Ageing Quad Panels - How Old is OLD?

by Gary Jacobson



MT May 15, 2002
Yesterday I received an email from Gary Jacobson with some very interesting information on the ageing process in ESL membranes, I asked for permission to publish it and since this was OK with Gary, here it is.

First a comment from me:
We have the same experience with ESL-63. We have listened to several pairs of these and the sound differs from pair to pair, and there is even more difference between new repaired panels and old ones. We have also seen evidence of the cracking Gary describes further down in his article. Below is a picture of a broken ESL-63 panel where you can clearly see cracking along the borders of the stator segments.

ESL-63 element with polyester cracking


G'Day Mats,

I thought I would send you a copy of a recent post to Audio Asylum and the ESL circuit although the latter may not have appeared just yet. This is a short informal note summarising the results of some comparative testing we've done with old Quad diaphragms (7 to 30 yrs old). It indicates that the panels although still OK are not (probably not surprisingly) producing the sound they did when manufactured.

Ageing Quad Panels - How Old is OLD?

This informal note brings together some interesting anomalies of observation with old and new original model Quad panels, with proposed and tested mechanisms for those apparent paradoxes.

Background

It is almost universally assumed that an old Quad panel will lose its "zip" because of coating loss, or arcing damage. If not this, then output certainly falls off if the EHT supply diodes or capacitors start to fail. We "know" this, don't we? Well, under some conditions it is certainly known that an EHT can fail "high" with an output of 8 to 10 kV(!) and partly, or completely "lock up" a treble or bass panel such that the output is lowered substantially or even completely absent. This is an apparent paradox - plenty of HT - little or no sound - but a fact, measurable, explicable. Another thing we *know* about the original Quad panels (the youngest is 7 or 8 and the oldest is going on 45 years) is that if not arced or otherwise abused, and they sound 'OK' in general, that they are a good reference, and much as  they were, all those years ago - right? It is very easy to be seduced by this idea, since they still sound superb and superior to most modern speakers (sorry other planar lovers), assuming a good HT supply and a fairly reasonable layer of coating is still present - or we'd be getting less output, wouldn't we? Or is there a paradox in this home spun "truth" also. There is - here's the story.

Introduction

About 10 months ago, I was thinking along exactly the lines mentioned above, when I happened to pick up a patent document (non-Quad I might add). The thought suddenly struck me (as in lightning) that this aforementioned idea was utter rubbish and totally assumption. Had David Williamson or Peter Walker been present, I suspect they would have *struck* me somewhat prior to that! A good slap 'round the ear - 'wake up lad!' The only way to verify actual performance electrically and otherwise, is to measure it in some fashion - preferably reproducibly.

So, do the old, seemingly *good* panels, produce the same flat, smooth frequency response as per specification? Answer: The same? NO, Flat and Smooth - YES (with qualifications). Well, that *was* helpful, wasn't it?

Having measured about 10 of the treble panels of 'untouched' vintage that we were able to lay hands on, we have found one (I'm *sure* of) that meets (nearly) the original specification. I consider that a minor miracle, if nothing else. Mind you, put in to service they produce music with alacrity, aplomb, fluidity and balance that one sometimes thinks only a Quad panel can - how can they be faulty? Clearly, we are not referring to gross discontinuities or faults here.

End of Fantasy - Enter Reality.

What Do We Hear?

When an old, but otherwise 'perfect' example of an original Quad Treble Panel is played beside my rebuilt (by me, for me, only me, right here, not for sale - OK?) panels, most of the 'new' panels sound identical in all important musical aspects, but not as 'forward' or 'bright' in balance. Listening carefully, you can certainly hear the extension to the highest frequencies of human hearing; the mid range is of perfect tonal accuracy, but when you play those old originals, with the same source and in the same speakers you get that little bit more 'presence' and 'forwardness' and 'sparkle', call it what you like. It is a very attractive sound and very easy to listen to, and not in any way fatiguing, as is the 'brightness' of some (even quite good) moving coil tweeters. I was reminded of a comment made by Peter Walker at a demonstration in the 50s that, "...if you 'noticed' the treble or the bass, then something was wrong, as you do not 'notice' it at live performances.."I am drawn to the conclusion, or rather, have known for some months that the old panels are indeed "wrong" in comparison to what they would have been factory-fresh. I can measure it. I can hear it. Others can certainly hear it...and...we have...

