Quad ESL 57 – crossover/phase issue?

mattstat

Member
2009-10-19 7:47 pm
3, When both speakers are measures together there seems to be some cancellation at certain frequencies.
I don't use REW, so can't look at your files. And I'm not sure what kind of measurement you're referring to here, but if you are measuring both speakers simultaneously with a correlated signal and the mic at the listening position, it's normal to have interference/cancellation between the speakers at higher frequencies. They would both have to be EXACTLY the same distance from the mic at all parts of the midrange/treble panels to prevent phase cancellation at some frequencies.

To me, the potential wiring/phase issue is the largest one to be concerned with. There are a zillion other things that might be a problem, but focusing on them seems to be making things more confusing (and most of them are probably fine anyway). If you aren't used to making measurements and dealing with electrostatics in general, it can be overwhelming trying to work on every idea at once.

I think you can probably get this sorted, but it will require a methodical approach focusing on one thing at a time. Start with the biggest/easiest to fix. Get that resolved to everyone's satisfaction, then move to the next thing if that wasn't it.

Multiple people here want to help you, so take your time or take a break if you need to, and then come back to it. Sometimes a few days away from a problem helps.
 
Rodney, if you are concerned about one or more panels out of phase, but not about decreased sensitivity, here is what I would do (apart from physically checking the wiring):
Disconnect the signal wires from all panels sequentially, except one (left bass, right bass, middle, treble). Isolate well the disconnected wires. This way only one panel will work at a time. Do the same on the left and right speaker. You can test the bass panels mirrored (left speker left bass vs. right speaker right bass). Then place them about 2 meters apart, and listen to them from the centre also from 2 meters distance (equal side triangle). Now it will be clear by hearing if one panel is out of phase or less sensitive.
 
I am very willing to help out however I can,

Hello Sheldon,

I appreciate your willingness to help.

I've done some very close-up measurements (at 1cm). My final measurements I promise! They just might throw some light on the problem. Of course, my home measurements might not have much value. REW files attached. (in two parts as one file was too big)

Even so, it seems to me there is quite a lot of inconstancies between the panels, particularly the tweeters.

I can't understand why the tweeters are not rolling off at the bass end. Perhaps there's leakage from the bass panels. Despite being measured at 1cm from the grill.

To be fare to the OTA panels, we really don't know just how consistent the original Quad panels were. Having said that, I do think that my speakers are not up to scratch and do need attention.

At the end of the day, how many Quad ELS owners out there have taken measurements of their speakers...goodness knows what they might discover. Sometimes it's better not to know!

Regarding you helping me further, I really don't know how we can manage this. I don't have the confidence or will to start disassembling them in order to test each individual part.

Thanks, Rodney.
 

Attachments

  • Close-up tests LEFT Quad 13.1.22.mdat
    5.6 MB · Views: 12
  • Close-up tests RIGHT Quad 13.1.22.mdat
    5.6 MB · Views: 6
This is not directed at you, Dr. Stokes.
You're one of the very few respected authorities on this subject, and I'm grateful for your contributions.
My apologies if you feel offended.

Oh hell no, I don't feel offended at all. We are all good here. I'd love to see more watchmakers in the area, and less slap and dash folks.

Don't forget that this is a hobby of mine, I try not to take myself or my work too seriously. I hope I didn't come across as offended. It's hard to get meaning in these forums especially across cultures too.

Sheldon
 

rayma

Member
2011-04-29 8:37 pm
Perhaps the audio transformer is an early one when MONO was in common use and Quad didn't pay much attention to polarity?
Peter Walker described their quality control process. Face the test speaker toward the reference speaker,
a couple of feet apart. Place a mic in the center. Play a mono signal to both. The difference picked up by
the mic should be very small. This would not work if the test speaker's polarity was wrong either overall,
or even wrong for just one of the panels.
 

wout31

Member
2012-02-13 7:59 pm
Limburg
Let me start by saying that I'm not a watchmaker, but I don't think of myself as a slap and dash person either.
Second let me say I have done many REW measurements on ESL-63 and very little on ESL-57. Never have done these close mic measurements.
Also, reading measurements and drawing conclusions is a totally different challenge from making the measurements.
Having that said, let me explain what I think I see.

First, I see 2 discrepancies in the tweeter panel of the left speaker. There are 2 large dips in the SPL. One in the low frequencies and one in the high.
The first is at 250 Hz, down 20 dB, the second is at 3.8 kHz and is down 25 dB.
You can also see the difference in the tweeter panels in the step response. The one in the left speaker has a very large overshoot.
Very very careful conclusion: The tensioning of the Mylar on that panel is too low?

Second, I see much more energy in the right speaker right bass panel over a much larger frequency range than in the other 3 bass panels.
Don't see much difference in the SPL on the bass panels, but there is definitely something going on in that right speaker right bass panel.
In general all bass panels are not very much in line energy spectrum wise I think, but the right panel in the right speaker is off the most.
No conclusion here. I see the differences, no idea what might cause them.

tweeters.jpg


step wteeters.jpg


energy.jpg
 

Zung

Member
2005-04-30 7:30 pm
Geneva
From your step response, it's clear the tweeters are in phase, but the L one has about 2x the output of the R one (3dB).
Assuming the mic to tweeter distance in both cases is the same, and considering OTA's meticulous craftmanship, there's a possibility the 2 tweeter panels are different, as is the case with the bass panels.
Can you post pics of both speakers, naked?
 
Can you post pics of both speakers, naked?
First,, I wish to thank you all for your time taken analysing my measurements.

Regarding "stripping" my speakers. Presumably that means removing the front grills as well as the back. I can try that but I believe the front grills are fixed with staples? How easy is it to re-staple them back? I think I have a heavy duty staple gun somewhere. Will we gain much by checking the tweeter and bass panels I wonder?

Rodney
 
I'd recommend against removing the front grills, it's a big messy deal. Many heavy staples.
Hi rayma, despite the speakers having some issues, they still sound very nice indeed, at least to my elderly ears:D
It may well be that my pair of speakers has an assortment of drivers from different batches.
Even if I bought another pair on ebay, they most certainly would need restoring and that would mean sending them up to OTA. Maybe they would make a fantastic job, we just don't know. A newly restored pair most probably would be fitted with up-to-date drivers from the same batch.
We don't know how long these speakers have been by Steven at OTA. Strange that he wasn't aware that they had to be played with one speaker connected out of phase! If he did, it wasn't mention when I bought them.

Rodney
 

Zung

Member
2005-04-30 7:30 pm
Geneva
If you have to go to the hardware store, here's your shopping list:
  • Stapler
  • Staples, 10mm deep (good for the front grill as well)
  • Electric screw driver if you don't already have one
  • Screws, M2.9x16mm, self-tapping wood screws, stainless steel, Philips head, to replace those you mess up
  • Wood glue to fill in the holes of the previous screws
Me, I never bothered to put the felt back and have been living happily since a quater of a century. But this is controversial, I know.