Nice speaker!👍
At first glance, this is a nude Audiostatic.🙂
The alu bars are good mechanically, but they can make a little electric shock, if they aren't earthed, and they give parasitic capacitance.
The toroidal transformers have relatively little stray inductance, but they have a lot of interlayer capacitance, this is often bigger than panel's, especially at 4 trafos. This causes early high-frequency roll-off (or peak, if the resonance isn't damped) and the resonance point is shifted into audio range. If we damp the resonance well, the roll-off begins at the mid-high boundary. I suggest the "conservative" trafos with EI, double-C or UI cores and with simple winding, but thick insulators.
At first glance, this is a nude Audiostatic.🙂
The alu bars are good mechanically, but they can make a little electric shock, if they aren't earthed, and they give parasitic capacitance.
The toroidal transformers have relatively little stray inductance, but they have a lot of interlayer capacitance, this is often bigger than panel's, especially at 4 trafos. This causes early high-frequency roll-off (or peak, if the resonance isn't damped) and the resonance point is shifted into audio range. If we damp the resonance well, the roll-off begins at the mid-high boundary. I suggest the "conservative" trafos with EI, double-C or UI cores and with simple winding, but thick insulators.
Hi,
... in addition to the above, the 2Ohm might be just too high, forming a distinct voltage divider at the upper bandwidth limit.
The higher the transformation factor of the transformers, the lower the impedance minimum at the resonance point.
Especially towards the upper end where one tries to reduce the burden on the amplifier with the serial resistor, one should take care to not use a too high value for R here.
The phase shift at and around this impedance minimum is low and the power demand isn´t that high, which makes this impedance minimum typically a load the amp can deal with comfortably.
Check the impedance plots, not just the amplitude response.
jauu
Calvin
... in addition to the above, the 2Ohm might be just too high, forming a distinct voltage divider at the upper bandwidth limit.
The higher the transformation factor of the transformers, the lower the impedance minimum at the resonance point.
Especially towards the upper end where one tries to reduce the burden on the amplifier with the serial resistor, one should take care to not use a too high value for R here.
The phase shift at and around this impedance minimum is low and the power demand isn´t that high, which makes this impedance minimum typically a load the amp can deal with comfortably.
Check the impedance plots, not just the amplitude response.
jauu
Calvin
To be honest, they are definitely NOT a copy of Audiostatic.
I must admit, they look very similar, but our construction was completely developed by us. Coincidentally we had the same thoughts when thinking about our frame of the ESL. Anyway, there are some differences. The sides for example are made of GFK not aluminium like Audiostatic does. Also we simply use a flat material in contrast to Audiostatic which are using angled or U-shaped material. I am sure our construction has the same stability and is way lighter plus it is non conductive, except the cross bars.
Anyway, we don't mind if they "look like Audiostatic". We both know what we have build. That's the only thing that counts...
Thank you for the hint with the trafos. Maybe we give them a try, but what i've seen is that a lot of people, including Calvin, are using toroids. We also already bought them, so first we want to use the toroids.
At the moment we are mounting the remainig three panels (2 for Dani, 2 for me).
We still are not sure about the bass region. I'd use a Ripol and are already in contact with Axel Ridtahler. Maybe we can find a good solution.
The idea of Calvin with the Dipol coloumn is also nice. Another solution would be a TML...
Hi,
Bolserst, Golfnut et al have discussed the parameter sets of suitable audio transformers for different panel constructions.
See for example this thread, #24 following .... or this thread about LF-capable transformers, or eventually the Step-up transformer design Thread.
I used standard off-of-the-shelf 80-150VA toroids (230V/6V for ~1:70, 230V/9V for ~1:50) for my curved non-segmented panels ... here indeed larger cores result in higher bandwidth.
With the flat segmented panels smaller cores and higher transformation factors with 4-8 trannies worked better.
Still though the LF X-over freq shouldn´t fall much below 200Hz with these kind of power trannies.
Partnering a dipole bass with the panel requires the panel to reach low enough in freq.
400-500Hz as ML did (or still does?) is only required -and imho sensible- for small sized panels .... as it gives up alot on that proverbial ESL sound fascination.
With sufficiently sized panels I regard a X-over between 150-250Hz as almost optimal.
A Ripole wouldn´t reach that far up ... so I copied the conceptual idea from ML´s Statement ii to use a filler dipole tower and eventually a dedicated subwoofer below ~50Hz.
It certainly increases the effort considerably, but bass-to-midrange integration is not only as seamless as with an FR-ESL, it also comes with an improved linearity in amplitude response -missing on that anemic touch due to the typical 100-200Hz dip.
Besides, the dynamic range then rivals that of horn systems and the attac/speed even surpasses.
I also combined Ripole subs up to 80-100Hz with my flat segmented panels which could be easily driven down to 100Hz.
This already improved dynamic range and sound quality considerably, just not as extreme as with the dipole tower + sub.
jauu
Calvin
Bolserst, Golfnut et al have discussed the parameter sets of suitable audio transformers for different panel constructions.
See for example this thread, #24 following .... or this thread about LF-capable transformers, or eventually the Step-up transformer design Thread.
I used standard off-of-the-shelf 80-150VA toroids (230V/6V for ~1:70, 230V/9V for ~1:50) for my curved non-segmented panels ... here indeed larger cores result in higher bandwidth.
With the flat segmented panels smaller cores and higher transformation factors with 4-8 trannies worked better.
Still though the LF X-over freq shouldn´t fall much below 200Hz with these kind of power trannies.
Partnering a dipole bass with the panel requires the panel to reach low enough in freq.
400-500Hz as ML did (or still does?) is only required -and imho sensible- for small sized panels .... as it gives up alot on that proverbial ESL sound fascination.
With sufficiently sized panels I regard a X-over between 150-250Hz as almost optimal.
A Ripole wouldn´t reach that far up ... so I copied the conceptual idea from ML´s Statement ii to use a filler dipole tower and eventually a dedicated subwoofer below ~50Hz.
It certainly increases the effort considerably, but bass-to-midrange integration is not only as seamless as with an FR-ESL, it also comes with an improved linearity in amplitude response -missing on that anemic touch due to the typical 100-200Hz dip.
Besides, the dynamic range then rivals that of horn systems and the attac/speed even surpasses.
I also combined Ripole subs up to 80-100Hz with my flat segmented panels which could be easily driven down to 100Hz.
This already improved dynamic range and sound quality considerably, just not as extreme as with the dipole tower + sub.
jauu
Calvin
I know it, because I follow your work with sympathy. I wrote, "at first glance". The Audiostatics are beautiful speakers, isn't tragedy a little visual similarity. But the alu bars can collect the spreading charge from membrane, same as Audiostatic (I got a lot of mini shocks for my better memory 🙄 )
Congrats for your speaker, nice work!
Last edited: