Quad ESL-63

Improved Stability.
Let us start with the easiest tweak of them all. The membranes of the Quad ESL-63 are large and the speaker is not heavy, it is therefore easy to imagine what happens when for instance a bass drum makes the membrane move forwards. Yes, that's right, the speaker will move in the other direction (Newton's third law). The easiest way (and best I think) to minimize this problem is to put some weights on the top of the speakers. I used HiFi magazines when I tried this, but when I bought my Gradient SW-63 there was no longer need for this, as the ESL-63's are firmly hold in place by the SW-63 and relieved of the lower frequencies. This tweak improves the quality a great deal, giving better dynamics and bass.
Newton's Third Law

Rubber Feet (resonance damping).
From Renze de Vries in the Netherlands I got this email about feet for the ESL-63/SW-63 combination:

I think I found the definitive solution to put the ESL/SW-63 on its feet!
As you know, the Gradient SW-63 comes with two sets of feet to place the loudspeaker combination on the floor, a couple of spikes to be screwed and locked into the bottom plate and and a set of small rubber pads to be sticked to it. I never used the pads, because I don't like rubber (instead I tried out larger rubber feet used to keep washing machines from walking through the kitchen - with the expected detrimental result for the sound).

The use of spikes didn't satisfy me either. As all cones and spikes, they resonate along with the device they are attached to, adding their own resonance frequency of course and, by doing so, modifying the total resonance spectrum. The results can be more or less in the desired direction (too fat bass for instance sounding thinner, and thus appearing more tight, as was the case when I used them) or they can be bad, but whatever the results may be, you are juggling resonance(peaks) instead of getting rid of them altogether. So, the best solution is to use real damping devices for feet, real vibration absorbers.

Sorbothane pucks or hemispheres are well known and working quite well to handle this job. But alas, as they are necessarily soft, there's a limit to the weight you can put on them. For the ESL/SW-63, with a total weight of 37 kilo's (about 81 lb) each, one would need the dire amount of 24 Audioquest Big Feet, setting you back about 1300 Dutch guilders.

But here's the perfect solution: Vibrapods. They are soft vinyl pucks, to be placed under whatever piece of equipment you like, delivering just about the same results as Sorbothane, and what's more, able to carry more weight. The Vibrapod Company (St. Louis, USA) has 5 weight ranges, from 1 to 11 kg per piece, so the highest numbers, 4 and 5, should be just right to do their job under the ESL/SW-63 combination. And they do!

I placed 2 number 5's and 2 number 4's under each loudspeaker (the 5's in front, where the speaker is heavier) and yes, everything the raving Pod-reviews promise you is there: blacker background, easier presentation, more dynamics, sweeter highs, more precise and tuneful bass, more midrange detail. Violin tone, for instance, is cleaned up, contrabass easier to follow, grand piano more solid. One simply listens with more ease. And all this for the beautiful price of 6 US dollars (4,75 Britsh pounds) per pod.

Any downfalls with the Vibrapods? Not that I know of, or it should be the decreased firmness with which the loudspeaker combination stands on the floor. Because of the soft feet, one can move the top of the speaker slightly by hand. So this is not the most rigid loudspeaker placement. But who says that's the right way? In my experience, this Vibrapod solution works better than the rigid approach I tried with slabs of bitumen and concrete blocks.

Here's the Vibrapod Company's site, where you can find all information, distributors, reviews and all: http://www.vibrapod.com

Best Regards,
Renze 

Improved Input Frequency Correction.
Directly on the cable connectors, there is an 1R5 resistor in parallel with a 220uF capacitor (remove bottom cover to find it). The capacitor is electrolytic, which is not very good since almost all frequencies in the mid and treble passes through it. Replace this with a high quality polypropylene capacitor, and while you are at it, replace the resistor as well with a high power non inductive type. Polypropylene capacitors with the required value are large, so it will be hard placing it inside the Quad. I have placed mine outside, and replaced the network inside with a straight wire. I have used only 100uF (250V Solen) in parallel with 1R5 and it works fine. With this easy tweak you will get a greater clarity in the mid and a much more crisp treble. Some run their speakers without the correction network, I have tried it and in my living room, with the Quad's placed in the middle of the room the result was a little too prominent midrange for my taste.