A Mechanism

Plastic film undergoes a process of 'work hardening'. Flexible though it was when fitted, the polyester in these diaphragms was heat treated to 'set' the polymer, and was then stretched to the correct (!) tension.  This is to avoid polymer 'creep'. If not done, the diaphragms would present with all kinds of odd tensions within a few months of operation. Some folks overdo this and *completely and totally* heat shrink the film as much as they can to get the tension in it. That's absolutely wrong, not the way Quad did it, but a long and separate story also. It gives an uncontrollable, unrepeatable tension, for one thing.

If 'work hardening' is a new phrase to you, just remember what happens when you flex a thin piece of plain copper back and forth like a hinge a few times - crack! That's the general idea.

Now, clearly, with flexible films this is not going to happen in minutes, hours, days, months or even a few years. However, over the lifetime of these older panels 5 years and more, there is time for this to occur. It is slow, accumulative and virtually unnoticeable as it is occurring, until you compare it with a panel built to original specifications with new material. The high voltage environment certainly has a part to play in polymer hardening, and so does another factor that I'll mention later. The major culprit in most of the panels seems to be those 500 to 20 000 Hz vibrations. Remember, a lot of the original elasticity was deliberately removed by heat treatment to avoid polymer 'creep' prior to fixing the diaphragm to the stators. This is also true of the '63 and the 900 series, by the bye. Yes, I am implying that the same thing is happening in your later model Quads also.

Evidence

All very nice - you say. How do you measure this process? Directly, we can't; but we can, and have, taken old diaphragm material out of panels and subjected it to tensile strength tests to failure. We have done the same with several modern plastic films of the same chemical composition.

What do we see?
The older film exhibits a typical Hooke's Law region (linear stretching), and then sudden catastrophic failure with a very sharply defined failure point - not unlike the failure of gray cast iron - qualitatively! Of course the magnitude of the forces involved is far less. New film of similar composition and gauge from (in some cases) the same manufacturer; fails with a slight hiccup in the stress/strain curve with a much broader, but definable, failure point (heat setting aggravates this) but not as sudden and sharply defined as with old films. This tells us that there is non-linear stretching continuing as part of the failure process, for what it's worth. I am not advocating driving your Quads to mechanical destruction!

In simple terms, this means the old film exhibits all the mechanical properties of a stiffer, more brittle material. Which after all those years in HT fields and exposed to that much persistent vibration; it is; in a manner of speaking. We can go a bit further, and take a sheet of new polymer, expose it to many thousands of hours of audio frequency vibrations, remove it and measure a very, very similar effect. The assumptions are for the purposes of this test that one listens about 2 hours a day, 5 days a week, for about 10 months of the year. The test can then be run 24 hours a day, so we can squeeze about 2.5 weeks of 'normal' operation into a single day, and simulate 10 months in a little under 20 days. Over the last 8 to 9 months we've looked at about 8 to 9 years of running time.

Do we measure any 'hardening' effect in action? Yes, we do.

Would a totally unused Quad show this 'work hardening' of the membrane, just sitting around somewhere in storage? I really can't see how it could, but if you've got some unused panels lying about, there's one way to find out - anyone? Actually, I have a proposed mechanism for those situations also - see 'Other Matters' later in this note.

In 'reality' where some of us live when not at the Asylum or similar institutions, we probably don't play our speakers *all*that*much*, *all*the*time*, but they do tend to be switched on with the film immersed in EHT all the time. So, we have a whole range of vintage panels, quite untouched, not arced, properly supplied with HT and music signal, which all sound subtly different from each other, and not so subtly sometimes! They all, in turn, sound different to new panels built *to* *the* *original* *specification*!! That is, built, not to match the sound of an old panel, but to the specification as used on the Quad factory floor. Quad as of yore, not IAG Inc. as they are today.

A Significant Measurement

The "sounding different" is caused by a perfectly simple and quite significant mechanism for *any* speaker. The moving diaphragm has altered in stiffness! This produces, in a Quad ESL of 8 or more years age, which has experienced regular use, a 'lift', if you want to use an audiophile word, of between 1.2 and 1.8dB in the mid range frequencies and higher. There can even be, in theory, a difference between the HF and mid-range strips since they are exposed to different rates of vibration (duh!) and different amounts of vibrational energy. HEre, then is the apparent paradox, and olde panel that sounds louder!

Other Matters

1. Temperature effects were not considered in these tests - i.e. the tests were performed at constant temperature. In the real world temperature fluctuations of even ordinary magnitude make a difference to the flexibility of polymer films. I do not suggest that we thrust a Quad into the nearest snowdrift, but, if you've got a nice big refrigerator or cold room?

