Ah D&D, many misspent hours with that! 
(..my favorite is still 2.5 rules, though 3.5 is good and in some instances better mechanically.)
Btw, don't worry about the other thread.
The Fostex should average about 89 db on-baffle and placement proposed within it's usable bandwidth (..it should actually be a bit higher than this because of diffractive gain from the baffle below 1.8 kHz). The Gradient will of course be "padded"-down with baffle-step losses and will likely be a bit generous in spl in the lower-midrange and upper-bass (classic slight downward "tilt" from 150 Hz). There will likely be a bit of a presence "dip" (-1/1.5 db) in the freq. response around 1.8-2.6 kHz from the average, but that just aids in the depth perspective. Both filters will have insertion losses (and in the Gradient's case - is required), so the final target might be slightly less than 89 db.. but I'm hoping it will be close. (..it also depends on IF you do a more traditional tweeter inclusion: with enough lower freq. bandwidth from the tweeter, average "eff." could actually exceed that 89 db.)
Also, don't forget that the outlined design isn't a dipole - rather "open-back" for at least the midrange to mid.-treble: and it's not about rear leakage (out-of-phase spl from the rear of the drivers), but rather achieving an "enclosure" that is essentially non-resistive (that doesn't add to the compliance of the driver and alter excursion).
(..my favorite is still 2.5 rules, though 3.5 is good and in some instances better mechanically.)Btw, don't worry about the other thread.
The Fostex should average about 89 db on-baffle and placement proposed within it's usable bandwidth (..it should actually be a bit higher than this because of diffractive gain from the baffle below 1.8 kHz). The Gradient will of course be "padded"-down with baffle-step losses and will likely be a bit generous in spl in the lower-midrange and upper-bass (classic slight downward "tilt" from 150 Hz). There will likely be a bit of a presence "dip" (-1/1.5 db) in the freq. response around 1.8-2.6 kHz from the average, but that just aids in the depth perspective. Both filters will have insertion losses (and in the Gradient's case - is required), so the final target might be slightly less than 89 db.. but I'm hoping it will be close. (..it also depends on IF you do a more traditional tweeter inclusion: with enough lower freq. bandwidth from the tweeter, average "eff." could actually exceed that 89 db.)
Also, don't forget that the outlined design isn't a dipole - rather "open-back" for at least the midrange to mid.-treble: and it's not about rear leakage (out-of-phase spl from the rear of the drivers), but rather achieving an "enclosure" that is essentially non-resistive (that doesn't add to the compliance of the driver and alter excursion).
There will of course be some variation in Fostex samples:
1st is spec. sheet
2nd is independent w/ 1 kHz at about *86 db w/OUT baffle effects and 88 db at 3-5.5 kHz where you'll get some oscillation from diffraction (higher and lower spl). Diffractive gain below 1.8 kHz should raise the response to almost match that 3-5.5 kHz region.
*compensate by dropping the pressure 10 db from what's shown.
THEN there is the contribution of the Gradient and Tweeter (if any) that will "lift" (increase spl) that result a bit, and insertion losses that will lower it some as well.
Note: midranges are typically lower in spl than everything else (..bass & tweeter) where a summed-response of all drivers usually results in a higher spl average than the midrange. Of course that won't be the case IF you don't decide to add a tweeter of sufficient bandwidth, BUT you might just subjectively prefer the result (despite the slightly lower spl).
1st is spec. sheet
2nd is independent w/ 1 kHz at about *86 db w/OUT baffle effects and 88 db at 3-5.5 kHz where you'll get some oscillation from diffraction (higher and lower spl). Diffractive gain below 1.8 kHz should raise the response to almost match that 3-5.5 kHz region.
*compensate by dropping the pressure 10 db from what's shown.
THEN there is the contribution of the Gradient and Tweeter (if any) that will "lift" (increase spl) that result a bit, and insertion losses that will lower it some as well.
Note: midranges are typically lower in spl than everything else (..bass & tweeter) where a summed-response of all drivers usually results in a higher spl average than the midrange. Of course that won't be the case IF you don't decide to add a tweeter of sufficient bandwidth, BUT you might just subjectively prefer the result (despite the slightly lower spl).
Attachments
Last edited:
Had much fun for many years with Pen&Paper RPG, AD&D 2nd for the longest time, now this period is behind me unfortunately.Ah D&D, many misspent hours with that!