I have received some feedback on this tweak from Wesley in Hong Kong who has tried it in his very good set-up.
He thought that the clarity of the sound was improved, but he also thought that the soundstage was a bit too up-front. I recommended him to add another 100uF polypropylene capacitor in parallel making it a total 200uF in parallel with 1R5. A few days later I received an email stating that this extra 100uF capacitor is a must. He wrote:
I decided to add another 100uF cap. Not only was the presentation too up-front, but the sound was getting a little congested after a few days of listening.  Remember I mentioned that the midrange was better focused?  I was wrong. Actually overall, the sound was much, much more detailed, but the up-frontness of the sound made for example a vocalist too in-your face of a sound.

I decided to try myself, in my system some recordings sound a bit too up-front also. My conclusion after some days of listening is that with 1R5//100uF the midrange is more detailed and up-front, with 1R5//200uF it is a bit more relaxed and laid back. Which combination you prefer may come down to your personal taste and/or the speaker positioning. I prefer the 1R5//200uF combination while the soundstage is more natural as is the voice reproduction, the treble is also a bit better defined and this also makes the bass reproduction more firm. I do however miss some of the directness of the sound with the 1R5//100uF link, maybe I will try a middle way later, like for instance 1R5//168uF. This correction link seems to be the right place to do the fine-tuning of the sound from your ESL-63 loudspeakers, and even if the used capacitors are not cheap they are much cheaper than 'high-end' cables.

More feedback on this tweak, this time from Renze de Vries in Holland:
I  had the capacitors and resistors in the input correction circuit replaced by two (big, indeed!) 100uf polypropylene capacitors and a high power metal film resistor. Happily, we succeeded in keeping it all inside - there's room for one capacitor right next to the old spot, while the other one was placed  in the corner near the transformer. We secured them  tightly with strips (near the transformer are two grounding posts, solder a wire between them and hook up the seatbelt).

But how does it sound? Well, first of all I have to say it isn't exactly what I expected. Maybe I expected something more spectacular. Well, it isn't spectacular, it's music.

For instance, a crisper treble? Not with me, I had a very crisp treble already, and now I have a more smooth, less distorted treble. A lot of digital glare and other nasties are successfully weeded out. Another result I didn't expect but which revealed itself from the outset, is better, more extended bass. A grand piano sounds definitely more firm, a harpsichord less like a toy (as it should be of course, listen to a harpsichord live, it almost always sounds softer and grander than you would expect). And then there's the midrange, which opened itself up after a few days of playing. In my experience that's where the modification really takes effect, in the midrange: more definition, more detail. Focus is improved considerably too, as is (of course) the exact location of instruments in space.

Summing it up, I would say that this modification contributes less to some special area of sound reproduction (a more detailed treble for instance), as to an overall effect: improved resolution giving more definition over the whole frequency range, nothing excluded or overaccentuated. And that's good.

I think Renze has probably described the increase in performance with this tweak better than I have done previously. When I wrote this page this was made from memory and since it was a couple of years between the tweaking and the site construction I may not have remembered everything correctly. From a theoretical standpoint this tweak should lower the distortion in the midrange and treble, and also make the sound 'faster'. This is because of the non-linearity and polarity of the electrolytic capacitor, and when you think about it the result ought to be pretty much what Renze has described. The improvement in the bass when the treble is improved is a side effect I have noticed and described previously on other tweaks.

More feedback on this tweak, from Graham Horwood:
As far as the ESL is concerned I did try your middle way with the input circuit by using a 150uf Solen with the 1R5, but found that still too in my face. So I have now paralleled that with a 56uf and its perfect. They both fit inside by the way, the 150uf alongside the transformer and the 56uf  in the original spot. Tight but OK.
Once again many thanks for your guidance and web site tweaks. They are the best improvement I have ever achieved (MT: The other tweak is the Gradient filter by-pass).

Removing Metal Grilles.
This tweak is a little more tricky and a little harder to undo (if you want to sell your speakers, the buyer will certainly want them in original shape). To be able to remove the metal grilles you must undress the speakers. First remove the top plate by pushing it gently to the right, then loosen the fabric and gently undress the speaker. Remove the grilles as described below and be careful not to bend them if you want to be able to re-fit the grids. This tweak improves the clarity, especially in the treble. Even more details can be heard.

