rear of the chassis, using thick wires to form an ideal star-grounding point. The safety earth from the IEC inlet is also connected to this — all is exactly as it should be! The RMP2 passed the ‘Windt Hummer’ test for pin 1 grounding problems with flying colours.
Operating the RMP2 is very straightforward. The rotary gain switch spans a range of +24 to +80 dB, with most steps being 2dB increments — although the first and last are 6dB, and the penultimate one 4dB. The input-impedance switch positions are marked as low, high, and medium because they can be customised to order, but the standard values are 270Ω, 30kΩ and 2.7kΩ.
The 2.7kΩ setting is typical of most conventional preamps and will work well with pretty much every mic. The idea of a high-impedance option is to minimise the electrical loading on the capsule, and thus maximise the transient performance and output level. However, 30kΩ is much higher than offered on most other ribbon preamps, and a potential downside of such a high input-impedance is that, for some mics, it may result in ’under damping’ the mic’s output transformer. This could potentially cause an exaggerated, or even harsh, high-frequency response due to uncontrolled ultrasonic transformer resonances. I’d prefer to see a lower ‘high’ option — somewhere around 10kΩ, as on the fixed-impedance version.
At the other end of the scale, the low-impedance setting is really provided for use with vintage mics specifically designed for operation within the (long obsolete for microphones) matched-impedance environment. In such circumstances the output level will be around 6dB lower than when used with a medium or high input impedance, and there’s also a theoretical 6dB noise penalty. With most modern passive ribbon mic designs this low-impedance option may also ‘over damp’ the transformer, typically resulting in a thinner, bass-light sound, possibly with some transient distortion, too.
If the side-effects from different impedance options sound worrying, rest assured that no damage will occur to the mic whatever the impedance setting. The options simply change the performance and tonal characteristics, allowing you to experiment and find a setting that provides the ‘best’ sound for a particular mic and recording situation. The only practical precaution I’d recommend is to dim or mute the monitoring when switching impedances, to avoid audible ‘thumps’. In general, I’d recommend using the high-impedance option unless the extreme high end seems too prominent, in which case the medium option will probably work better. Having said that, at very high gain settings I found that the medium-impedance option gave the lowest background mains hum levels.
The published technical specifications for the RMP2 are generally very impressive. The stated maximum output level is “more than +22dBu” into impedances over 5kΩ (from an output impedance of 50Ω). Distortion (THD+N) is given as less than 0.002 percent with 60dB of gain, and the equivalent input noise (EIN) is below -129dBu (with a 150Ω source, 20kHz bandwidth, and 66dB of gain). All are very good figures.
My test-bench measurements, using an Audio Precision test set, showed that the maximum output level is actually +28dBu, which is even more impressive. It means that the recorder, A-D converter, or DAW interface will always clip before the preamp, so it’s really not necessary to have meters on the preamp — the destination device, which will always have meters, will always be the critical element when it comes to setting levels.
The THD+N figure I obtained (with 60dB gain and a +4dBu output level) was 0.07 percent, and only by boosting the input level to generate an output of +24dBu could I improve the THD+N figure to 0.008 percent. Although these figures are not quite as good as the published specs, they still represent an impressive performance, and especially so for 80dB of gain (with a +24dBu output level), which achieved an excellent THD+N figure of 0.08 percent.
One of the more unusual specifications to publish is the variation in bandwidth as the gain is increased. This phenomenon is to be expected: all amplifiers are defined by a value known as the ‘gain-bandwidth product’, so the more gain delivered by the amp the narrower the bandwidth over which it can operate effectively, and vice-versa. In the RMP2’s case, the published specs claim a bandwidth extending beyond 1MHz at low (<40dB) gain settings, reducing to around 300kHz at 60dB of gain. This is equal to or better than much of the competition over that ‘traditional’ gain range.
Of course, the RMP2 is unusual in delivering 20dB more gain than most, yet the manual only quotes the -3dB bandwidth limit at 74dB of gain (70kHz), and not for the maximum of 80dB. In testing this myself I found the -3dB point falls to a still-very-acceptable 40kHz at 80dB of gain. For all gain settings below 68dB, the response was within 0.1dB between 10Hz and 20kHz.
A Fast Fourier Transform plot revealed some mains frequency harmonic residuals in the output signal. With 60dB of gain (and using the medium-impedance setting), the 50Hz fundamental and 150Hz third harmonic components were below -80dBu, and even at the maximum 80dB of gain they remained below -60dBu, all of which is pretty good. However, switching to the high-impedance mode caused the mains hum levels to rise by a further 20dB, rendering them audible in some circumstances. I presume this is as a result of increased sensitivity to radiated fields within the unit itself when configured in the high-impedance mode.
For testing purposes, I used the RMP2 mostly with a pair of AEA R92 passive ribbons, which are designed for close-up applications with loud sources. I also used a pair of vintage AKG D224E ‘two-way’ moving-coil mics, which are comparable to a small-diaphragm capacitor mic in terms of sound quality. I also borrowed a couple of classic (and very low output) Coles 4038 ‘BBC’ ribbons for the ultimate testing, and I compared the RMP2 directly against my own AEA TRP dual-channel ribbon preamp.
Overall I found the RMP2 to be a joy to use. It looks very elegant, and the controls all feel solid, professional and reliable. The internal construction is to very high standards and well engineered, so this product should give decades of faithful performance.
The RMP2 was designed following the old adage of being ‘a wire with gain’ and I could detect no recognisable sound character — it simply makes the mic signal bigger, without adding anything or taking anything away. The noise floor is both very low and smooth, even at high gains, and the unit sounds very open and clean, with fast transients and masses of headroom. It really is totally transparent and effortless-sounding, allowing the mic to shine through without imposing any restrictions or character of its own. The RMP2 is a genuinely top-quality mic preamp.
I found only one minor disappointment: a (barely audible) rise in the level of background mains hum, mostly in channel 2, when operating at extremely high gain levels — but only when in the high-impedance mode. I discussed this with the designer and he has since added some internal shielding around the internal mains transformer to minimise the interfering magnetic field, which appears to have cured the problem completely.
If I were to be really picky, I’d also comment on the lack of a high-pass filter. Ribbon mics are typically tuned to resonate somewhere around 10Hz, and they are consequently extremely sensitive to sub-sonic vibrations. Although the RMP2 has a lot of headroom I would prefer to remove the worst unwanted sub-sonic contributions as early in the signal path as possible, to avoid having to leave unnecessary headroom at the conversion stage. Perhaps a switchable second-order high-pass filter turning over at, say, 40Hz — or something like that — would be a useful addition for many users.
Nevertheless, David Cicero has made a fine job of the RMP2, which is a very attractive and welcome addition to the market place, and I look forward to Integer Audio’s future products with eagerness.
The most obvious direct competition is in the form of AEA’s two ribbon preamps, the compact TRP (which costs a little less) and the more elaborate RPQ (which costs a little more) — as well as its 500-series rack equivalent). However, several other mic preamps have ‘ribbon modes’ that increase the input impedance and disable phantom power, and several more offer similar levels of extreme gain. Possibilities include True Systems’ P-Solo ribbon mic preamp, several AMS-Neve 1073 variants, Focusrite’s ISA preamps, and several Grace Design preamps.