McIntosh and Autoformers. . .?


What is an Autoformer, and what kind of difference will it make to the sound? I heard some B&W 803's the other day with the McIntosh 402, and it sounded absolutely unreal! I was so impressed, I am thinking of dumping my current SET gear, and going with McIntosh and B&W's.

A freidn tells me that I only want the new McIntosh stuff with the Autoformers. I don't know why. Will the 202 have similar sonic characteristics to the 402? What about the 6900?

Thanks!

B
hantrax

Showing 4 responses by nealhood

Germanbox, your speakers and amp combo would appear to fall smack in the middle of the grey area with regards to sonic improvement with the addition of an auto-transformer. My experience was with the MA-1 Mk II, not II.2. Yet I would speculate that the benefits with the AT on my Maggie 3.6s would be the same whether the MA-1 version was the II or II.2.

My friend, you have an outstanding system to say the least. You deserve the best that it can offer, whether this means adding the AT or doing nothing. So, for the comparatively small price of Paul Speltz's "Zero" transformers, my recommendation is to contact him and work out an audition arrangement. Or, you can opt for Ralph's more expensive (buy probably no better) "Music Z" auto-transformer. Fit, form and function is basically the same. Then you can provide us with a follow-up report and further shrink that grey area where the AT works well.

The autotransformer is a single winding coil with taps on a core of steel (generally). The taps allow impedance transformation by the ratio of turns being used. So one is able to connect, lets say, an output stage with output impedance of 1K ohms to an 8-ohm speaker with minimum loss. There are some losses and residual distortions associated with ATs. Hysteresis and eddy current losses define the core power losses and the core linearity will contribute a small amount (negligible in most cases), of distortion. This is generally only an issue at very low frequencies where the signal drive is often high and reactance is low therefore more current, therefore more volt-seconds are produced in the transformer driving the core closer to saturation.

An autotransformer does not block DC. As a matter of fact, residual DC on the output of an amplifier is detrimental to the AT operation as it will saturate the core. This residual DC creates constant flux in the core - then the dynamic AC output either adds or subtracts from this residual flux. When additive, if the flux goes above its knee limit (determined by the amount and type of steel in the core), then it saturates and the transformer then approaches a short circuit. Needless to say most amplifiers do not like short circuits. In balanced operation though, assuming the same residual DC from both circuit halves, the flux will cancel in the core so this I guess we could say that it blocks DC given the assumptions that the residual DC in each half is equal.

Even with the negatives of inserting another piece of iron in series with the signal, the advantages of an AT should outweigh the disadvantages. It is easier to design a power circuit with impedance higher than 4 or 8 ohms due to classic technology. Optimizing a design and living with the resultant impedance then adding a well designed AT should just about always be superior to designing a low output impedance stage using global feedback and all the other tricks of the amplifier trade. The AT just makes life easier on the output stage. The output transistors handle less current therefore produce less dynamic signal distortion due to changes in beta (gain) as more current is passed through them.

Most tube amplifiers incorporate an output transformer as part of their design due to it being virtually impossible to design classic tube output stages with low output impedance. So full transformers are used to effect the required impedance transformation. Assuming the output transformers are well designed for audio applications, I don't think the AT is necessary. However, the AT is almost mandatory for OTL amplifiers. And, it really makes a difference here. Listen to an Atma-shpere with the AT on a pair of Magnepans, or most any other high end speaker, and you will be willing sell your wife to buy one.
You are right Zaikes. With speakers above 8 ohms there is a lesser need to consider the AT. I am not sure just where the crossover point of diminishing returns lies but I would probably pick 10-12 ohms. So even with an 8 ohm speaker I personally think the benefit of an AT could be considerable.

Even though McIntosh claims 35db feedback in their amplifier the fact is that, with a relatively low impedance speaker such as the Magnepan, the driver circuits will be working harder, pushing themselves closer to their limits so to speak, as well as the output transistors. Feedback aids this process. With the larger operating envelop and higher output stage currents we get negative attributes such as higher transient modulation distortion and, for lack of a better term, some amplifier signature.

With the AT in the McIntosh, I would not think the driver and output circuits will not be operating as widely in their envelop. Less driver swing, less output stage current, thus less dynamic distortion and less of a "signature". Keep in mind that some of these distortions are hard to characterize and I am not an expert. The propeller head on Audio Asylum may be good site for further discussion.

I see that Paul Speltz has posted on this thread. I highly recommend his advice on amplifier/speaker compatibility as he not only markets the "Zero" AT but also has much experience in their application. There is a couple reviews of the "Zero" in press; one on the "TNT" audio site and one in the January edition of AudioXpress magazine. I think its January, could be off a month though.
Good news, from my perspective anyway. I don't get bowled over by cables. Just keep then short as possible and use any decent 10 or 12 awg shotgun cable (available at your local Home Depot or Radioshack). Longer runs may require more attention but, with monoblocs you should be able to keep them relatively short.

I have not heard the dramatic differences between cables as I do between components. There have been very slight changes but, it seems that whenever I was happy with a given system, I was also happy with it when different cables were employed. They just don't make or break a system in my opinion.

I know the merits of cable selection can be a subject of heated debate and I am going against the grain here so you need to keep an open mind. However, to get started forget about the expensive stuff. In your system even lamp cord will perform excellent.