Happy 4th. Sparks and Sparklers but no sparks from my ESL speakers...
Inspired by reading the recent coffee-table book on QUAD and after a year of thinking about it I figured out how to drive the QUAD 57 from my direct drive amplifer! As QUADAPHILES know that speaker has rather special requirements to both protect it from over voltage (sparking) and is somewhat amplifier fussy. The RM-10 has been popular among QUAD owners and it is widely known that I was using that speaker when I designed the RM-10. Some years later I got a pair of 63's and added some banana jacks on the back so I could direct drive them from a high voltage OTL amp. It's a very easy thing to do and I am willing to post a how to if there are interested parties who have the skills to solder and play safe around high voltages. Actually as long as any ESL is not playing music it is quite safe. The only high voltage that is present at that time is the polarizing and getting shocked by that is only an annoyance, not a danger as it is DC and of low current. It's the high voltage AC from the music that one must not come in contact with. For my pair I installed 4 jacks so I could jumper back in the standard drive electronics easily.
My speaker, the Acoustats and most DIY ESLs are simple 3 wire systems. Those being the front, back and diaphragm wires. There are dozens of articles on how to build your own panels. In fact ESL panels are the easiest speaker to build needing only moderate woodworking skills and easily obtainable materials. By contrast, a magnetic planar or cone driver is almost impossible to build in one's garage shop. The difficulty the DIYer encounters is the drive electronics. The often given solution, reverse connected output transformers being the most popular, is just horrible. Those transformers never have high enough ratio or peak voltage rating. Usually the DIYer finds he has to get a several hundred watt amplifier to get the voltage and then the transformer can't take it. The best solution is the direct-drive amplifier, but those are not for the unskilled to design or make, evidenced that only Acoustat and Beveridge made them and Acoustat quickly converted to a poor sounding substitute transformer box (I worked on one, it's really horrible and doesn't sound nearly as good as a direct drive amp). Acoustat quit making their amp because they wanted wide appeal and their amp was giving them a lot of trouble. Beveridge's amp was better but sat under the speaker making tube replacement difficult. Both were build on circuit boards, which was a bad idea. Both were built at the lowest possible cost as both designers were more concerned with the speaker, not the amplifier. They both had solid state front ends, the Acoustat having an OP AMP. I on the other hand have a lot of concern about amplifier sound and reliability. My amplifier is all tube and has balanced and unbalanced inputs.
The world of ESLs is largely divided into two categories. Here I am just discussing the ESL itself, not considering any cone drivers that may be employed.
1. Full range drivers that have large area and that area produces all frequencies. These tend to be very directional (the Innersound is a prime example) unless they are curved (Martin Logan) or approximated into a curve (Acoustat, Sound Labs). Most DIY designs are full range and either curved or flat, the latter being much easier to build. All are less efficient than multi way as they have high capacitance and will sound very, very bright unless EQ is applied. They all have EQ and this is best applied in DIY situation by a separate graphic EQ which can also correct for the room. This what my customer with the Acoustat Eight (8 panel speaker) does with great success. At 2 KHZ, the response is pulled down 10 dB and the curve is a gentle smile on both ends. This results in flat response in the room. Beveridge and Acoustat both employed EQ in their amps and Beveridge was very particular about his. Many DIY builders do nothing about EQ and have some very strange sounding speakers that could be easily corrected with a descent graphic EQ that can be had for $500. Yamaha and DBX make these both having good sound, balanced XLR output and can drive long lines to the amps that typically sit close to the speakers.
2. This second class of ESL speakers is multi way using panels of different dimensions and often different voltages to obtain line source dispersion. By adjusting the panel size and shape flat response can be obtained without the need for EQ. All the QUAD speakers do this having much larger area for the bass than the treble. The 57 is a 3 way line source where the 63 is a two way point source. I prefer the 57.
The 63 is an easy drive as it is low capacitance (much lower than the Beveridge and Acoustat) and has a single drive voltage. Only 3 wires are needed from the amp to the speaker and test lead wire is quite suitable and safe. The 57 is quite an different issue as it uses two drive and two polarizing voltages which are 1500V for the midrange/tweeter and 6,000 V for the woofer. These are provided in the speaker by the polarizing supply and the audio input (step-up) transformer which is quite complex and contains some crossover components which often go bad. The polarizing supply almost always needs to be replaced if it is original. With my direct drive amp these components can be removed (and sold for good money) as all you need are working panels.
My direct drive system provides both polarizing and audio voltage that will never overdrive the speaker. A common problem with the 57 is that it was intended to be used with the QUAD 22 amplifier which was 15 watts at 16 ohms which turns out to be 30 watts at 8 ohms or a peak of 30 volts which is just what the RM-10 provides, making it the safe choice for standard 57s. People get in trouble with bigger amps as they have higher output voltages. Clamping devices have been created for the 57 but many think they impair the sound of this very clean speaker.
Here is the run-down on what my direct drive amp can do. It can be made at any output voltage and I have a high and low current version because some ESLs draw very high currents due to high capacitance and some don't. This is just the same as the fact that speakers can be 2,4,8,16 ohms and anywhere in between. ESL speakers can be low capacitance, mine are just 100 pf. Beveridge model 2's are 4500 pF. That's a 45 to 1 difference, a much larger range than we see in cone speakers. In addition being capacitive the impedance varies inversely with frequency being lower at higher frequencies. When Beveridge went to transformer drive the result was a speaker that went from 100 ohms at 100 HZ to 1 ohm at 16 KHz. We had to find amplifiers that would deliver over 40 amps of current. These speakers were not suitable for most conventional amps tube or transistor. Roger Sanders makes a solid state mono amp that delivers 2000 watts at a cost of $8000 per pair.
Acoustats are about 300 pF per panel but the good ones have at least 4 panels per side making them 1200 pF. That still puts them in the high range. However if you want to make a small rather directional ESL you can lower that capacitance but a woofer will be required. If there is interest I will direct DIYers to the construction designs I find most appropriate or provide some easy to build suggestions. Acoustat transducers appear to be somewhat available from people who have scrapped the speakers. The panels are almost always good. The problem has always been driving them.
The Quad 57s are only 200 pF and the 63's are low also. In general the ESLs that use full range panels have the highest capacitance and are therefore least efficient. The multi-way ESLs (that means mulit-way in the electrostatics themselves) tend to be low capacitance and much more efficient. The extremes are the QUAD 57 needing 15 watts of drive and the Beveridge needing 1,500. Trumpet music is most demanding. Being a little in disbelief about the 1500 watt number I measured a half amp at 3000 volts. No wonder these speakers can't be driven any other way. The QUAD on the other hand needs the voltage of a 15 watt amplifier (at 16 ohms) but from a 4 ohm tap. In my experience that although the QUAD 22 amp was made to drive the 57 speaker, it was rather rolled off at the top due to the fact that the speaker impedance did fall to about 4 ohms. Not everything QUAD did was exactly as they said, contrary to the rather perfectionist philosophy proposed in the QUAD book.
One nice thing about my direct drive system is that by changing just 2 capacitors in my crossover one can set the brightness of the speaker to his desires either more or less than the standard by as much as 12 dB in either direction. Far more than one would need.
Please respond to the poll so I can get an idea of your interest and have a happy 4th.
Poll