Fat recording console, circa 1944

For my 200th post, I thought I'd offer something fun. Here's an article from a 1944 SMPE (predecessor of SMPTE) Journal describing a 20-channel re-recording console built for MGM Studios. Among its many interesting features is the use of mixing transformers, resulting in only 23dB of loss from the summing of 20 channels. Please right-click on the following link and select "save target as."

2Meg PDF

In those days, sound-for-film was the vanguard of audio recording technology. In fact, the earliest "pro audio" textbook of which I'm aware is Motion Picture Sound Engineering from 1938, which was co-authored by Harry Kimball, one of the authors of the article posted above. This book is referenced in a number of later works, including Tremaine's Audio Cyclopedia. Kimball's contributions to the book include a couple of seminal chapters on the design of equalizers, a subject which is also covered (in less detail) in the paper about the MGM console.

After much searching, I was lucky enough to obtain a clean copy of this book a few months ago. I intend to scan and post selected chapters when time permits. Since the book was published 66 years ago, I don't believe I'd be violating any current copyrights.

Thanks to Tom F. for the scan of the SMPE article, and to PRR for the webspace to host it.

Thanx Dave :thumb: That's a good read...

[quote author="NewYorkDave"]

In those days, sound-for-film was the vanguard of audio recording technology. In fact, the earliest "pro audio" textbook of which I'm aware is Motion Picture Sound Engineering from 1938, which was co-authored by Harry Kimball, one of the authors of the article posted above. This book is referenced in a number of later works, including Tremaine's Audio Cyclopedia. [/quote]

Hey,
I seen those pics in the Audio Cyclopedia, it is definitely one of the gems in the goodOl' 621 at college.

Pretty cool they totally had forseen Flying Faders and Total Recall way back in '44 :shock: and Euphonic like control surface with distant processing racks as well..

Transformer summing looks very interesting and I would love to know about this. Is this something you're going to use in your DIY mixer, Dave?

[quote author="sismofyt"]Pretty cool they totally had forseen Flying Faders and Total Recall way back in '44 :shock:

T[/quote]
They not only predicted them, but clearly expected them to be implemented in the near future. This was only about 20 years after the beginnings of commercial radio (!!!) and innovation like that was taken for granted.

How about those foot-pedals ?

-Eric

[quote author="sismofyt"]Transformer summing looks very interesting and I would love to know about this.[/quote]

The principle of operation is essentially the same as the "hybrid coil" that has been used for many decades to separate send/receive audio from a two-wire telephone circuit... except in the case of the mixing transformer, it's used the other way around, to combine two signals with minimum interaction. If you have access to a copy of Tremaine's Audio Cyclopedia, second edition, there's several pages in the Transformer and Sound Mixer chapters that go into some detail on mixing transformers. There's also a ton of references about hybrid coils on the web.

The main attraction of hybrid coil mixing, at least circa late '30s and early '40s, is that it can be arranged to have half the loss of conventional resistive mixing (e.g., 3dB of loss from summing two inputs as opposed to 6dB). When gain was more readily available, as amplifiers and tubes became more affordable, coil mixing ceased to be cost-effective. As early as 1953, the Radiotron Designer's Handbook, 4th edition, offered only a very brief explanation of transformer mixing and commented, "coil mixing appears to be very little used." I can verify from my own research that coil mixing is not often seen in mixing console designs even from 50 years ago.

Is this something you're going to use in your DIY mixer, Dave?

Click to expand...

Nope. First of all, my design is not a constant-impedance type. Coil mixing can only be used with defined impedances on all ports; and even then, it's hard to maintain good isolation (trans-hybrid loss) between the ports across a wide frequency band. Telephone hybrids have it easier, since the frequency range of interest is restricted--but even then, the adjustments can be critical. (As a broadcast engineer, I've wrestled with my share of hybrids, so I know).

In a mixing application, it's much more cost-effective to provide a few extra dB of makeup gain than to mess with expensive specialty transformers. And, as much as I like iron, cascading a bunch of transformers unnecessarily will not do good things for distortion and frequency response figures. Coil mixing is a very cool idea, on the face of it, but the benefits don't seem to outweigh the drawbacks in 2004.

Having said that, if one wanted to use coil mixing nowadays, there seems to be a few options available. The UTC LS-141 was commonly used in this application, and here's a nice pair of 'em on evilBay:
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=3823531085&sspagename=STRK%3AMEWA%3AIT&rd=1

Sowter and Lundahl make hybrid transformers that might be adaptable to the purpose:
http://www.sowter.co.uk/specs/8356.htm
http://www.lundahl.se/pdfs/datash/6702.pdf

Edcor makes a line of "coupling" (two-channel mixing) transformers, but they offer very little on the way of specs, certainly none regarding isolation between inputs, and that doesn't inspire a lot of confidence on my part. Still, for $8, it might be worth buying one to test out on the bench. Note that these are unshielded and therefore not suitable for low-level signals.
http://www.edcorusa.com/transformers/audio/coupling.htm

Thanks Dave,

That's a fascinating and early audio engineering article!

Yea, I was interested in the transformer-summing too. It's an obvious solution when active-amplification was bulky and expensive! The EQ information is great too.

:thumb:

Mark

Thank you for that information, Dave. :thumb:

I don't understand the difference between an ordinary trannie with two primaries and that UTC. Is there a difference? For mixing say ten signals/channels together could you use a trannie with ten primaries?

The WE console looks like it's doing some pre-summing with more trannies as well.

I'm curious as to how that would sound!