From the Quad ESL-63 service manual:
Remove the wooden top from the loudspeaker by sliding it to the right (viewed from the front). Detach the grille cloth from the velcro hooks as shown in picture and pull the cloth down taking care not to ladder it.




From the Quad ESL-63 service manual:
The grilles are taped top and bottom to prevent rattles and once the tape has been removed it is possible to prise the front and back grilles from their retaining slots in the side extrusion. Care should be taken when handling the grilles as it is relatively easy to distort the edges beyond repair.

Removing Dust Cover.
This tweak will possibly shorten the lifespan of the speakers a little, as the charged membrane will be in direct contact with the polluted air. In order to get in contact with the dust cover you must undress the speakers in the same manner as with the previous tweak. Follow the instructions from the service manual below. As you can see the film is dampened in the middle, which effects the radiation (the ingenious delayed rings). This tweak improves clarity, but more importantly, it makes the speaker a better point source and therefore improves the imaging qualities. It also improves the dynamics. If you perform this tweak it is advisable to make some sort of cover to put on the speakers when they are not used since the membrane is charged and attracts particles in the air.

From the Quad ESL-63 service manual:
The dust cover is fitted to a clip-on frame and is removed by inserting a screwdriver at either of the top corners and gently lifting a corner free. The screwdriver is then slid down between dustcover and side frame. This should be repeated in each of the other three corners after which the dust cover can then be withdrawn.




The copy of the Quad ESL-63 service manual I have is not readable in the part that describes the picture above, but I think the basic instruction is:
Loosen the metal rods at the top and remove the dustcover frame gently without damaging the mylar film and then store it a safe place.

 

Parts Connection Quad ESL-63 Tweak Kit.
The Parts Connection sells a tweak kit for Quad ESL-63. I do not know what it contains or if it is good, but ask them if you are interested.

Crosby Modification for Quad ESL-63.
I received a nice email from JEHAN TITUS describing a modification of the Quad ESL-63 made by Crosby Audio Works. I have tried to find information about this modification, but all I found was that it was reviewed in The Absolute Sound 94,111. Below is the description JEHAN TITUS made in the email.

Email: Hi Wonderful page. I was hoping someone would post something about the ESL63 soon. There is another commercially available mod called the CROSBY MOD from Crosby Audio Works in California. It was extensively reviewed in T.A.S. several years ago and it works. It is basically all the above mods except that they stiffen the frame and use a thinner dustcover, the parts are upgraded ( MIT internal cabling etc). But no protection circuitry bypass or mains filtering is attempted. -hope this will add to your list of tweaks! I have had this done to mine and the improvement from standard is quite noticeable. ps; the biggest improvement in the treble however was when I switched from using a pramp with an alps pot (ARC) to a stepped attenuator try it! thanks again, JEHAN TITUS
 

I searched the net a little more and came up with this link to CAW.

Crosby Audio Works

Power Line Filters.
The Quad ESL-63 benefits from the use of power line filters. I got better imaging when I applied this tweak.

Internal Cabling.
The cables used internally in the speakers are not of the highest quality, so replacing these with higher quality cables might be worth trying. I have not tried it yet.

Improved High Voltage Stability.
This is another tweak I have not tried yet. I read on SDS Labs homepage about someone (Anders Enqvist) who has tried putting a large capacitor on the high voltage supply. This is an interesting tweak, as the power line filtering gives good results.

Disable Overload Protection.
The overload protection system may influence the sound, so it might give some positive results to disable it. I have not tried, as I think it is a good insurance to have it. My speakers are around 15 years and in perfect shape, and will possibly last for another 15 years or more without service.

Improved Mechanical Stability.
Improving the mechanical stability, and applying damping material is never wrong. I have not made this yet.