Thanks, I do not, it challenge the little I know.Btw, don't worry about the other thread.
I have read a bit about the baffle-step loss (and gain) since you mentioned it before and I think I understand it a bit better but I still don't get how to predict at what frequency it start to roll down (and increase a bit just before) when it is well explained it goes too much in math for me, so I'll rely on VituixCAD to do that math for me
Insertion Loss of filters is a new thing for me, most often it's said that the simplest crossover is the better, but difficult to find a article that is simple enough to read that explains it, most often it's related to antennas.
I am better to understand graphs than pages of obscure text anyway.
About that "traditional tweeter inclusion" :
I don't know what is traditional about tweeter
If I understand well you mean The Fostex would be more efficient if it had only to reproduce a shorter band of frequencies adding a tweeter that goes low as you say?
You did good to remind me the difference between the two I was a bit lost there, so by example is a small sealed enclosure considered the most resistive compared to any other alternative? The trapped air volume inside restrict the movement of the membrane? when a non-resistive is better suited to driver ?
Sorry lots of questions
There will of course be some variation in Fostex samples:
1st is spec. sheet
2nd is independent w/ 1 kHz at about *86 db w/OUT baffle effects and 88 db at 3-5.5 kHz where you'll get some oscillation from diffraction (higher and lower spl). Diffractive gain below 1.8 kHz should raise the response to almost match that 3-5.5 kHz region.
*compensate by dropping the pressure 10 db from what's shown.
THEN there is the contribution of the Gradient and Tweeter (if any) that will "lift" (increase spl) that result a bit, and insertion losses that will lower it some as well.
Note: midranges are typically lower in spl than everything else (..bass & tweeter) where a summed-response of all drivers usually results in a higher spl average than the midrange. Of course that won't be the case IF you don't decide to add a tweeter of sufficient bandwidth, BUT you might just subjectively prefer the result (despite the slightly lower spl).
That is very clearly explained, thanks
I will try without a tweeter but then I'd like to reserve me the opportunity to add a tweeter without having to redo the cabinets, I thought that maybe I should separate the tunnel part from the ported part to allow me to just adjust the top in case it is needed to reposition to accommodate the change.
What worry me a bit is the power that the Fostex will tolerate 5W does not seem much, but I suppose not having to play the lower frequencies change the power that it will receive.
I usually listen to music around 1/4 of my integrated amplifier volume here are the specs so you have an Idea what is my reference.
Koda KD-261 (not very clean when not playing)
Peak Music Power (PMPO) 300w x 2 (8 ohms)
The output power of distortion
limitation (rating condition)
≧100W x 2 (8 ohms)
THD 0.7% (20Hz~20kHz)
S/N Ratio
Frequency Response
≧86db (A-weighted)
20Hz~50kHz +/-1.5dB
Rating load impedance L: 8ohms R:8ohms
Rating Power consumption 400W
Cabasse Jersey MT221 (not very detailed in the lower and top frequencies)
Ways : 2
Drivers : 1 x DOM20 - 2 x 17T15
Efficiency : 91 dB
Crossover : 3 200 Hz
Frequency Response : 45 - 20 000 Hz
Nominal Power : 100 w
Peak Power : 700 W
One other thing; how did you calculate the loss in dB from specs 2.83V/1m from the other 30cm 2.83V/30cm for the Fostex measurements ?
I have continued to experiment/practice with REW measurements and one thing is still completely unclear to me is how to establish consistent levels
In the the document from Texas instruments Audio Characterization Primer or the Guide from Kimmosto it's not evident for me.
I have tried hooking up a multi-meter to measure the output of the amplifier but it's clear that playing anything make the voltage very unstable, I have read that I should have a resistor between the two leads but I'm cautious about what I read on Internet.
Any good way to make sure that I am always at the same level for the measurement?
I always turn down the potentiometers of the UMC204 and also on the T.Amp before I turn it off, don't want any bad surprises when I have to manipulate the Drivers, but then it's difficult to be back at exactly the same level.
In the the document from Texas instruments Audio Characterization Primer or the Guide from Kimmosto it's not evident for me.
I have tried hooking up a multi-meter to measure the output of the amplifier but it's clear that playing anything make the voltage very unstable, I have read that I should have a resistor between the two leads but I'm cautious about what I read on Internet.
Any good way to make sure that I am always at the same level for the measurement?
