Moving Magnet Cartridges Revisited
Like most hi-if enthusiasts who grew up in the age of the LP, my first pickup cartridges were moving-magnet (MM) types. But I have been a convert to the more expensive moving-coil (MC) models since before records went out of fashion the first time; over 30 years ago. My conversion to using this more fiddly technology was due to my appreciation of three principal advantages of MC cartridges: lower noise; a sweeter top-end; and a better stereo image.
Although the resistance of an MM cartridge would lead you to believe that it is a low-noise source, in practice, the source impedance is inductive which means signal to noise degrades with frequency. On the other hand, provided a good head amplifier or transformer is used, the MC type cartridge affords an "inky black" noise-floor which MMs are never able to match.
Compared with the moving-coil type, MM cartridges always produce a slightly uncontrolled top-end which tends either to be a bit muted and "soft" or is "splashy" and exagerrated.
Better stereo image
The third benefit of a moving-coil cartridge is that they nearly always produce a better, more three-dimensional stereo image. Quite what causes this difference is difficult to say. MC cartridges do tend to produce better channel separation than MM types and perhaps this is the beginning and end of the story. However, from an objective standpoint, measurements confirm that MM cartridges they have adequate separation to give acceptable stereo image. Yet the stereo image always seems papery and flat from a MM type, compared with that obtainable from a good MC.
Frankly, I was quite happy in my settled beliefs concerning phono cartridges and have only recently challenged them whilst working on the software product Stereo Sauce with my colleague Alastair Macmaster. As part of this project, we developed phono equalisation in software; the idea being to capture needle-drops unequalised "directly from the pickup" and equalise them in a subsequent software process.
Even our initial results convinced me of the superiority of this approach. But equalisation has turned out only to be the beginning. Subsequent work has revealed that robust models may be employed to reduce the worst "clicks and pops" without damaging the audio (certainly the biggest drawback of analogue LPs). Moreover, much of the harmonic distortion generated in the replay of LPs may be calculated and subtracted from the audio - greatly enhancing the clarity of the medium; especially towards the centre of the record. The distortion mechanisms of LP replay are discussed here.
But one fault in the reproduction of records which cannot be corrected in software is poor tracking, this being the phenomenon in which the stylus parts company with the groove wall due to it being too inert to accelerate to the velocities demanded by a heavily modulated groove. The brief pops and clicks which plague records may be concealed in software because these disturbances last only a few samples at a time, but poor tracking can cause the stylus to leave contact for a period much too long to correct. For example, in a typical failed tracking test, the stylus leaves the groove for about a third of a cycle on a high-velocity 300Hz signal. That's a period of 1mS. At 96kHz sampling, 1mS is represented by 100 samples. Under normal circumstances of audio replay, it's anybody's guess what those 100 samples ought to have been. In modern data-processing terms we can think of poor tracking as "read-errors" which are impossible to correct or conceal. (See illustration right.)
And frankly these read errors are the least of the issues. For every time a stylus fails to accelerate fast enough and loses contact with the groove wall, it also fails to "move out of the way" of the next bend in the plastic groove. The result is damage to the record. When mistracking is present it isn't just the data that's getting trashed!
I have therefore developed a real aversion to poor tracking in phono cartridges and this provoked a more serious and fundamental study of various types a models of phono cartridge to assess their performance in this regard. The results, to my great surprise, revealed that MC cartridges do not (as a general rule) outperform MM types in terms of tracking. I had wrongly attributed the sweeter top-end to better tracking from the MC type cartridge. In fact, I discovered that some very expensive MC cartridges are actually very poor in terms of tracking. And some very modest MM types are exceptionally good.
The dissection of the mechanical assembly of typical, modern cartridges gives us much insight into this. A contemporary MM type (the AT95E) has two tiny, rare-earth magnets mounted perpendicularly at the top of the cantilever as illustrated in the figure (left). One can hardly imagine a lighter, more unecumbered electric motor part. By comparison a typical MC cartridge has coils and connecting wires clustered about the stylus cantilever which add to the inertial mass.
(The more gubbins symmetrically disposed around the cantilever in the MC type also explains why these cartridges often require a greater vertical tracking angle than the MM type. And a high value of VTA is bad news as explained here.)
I therefore turned my mind to the idea of electronics to compensate for the various shortcommings of the moving magnet type cartridge so that I might employ it for its better tracking. I discovered quickly that he "villain of the piece" in the moving magnet cartridge is the inductive nature of the pickup's impedance. It is this impedance which causes all the problems with this type of cartridge:
- Firstly, the cartridge must be terminated with a high resistance or the high frequency response is unacceptably attenuated. As frequency increases, the increasing reactance of the cartridge fails to damp the electronic noise due to these high termination resistances, and this sets a lower bound on the noise performance of the moving magnet cartridge.
- Secondly, the more ragged top-end (which I had falsely attributed to poor tracking) is due to the capacitance of the connecting cables interacting with the inductance of the cartridge to produce a resonant network with a frequency response peak in the audio-band where the impedance of the pickup inductance and cable capacitance are equal and opposite.
- Thirdly, the higher impedances of the moving magnet pickup circuit compromise stereo performance due to crosstalk between the right and left signals in the unscreened wires of the cartridge and tonearm.
Several circuit arrangements were tried to try to address this troublesome reactive source impedance and thereby isolate it from downstream capacitance and resistance. Negative impedance converters seemed to provide a possible solution, but these circuits are beset with stability issues unless certain, specific criteria are met.
A current-input stage was also designed. This involved the MM cartridge feeding the low-impedance virtual-earth point of a feedback amplfier arranged so that an internal "dummy" cartridge formed the feedback leg. The advantage of this topology being that where no voltage is present (ie. at the virtual earth) the cable capacitance can have no effect: and that the termination impedance is virtually a "short-circuit". This approach did yield some improvement as well as a workable prototype.
But an entirely more successful strategy was discovered to be a small amplifier fitted to the back of the cartrdige itself and that is the approach which we eventually pusued. I set the the requirements for such an amplifier as follows:
- It must be small, because space is very confined
- It must be light, so as not to upset the fundamental LF resonance of cartridge compliance and arm mass
- It must not require any modifications to a standard turntable wiring and must therefore derive its power via the audio cables (as does a phantom-powered microphone).
The first two requirements are straightforward and self-evident. The third is more unusual, and more taxing. It is the sucessful fulfilment of this self-imposed requirement which is the principal reason for developing this design into a commercial product.
After all, it would be trivial to arrange for a small amplifier stage in the headshell if extra wires could be fed down the arm to provide the ciruit with power supplies. But this would preclude this technique from the vast majority of turntables without extensive and delicate re-work. The approach described here may be applied to any turntable: past and present.
LinksPhædrus Audio Home Page
© Richard Brice 2016. All rights reserved. Newnes, CYP other information by permission