Positioning of the Speakers and Damping the Reflections.
The positioning of the Quad's is of great importance even if it is not really a tweak. Most important is not to place them too close to the wall behind and to keep the distance from the listening position equal to both speakers. After this it is really a trial and error process, make small adjustments to the position and try different angles. Some basic knowledge about acoustics is certainly not in the way when you are trying different positions. Thanks to the dipole design the Quad ESL-63 may be placed quite close to the sidewalls without too much reflections. Damping on the floor and on the walls will reduce reflected sound and enhance the musical experience. Use mirrors to place the damping material in the right position, it is easy, just place a mirror on the floor and when you see the center of the Quad ESL-63 in the mirror from the listening position you have found the correct position for the damping material. The same method can of course be used to find the position for damping material on the walls. The ceiling reflex is usually of no importance with the ESL-63's. On the floor it is most convenient to use thick carpets, and on the walls curtains with some damping material applied behind them should not be too ugly. Some damping behind the ESL-63's may also be applied to reduce the reverberation in the room, and if the speakers are positioned close to the wall behind this is of great importance. A clever way to apply damping behind the speakers is to build a free standing frame and apply some damping material (for example Acousto Q) hanging loose within the frame, this will make it easy to remove the damping material when the speakers are not used and it is also easy to alter the position and angle of the frame. If you apply damping material you will get more detail and better soundstage focus from your beloved Quad ESL-63's (or any other speaker).


Andy's Tweaks
From Andy I received an email with some tweaks.
Andy's own words:
I have tried all of these and they are all OK

Remove the protective device that prevents overvoltage, only if you have a low power amp of course. Replace the 3R3 resistors on the pcb by non inductive ones from M Percy USA. Build them on a separate mini pcb preferably PTFE and at the same time replace the 1mfd mylar cap by a MIT PP & Tin. However well run in the cap as otherwise it seems somehow to set off the protection circuit.
Replace the EHT supply by one using GP02-03 diodes (you will need less of these than the original circuit) and use PP caps in the multiplier chain. Not a huge improvement this but worthwhile.
Add a few single hole ferrite beads on the wire from the EHT unit to the diaphram. Type of ferrite is important. 6 hole beads seem to close down the sound we tried Phillips 3S1 type and the result is incredible.
Replace the connectors by decent binding posts.
Play without the wooden top and with the cloth removed.

My comments (MT):

Be careful. I have already mentioned this tweak above and I will try it sometime in the future. Better resistors may help.
I don't understand what Andy means when he states that you will need less of these, it is a voltage doubler and if one diode is removed the voltage will only be 2.75kV. PP caps may be worth trying but note the very high voltage (5.25kV), and maybe a different type of diode will make some difference, I don't know.
Normally the current flow in the EHT supply is extremely low but could be worth trying anyway.
Definitely.
The cloth definitely adds something (or rather subtracts) and if you are happy with the looks, do play without them. The top cover can create an unwanted reflection and I have heard people mentioning this tweak, stating that the stage height is improved. Try it yourself and if you are happy with it, please let me know.

Gradient SW-63

Replacing Gradient High Pass Filter.
This tweak may not seem to be such a major one as it is, but you will not experience the true magic possible with this loudspeaker combination if you don't try it.

When I measured the speakers I noticed that the ESL-63 and SW-63 was nearly 90º out of phase at 115 Hz, no matter how they were connected. This was not a surprise really, as the the low-pass filter for the SW-63 is third-order, and the high-pass filter is second-order. To get them in phase the high-pass filter should be first-order or third-order. I measured a little and placed a high quality polypropylen capacitor in series with the main amplifier input, giving a first-order 115 Hz high-pass filter. The result was AMAZING! Suddenly it was very easy to find correct phase and adjust the level. The bass was a lot more homogenic and full of nuances and flavors, I actually lowered the subwoofer level quite much, with the interesting result of a more firm and heavy bass. The midrange and treble is now exceptional with no coloration from the filter.

With this tweek you get correct phase at the crossover between the ESL-63 and the SW-63. The 3rd order high-cut filter for the SW-63 alters the phase 270º and the low-cut filter for the ESL-63 alters the phase 90º, which means that the speakers are 180º out of phase at the crossover (if the amplifiers does not inverse the polarity). This is easily corrected, simply by switching the connections at either the ESL-63 or the SW-63. When the tweak has been performed you will easily hear which connection is correct, if the connection is not right there will be less bass. That is because when the loudspeakers are running out of phase in the crossover region, the SW-63 will 'eat up' the bass produced by the ESL-63 and vice versa.