I always turn down the potentiometers of the UMC204 and also on the T.Amp before I turn it off, don't want any bad surprises when I have to manipulate the Drivers, but then it's difficult to be back at exactly the same level.
I have read a bit about the baffle-step loss (and gain) since you mentioned it before and I think I understand it a bit better but I still don't get how to predict at what frequency it start to roll down (and increase a bit just before) when it is well explained it goes too much in math for me, so I'll rely on VituixCAD to do that math for me
Yes.. and no.
Rely on VituixCad for generating the components, but do NOT rely on it for an "auto" outcome.
In other words: change the component values (for parallel crossover: Inductor + Resistor) from that "auto" outcome to see different effects.
THAT's critical, and really - it's what makes modeling so powerful: allowing you to quickly see a range of options and their effects.
As far as getting a "better" (or perhaps a more "elegant") solution is something you'll learn about as you go, and something that can be asked about once you've actually got a good measurement and put that into VituixCad.
It was Tim's (educated) guess, and actually I think he spec'ed -10.5 db, not -10 db.
Of course at about double the distance (30 cm to 60 cm) the "average" would -6db loss. 60 cm to 90 cm would be -3db. 90 cm to 180 cm for -6 db, but it's only 10 cm (from that 180), so about a 5+% loss of -6db or slightly more than -.3 db. Total around -9.3+ db..
Critically though, the testing was NOT done to find a 1 watt 1 meter response, but rather just the linear response and axial response when compared to other drivers tested in the same manner (and even to show overall pressure differences between those drivers). In other words I'm not confident in the outcome presented.. but then again, I'm not confident in Fostex's either.
I'm not even confident in the merged response on it's own, and in that respect I am reasonably confident in Fostex's overall response (..how it behaves relative to the average below 1.5 kHz).
Basically clear as mud!
In any event you still need to make your own measurements, because relying on anyone else's data can get you into "trouble".
I'm always worried about that.
But you've got modeling for that: seeing how multiple high-pass filters work-out (or don't) for the design.If you've got to compromise on the design by limiting bandwidth of the driver to 600 Hz up, then you do that even though floor-bounce will induce some suck-out at particular distances from the loudspeaker that are "less than optimal".
Nothing is "fixed" at this point.
This one is beyond me..
I use an spl-meter (and different programs). Best to ask REW questions in the official REW forum:Official REW (Room EQ Wizard) Support Forum | AV NIRVANA
Oh, and if you get a good *answer please provide a link to it here.
*well other than the obvious: "don't touch the dac and amp".
(..and also Operating System output level changes.. like from a DAMN UPDATE!)
Last edited:
Yes.. and no.
Rely on VituixCad for generating the components, but do NOT rely on it for an "auto" outcome.
In other words: change the component values (for parallel crossover: Inductor + Resistor) from that "auto" outcome to see different effects.
THAT's critical, and really - it's what makes modeling so powerful: allowing you to quickly see a range of options and their effects.
As far as getting a "better" (or perhaps a more "elegant") solution is something you'll learn about as you go, and something that can be asked about once you've actually got a good measurement and put that into VituixCad.
Yes I understand.
I can't wait to do those measurements but I have a lot of things to arrange before, my biggest issue at the moment is not having a workshop where I live, I am in the process of finding a solution to that but it isn't fast.
It was Tim's (educated) guess, and actually I think he spec'ed -10.5 db, not -10 db.
Of course at about double the distance (30 cm to 60 cm) the "average" would -6db loss. 60 cm to 90 cm would be -3db. 90 cm to 180 cm for -6 db, but it's only 10 cm (from that 180), so about a 5+% loss of -6db or slightly more than -.3 db. Total around -9.3+ db..
Critically though, the testing was NOT done to find a 1 watt 1 meter response, but rather just the linear response and axial response when compared to other drivers tested in the same manner (and even to show overall pressure differences between those drivers). In other words I'm not confident in the outcome presented.. but then again, I'm not confident in Fostex's either.
I'm not even confident in the merged response on it's own, and in that respect I am reasonably confident in Fostex's overall response (..how it behaves relative to the average below 1.5 kHz).
Basically clear as mud!
In any event you still need to make your own measurements, because relying on anyone else's data can get you into "trouble".
I see the principle, so in fact I could simply do the same when i do my own measurements; "linear response and axial response"
I don't have to worry about really measuring the sensitivity (dB) of the speaker as long as I do all the drivers measurements at the same level and at the same distance.
