MC phono stage without transformer?


A newbie question:

I read a lot of 'reservation' about using an external MC step up transformer to increase the gain of an MM phono stage. But as I searched around for MC phono stages, I noticed that a lot of these actually have internal step-up transformers, some of these transformers are exactly the same as what some people used to make their external step-up.

So if transformer is no good, I should really be looking for an MC phono without the tranformer? Do these exist though?
viper_z

Showing 2 responses by kirkus

While I very much respect Raul's and Atmasphere's comments on the subject, and I agree that the subject is a complex one . . . it isn't terribly ambiguous, and there are some simple facts to the subject.

First, transformers and inductors are eschewed components in undergraduate EE design practices, and this is for sound, logical reasons. But their potential shortcomings have been very well researched over many decades, and they can be measured . . . and it is indeed possible to transparently apply a high-quality transformer to the task of a MC input stage.

Second, since when is an audio transformer an 'economy' or 'low-budget' component? Yes, a designer can save some time and effort by slapping an off-the-shelf transformer in front of an existing MM phono stage and calling it done . . . but if the product is designed in this manner, the transformer's shortcomings are likely to be only the tip of the iceberg. This is a problem of methodology, not philosophy.

Third . . . in a low-noise circuit, input stage noise will dominate, if the designer has done his/her job correctly. And low-impedance transducers (like a low-output MC cartridge) simply do NOT deliver their best noise performance when coupled directly to an input stage, especially one that uses JFETs or vacuum tubes. Now I understand and respect a design approach that tolerates a mismatch here to acheive other objectives, but in terms of noise performance for i.e. a 5-ohm phono cartridge, a well-executed transformer/autoformer input topology will ALWAYS perform best.

And finally, the biggest issue with improper cartridge loading has very little to do with the transformer or the lack thereof, but with the lack of standards and consistency among cartridge manufacturers. There is simply too much variation in cartridge design and manufacturing methods to be able to allow EVERY cartrige to work its best with a single type of phono preamp . . . a simple gain adjustment and loading resistor can only do so much. Kinda like finding ONE amp that will bring out the absolute best from EVERY speaker.
I would like to bring up a couple of related side issues for comment . . .

First, I think that a division must be made between the use of external vs. internal transformers. Obviously, the proper loading of the transformer secondary is extremely critical, and if it is an external unit . . . than I can't see how the real-world loading can be anything but indeterminate. It also seems that external units pretty much all have very high ratios . . . for i.e. 30dB gain. This makes for an inherently compromised transformer design, and additionally puts the entire chain in a sub-optimum gain structure. An interesting comparason would be to an active MC head amp . . . an arrangement that I have found quite unsatisfactory with the few that I've used and measured.

Second, a transformer simply cannot offer anywhere near the level of flexibility in cartridge loading that an active design can . . . there are certainly a great many cartridges (including all MM types of course) for which a transformer is simply unsuitable. I must say that I am greatly impressed with the thoroughness that Raul has approached this subject, and we are quite fortunate that he freely shares it on this forum.

I'm currently in the evaluation stage of a new phono preamp design, and the main topological choice was between a direct bipolar input with four MPS-A18s in parallel, and using a low-ratio Jensen JT-346 with a 990 opamp. I chose to use the transformer/990 because it was about 5dB quieter (5 ohm source), had somewhat greater headroom, and inherently blocks the input bias current - in addition to the usual transformer strengths of ground loop isolation and RFI rejection. That's not to say that the direct bipolar approach didn't have its strengths . . . bandwidth and LF linearity were of course among them. And I did have to make a decision to make the unit incompatible with many cartridges in order to use the transformer topology.

In the end, there are pitfalls with both approaches, and I'm confident that my success or failure with this will be overwhelmingly dependent on my ability to carefully optimize all the circuit and construction parameters . . . and very little to do with my "reactive component dogma".