Driving ESL headphones - some questions
I am pretty new in ESL designs and recently played around with some DIY-transducers.
My very first experiments were driven by the simple "step up transformer"-method
that soon blew my test amp because I wanted some more bass out of that (too small) transformer :o
Nonetheless, now I have got some spare time and so I want to improve my design by building a direct drive system.
But there's a question that I can't find a simple answer to it:
is there any difference in driving the ESL's stators rather than the diaphragm?
Most of the designs I have seen by now bias the diaphragm and drive the stators phase-inverted.
But to me both driving options seem equal :confused: Plus, driving the diaphragm rather than the stators only needs one amplifier.
So, can anybody tell me the difference or (dis-)advantages?
It has been shown that by using a very high resistance coating on the diaphgarm and driving the stators lowers the THD especially for the lower frequency's.
This is called constant charge mode.
There have been many discussions on the ups and downs of using a direct drive amplifiers in these threads.
The main and biggest issue involved is the use and safety of such high voltages as they are in the order of 10 to 20 times more than what is required to drive a smaller sized headphone unit.
All else aside, I am all for designing a DD system for ESL's as I have been pursuing this goal for the better part of the last 10 years or so.
But, DO BE VERY AWARE of the level of voltages involved, and, ESPECIALLY the HIGHER CURRENTS at those increased voltages that is Required!!
The biggest issue is finding output devices with good characteristics at audio frequency's using such high voltages.
There are some out there but it takes time to sort through them all, as I have created a few lists in these thread already.
The second issue of concern is the amount of driving voltages needed to drive such a system, as it can take as much as 10KV peak to drive a system to its fullest potential.
But a workable system can be realised using amp running on 2Kv to 3Kv power supply as well using a push pull or bridged(BTL) using a Class A SE amplifier on each stator.
Creating a amplifier that uses a bipolar power supply at these type of voltages would be quite an undertaking, but not impossible to do.
More on that later. :)
This will allow a voltage swing across the stator that is approximately double the supply's voltage at 4Kv to 6Kv peak.
This much voltage is probably ample for a large panel ESL to SPL's that would put you out of the room.
It would work also for smaller sized one as well, but at a lower SPL and but still be plenty loud enough to listen to, as it would have a lesser amount of surface area.
I have been specializing in the building of smaller desktop panels and I have gotten them to get as high +105db and more at 1 meter using such voltages.
It takes using a very high bias voltage on the diaphragm at about 7Kv to 10Kv or so to do it.
Using such a high bias voltage increases the sensitivity +6b for every doubling of the Bias voltage.
This greatly reduces the required drive current and power demand from the amplifier and only took about 3KV to 6Kv peak on the stators to realize such SPL's at 100db and above.
Stator coating materials become a big issue in the design of the panel at this point to keep it from arcing and burning up!!
Yes, I have burned a few and fire now becomes the main issue on the extreme side of things. ;)
I do believe that such a system can be realized as long as all factors are taken into consideration.
It takes less power to drive an ESL at the lower frequency's than it does to drive them at the highest frequency's.
Besides the action of Dipole Cancellation working against you, ESL's can produce bass providing that they have enough displacement as well.
But in most cases it is the cost and size of the step-up transformer design is the limiting factor on low frequency performance.
This has been discussed time and time again.
I think that the best of both worlds could be combined to produce a decent full range system with using a D-D amp for the lower frequency's and a transformer for the higher frequency's.
This is a concept that I have been thinking about in more recent years.
Much like what has already been done with commercial interfaces that use the technique of having two step-up transformers but only eliminating one with a D-D amp.
Using the D-D amp for only the lowest frequency's greatly reduces the power requirement of the D-D amplifier itself, While increasing the ESL's performance for the lower frequency's without the need of some costly and heavy iron. ;)
Just a few thoughts FWIW. ;)
P.S. Sorry for going a little OT as your questions were more geared for headphone ESL drivers, Therefore the drive requirements are not as extreme but the concept is the same for the most part.
I have posted some examples of ESL amp designs in these threads suitable for small drivers as well in the Headphones forum.
I just get very excited as I am very interested in such things!! :)
1) Distortion. ESLs made with highly conductive diaphragms have much more distortion than those made with highly resistive diaphragms. These two methods of driving ESLs are often called "Constant Voltage" and "Constant Charge" respectively.
More information on the difference between these methods here:
2) Stability. ESLs made with highly conductive diaphragms have a non-linear drive mechanism that increases the force on the diaphragm the closer it gets to either stator. This acts like a non-linear stiffness term that often results in the diaphragm collapsing and sticking to one of the stators.
Plots of the nonlinear behavior here:
3) Arcing. ESLs made with highly conductive diaphragms have the potential for high-current arcing between the diaphragm and stators. Burning small holes in the diaphragm from the little lightning bolts is the best case. In more spectacular arcing events, the diaphragm and/or stators can catch on fire. Not exactly what I want to have next to my ears.
commercial current production ESL headphones from Stax, Koss are constant charge, balanced Stator drive
the balanced drive is another advantage that gives 2x the Vswing for a given supply/output device Vrating
head-fi.org and Head-Case.org both have diy sections with ESL headphone and amp projects
I've been making many ES headphones and amplifiers in the past few years. If you are interested in making your own, here is the link
At closer look there are distinct differences, most of which have been listed in the threads above.
It might help to understand that the working principle of a symmetric constant charge (cc) ESL differs fundamentally from the constant voltage (cv), driven membrane ESL.
The differneces regard the treatment and functioning of the membrane.
With the cc ESL the two stators span up a signal dependent electrical filed. The membrane is the support for a charge, on which the field can at upon.
One could say that the field grabs the charge and tosses it around.
The forces rise with increased field strength and amount of charge.
As the field can be made very homogenous and the mass of the charged membrane doesn´t count at audio frequencies, he electrostatic drive is perfectly linear.
The membrane is coated wth a high-ohmic compound to keep a charge on the membrane and stationary over the membrane area.
Due to this high impedance a possible arc over takes up much less energy, that may not get hot enough to melt a hole into the membrane, let alone that it could burn the whole.
The cv ESL is basically a sandwich of two single-ended ESLs where the stators are charged up with opposite polarities.
From its single ended nature follows, that the membrane must be seen as a moving stator.
It takes up an active part of the drive system.
As the signal voltage is applied to the membrane its coating must be low ohmic.
This means that signal dependent charging currents flow.
The high voltage stresses the coating and membrane contact already, but while the currents are negligibly small and quite constant with a cc ESL they are much larger and varying with an cv ESL.
This has been a source of unreliabilty and breakdown (it took Final off of the market).
In case of a arc over, the spark can devolp much more energy and heat, thereby melting larger holes in the membrane or burn it down completely.
Beeing of single ended nature the driving forces are not linear, but vary quadratically.
Sandwiched together this non-linearity cancels to a first degree, but still remains higher than with a cc ESL.
thanks for all your answers, I see there's a lot for me to learn about ;)
Great! Exactly what I wanted to know ;)
Just searched around a bit for your posts and... I'm more than very impressed of your designs :) There are some details I'll have to copy when building my "final design" :cool:
Hopefully not this time...
Now I'm going back to do some more research and maybe in some time I'll present my result :)
|All times are GMT. The time now is 08:10 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright ©1999-2015 diyAudio