Obviously I would have an issue to align Far-Field with Near Field if I am at the same level but that is essentially needed for the lower frequencies.
Unsurprisingly my "math" was off.. and a distance attenuation calculator results in about -10.5 db:
Distance Attenuation Calculator - Omni
-oh well.
Distance Attenuation Calculator - Omni
-oh well.
Unsurprisingly my "math" was off.. and a distance attenuation calculator results in about -10.5 db:
Distance Attenuation Calculator - Omni
-oh well.
Not so far, nice site by the way
I felt very alone earlier while trying to do some measurement with REW, I tried to take a measure with the voltmeter with a 1khz signal then tried to play with the calibration in the generator, and when I was measuring my level was so low, so I thought immediately that I was messing with the wrong stuff and couldn't get it back to the previous setting, so I ended up resetting the whole REW preferences.
It's only after that I was redoing the level calibration that I notice my mike input wasn't very much responding to my changes, I realized I completely forgot to enable the shadow power for the mike on the audio interface...
Hi ScottG,
Hope you have a great day
It is a holiday I had some more time to set up again the REW and do some tests
Here is some parameters for the tests :
- Each Driver is on the floor near the middle of the room on a 3x3 egg crate foam stuff
- Mic tip is about 30cm distance, 6.5LF about 1cm lower than the AXP-08, and FF85WK at about the same height as the AXP-08
- Soundcard and Mic are calibrated
- Level of reference at about -18dB, the Mic between 1 & 2dB lower
- dB calibration at about 70db (Had to use my phone) and I set it to Z-weighted
- The Graphs on the screenshot are in the same order as the tests on the left, AXP on top, 6.5LF in the middle, and the FF85WK at the bottom
- The Red, Orange, Pink Graphs are the one that are left without adjustments, all the others are stopping at 14ms on the right of the first impulse
- I didn't touche anything during the measurements besides switching the drivers
Conclusion
- I think I start to understand the basics of the software.
- It is to suppose that my dB calibration is not accurate and I have no idea yet how to get the voltage to the drivers
- I see that the AXP-08 & FF85WK have kind of the same shape but not the 6.5LF I have no idea why, it could not be the convex dust cap on the 6.5LF is it ?
- There is a serious gap in level between each driver, the AXP-08 is near the limit of the REW headroom
- I am curious to see the difference on a IEC Baffle
Any Idea if you see something wrong in how I did it?
Why the 6.5LF have a different shape? Here is the DATSV2 measurement :
D 126 [mm]
Re 3.253 [Ohms]
Fs 84.79 [Hz]
Zm 25.61 [Ohms]
BL 4.911 [N/A]
Qms 5.28
Qes 0.7684
Qts 0.6708
Vas 7.197 [liters]
L10k 0.2055 [mH]
n0 0.5443 [%]
dBSPL 89.46 [1W/1m]
Ms 10.69 [grams]
Cms 0.33 [mm/N]
Thanks
Hope you have a great day
It is a holiday I had some more time to set up again the REW and do some tests
Here is some parameters for the tests :
- Each Driver is on the floor near the middle of the room on a 3x3 egg crate foam stuff
- Mic tip is about 30cm distance, 6.5LF about 1cm lower than the AXP-08, and FF85WK at about the same height as the AXP-08
- Soundcard and Mic are calibrated
- Level of reference at about -18dB, the Mic between 1 & 2dB lower
- dB calibration at about 70db (Had to use my phone) and I set it to Z-weighted
- The Graphs on the screenshot are in the same order as the tests on the left, AXP on top, 6.5LF in the middle, and the FF85WK at the bottom
- The Red, Orange, Pink Graphs are the one that are left without adjustments, all the others are stopping at 14ms on the right of the first impulse
- I didn't touche anything during the measurements besides switching the drivers
Conclusion
- I think I start to understand the basics of the software.
- It is to suppose that my dB calibration is not accurate and I have no idea yet how to get the voltage to the drivers
- I see that the AXP-08 & FF85WK have kind of the same shape but not the 6.5LF I have no idea why, it could not be the convex dust cap on the 6.5LF is it ?
- There is a serious gap in level between each driver, the AXP-08 is near the limit of the REW headroom
- I am curious to see the difference on a IEC Baffle
Any Idea if you see something wrong in how I did it?
