If you go here there is a waterfall response for the JX53, the previous driver to the JXR6. Although not identical, it should be similar above 200Hz. The Jordan site doesn't have a waterfall for the JXR6 yet. Having listened to the 53 and the 6, the JXR6 is slightly more delicate in the HF and has a much better LF output (mine have a measured Fo of 65Hz).
It seems that older driver was smoother:
http://www.hifisound.de/oxid/out/oxbaseshop/html/0/test_pdf/EJJ-1110856.pdf
As I see they offer electrostatics quality. What interesting me the most is how to cancel small pressure changes in the step response at transmission times ca. 100 and 400 microseconds or 8000 and 16kHz.
Does anyone tried non-electrical corrections? For example if you build rotationally symmetrical baffle like a disc you can build in the shape bulges or cavities that will change a pressure fall but don't know if they can smooth frequencies >1kHz.
Do you know maximum continuous SPL of this small single drivers?
http://www.hifisound.de/oxid/out/oxbaseshop/html/0/test_pdf/EJJ-1110856.pdf
As I see they offer electrostatics quality. What interesting me the most is how to cancel small pressure changes in the step response at transmission times ca. 100 and 400 microseconds or 8000 and 16kHz.
Does anyone tried non-electrical corrections? For example if you build rotationally symmetrical baffle like a disc you can build in the shape bulges or cavities that will change a pressure fall but don't know if they can smooth frequencies >1kHz.
Do you know maximum continuous SPL of this small single drivers?
Having a circular, hard-edged baffle would create more problems than it would "solve," in my estimation at least.
Do you have any links to projects using such a method to correct for frequency abnormalities? I can >>maybe<< imagine it being successful for cheap drivers, but I think you are splitting hairs with high-quality drivers like the Jordan and Bandor.
Do you have any links to projects using such a method to correct for frequency abnormalities? I can >>maybe<< imagine it being successful for cheap drivers, but I think you are splitting hairs with high-quality drivers like the Jordan and Bandor.
Linkwitz has a nice spreadsheet showing the relationship between driver area, frequency, excursion, and SPL:Colin said:
http://linkwitzlab.com/spl_max1.xls
jzagaja said:
I suspect the surround formula changed. Metal cones are sensitive to surround formula and weight. The JX53 used to have a spider that could help in the damping process, but add to the effective mass. I think you will find that they will actually sound much better than what the graphs tell you. But one you get to listening to them, it might be still possible to slightly tune it by using tooth paste, the kind that drys hard instead of the gel type. Normally you want to keep the dry weight within 0.2 grams. Two areas that might help is some on the cap and some on the outer edge of the cone. It can be applied using a cotton swab.
Consider that the acoustic wave front is propagating away from the transducer/baffle at the same time it is propagating across the baffle. Thus, when dealing with a flat baffle, if the original phenomena are generated by the cone and dust cap (as I predict those evident in the Hobby HiFi test are) that there will be a leading edge containing the error ahead of the outward propagation from the "corrective" baffle features.
The sum is to increase the complexity of the step or impulse response while not actually compensating for the original error.
Also, while I judge the Hobby HIFI tests to be more representative of performance than the Jordan tests, there are still a couple of things that give me pause. If you want to continue with your experiments, I suggest that you acquire the necessary instrumentation to allow you to do your own before and after tests.
For the type of work you propose, I believe that impulse responses will be easier to work with than a step response. In graphing, the DC component of the step response tends to swamp out the high frequency information. An impulse response should more graphically show the high frequency components of a transducer's transient response.
The sum is to increase the complexity of the step or impulse response while not actually compensating for the original error.
Also, while I judge the Hobby HIFI tests to be more representative of performance than the Jordan tests, there are still a couple of things that give me pause. If you want to continue with your experiments, I suggest that you acquire the necessary instrumentation to allow you to do your own before and after tests.
For the type of work you propose, I believe that impulse responses will be easier to work with than a step response. In graphing, the DC component of the step response tends to swamp out the high frequency information. An impulse response should more graphically show the high frequency components of a transducer's transient response.
Colin - you are right about the room effect. We can also discuss opposite points of view e.g Moulton vs. Geddes but in fact the first pressure changes that aren't room dependant are IMHO important.
Colin and Dumbass - please read US patent 4268719:
http://ucz.tme.szczecin.pl/~jacek/temp/4268719.pdf
MarkMcK- in the paper above Mr. Manger is using correction via bulge. He wants step response between reversal points to be monotonic and linear. Why? As I know a perfect speaker should reproduce an impulse as a doublet with some ringing or overshot. Correlated step response is then e-function with some undershot.
In the opposite to the design described above is 2" compression tweeter in a waveguide and 18" woofer. Impulse and step is very clean but decays quicker and with higher undershot:
http://ucz.tme.szczecin.pl/~jacek/temp/Summa.gif
At the end - can we count for DSP? I think no, see equalized speaker/room (1D measurement):
http://ucz.tme.szczecin.pl/~jacek/temp/ComparisonStepResponseCorrected2Uncorrected.png
Colin and Dumbass - please read US patent 4268719:
http://ucz.tme.szczecin.pl/~jacek/temp/4268719.pdf
MarkMcK- in the paper above Mr. Manger is using correction via bulge. He wants step response between reversal points to be monotonic and linear. Why? As I know a perfect speaker should reproduce an impulse as a doublet with some ringing or overshot. Correlated step response is then e-function with some undershot.
In the opposite to the design described above is 2" compression tweeter in a waveguide and 18" woofer. Impulse and step is very clean but decays quicker and with higher undershot:
http://ucz.tme.szczecin.pl/~jacek/temp/Summa.gif
At the end - can we count for DSP? I think no, see equalized speaker/room (1D measurement):
http://ucz.tme.szczecin.pl/~jacek/temp/ComparisonStepResponseCorrected2Uncorrected.png
After doing a quick check with soundcard loop back I see that a perfect step response in the time transmittance of interest should not decay quickly!
http://ucz.tme.szczecin.pl/~jacek/temp/PerfectStep.gif
The question is what will happen if it does?
Waterfall of Mr Geddes Summa (exceptional response for the horn loaded speaker):
http://www.nutshellhifi.com/wtrfall6.gif
http://ucz.tme.szczecin.pl/~jacek/temp/PerfectStep.gif
The question is what will happen if it does?
Waterfall of Mr Geddes Summa (exceptional response for the horn loaded speaker):
http://www.nutshellhifi.com/wtrfall6.gif
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.