Autoformer vs Speaker impedance Curve


Autoformers vs speakers with wild impedance curve swings (for instance; MC601 amp paired with B&W 802D3 speakers).

There’s a wealth of information about tube amp audio transformers interaction with speaker impedance, but I can’t find anything regarding Autoformer and speaker impedance/phase curve relationships. 

Can any techies enlighten me? 

Thanks!

(I tacked a similar post onto the end of a 10 year old thread but thought I might get a few more hits with a new thread.  Sorry for the redundancy)



73max
You may want to ask McIntosh or a big McIntosh dealer like Audio Classics about the function of their autoformers.
An autoformers is just a transformer where the primary and secondary share a terminal. It doesn't make them behave radically different than a normal transformer. Pretty much everything that applies to trasformers is true for autoformers. 
Thanks all!

I’ve looked at Mac’s Autoformer benefit explanations. My takeaway: better matching of amp to speaker impedance. Ie, 4 ohm terminals for 4 ohm speakers, 8 ohm terminals for 8 ohm speakers, etc. The catch is these are “nominal” loads. 

What isn’t addressed is, what happens “electrically and sonically” with Autoformers when 8 ohm speakers, like 802D3s, dip down to 3 ohms or exceed 20 ohms???  

With typical output transformers (common with tube amps)  speakers with wild impedance swings (and phase) present a challenging load to the amp and fidelity can suffer.

Can the same be said of autoformers? Or, are these electrically different enough from transformers as to not present the same issues?

Can any of you electrical engineers or techs explain the relationship between autoformers and large speaker impedance swings? 


Large impedance swings and severe phase angles do exactly what you’d expect; they produce dips and humps in response. McIntosh is about the only solid state manufacturer to use output coupling transformers. If you’re driving benign, modestly reactive loads, all is peachy. But you throw some significantly reactive speakers at it, not so much. The autoformers in a SS Mc mainly pad the output devices from the messy business of truly gripping the speakers for the sake of letting the gain devices behave nicely. Not the best approach in my opinion and many others. Tubes are a bit different in that the transformers are converting fairly high voltage into current to drive a speaker. Mc autoformers aren’t doing the same job since transistors operate at voltages and currents loudspeakers are quite happy with. My Focals are directly connected to the drain pins of the amp’s output MOSFETs. It works well.
What isn’t addressed is, what happens “electrically and sonically” with Autoformers when 8 ohm speakers, like 802D3s, dip down to 3 ohms or exceed 20 ohms???

With typical output transformers (common with tube amps) speakers with wild impedance swings (and phase) present a challenging load to the amp and fidelity can suffer.

Can the same be said of autoformers? Or, are these electrically different enough from transformers as to not present the same issues?
With either an autoformer or output transformer the operation is similar, so we don’t have to separate them.

Transformers **transform** impedance. To do this, it goes both ways- if a lower impedance as you suggest above is on the 8 ohm tap, a lower impedance is thus also seen by the output devices.


This is exactly the same as if there were no output transformer at all- think about it- if an amp has no output transformer, it too will see a varying load.

How the designer set up the amplifier with the transformer plays a big role. The reason you do this BTW is to reduce distortion- all amps have higher distortion playing lower impedances, so if you can raise the overall impedance seen by the output devices distortion will be reduced. This is why Mac uses autoformers.

Now there is the issue of the varying load, and a lot depends on what the designer of that load expects. You might want to read the article at this link:
http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php

ESLs have an impedance curve that typically varies by about 10:1 over their range. Yet it works better for them if the amp can make constant power rather than constant voltage. A lot of box speakers don’t behave this way- for them, constant voltage over the range is more important, this in order to control the normal resonance and resulting impedance peak usually in the bass region near the speaker’s cutoff. To accomplish the latter, enough loop negative feedback applied will cause any amplifier, tube or solid state, output transformer or not, to behave as a voltage source and the resulting output will be constant voltage. This is why this comment:
Large impedance swings and severe phase angles do exactly what you’d expect; they produce dips and humps in response. McIntosh is about the only solid state manufacturer to use output coupling transformers. If you’re driving benign, modestly reactive loads, all is peachy. But you throw some significantly reactive speakers at it, not so much.

- is only partially true- if the amp is acting a voltage source, the above statement is false. Specifically in the case of MacIntosh, the statement is false; Mac led the way in the late 1950s towards developing the idea that speakers be ’voltage driven’ (IOW act like a voltage source) and I don’t think they’ve backed away from that, since most of the industry has followed their lead.