 
I have received an email from one of my visitors who has performed this tweak, and he seems to agree with me almost entirely about this tweak. Here is some of his comments:
The amps (Quad 606s) are phase correct at their outputs but introducing the caps inverts the phase of the outputs with respect to the subs. I reversed the ESL63 connections at both speakers to bring them back into phase with the subs. Result: The bass at the 16ohm setting is now in line with your reported experience. I have not yet tested it at 4 ohms.

You are right about the midrange and high frequencies - clearer cleaner and fuller. The total sound is improved dramatically over the gradient filter's performance.

I am enjoying the system more than ever and finally feel that I can retire from the audiophile's constant insecurity about his system and the need to fool with it.
ESTG/

More feedback on this tweak, from Graham Horwood:
I have just completed bypass of Gradient filter by adding a Hovland Musicap @ .33uf and the result is truly remarkable, thank you.

For your info which may help others, I have fitted two Musicaps inside the Gradient filter box, but not connected to the circuit board, by adding four RCA sockets ( two in / two out ) by drilling the rear plate, there is room.
I then linked the hot terminals for each pair using the Musicaps and linked the negatives with silver wire and it all fits neatly inside the Gradient filter box. This way there is no need to delve into either the pre or power amps with possible warrantee infringement. All that is then needed is an extra 2 pairs of interconnects. My pre amp incidentally has two outputs per channel so one feeds the musicap and the other the Gradient filter for the SW.

The sound I am now getting is so much sweeter and yet the bass is firmer and more rhythmic. Very close to being there which is what it is all about. I suspect I am getting two benefits.
1. better phase response as you said thus better bass
2. as my power amp has a low 5K impedance ( Linn Klouts) it is now seeing the pre amp load direct rather than through the higher output impedance (1K) of the Gradient filter and also distortion is lower hence the sweeter hi end.

Once again many thanks for your guidance and web site tweaks. They are the best improvement I have ever achieved (MT: The other tweak is the input filter network tweak).

More feedback on this tweak, this time from Renze de Vries in Holland who has just bought a pair of SW-63s:
Bought the beasts! So there I was, several Dutch guilders lighter and full of expectations. Hooked them up standard and what do you think: rubbish. Very slow and ill defined bass, very coloured and booming. After a week I considered selling them again: better thin bass than irritating bass. But of course, I didn't follow your advice yet, so I waited. Yesterday I bypassed the high filter for the ESL-filter with three parallel 4,7 nF capacitors (14,1 nF, just what I needed for my 100k impedance amp) and from that moment you know the story. I mean, it's exactly as you've written. Very easy to find the right phase between the two speakers (couldn't hear it at all before!), very transparent soundstage, very fast and tuneful bass (indeed I lowered the level considerably too), spectacular refined mid and treble and so on, and so on.
It makes me wonder: why could one build a filter like that, with built-in problem, selling it for a considerable amount of money, that makes it impossible to hear the true qualities of this speaker combination? It isn't good at all. I's a shame, really.

Altering Gradient Low Bass Filter Characteristics.
I received the Gradient filter schematics on email from a web friend and I have made some simulations on the SW part. The simulations looks exactly as the measured frequency response I used for making this tweak, it was a trial and error process with altered capacitor values and measuring of the response. Now I know how it works, the subwoofer filter is a three stage construction and I will call the first stage HP, the second stage LP1 and the third stage LP2.



The modification I have made is to alter the value of one of the two (per channel) 100nF capacitors to 147nF. This is an easy operation, just solder 47nF capacitors in parallel with the 100nF on the solder side.



The response above is the original Gradient SW filter response.




The response above is the modified Gradient SW filter response.




The response above is the original Gradient SW HP filter response.




The response above is the original Gradient SW HP filter response.




The response above is the improved modified Gradient SW HP filter response.




The response above is the original Gradient SW LP1 filter response.




The response above is the original Gradient SW LP2 filter response.




The response above is the original Gradient SW filter response compensated for the 6dB slope due to the open baffle construction.




The response above is the modified Gradient SW filter response compensated for the 6dB slope due to the open baffle construction. As you can see the 6dB slope needed does not function perfectly. I tried altering the values and below is an improved filter modification with its 6dB slope compensated response, looks quite good.




