Fostex FX120 : the second part...
Since I had exams I didn't have time to spend on DIY audio these days.
Now I'm on holiday and after having finished my Home theater system (I'm writing a part on it for my website) which works great, I thought I'd replace, as I foresaw it, my TQWT's for a better enclosure.
Well, I ran some simulations once more with MJK's outstanding MathCad sheets, and drew another TQWT design (I found it to model better than TL's or straight lines).
I built the prototype with a friend, sealed it perfectly and tested it.
Though 40Hz passes perfectly well, there is a lack between 100Hz and 50Hz that wasn't that deep in the simulation. Bass was rich and complex, but had no authority, consequence of the small size of the unit. Moreover, BSD decreases the bass output.
I then remembered something : Fostex didn't give the BL factor in its datasheet, and I estimated it to be 4.3 according to the efficiency of the driver. And then, I used CAAD4.1 to model a bass reflex enclosure, entered the driver parameters and was surprised to see that the BL calculated by the software from Fostex's params was not 4.3 but 5.16 !
Here was the key of the problem, and when I ran the simulation again, it became much more accurate, thanks to Martin's great work.
Well, though this enclosure is definetely better than the first one, I cannot accept such a frequency response !
I played arround with different enclosures types, but none of them can bring my FX120s down to 40-45Hz without linearity issues. Even at 50Hz, linearity is correct but not perfect
Double bass reflex looked really promising but a dip arround 100Hz makes it unusable. Does anyone has some tools to simulate a double chamber reflex, with 2 ports on the outside instead of one in MJK's sheet ?
Back loaded horns simulated using MJK's sheet gave me excellent results (peak impulse response, 30Hz...) but after 200Hz, the response presents some +10/-6dB spikes that are unacceptable. Judicious damping may correct the problem ?
Well, after all that, I decided to use the TL sections sheet, and filling it was really long. I thought increasing the surface of some sections of the TQWT judiciously and reducing some others to fill the lacks and damp the peaks could do the trick but right now I didn't obtain anything better. :smash:
The last idea I wish to explore would be a double TQWT/BR enclosure : a well tuned TQWT followed by a BR resonator or a TQWT including a resonator could maybe be able to bring the response down to 40Hz, but I doubt anyone has ever tried it. I suppose the DBR sheet isn't accurate enough to make the difference but maybe I'm wrong ! :confused:
And then my philosophical point of view : MLTQWT's are in fact only BR which have judiciously used standing waves issues :rolleyes:
Any comments welcome !
The simulations : the first one with the real BL of the driver and the second with the bad one
I doubt that this small driver can ever be linear down to 40Hz.
You should concentrate on the 50 to 60Hz region, that's where you should tune the BR.
You do not mention what 'Sd' you used or whether you used published or measured specs, so WRT the 'BL' value, when I use published specs and a reasonable assumption for 'Sd', I arrived at 83.48 cm^2, 6.5133 N/A.
Pushing a driver this far below Fs requires lots of Vb and do not recall trying to get this far below without a huge horn, but now that we have MJK's WS, I plugged some Voigt style pipe numbers in and came up with a much better sim than I thought I would, though whether or not it will perform 'close enough' to prediction remains to be seen.
L = 114.7"
So = 1.567"^2 (pipe physically comes to a point)
SL = 157.08"^2
rp = 3"
lp = 2.28"
density = ~1 lb/ft^3 at pointed end only, with only some felt on one of the parallel side walls to damp any standing waves, so all damping experimentation should in theory only be done in the first ~23" of the pipe, though if you want to roll off the bottom more to improve transient response or better blend to the room, then add some on the 'floor' of the pipe.
Xo/zdriver = 0.682/~78.2" - IOW driver is up from the bottom ~36.5".
Obviously, due to its length it will need to be folded, so something like Bob Brines' Weems pipe variant, except with just a divider terminating at the bottom with no gap (to create a 'point') and the driver is so far down the pipe that it will be on the same face as the vent, which needs to be right at the bottom for lowest Fp: http://www.geocities.com/rbrines1/im...olded_Pipe.gif
FWIW, adding 4 ohms series resistance (Re = 11.3, Qes = 0.728) flattened it out and damped the slight peaking at Fp.
Anyway, if you build it, please let us know what tweaking you did and how well it worked, or not. For sure it will not play very loud, but in a small room may be adequate.
-2db at 50Hz.
I tend to agree with carlosfm, you can only push a 4.5" driver so far. The fx120 doesn't have a very large x-max either. Maybe you can push it so it will play 40Hz, but at the expense of power handling, and dynamics. Everything is a compromise.
Thanks for all these suggestions ! :cool:
I'm not obcessed by 40Hz, but I noticed that almost every music I listen to goes down to 40Hz, and some even go down to 30Hz.
If I want to bring this speaker down to 40hz, it is because I cannot have a subwoofer (too expensive, and I don't have a good place for it anyway).
Dynamics are not really important because I never bring my speaker to its +-2 mm Xmax, except when I wanna listen to music out of my bedroom :D. Anyway, the woofer displacement cannot be worse than on an open baffle.
I'm almost sure now that BR and DBR cannot bring this speaker down to 40Hz without linearity issues, and TLs don't look so promising unless they are severily stuffed and really large.
Well, if I'm condamned to have a big box, I think I will look for a Back loaded horn enclosure that allows an almost infinite bass extension and an amplified peak response which, if I understand, can be equal to a bigger driver's in a conventional enclosure.
Apart from Martin's BLH sections sheet, is there another tool for simulating this enclosure, cos' I didn't find any on the net ?
Here is the best modelisation I obtain for a BLH.
Do you think it's possible to damp out all these spikes by putting absorbing materials in the horn itself ?
-2db at 50Hz is not too bad, is it?
That's simulated, in-room response will probably extend a little more, as long as it isn't a big room and you take care to position them.
yes it would be perfect with a sub. Anyway it is the solution I will be condamned to if my back loaded horn project fails.
Room resonnances might help, and I know that my current TQWT's only go down to 70Hz but one of them is in a corner and extends down to 45Hz, with an equalizer.
Thaks for your nice simulations !
I choosed this driver because at this time I didn't know much about speakers and a friend who heard it told me it was outstanding for its size, which is right because I didn't hear anything better concerning sound quality, though bass cannot extend really low.
What I'd like to know right now is if my BLH simulation is accurate enough to start building a new prototype, because cutting wood is expensive and boring
Your coupling chamber is 32 inches long!
Looking at your geometry, I believe you have designed a TL and not a back loaded horn. I have serious doubts that this design will work and it is probably not worth building. Don't believe everything you simulate, you need to ask yourself if the model geometry and calculated results make any sense and if they fit the assumptions that went into deriving the equations.
|All times are GMT. The time now is 07:19 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright ©1999-2017 diyAudio