Why the 6.5LF have a different shape? Here is the DATSV2 measurement :
D 126 [mm]
Re 3.253 [Ohms]
Fs 84.79 [Hz]
Zm 25.61 [Ohms]
BL 4.911 [N/A]
Qms 5.28
Qes 0.7684
Qts 0.6708
Vas 7.197 [liters]
L10k 0.2055 [mH]
n0 0.5443 [%]
dBSPL 89.46 [1W/1m]
Ms 10.69 [grams]
Cms 0.33 [mm/N]
Thanks
Last edited:
And by the way John Mulcahy (REW) found the issue I had with voltage measurement and I feel stupid again, I had to press a button on my multimeter when in the right mode to switch from DC to AC, and also it is better to make that measurement with a signal to the amp that is close to the grid mains, in my case 50Hz
But now concerning the SPL measurement in REW, when I set my amplifier at about 2.8+V the level are lower and so does the dB measured, I have yet to sort that out.
But now concerning the SPL measurement in REW, when I set my amplifier at about 2.8+V the level are lower and so does the dB measured, I have yet to sort that out.
-yeah, this is something for the REW forum.
..the top-end of both the Fostex and the Gradient look wrong, and at least in the case of the Fostex (because of its smaller diameter), it shouldn't be a near-field effect at 30cm.
I would concentrate (to the exclusion of everything else) on getting the Fostex looking right first using Tim's measurements for those higher freq.s (3 kHz up) because he did the same 30 cm distance for those freq.s.. (..and 10 cm for <3 kHz.)
..anyway, I've got my own project struggles right now (..with amp's mysteriously "cutting-out" ) while trying to test the bass-shakers in my HT setup. ..and then a LOT more to do from there.. so it might be a bit (serveral days) before I'm back here to the forum. Again though, the REW forum is the place to be and it "sounds" as if you are getting good help (..which is one of the reasons I mentioned REW initially, as opposed to Arta.)
..the top-end of both the Fostex and the Gradient look wrong, and at least in the case of the Fostex (because of its smaller diameter), it shouldn't be a near-field effect at 30cm.
I would concentrate (to the exclusion of everything else) on getting the Fostex looking right first using Tim's measurements for those higher freq.s (3 kHz up) because he did the same 30 cm distance for those freq.s.. (..and 10 cm for <3 kHz.)
..anyway, I've got my own project struggles right now (..with amp's mysteriously "cutting-out"
) while trying to test the bass-shakers in my HT setup. ..and then a LOT more to do from there.. so it might be a bit (serveral days) before I'm back here to the forum. Again though, the REW forum is the place to be and it "sounds" as if you are getting good help (..which is one of the reasons I mentioned REW initially, as opposed to Arta.) -yeah, this is something for the REW forum.
..the top-end of both the Fostex and the Gradient look wrong, and at least in the case of the Fostex (because of its smaller diameter), it shouldn't be a near-field effect at 30cm.
I would concentrate (to the exclusion of everything else) on getting the Fostex looking right first using Tim's measurements for those higher freq.s (3 kHz up) because he did the same 30 cm distance for those freq.s.. (..and 10 cm for <3 kHz.)
..anyway, I've got my own project struggles right now (..with amp's mysteriously "cutting-out"
Thanks, don't worry about it I've got to spend some time to "play" with REW.
If I could help you I would, but you would spend more time trying to explain me than actually working on those problems
I'll load balance my questions toward REW if needed, and see if they are still in the mood
-yeah I'm still not sure what was up with the amp, but it seems to be working now. So that's good.
Turns-out the front "legs" of the platform (for the bass shaker) I built for each of the HT chairs was screwing-up the power-recline. "Mostly" fixed one (hack-job - literally), and now have several more to go.
..and a few minutes ago my dog was run-over (..got out while cleaning-up the mess on the HT side of the house), so that's not so great.
For me, a less than satisfactory weekend..
I'm hoping yours went a LOT better!
Turns-out the front "legs" of the platform (for the bass shaker) I built for each of the HT chairs was screwing-up the power-recline. "Mostly" fixed one (hack-job - literally), and now have several more to go.
..and a few minutes ago my dog was run-over (..got out while cleaning-up the mess on the HT side of the house), so that's not so great.
For me, a less than satisfactory weekend..
I'm hoping yours went a LOT better!