Here's a schematic of the quad esl-57:
http://www.quadesl.com/pdf/quad_schematic.pdf
Here's some more info:
http://www.flashbacksales.co.uk/acatalog/manualpics/quadman.html
Hope it can help you in some way, I just happened to remember these from my earlier inquiries....
http://www.quadesl.com/pdf/quad_schematic.pdf
Here's some more info:
http://www.flashbacksales.co.uk/acatalog/manualpics/quadman.html
Hope it can help you in some way, I just happened to remember these from my earlier inquiries....
ESLs are funny.
They work out to be more or less bandpass filters of a sort.
You can get high sensitivity/output but narrow bandwidth (like a high Q filter), or you can get wide(r) bandwidth with less sensitivity/output (like a low Q filter).
The way that many designers solved the problem was to break up the frequency range and drive each section with an optimized signal = transformer with the right turns ratio for the cell/freq range.
If you look at the Quad schematic, and many others, there are ways of trying to do this bit using all one single "transformer" wound on a single core...
For DIYers it makes sense to also consider the Acoustat method, which used two separate transformers, driving ONE cell to accomplish the same thing!
Other than that, you have to apply EQ in the normal sense, or match for wide bandwidth, which reduces the effective sensitivity of the speaker... and/or like the MLII for example gives you a screamingly low Z to drive...
Typically direct drive circuits will need some EQ in the electronics to ge a flat response... fyi...
_-_-bear
They work out to be more or less bandpass filters of a sort.
You can get high sensitivity/output but narrow bandwidth (like a high Q filter), or you can get wide(r) bandwidth with less sensitivity/output (like a low Q filter).
The way that many designers solved the problem was to break up the frequency range and drive each section with an optimized signal = transformer with the right turns ratio for the cell/freq range.
If you look at the Quad schematic, and many others, there are ways of trying to do this bit using all one single "transformer" wound on a single core...
For DIYers it makes sense to also consider the Acoustat method, which used two separate transformers, driving ONE cell to accomplish the same thing!
Other than that, you have to apply EQ in the normal sense, or match for wide bandwidth, which reduces the effective sensitivity of the speaker... and/or like the MLII for example gives you a screamingly low Z to drive...
Typically direct drive circuits will need some EQ in the electronics to ge a flat response... fyi...
_-_-bear
There's really two "eq's" here. There's the impedance EQ that some people add which keeps the panel from looking like a dead short at high frequencies. I use tube amps, and they don't care about such things. The big EQ that you need to worry about is pressure feedback. This is the loss of output (6 dB/octave) when wavelengths get longer than the shortest ESL dimension will make an un-eq'ed panel sound like a tweeter. My 48" x 20" panels start to roll off below about 700 Hz. I use a shelving EQ in my active crossover to compensate for this. So my gain below 700ish Hz is increasing at 6 dB per octave. This makes the panel exssentially flat at the expense of increasing excursion as frequency drops. I cross over to a cone at 130 Hz, so it's not a big deal.
But no matter what you do with an ESL, short of putting it in a GIANT box to capture the back wave, will have to cope with the pressure feedback. Segmented designs, bring more panel online as frequency drops. when done correctly, this forms it's own shelving EQ with a 6 dB per octave rise to counteract the fall. but you can only add so much panel, then the whole thing starts to roll off.
That's the vexing problem with a dipole, maggie, quad, and even the one you make.
Sheldon
But no matter what you do with an ESL, short of putting it in a GIANT box to capture the back wave, will have to cope with the pressure feedback. Segmented designs, bring more panel online as frequency drops. when done correctly, this forms it's own shelving EQ with a 6 dB per octave rise to counteract the fall. but you can only add so much panel, then the whole thing starts to roll off.
That's the vexing problem with a dipole, maggie, quad, and even the one you make.
Sheldon
maudio said:
Good point, I meant to talk about that. I tend to run my ESL’s with cone woofers, so my crossover point is higher than that natural resonant point of the panel (as built with point supports). But your point is well taken, and possibly could be harnessed to compensate for the pressure feedback as well.
A “U” shaped panel or baffle would still suffer from pressure feedback. The only way to get around it is to put the ESL in a box, but the Vas is huge, incomprehensibly huge, so that’s not really feasible unless you are cutting holes in your walls and using a second room to capture the backwave(s). You can change the pressure crossover point with “wings” like in the “U shaped concept”, the longer the path around the panel, the lower that feedback crossover point starts.
Sheldon
The resonance peak
that's not the way to go if 'best possible quality' is the goal. Allowing a peak in the lower freq range means a Q larger than the desired 0.5..1, probably even somewhere as high as 5-10
Not a good thing for clean bass reproduction. Unfortunately a lot of commercial esl's do so, could be a good explanation for the bad reputation esl's have in the lower frequency range
I prefer to cross the esl to a low Q cone speaker (in dipole) at least one octave above resonance and place a notch on it for good measure Does require some substantial electronics though
that's not the way to go if 'best possible quality' is the goal. Allowing a peak in the lower freq range means a Q larger than the desired 0.5..1, probably even somewhere as high as 5-10
Not a good thing for clean bass reproduction. Unfortunately a lot of commercial esl's do so, could be a good explanation for the bad reputation esl's have in the lower frequency range I prefer to cross the esl to a low Q cone speaker (in dipole) at least one octave above resonance and place a notch on it for good measure
Does require some substantial electronics though
maudio said:The resonance peak
that's not the way to go if 'best possible quality' is the goal. Allowing a peak in the lower freq range means a Q larger than the desired 0.5..1, probably even somewhere as high as 5-10
I prefer to cross the esl to a low Q cone speaker (in dipole) at least one octave above resonance and place a notch on it for good measure
I was thinking that a mesh behind the rear panel much like the Quad ESL63 would reduce the Q of the resonance, then it could be used for bass augmentation. Your point is well taken though, ESLs do not do bass well, so why fight it? I cross over to a TL woofer at 130 Hz, and it works well.
My resonance is about 2 octaves below my crossover point, and I’m crossing 4th order electrical, so I never added a notch to get rid of it, but it may be a good idea to try. The problem with my 40 Hz resonance and trying a notch is that area is so polluted with room modes, I may never know if the notch did much good.
Sheldon
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.