2. What about a sealed, boxed, untouched Quad Treble Panel? Glad you asked. I did. We had one, vintage 1981-2 as far as we can glean. Perfect condition to the naked eye, not exposed to all that vibration for all those years, but storage history was uncertain. The box looked pretty good. OK, I can apply the bootstrap and Tuke's JackKnife to 10 panels' results; but one panel? That's pushing your luck. Anyhow, what did it show on measurement/audition? It's different from all the others and not in a subtle way either, when musical refinement is your goal. I'm not going to say it was faulty because it measured well across the pass band of the panel. So? A physical chemist has pointed out to me that the soluble nylon layers and the polyester layer do not bond intimately (hey - I knew that). Well enough, it seems, to last for a dozen years or more, but eventually, the nylon 'falls off', so to speak. There are so many absolutely *necessary* reasons for using nylon on this particular speaker that Quad didn't consider anything else, until the material became unavailable to them in commercial quantities. When encouraged to flex and bend together when 'young' the three layers assume virtually identical expansion characteristics; and stay so, if 'exercised' regularly. As they become older, if not used at all, the layers exhibit different expansion rates. Particularly nylon, which is hygroscopic and expands and contracts by absorbing atmospheric moisture over the years - the dust covers are not perfect. Indeed, British Patent (GB 815,978) for the speaker stipulates 38% to 95% R.H. to achieve the necessary conductivity on the diaphragm. Moisture adsorption is essential, in fact. To appreciate this effect, spread a thin film of some innocuous glue on your skin - they flex at different rates - very different to a speaker diaphragm of course, but you get the general picture. Again, a subtle stiffness differential is proposed as the mechanism that alters the sound of the panel from brand new, factory specification. This, if proved, would be additive to the previous mechanism, as observed, since the absorption of molecular water is required for the original panels to work properly.

3. Duplicating the "old" panel sound, is, of course, *entirely* possible with modern materials. You simply add a couple of ounces to the tension (above specification) and apply just about any coating you like and voila (!) you can get a good match, 1.2 to 1.8 dB 'lift' and all to an old 'perfect' panel - been there, done that. How can I seriously advocate a 20 to 40 year old panel, however apparently 'pristine' as a *reference* when I know (now for sure) that even a 7,8 or 9 year old is suspect? I cannot.

4. How long before this effect is significant? You'll never hear it in isolation. It occurs far too gradually for that. However, if you compare, as we have done, a factory spec., brand "new" panel, 6-micron, nylon coated with an olde panel you will hear the difference. If your speakers sound a bit "off" suddenly, then it's something else that's causing it, not this effect. Look elsewhere. You can live with this quite easily, and I would too. Just don't put your head in the sand and think that because you've never arced the speaker, and never had a bad word to say to it that you've got the sound that was produced by the original production Quad ESL - you haven't - that's all.

Conclusions

Old Quad Treble panels sound different in sound from each other and from modern equivalents rebuilt to the *exact* original specifications - 6 micron polyester coated with nylon, tensioned precisely - because of a slow increase in stiffness of the diaphragms of old panels with time. The 'paradox' of an increase in the mid range output and some higher frequencies is explained by this increase in stiffness. Higher output in any ESL is related directly to higher membrane tension (Kellogg, Shorter, Hunt, Williamson, Walker, Janszen and many others) The proposed mechanism for the increase in stiffness is the polymeric equivalent of 'work hardening' such as occurs in metals, though *not* an identical mechanism.

Aaarrgh!!!   I just want to listen to My Music - OK?

You can rebuild a panel that can *match* the sound of an original, old panel of 'perfect' vintage condition. The sound is quite wonderful. Everything you expect from a Quad panel will be there, but actually subtly altered as described in this note. It is a bit brighter, and more forward. The olde panel is *not * a reference in this sense, and people shoul dnot try to pretend that it is. I used to count myself in that brigade, so this is a bit of self-criticism. You can rebuild a Quad panel to the original factory and patent specifications (6-micron, nylon coated, correct tension). It will sound utterly superb. It will not be as forward or as bright as the old 'run in' panels. You will *never* hear this other than in a side-by-side comparison. The slight exaggerations (if one may say that) of the olde panels are absent. The sound is truly mellifluous, with all the depth, perspective and precision that one expects from a Quad panel. The 'magic' is really there.

Which one do you want to listen to?  I don't know. It's a free world.

(c) Gary Jacobson, May, 2002

The Quad ESL (Gary's web site)

  

This site uses frames, but if you have found this page via a search engine and do not see a frame on the left side of this page, you can use the home button to get to the main frame of the MT Audio Design site.