I've been modeling the Usher 8945k in LSPCad and comparing it to known enclosures. So far everything matches... except the enclosure I'm trying to copy. 
I started with Dennis Murphy's version. I put the volume into LSPCad, used the same port frequency, and other than a discrepancy in port length, everything matched. The -3dB point was the same as Dennis had measured.
Next I modeled the Euphase enclosure with the same results. Everything lined up.
Both of those guys are using a box volume of 19-25 liters, with a port frequency of 39hz.
But... both of those designs have a -3dB point of around 45hz. Usher's CP-737 has a published response of 35-20khz, +/-3dB. I can't get ANY enclosure size to model anywhere close to that. D'Appolito tuned the enclosure to 30hz, but didn't mention the box volume. I made an educated guess based on external dimensions of the speaker and came up with around 75-80 liters of internal volume. Nothing I do gets it close. The only thing I'm not taking into account is the baffle step compensation that D'Appolito does with the 2nd sub, but that shouldn't have any effect on low frequency extention... right?
Is somebody else willing to fool with this and see what they get? I've got the crossover for this 2.5-way, but I can't figure out the enclosure volume.
I started with Dennis Murphy's version. I put the volume into LSPCad, used the same port frequency, and other than a discrepancy in port length, everything matched. The -3dB point was the same as Dennis had measured.
Next I modeled the Euphase enclosure with the same results. Everything lined up.
Both of those guys are using a box volume of 19-25 liters, with a port frequency of 39hz.
But... both of those designs have a -3dB point of around 45hz. Usher's CP-737 has a published response of 35-20khz, +/-3dB. I can't get ANY enclosure size to model anywhere close to that. D'Appolito tuned the enclosure to 30hz, but didn't mention the box volume. I made an educated guess based on external dimensions of the speaker and came up with around 75-80 liters of internal volume. Nothing I do gets it close. The only thing I'm not taking into account is the baffle step compensation that D'Appolito does with the 2nd sub, but that shouldn't have any effect on low frequency extention... right?
Is somebody else willing to fool with this and see what they get? I've got the crossover for this 2.5-way, but I can't figure out the enclosure volume.
Remember that sometimes +/- 3db is
a cheap way of actually saying - 6dB.
And that the -6dB point is the commonly
accepted quoted bass extension limit.
🙂 sreten.
a cheap way of actually saying - 6dB.
And that the -6dB point is the commonly
accepted quoted bass extension limit.
🙂 sreten.
That's a possibility, but that means D'Appolito fudged the response plot in the magazine publication too.
http://soundberry.com/0_000080.html
According to that same publication, the response was flat within +/- 1.2dB from 200hz to 20khz. That impressed me. 🙂
http://soundberry.com/0_000080.html
According to that same publication, the response was flat within +/- 1.2dB from 200hz to 20khz. That impressed me. 🙂
1.) Did you include series resistance in your calculation? Series resistance will raise Q and require a larger box and lower tuning..... Resulting in lower sensitivity and a lower cutoff....
2.) Always suspect measurements under 200Hz....
3.) The review measurements appear highly smoothed....
2.) Always suspect measurements under 200Hz....
3.) The review measurements appear highly smoothed....
Jim85IROC said:
Without actually putting the numbers in, I'd say it's clever specsmanship at best. For bookshelf designs, if the F3 is much below 50Hz the linear excursion limited power handling is usually rotten.
Scott Hinson
Ron E said:
Agreed, always suspect measurements under 200Hz, unless the measurement conditions including environment, equipment and method are specified.
According to the screen caps the measurements are smoothed at .17octave. They do look more than that, so it could be a relatively short MLS sequence which will do the same thing.
Scott
Jim85IROC said:
I wouldn't be all that impressed. With modern measurement gear and computer optimization of crossovers it's relatively easy to achieve this kind of performance in a few hours or even less.
The interesting is the behavior around ~150Hz. If I didn't know any better I'd say it was either horrid port pipe resonance, or more likely badly scaled and spliced nearfield/farfield plots.
Scott Hinson
Jim85IROC, my understanding is, what is confusing you is that you are not taking into account that this is an EBS alignment and baffle step correction is done hand in hand with the EBS alignment.
Some crude calculations: the serial resistor at tweeter is 2 ohms. Tweeter Re is 5.6 ohms. That more or less means tweeter output is attenuated 2.5 db. Tweeter sensitivity is 88db, that means system target sensitiviy is 85.5db. Woofer sensitivity is 88db, -6db baffle step + 6db from parallel drive of two woofers gives you 88db woofer sensitivy.
I don't know the box volume, but as you guess say it's 70lt. With 30Hz tuning alignment, it gives a shelving starting at 200Hz till 38Hz, about -3.5db or so around 38Hz, which makes sensitivity 84.5db around 38Hz, which is -1db from system target ot 85.5db. Considering the -3db point relative to overall system sensitivity, 35Hz is not very unexpected.