As in the previous modification this is an easy operation, just solder a 47nF capacitor in parallel with the 100nF capacitor on the solder side and put a 47k resistance in parallel with the 22k resistance.


Above you can see the Gradient filter simulated (left graphs), note the bad phase correlation that I have described on the 'Quad Tweaks' page. My tweaked Gradient filter is simulated in the graphs beside (right) and the phase difference is now approximately 180 degrees. You just have to switch the polarity on either the subwoofers or the ESL-63s to have correct phase. The damping of lower frequencies in my tweaked version is not too impressing, still this version sound much better and the small amount of damping also makes the ESL-63s increase their performance a lot.

Result:
I have altered the filter as described above and even if nothing is perfect, this is the closest approximation I have yet heard. The bass is now awesome, giving a flat response down to almost 20Hz. There is nothing flabby or indistinct about the bass, it is in fact articulate and firm. The soundstage is improved and so is the treble, this is a rather funny effect of an improved bass and I have heard this phenomena before. It is probably the transients that improve, remember that a perfect pulse consists of ALL frequencies. I am very happy with the result and this filter tweak is highly recommended. It is easy to perform and equally easy to 'de-perform'. The maximum SPL is decreased a little, but I think the improvement in quality makes up for the loss in quantity.

Feedback on this tweak, from Renze de Vries in Holland who has just bought a pair of SW-63s and decided to try altering the filter (some comments here are about the filter bypass tweak):
Went down to 20 Hz and yes, this is absolutely magnificent. Soundstage increased a little, although I think my room and the way the speakers are placed (on the short side) puts a limit on this, visually and, along with it, aurally.
But the bass. Very uncoloured, very dry. It makes the impression of less bass, because if there's no information down there the speaker will let you know. I played a very close miked contrabass solo and thought I missed something. Till Gary Peacock, that's the one, touched his lowest string. I never knew this solo was played on three strings most of the time, but the ESL-SW63 combination informed me about it without a doubt.
Resolution in mid and treble increased too. As capacitors need time to settle down over time,  I had one very bad day with the new capacitors in the filter bypass: very thin, shrill, unpleasant, distorted sound. And the next day: alright. The speaker informed me about new components in the signal path like never before.
The background: very quiet, very 'black'. Transients are suddenly there, with all the speed I could wish for. And talking about speed, this speaker is fast as a race car now. I played with the pause button for a while, just for the fun of it, because when I let the music in again, it started so utterly accurate and fast. Acceleration, I never knew I was short on this.
So thanks for the good advice, this is landmark bass performance indeed. I think I am the happy owner of the best  loudspeaker money can buy.
... and very neighbour friendly too. It's almost  03.00 am now and I'm playing with volume down and everything's there, I mean here. Even church organ, with the neighbours fast asleep. How little do they know, I mean hear.  

Altering the Input Impedance of the Gradient Filter.
As I use a tube preamplifier and as such items does not like low impedances, I have altered the input impedance of the Gradient filter from 22k to 500k. The improvement was quite substantial in treble quality and soundstaging. I recommend this tweak if you are using either a passive preamplifier or a tube preamplifier. If you use a solid state preamplifier it will probably not make much difference.
Having been listening to my system after this alteration for a week I am surprised of the magnitude of improvement. I knew that it would make a difference, but not that it would be as large at it has turned out to be. The whole scenery opens up and there are a lot of details I have never heard before. Voices appear even more natural than before and everything is crystal clear.
The circuit used by Jadis in the preamplifier output section (ECC82 common cathode stage with no by-pass capacitor) has an output impedance of several kilo-ohms. When loaded with low impedance the tube loadline is altered considerably and this effects the sound.

Quad ESL-57

Stacking Quad ESL.
I once attended a demonstration of the famous Mark Levinson stacked Quad system, using stacked Quad's (of course) with a Decca Kelly Ribbon in the middle to improve vertical radiation and two VERY large subwoofer systems. This system was one of the best I have heard, although the subwoofers were far from free of box sound.

Double Quad

I am no expert on the original Quad ESL, so I will just put a link to experts here.
The Quad ESL (Gary Jacobsons website)
Sheldon's Audio Designs (SDS Labs)

There is also a lot of interesting material on the ESL Circuit site.

The ESL Circuit