Although the patent is irrelevant to this topic...the patent on this design expired in 2002.
Is this step up transformer design a good design that effectively and practically allows for full range frequency response to be achieved by a large electrostatic panel without extraordinary stress to the amplifier?
https://www.dropbox.com/s/99njq7hohpkcvgh/OldPattentFull%20Frequency%20Response.pdf
Is this step up transformer design a good design that effectively and practically allows for full range frequency response to be achieved by a large electrostatic panel without extraordinary stress to the amplifier?
https://www.dropbox.com/s/99njq7hohpkcvgh/OldPattentFull%20Frequency%20Response.pdf
Hi,
due to the falling impedance of ESL with rising frequency -factor of 1000:1 over 20Hz-20kHz- the impedance seen by the amplifier will be non-constant with quite wide variations.
The impedance response of the panel leads to a falling voltage and rising current requirement of the panel.
The two transformer circuit is a try to better match the panels impedance response to the amplifiers requirements.
It means less stress on the amplifier.
At the same it is a form of equalization to shape the amplitude response.
Similar techniques have been to use a transformer with taps -see the Quad 57- or a Autoformer ahead of two identical transformers -see Audiostatic.
The Strickland patent has imho the advantage to allow for different transformer styles at the same.
Say a simpler to build, cheaper EI core for the Bass (with high U), where higher losses and stray inductance (lower bandwidth) may be rather positive, and a lowloss low stray inductance wide bandwidth transformer with considerably lower U for the midhighs.
btw. these techniques also mean, that the term Fullrange-ESL doesn't apply any more.
Analogous to dynamic speakers, such designs should correctly be named 2-way ESLs, as the multiple transformers and associated circuits are nothing else but kinds of Xovers.
In case of a Quad 57, or a ML CLX it becomes even more obvious as those designs even use two different panel drivers
jauu
Calvin
due to the falling impedance of ESL with rising frequency -factor of 1000:1 over 20Hz-20kHz- the impedance seen by the amplifier will be non-constant with quite wide variations.
The impedance response of the panel leads to a falling voltage and rising current requirement of the panel.
The two transformer circuit is a try to better match the panels impedance response to the amplifiers requirements.
It means less stress on the amplifier.
At the same it is a form of equalization to shape the amplitude response.
Similar techniques have been to use a transformer with taps -see the Quad 57- or a Autoformer ahead of two identical transformers -see Audiostatic.
The Strickland patent has imho the advantage to allow for different transformer styles at the same.
Say a simpler to build, cheaper EI core for the Bass (with high U), where higher losses and stray inductance (lower bandwidth) may be rather positive, and a lowloss low stray inductance wide bandwidth transformer with considerably lower U for the midhighs.
btw. these techniques also mean, that the term Fullrange-ESL doesn't apply any more.
Analogous to dynamic speakers, such designs should correctly be named 2-way ESLs, as the multiple transformers and associated circuits are nothing else but kinds of Xovers.
In case of a Quad 57, or a ML CLX it becomes even more obvious as those designs even use two different panel drivers
jauu
Calvin
Last edited:
but isn't true that the these dual transformer designs can be used on a panel that only has one wire for the front stator and one wire for the rear stator, meaning that the combined output of both transformers feeds the entire panel instead of one transformer for one section of the panel and the second transformer for the other section of the panel.
You would try to find an Acoustat interface.
Typically they can be found ebay every once in a while.
Just Google "acoustat mk 121"
This is what it looks like,
https://www.google.com/search?q=aco...74CYCA&ved=0CCkQsAQ&biw=1280&bih=865&dpr=1.25
Else you would have to custom make them.
jer
Typically they can be found ebay every once in a while.
Just Google "acoustat mk 121"
This is what it looks like,
https://www.google.com/search?q=aco...74CYCA&ved=0CCkQsAQ&biw=1280&bih=865&dpr=1.25
Else you would have to custom make them.
jer
Last edited:
Hi,
if one hasn´t tried, a true recommendation can´t be given.
It needs thorough testing which trannies and which combination of transformation factors could be successfull.
As the MLs are non segmented panels, the midhigh transformer should be a low stray-inductance type, for example a toroid.
I don´t know the construction of the Acoustat mid-highs transformer, but as Acoustat used segmented panels the transformers will probabely not be optimized for low stray-inductance as the requirements in this regard are much lower for segmented panels.
Beeing a rather large sized panel with a capacitance in the 1-2nF range a transformation factor 1/30 to 1/50 should suffice.
A pair of 12V/230V or 6V/115V power toroids (50VA-80VA) might be a good start.
As lows-transformer EI-core power trannies may suffice as long as their stray inductance related resonance stays well above the Xover frequency.
I´d start with a multiples of EI-cores with 1/50 to 1/70 transformation factor.
This could be 2x 6-9V/230V or 4x 6-9V/115V.
Choose the wattage so that the pimary inductance is high enough for the specced 150Hz bandwidth limit.
It should be noted though that the ML panels typically feature a high base resonance frequency >200Hz and are crossed over at >400Hz.
This would allow for smaller wattage transformers.
Typically the curved panels exhibit a slightly rising response up to 1-4kHz and slightly falling off above.
This suggest to set the X-over frequency to ~800Hz-1kHz.
So far this is of course just mind-googling as it hasn´t been tested as hybrid panes usually work very well with a single transformer.
jauu
Calvin
if one hasn´t tried, a true recommendation can´t be given.