I think what is going on is that, the designer sacrificed efficiency in favor of bass extension. He could have shooted for 88db efficiency, with a none EBS alignment, which would give 45Hz~50Hz -3db.
This is my understanding of it, would be happy to be corrected if I am mistaken.
Also I would point out that the tweeter on CP-737 is angled. If you are trying to copy it, you should be able to do the same, otherwise time alignment will be off.
Some crude calculations: the serial resistor at tweeter is 2 ohms. Tweeter Re is 5.6 ohms. That more or less means tweeter output is attenuated 2.5 db. Tweeter sensitivity is 88db, that means system target sensitiviy is 85.5db. Woofer sensitivity is 88db, -6db baffle step + 6db from parallel drive of two woofers gives you 88db woofer sensitivy.
I don't know the box volume, but as you guess say it's 70lt. With 30Hz tuning alignment, it gives a shelving starting at 200Hz till 38Hz, about -3.5db or so around 38Hz, which makes sensitivity 84.5db around 38Hz, which is -1db from system target ot 85.5db. Considering the -3db point relative to overall system sensitivity, 35Hz is not very unexpected.
I think what is going on is that, the designer sacrificed efficiency in favor of bass extension. He could have shooted for 88db efficiency, with a none EBS alignment, which would give 45Hz~50Hz -3db.
This is my understanding of it, would be happy to be corrected if I am mistaken.
Also I would point out that the tweeter on CP-737 is angled. If you are trying to copy it, you should be able to do the same, otherwise time alignment will be off.
I don't see any listing or links to the Thiele-Small specs of the speaker, which would certainly be nice. The link that was given for the frequency response of an enclosure is not working, at least for me.
So, without any convenient means to see the specs, I will say only this: your box is tuned to 39 Hz, but you want an F3 of 35 Hz.
That rarely happens.
Generally, well designed commercial boxes use a woofer with a Qts of 0.38 or below, put into boxes that are equal to Vas or less. The F3 is above the tuning frequency. So I suspect that something with this situation is amiss right there.
The bass is snappier the smaller the box is compared to Vas, and the smaller the Qts.
You can build a box with a Vb larger than Vas, a Qts above 0.38, and an F3 above Fb, (tuning frequency). But for the reasons given, they are not generally recommended.
A misprint, perhaps?
So, without any convenient means to see the specs, I will say only this: your box is tuned to 39 Hz, but you want an F3 of 35 Hz.
That rarely happens.
Generally, well designed commercial boxes use a woofer with a Qts of 0.38 or below, put into boxes that are equal to Vas or less. The F3 is above the tuning frequency. So I suspect that something with this situation is amiss right there.
The bass is snappier the smaller the box is compared to Vas, and the smaller the Qts.
You can build a box with a Vb larger than Vas, a Qts above 0.38, and an F3 above Fb, (tuning frequency). But for the reasons given, they are not generally recommended.
A misprint, perhaps?
Re: Re: why doesn't my box size match that of others?
I would suggest you to run the numbers. 19lt tuned at 39Hz has minimum 102 db excursion limited output from 42Hz on. I wouldn't call that rotten. In case of CP-737, two woofers have plenty of SPL capability, even tuned as low as 30Hz.
ScottRHinson said:
I would suggest you to run the numbers. 19lt tuned at 39Hz has minimum 102 db excursion limited output from 42Hz on. I wouldn't call that rotten. In case of CP-737, two woofers have plenty of SPL capability, even tuned as low as 30Hz.
Here's a linke to euphase audo's measured t/s specs, which line up very closely with other people's measured specs, as well as with Usher's published specs:kelticwizard said:
http://www.euphase.com/RawDrivers/8945-W.asp
I realize that I'm not going to get a 3-dB of 35hz when tuning to 39hz. My problem is that I am unable to model an enclosure that does it even when I tune at 30hz like D'Appolito's original design.
The EBS is something that I hadn't considered, and something I want to try to learn more about. Now that I'm reading a little bit on it, it does look similar to what my modeling had resulted in. Between baffle step and EBS, that 2nd woofer's role seems like it could integrate nicely, but it also seems that I either need to find the proper volume that D'Appolito used, or I need to find a way to model this 2.5 way configuration more accurately.
Right - there is no difference in the bass modelling of a 2.5
way and a 2 way speaker with parallel twin bass units.
🙂 sreten.
Jim85IROC said:
http://www.mhsoft.nl/spk_calc.asp#newqts
Indicates a 0.8 DCR of the inductors will change Q to 0.36.
It will also slightly change the sensitivity which
is not modelled by changing Q, by around -1dB.
Not that I'm saying Q=0.36 is the best value, I'll play with it.
🙂 sreten.
edit : shows Q stepped from 0.32 in 0.2 steps,
Q=0.4 = 2R/driver, 2 drivers, left scale = 20Hz.
Attachments
- Status
- Not open for further replies.