QUAD 63 (and later) Delay Line Inductors

Please keep in mind that the human body has nonzero dimensions and therefore has self-capacitance. Estimating that to be 100 pF, the value used for human body model ESD tests and approximately the self-capacitance of a sphere of 90 cm radius, you could already get a fairly unpleasant shock from touching one stator while playing loud music with a large treble content, even when you are otherwise perfectly insulated.

Example: suppose you are playing the record Deutsche Marschmuzik: Einzug der Gladiatoren, which has a peak rate of change of about 0.05 V/us per volt of peak output voltage (about 8 kHz required power bandwidth) after RIAA correction according to an old article of Otala's research group, at a level of 4 kV peak. The centre is earthed, so you have 2 kV peak per stator.

0.05 V/us/V times 2 kV times 100 pF = 10 mA peak current into your body

Needless to say, the current can go much higher if the amplifier should oscillate or if your supposedly insulating wooden floor is wet or if you accidentally touch something earthed with your other hand.
  • Like
Reactions: 2 users
The real risk is IMO the touch of front and rear stator simultaneously (especially the middle part of ESL63, because they driven by low impedance). At one handed touch the body capacitance's load current can be high when the slew rate is high (transients or high frequency at full output). The high frequency current has reduced ability of excitation to depolarise cells, as ions only vibrate, they cannot penetrate enough quantity via cell's membrane, cannot generate action potential. As second factor is the skin effect. However, the local thermal effect can burn.
I warn people that the HV will shock the cookies out of you but is current limited and you'd have to really work at it to die from it (think electric fence), but the audio signal through the transformers CAN kill you so don't mess with the speaker when it's playing unless you know exactly what you are doing.



  • 1715718704256.png
    34.3 KB · Views: 25
Reviving again with a question on the HV bias.

First the situation with the step-up xformer and my direct drive amp. Output on the stators with no signal is 0VDC, and the bias voltage on the diaphragm is ~+5kVDC give or take, or who you ask.

The new direct drive amp I am working on for double the output voltage will most probably have a no-signal output to the stators of 2kVDC.
Don't ask - all will be revealed in the fullness of time :cool:
Can I change the bias to -3kVDC to still have the same sensitivity etc as in the old situation?

I think that you are clever boy... Let me guess, you will bias the membrane to -3kV and let both stators rest at +2kV. This will give you a 5kV bias field like the original speaker. Then you will modulate both the stators as usual in differential mode, but you will also modulate the membrane as Harold Beveridge does in my 2SW2 speaker? Then you will increase the output and keep the power supply voltage down. If you will do this you need conductive membranes. And if so... you will introduce distortion because of lateral current in the foil itself as you know. This is mitigated in the Beveridge design by introducing a high resistance in the stators instead of the membrane. This could be done by new stator design of the quads of course... is that your idea as well?
Listening to the video about the X models of the Quad ESL there is a question bubbling up in the back of my head.

Are they now using the Beveridge methode and driving the diaphragm as well as the stators in these new X models?
That would explain the claim of changed transformers and the text of leading sound to the diaphragm.
That would also explain the claim for higher efficiency. That extra SPL is in that case most likely (much) higher distortion levels.
Any thoughts?
As far as I understood from what Google did when translating from English to Ukrainian in the form of subtitles, this sir said that they gave more rigidity to the design as a whole, which of course affected the reduction of distortion and increased efficiency, because we know that if the stator goes towards the diaphragm, it not only increases distortion, but also reduces efficiency.

He went on to say that they have redesigned the step-up transformer responsible for the high frequencies, which is quite positive.

He also mentioned in passing that the reliability of the membranes has improved, which is also expected, if you remember that they coated the ECL 63 with Graphite 33, which then slid off the membranes, this is also very good, it is enough to remember that Stax introduced a conductive coating on the membranes at the stage of making a solution for mylar, so the resistance of the membrane coating remained for the life of the film.
And then he said about the black color, that it does not distract from the full immersion in the musical material:D.
AX tech editor
Joined 2002
Paid Member
Misspoke, very unlikely in my book. Marketing .... in that case.
As mentioned before on this forum some people claim they play much louder. This claim is not backed up by any info on the Quad website so far.
On the website there is only "Coming soon"
If you look at his body language and the way he has to 'remember' the script, my money is on misspoke.
If indeed they went 'the Beveridge way' they would surely make a lot of noise about such a spectacular innovation! ;-)
But we'll probably never know for sure.
Unless you buy a pair @wout31!

The transformer can be made sectioned so that the internal inductance and stray capacitance is improved to extend the high frequency. Good.
Reliability of the frames is no rocket science really... Better glue, better or automated dispensing, maybe they also increase membrane thickness from 3 to 4,8um (I have tried that with success) and maybe another thinner coating to reduce moving mass.
Maybe they exchanged the electrolytic (finally) with polyprop?
But just let this sink in to your mind... it came out 1981 and it is still the same construction. Still it has some advantages that is not seen anywhere else. You can play a square wave and monitor it with a scope and during the reproduction you can move the microphone and still have a nice square wave. Hands down to the genius Peter Walker!! Bigges news was the ambient light feature and stealth black. Need to rebuild my set!
  • Like
Reactions: 1 user
We've been through three major (and one minor) revisions of the modern quads. And nothing substantive has changed electrically or with the panels. I'd be shocked if this is anything more than marketing babble. So little has been re-designed electrically that when they went from the three PCB design of the ESL63 to a single PCB design of the 988, the circuit layout is just the three boards stuck together with no optimization. That has largely continued to this day.

Remember the delay line is a nearly pure resistive load on the step-up transformers so things like leakage inductance don't effect the high frequencies like it would on a standard single or segmented ESL. That's one reason if you are thinking of trying your hand at a scratch ESL built, the quad input transformers aren't the best thing to start with.

We will find out in time, but the only thing that I'd bet changed significantly is they added yet more lights to the base...

  • Like
Reactions: 2 users
AX tech editor
Joined 2002
Paid Member
Somewhat off-topic if I may.

I'm working on a newer version of my direct drive amp.
The idea this time is a pair of single-ended outputs, each formed by a HV tube loaded by an inductor to double the available swing for a given supply voltage.
For the inductive load I am looking at the secondary of a standard ESL63 step-up xformer.
An issue here is that the secondary will carry a DC bias of about 20mA DC.
Normally, in a single ended output transformer you'd put a air gap in the core to avoid saturation due to DC bias.
That is no option for a QUAD step-up transformer.

Any idea whether the DC bias will be a problem, and how I could test it?

You can put a good CCS to primary (I=20mA*step up ratio). You can measure the decrease of inductivity (at secondary). You can detect the shifted magnetic bias of core too with another trafo: The another, non DC biased one will be the source of high voltage, low frequency signal, which non distorted. If you connect the secondaries of transformers via resistor, you can follow at resistor with scope the magnetizing current and its asymmetry. This asymmetry is good indicator of bias problem. Of course, the decreased voltage-tolerance is detectable too.

Why not make a symmetrical dual choke to the nanocrystalline or amorphous double-C cores?