Phono stage tests

Article Index
Phono stage tests
All Pages




This is a measure of RIAA equalisation accuracy, shown in a graph here. LPs are cut with reduced bass and strengthened treble, to prevent groove excursions becoming too great, and suppress the hiss and crackle of vinyl surface noise. When played back bass must be boosted, and treble cut by exactly the same amount they were changed in the cutting process, as set by the RIAA curve. Phono stage frequency response is a measure of this, telling us how accurate equalisation is and what the resulting tonal balance of the phono stage will be. Rising treble means a bright sound balance and this common; it boosts apparent detail. Modern phono stages rarely, if ever, have falling treble and a warm sound.


The RIAA characteristic of a Rega Elicit amplifier, showing warp filter roll off at left.

Below 20Hz, in the sub-bass region, if the RIAA and derivative standards are followed gain remains constant down to 0Hz or d.c. Many phono stages simply follow this as it conveniently requires no additional filtering. However, this strongly amplifies disc warps, causing cone flap in reflex loudspeakers, something best prevented. The IEC issued an amended curve with additional 7950uS time constant but when used in replay equalisation this audibly lightens bass and is rarely used in phono stages. Many manufacturers use a compromise, and we measure gain at 5Hz to check for warp suppression, or lack of it.


We inject white noise in the phono input, through an accurate inverse RIAA network. Output is then displayed on our Rohde & Schwarz UPL spectrum analyser, as shown here, providing a picture that is published in our Measured Performance section of the magazine review.


The Rega Elicit, measured using a pre-equalisation network to show frequency response.

In some cases where low frequency signal-to-noise is an issue, we inject noise unequalised and use post-equalisation, applied within the UPL spectrum analyser. This gives a smoother looking curve.

The quoted frequency response is usually wider than the 20Hz-20kHz spectrum shown and is measured using a manually swept sine wave generator.




Cartridges produce a weak signal that needs amplification before an amplifier can deal with it.  Insufficient gain from a phono stage means volume must be turned right up on the amplifier to compensate and this might not be enough for full volume.

The benchmark amplification figure for Moving Magnet (MM) cartridges is x100, or 40dB, and for Moving Coil (MC) ten times as much, x1000, or 60dB in external phono stages.


In practice these are best seen as minimum values, especially if the phono stage feeds an amplifier whose line inputs have a low 400mV sensitivity, a common figure. An MM cartridge produces just a few millivolts at ordinary music levels (i.e. not peaks) so a phono stage with x100 gain will barely be producing 200mV and volume will have to be turned right up in this situation. Another nuisance here is the big change in volume when switching from LP to CD. Some amplifiers (e.g. Arcams) have adjustable input sensitivity so volume levels can be equalised to avoid this.

Ideally, around x200 gain (46dB) is need for MM cartridges and x1000 -  x3000 (60dB-70dB) for MC cartridges.

Phono stages with an output volume control (e.g. Icon Audio) can be used to drive a power amplifier direct. In this situation roughly x5 (14dB) more gain is needed, as most solid-state power amplifiers need 1V input to deliver full output, although many valve power amplifiers need considerably less, down to 300mV.


We measure gain at 1kHz. Input voltage is measured direct (i.e. p.d and not generator emf), to get a true input value. Output voltage is measured by a Rohde & Schwarz UPL spectrum analyser. Input levels of around 10mV are used for MM and 1mV for MC.



It is remotely possible, in a few circumstances, for a cartridge to overload a phono stage, causing distortion. We test for this possibility.

Generally, modern transistors and integrated circuits have enough headroom to easily avoid being overloaded by any cartridge, but there are some limitations.

Most external phono stages use integrated circuits, fixing maximum output swing to 10V, usually because they are limited by 15V supply rails. In some designs this sinks to 6V, possibly because the designer thinks this adequate (0.5V is enough to drive most amplifiers) and it makes for a cheaper power supply. When gain for MM can be freely set, however, a gain of x200 – quite common – would put such a stage into output overload with a modern Ortofon 2M Red or Blue cartridge, able to deliver 35mV peaks.

This tells us that with MM cartridges a phono stage should ideally be able to accept 40mV input before overload. Output swing usually sets the overload ceiling, and the input limit is the output divided by gain. So if the output limit is 10V and the gain x100, the input limit is 0.1V, or 100mV. This is more than enough for modern cartridges. However, if the output limit sinks to 6V and the gain used is x200, then the input limit becomes 0.03V or 30mV, in which case overload becomes a possibility. So overload isn’t likely, but it is possible.

None of this applies to valve phono stages. Their output swing is up to 45V, and valve phono stages rarely, if ever, go into overload.

Overload conditions are rare with Moving Coil cartridges, although not impossible. Maximum output from an Ortofon Cadenza Red, which produces a healthy 0.7mV at 5cms/sec rms, is around 3.5mV.  Feeding a phono stage with a low x1000 (60dB) gain gives 3.5V output maximum. It would be possible, using higher gain to exceed 6V, if rare.


We input a 1kHz test tone and increase level until output overload, seen as peak clipping, occurs, as seen on an oscilloscope, connected to the monitor output of our Rohde & Schwarz UPL spectrum analyser, The input applied is measured direct across the input terminals (so is not a generator emf) and the output monitored for distortion on the UPL so where soft clip occurs, the limit can be set by using 3% distortion criterion. This usually only applies to valve phono stages.



The noise spectrum of a phono stage follows the RIAA gain curve.


This is a measure of hiss and, with valve phono stages, of hum too. It tells us whether an annoying steady hiss or hum will spoil enjoyment of the music. With ordinary Moving Magnet cartridges, both transistors and valves are quiet enough for hiss to be inaudible at a typical listening position, say 3m from the loudspeakers, with volume up. A large number of commercial Phono stages use ultra-quiet ‘audio preamp’ silicon chips, purpose designed for such work.

This is not necessarily the case with Moving Coil cartridges, some of which have very low output (e.g. Goldring Legacy). Then hiss will be audible with all except the quietest Phono stages. Although many MC preamps also use silicon chips, which have adequately low noise, input transformers better exploit the low impedance of MC cartridges and give the lowest noise figures (e.g. Luxman E200).


We measure output noise, IEC A weighted, in mV or µV, across the audio band (A weighting defines a band around 1kHz). This figure is then divided by the gain to give an input noise value. Values of 0.3uV - 0.8uV for MM inputs are typical with this measurement method, when the input is shorted to ground.

It’s little known that MM cartridges produce quite a lot of noise, around 3uV IEC A weighted our measurements show, enough to swamp the input noise of phono stages, so cartridge hiss dominates in practice.

This is not the case with MC cartridges. They are virtually silent so phono stage input noise dominates.



Hi-Fi World, Powered by Joomla!; Hosted by Joomla Wired.