It needs thorough testing which trannies and which combination of transformation factors could be successfull.
As the MLs are non segmented panels, the midhigh transformer should be a low stray-inductance type, for example a toroid.
I don´t know the construction of the Acoustat mid-highs transformer, but as Acoustat used segmented panels the transformers will probabely not be optimized for low stray-inductance as the requirements in this regard are much lower for segmented panels.
Beeing a rather large sized panel with a capacitance in the 1-2nF range a transformation factor 1/30 to 1/50 should suffice.
A pair of 12V/230V or 6V/115V power toroids (50VA-80VA) might be a good start.
As lows-transformer EI-core power trannies may suffice as long as their stray inductance related resonance stays well above the Xover frequency.
I´d start with a multiples of EI-cores with 1/50 to 1/70 transformation factor.
This could be 2x 6-9V/230V or 4x 6-9V/115V.
Choose the wattage so that the pimary inductance is high enough for the specced 150Hz bandwidth limit.
It should be noted though that the ML panels typically feature a high base resonance frequency >200Hz and are crossed over at >400Hz.
This would allow for smaller wattage transformers.
Typically the curved panels exhibit a slightly rising response up to 1-4kHz and slightly falling off above.
This suggest to set the X-over frequency to ~800Hz-1kHz.
So far this is of course just mind-googling as it hasn´t been tested as hybrid panes usually work very well with a single transformer.
jauu
Calvin
A point of clarification here: Acoustat's MK-121 interface was NOT used to drive segmented panels. The output of the MK-121 interface is full range, driving all panel area equally.
Acoustat DID use segmented panels on the later Spectra series, but these do not use the MK-121 interface. The MK-2123 and MK-2146 interfaces used in the Spectra series, although still based on Jim Strickland's patented two-transformer design, are implemented in an entirely different way. In these interfaces, two identical transformers are used, one for each phase. Each transformer has one primary (driven in parallel but out-of-phase) and two secondaries, one for lows and one for mids/highs. The outputs of the secondaries are recombined in an RC network, much the same as the MK-121, but the output is further divided to drive the three segments of the panels. Because of the different panel areas at different frequencies, the equalization provided by the Spectra interfaces is quite different than the full-area MK-121. (For example, the highs are much 'hotter' to compensate for the much smaller high frequency radiating area.)
Therefore, using an MK-121 interface to drive a segmented speaker may not yield the desired results, due to the differing equalization requirements. This will all depend on the relative areas of the segments of your speaker. Nevertheless, an old MK-121 may be a better starting point than trying to 'roll your own'.
Perhaps it is just a matter of semantics, but I don't entirely agree with your analysis, especially as it applies to Acoustat's original full-area speakers. The traditional definition of an X-way speaker means that the electrical signal is divided X-ways by the crossover, and then fed to X individual (and typically different) drivers. While it is true that Acoustat's full-area speakers DO split the signal two ways through the step-up system, the signal is recombined before being fed full-range to an array of identical drivers, all connected in parallel. So, I think it fair to call the earlier Acoustats single-way, full-range.
For Acoustat's later Spectra series, where the signal was split two ways for the step-up system, then recombined, then split again for feeding to different panels areas, one CAN make the arguement that this is a multi-way speaker. The only difference is that instead of each driver being optimized for the intended frequency range (i.e. woofers and tweeters) all of the drivers in the Spectra series are identical.
Ultimately, since so many ESL systems are NOT capable of producing the full audio range without the aid of a conventional woofer, I think it is still fair to call ALL of the larger Acoustats 'full-range ESL's'. Certainly, that was Acoustat's intention of calling their larger models 'full-range ESL's', to distinguish them from other brands that were clearly not capable of full audio band reproduction using only an electrostatic driver(s).
Calvin: now that I have dumped on your parade (twice!) I wanted to thank you for all of your contributions to this thread. Although I AM the AcoustatAnswerMan, your technical comments typically exceed the capabilities of my rusty talents (I have been out of the audio industry since 1993). In fact, in many cases you seem to have a better technical grip on ESL's than I ever had. Despite hanging out with Jim Strickland for many years, more often than not he left my head spinning after one of his 'talks' (which could last for hours).
And thanks to EVERYONE for your contributions to this thread. My purpose in starting this thread was to help Acoustat owners keep their old speakers running, since after all, old Acoustats never really die. You cannot imagine the pride I still possess in being involved with such a wonderful product! Who knew that 20+ years after the demise of the company, I would still be assisting Acoustat owners. This is FUN and deeply rewarding for me.
Yes, dispersion is controlled (i.e. improved) by arranging the individual flat panels in an arc, with the convex side facing forward. The Spectra series' panels are arranged in a single plane, and dispersion is controlled electrically by means of the time delay introduced in the frequency-dividing network. 'Spectra' is an acronym for Symmetric Pair Electrically Curved TRAnsducer. Oddly enough, the Spectra's 'phantom' curve is convex on BOTH sides of the speaker, something that cannot be obtained by mechanical curving.
Oops, this is actually not the thread I started (I started "Acoustat Answer Man is Here"). Nevertheless, my comments and thanks still hold true.
Another important difference between the Spectra and other woofer-tweeter ESLs, is that all areas of the panel provide output at low frequencies, not just the "woofer" portion. Basically, the "woofer" portion acts like a woofer, but the "tweeter" section acts like a tweeter & a woofer. When it comes to ESLs and low frequenices...more area is better.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.