I haven't looked at it but almost certainly not.
Then again, if I can recreate the same frequency response, at the panels, under load, as the response with the transformers, it doesn't matter I believe.
Jan
Then again, if I can recreate the same frequency response, at the panels, under load, as the response with the transformers, it doesn't matter I believe.
Jan
I also think it does mean that making the DD amp frequency response the inverse of what I measure with the input transformer, is not necessarily correct.
It does not take into account that the source impedance of the xformer and the DD are different.
But I will follow Hans' suggestion to measure the acoustic response in both cases. UMIK-2 is on its way ;-)
Jan
It does not take into account that the source impedance of the xformer and the DD are different.
But I will follow Hans' suggestion to measure the acoustic response in both cases. UMIK-2 is on its way ;-)
Jan
Got some time today and setup the miniDSP with the equalizer and the roll-off above 12kHz to mimic the response at the panels in the original situation.
The speaker now sounds less thin and more balanced. Funny how much difference a few dB in the midrange can make. Donald Fagen's The New Frontier sounds really nice.
But I realized I have made an error in the sims and the measurement to determine the required drive current. I have fed one side of the panel with the output of the single channel prototype, grounding the other terminal of the panel.
That works fine, but of course the current requirement is only half of what it should be when the other panel terminal is driven in opposite phase.
Driving that opposite side of the panel model with an inverse version of the main output (simulator) shows double the current, as expected.
You live and learn, I guess.
Jan
The speaker now sounds less thin and more balanced. Funny how much difference a few dB in the midrange can make. Donald Fagen's The New Frontier sounds really nice.
But I realized I have made an error in the sims and the measurement to determine the required drive current. I have fed one side of the panel with the output of the single channel prototype, grounding the other terminal of the panel.
That works fine, but of course the current requirement is only half of what it should be when the other panel terminal is driven in opposite phase.
Driving that opposite side of the panel model with an inverse version of the main output (simulator) shows double the current, as expected.
You live and learn, I guess.
Jan
Does anyone remember the LF roll off of the ESL 63? I should have measured it when I had the setup, but maybe this is common knowledge here?
Jan
Jan
Jan,
Great to hear you are making progress and had to reduce HF as expected.
Going back in this thread you will find several examples of LF roll off measured on axis.
Hans
Great to hear you are making progress and had to reduce HF as expected.
Going back in this thread you will find several examples of LF roll off measured on axis.
Hans
Hi Jan,
I have linked to a scan I made of a test with meaurements in a local hifi magazine earlier in this thread. I also discovered I have the original test from Hifi-News 1981. It actually also have some nice measurements. I can make a scan if anyone is interested?
P.S. Given the voltages involved I actually considered if we need to have a daily "heartbeat" post from you 😀
Mogens
I have linked to a scan I made of a test with meaurements in a local hifi magazine earlier in this thread. I also discovered I have the original test from Hifi-News 1981. It actually also have some nice measurements. I can make a scan if anyone is interested?
P.S. Given the voltages involved I actually considered if we need to have a daily "heartbeat" post from you 😀
Mogens
Jan,
Be aware that JA did not simply take a simple on axis recording but averaged a number of different recordings.
So be careful in using this as a reference to tune your DSP.
Hans
Yes I saw your posts.
You know, I turn up the variac to get 'only' about +/-1500V for testing, instead of the full +/-2400V.
You really have to be careful not getting the idea that 1500V is much lower and therefor somehow not as dangerous. 😱
Jan
You know, I turn up the variac to get 'only' about +/-1500V for testing, instead of the full +/-2400V.
You really have to be careful not getting the idea that 1500V is much lower and therefor somehow not as dangerous. 😱
Jan
Here in cyan is the result of an on axis recording at 1 meter distance.
Quad 63 (and later) Delay Line Inductors
Hans
Quad 63 (and later) Delay Line Inductors
Hans
Thanks Hans, I was more concerned with the low end.
I guess I should roll of the low end to prevent excessive excursion of the foil?
Jan
I guess I should roll of the low end to prevent excessive excursion of the foil?
Jan
Last edited:
No I don't think you have to roll off anything at LF.
The ESL will do the roll off by itself.
Hans
The ESL will do the roll off by itself.
Hans
Doing so at LF may usefully increase power handling/avoidance of a strike though, overall?
Settling for 40hz, say, vs a DD amp's ability to pin the thing at ELF, simply from a recording transient or similar.
(from an amateur just watching this with interest ...)
Settling for 40hz, say, vs a DD amp's ability to pin the thing at ELF, simply from a recording transient or similar.
(from an amateur just watching this with interest ...)
For years I have used a 6 dB roll off for 63s. Always with a sub. If you look at the typical power distribution for music.( except hip hop) it's centered around 100 Hz. Removing the bass demands can significantly increase the max output. Also the panels have a resonance around 80 Hz as I remember that would be good to avoid.
That was my thinking.
@Hans: is that natural rolloff caused by the xformers and xover parts? I will not go through that of course.
Jan
Jan,
In the sim the transformers and xovers do not roll off until very, very low and they do not harm anything at all.
This must be very close to the real thing IMO.
Hans
In the sim the transformers and xovers do not roll off until very, very low and they do not harm anything at all.
This must be very close to the real thing IMO.
Hans