Amplified & Equalized

Written by Haden Boardman

Having generated a few milli- or micro-volts by our humble little cartridge, we now have to equalize by applying the inverse of the RIAA curve that was applied when the record was mastered, and then amplify the signal enough to drive our main amplifier.

Electronically (or even just electrically) , a phono stage is  very difficult  to engineer properly, which helps explain a veritable cornucopia of designs out there. Tubes or transistors,  op amps, head amps, SUTs, passive or active… it is a mind-boggling array!

Take a modest MM / moving magnet cartridge, with a quoted output of 5mv: our main amplifier really would like to see a minimum of 500mv, so our phono stage has to add gain at 100 times, at the same time adding the necessary RIAA curve. To further complicate matters, that curve is so severe that  it can wipe out the gain of an entire stage of the design.

Most designs, tube or transistor, use active circuits which use negative feedback.  I don’t want to get too technical, but some basics can be covered.  The cartridge signal is amplified by a circuit having very high gain, this may be two or three tubes, several transistors, or one or two Integrated Circuits (which is must be remembered can contain 40+ transistors in each).

Part of the high output of the preamplifier circuit is returned (fed back)  to the input of the preamplifier, and at the same time a small network of capacitors and resistors adjusts the level of feedback selectively at the bass, mid and treble frequencies to apply the RIAA equalisation curve. Less feedback in the bass, very high feedback in the treble. Feedback can also help reduce levels of distortion, which is no bad thing. But as the feedback is different at the bass and treble frequencies (more gain need for the bass frequencies) the amount of distortion also changes with frequency, which is not a good thing.

Generally speaking, transistors (including ICs, or integrated circuits) are not as linear as tubes, nor do they have as much gain per device. That’s why transistors are usually used with feedback to help counter these issues. There is also an argument that using any feedback can ‘flatten’ the sound, or make it less real. In High End tube circles, especially those using big fat DHT (directly heated triode) amps, feedback is generally sneered at!

One of the first RIAA circuits every printed, one by RCA, is a classic example of a no-feedback circuit, and one that has been ‘reinvented, time and time again, although it always really stays the same.  It uses a single tube per channel, but a tube that is actually a kind of Siamese twin, with a shared glass bulb, but with two little individual halves inside.  One half of the tube amplifies, and then the signal is fed through a small bunch of capacitors and resistors, which selectively equalise and apply the RIAA curve. This is then amplified again by the final half of the tube. With no feedback applied, the distortion is around 5% per stage. In addition, maximum gain cannot be made from the tubes used, and the first half will overload more easily in the treble, because the RIAA equalisation is providing it with a very low load.

Adding more tubes, and using cathode followers or SRPP (series regulated push-pull) principles, can help with this last problem, but not  with the distortion. SRPP uses both ‘halves’ of the tube as one, thus twice as many tubes are needed. The ultimate configuration of this kind of circuit uses an LCR filter; instead of the small resistors and capacitors, a series of inductors, capacitors, and resistors are matched and sealed in what looks like a transformer. The use of inductors can be seen as a benefit over just using resistors, but the costs are enormous. The simple capacitors and resistors of the basic circuit could cost as little as a few cents, but LCR RIAA filters cost a good few hundred bucks. Kouichi “Nobu” Shishido published such a circuit in the November, 1996 issue of the Japanese magazine MJ Audio  Technology , which was reprinted in Sound Practices magazine and became a defacto circuit.

There are of course lots of hybrid designs, mixing  feedback stages  with all or part of the RIAA done passively; confusingly, there are also ‘hybrids’ mixing tubes with solid state elements in active gain stages.

Personally, I use a tube unit which utilizes feedback. It uses the same Siamese tube RCA used, but set for higher gain, with feedback in which I invoke the RIAA curve. I am not alone, as such a configuration has been used by Stewart Hegeman, Marantz, Hafler, Dynaco, Fisher, Scott, Audio Research (early!) Conrad Johnson, Luxman, Croft Acoustics, and many others. Carefully built, it’s an excellent basis for design.

People do get bogged down in specs. Being anal-retentive about the RIAA curve is pretty irrelevant, given the variations in EQ used in mastering actual records. The more electronics you feed something through, the worse it gets, generally speaking. Amplifiers DO NOT make things better, they add gain, but always subtract ‘something’; thus, the more amplifiers, the more is lost. Keeping a circuit simple can aid fidelity, but make it too crude, and it can be counterproductive, as it will tend to create more distortion.

All this is fine for MM cartridges, but what about moving coil? These have outputs one tenth to one fiftieth  that of a moving magnet. To increase gain, it is either necessary to add an extra active stage, use a transformer, or in the case of a real budget phono stage cheat, reduce the negative feedback to create more gain (as well as more noise and distortion!).

Phono stages already create more electrical noise than most other parts of your hi-fi system, but to boost the MC signal it becomes even harder. Using a tube as a kind of pre-pre-amp has been done many times, but they are nearly always noisy.  Using a select mix of FET and transistors is probably the best active compromise, and the use of a poor quality IC is best avoided.

Step up transformers can be one of the best ways of achieving gain; even these introduce distortion, and good ones are scary expensive.  To maintain bandwidth complex winding techniques are used, with high grades of wire. To stop the transformer picking up other unwanted signals (think 0.001v signal to 120.000V mains supply) mu-metal and extensive magnetic shielding is needed.

Personally, I prefer transformers over other step-up devices in a high-end set up. But down on a budget, the vast majority of phono stages are purely active. To be honest, this is probably a lot better than using poor-quality transformers.

There are so many phono stages on the market now. There are hundreds of the little (and not so little) blighters. If on a tight budget it’s impossible to avoid the IC-based ones. But look at the wall-wart power supply, make sure it’s at least 18V or more, and it will at least not overload. And as far as I’m concerned , try to upgrade to tube ASAP!

Back to Copper home page

1 of 2