Fostex xmax and dynamic capabilities

[roffe]

These days there are a lot of sexy, small low effect amplifiers on offer. By low effect I am talking about the range of 10-20W per channel. (I personally want to try the Dared MP-5, but that is another story).
Obviously, one needs high sensitivity speakers to get decent SPL from these, and on paper I really like the Fostex FE167E and the larger FE207E.

The one concern I have is the low specified Xmax. I have always thought that a speaker must work within the Xmax limits to give decent dynamics, and that you'll get compressed dynamic once excursion of the driver goes beyond that.

Contrary to my beliefs, I read about DIYers giving high praise to the dynamic of their full range Fostex efforts.

So my question is: Are the Fostex full range speakers with their low specified Xmax keeping up a "near linear" dynamic outside the Xmax limits, or do they deteriorate to dynamic compression once these limits are passed?

I have searched the net (and diyAudio) for info about how Fostex drivers behave outside Xmax, but have so far come up short. Obviously, Xmech is considerably larger than the specified Xmax for these drivers.

Ref Fostex FE167E spec: http://www.fostexinternational.com/docs/speaker_comp/pdf/fe167erev.pdf

 

[Tweeker]

They tend to be heavily loaded into horns and resonators and thus dont need much excurion to deliver thier dynamics. This is key.

They dont seem to suffer too bad when linear xmax is somewhat exceeded.

 

[Dumbass]

People who are into compression drivers, horns, etc, will generally scoff at 8" full rangers, especially when they aren't mounted in gigantic rear-loaded horn enclosures.

For "the rest of us", the 167 and 207 can offer very dynamic, domestically-feasible performance.

Martin King's projects 2 and 5 would work well. (FE167E is drop-in replacement for 164 in project 2.)

http://www.quarter-wave.com/Project02/Project02.html
http://www.quarter-wave.com/Project05/Project05.html

If you like a folded TL, the project 2 whitepaper has alternative cabinet plans. If you like the tapered look, I believe FE207E would work fine in the project 6 enclosure.

I run FE207E in vented enclosures, like them very much:
http://i68.photobucket.com/albums/i6/heatherandsteve/Projects/0567e9db.jpg

 

[roffe]

Nice to hear you have built a successful bass reflex design. Do you use a correction circuit like in the "project5" that you're listing?
How loud do you think your speakers can play without going into audible dynamic compression?

 

[Dumbass]

Nice to hear you have built a successful bass reflex design. Do you use a correction circuit like in the "project5" that you're listing?
How loud do you think your speakers can play without going into audible dynamic compression?

I have about 3 ohms (can't remember exactly) series resistance because the 45 liter cabinets are slightly oversized for the drivers. (I think "max flat" bass reflex alignment calls for ~20 liters?) I bought a pair of those nice big green non-inductive resistors madisound sells.

I used to have a baffle-step circuit, but took it out eventually. My speakers are against back wall, so none necessary.

I've never felt like they were even close to distorting, but have used them in relatively small domestic environs. There is sometimes a bit of that "one note" quality that people don't about bass reflex enclosures -- but for reggae and electronic, that's sometimes what you want. Forces of Victory sounds great cranked. Acoustic jazz bass sounds quite good.

If I were starting from scratch with the same drivers, I'd build Martin's project 5, or a tapered version a la project 6.

Here's another guy who's happy with FE207E in simple vented cab:
http://db.audioasylum.com/cgi/m.mpl?forum=bottlehead&n=99712&highlight=fe207e&r=&session=

 

[holdent]

I built a Bob Brines (http://geocities.com/rbrines1/Products.html) FB-16 which is a BR design for the FE167E and I'm very happy with it. In retrospect I probably should've gone for FT-1600 MkII which are MLTLs for the FE167E -- normally you have to build stands for BRs so you might as well build a cab that sits on the floor and only costs a little bit more.

The nice thing about picking commercial design like Bob's is that
you don't have to worry about cab design or design of the BSC.

Martin King discussed the issue of Xmax in his Project 2 that Dumbass noted above (p. 34):
"The bass went very low in frequency but it was not the type of bass output that you can feel and hear. After all, these are only six inch drivers [the FE164, FE167and FE168 Sigma] , with a limited Xmax, so the amount of air being moved is not tremendous."

 

[holdent]

I modeled Martin's Project 2 using the FE167E (instead of the FE164) using Martin's MathCad models (www.quarter-wave.com) with very similar results. The FE167E really is a good drop-in replacement for the FE164.

Martin has been refining his MathCad models since the Project 2 design and you can now get the plotted SPL response for the system. This gives you an idea of how the cab will behave while interacting with a rear wall and the floor. Unfortunately the Project 2 design doesn't fair as well when the back wall and floor response are included.

This is what I got with rear wall set to 100 m behind the cab (the best modeled response):

An externally hosted image should be here but it was not working when we last tested it.

I think this cab is tuned too low (?) and the sensitivity below 150 Hz suffers - a trade off for additional bass extension? Clearly there is also a problem with the "double-humped" rise in SPL above 150 Hz.

 

[holdent]

continuing along in this stream!....

I believe a better design for the FE167 may be a straight MLTL with the following parameters:
L= 28 in Zdriver = 12 in Zport = 24"
So/SL = 6 in x 12.5 in rport = 1.5 in Lport = 3.0 in
stuffing density = 0.25 lb/ft3

This design has approx the same depth as Project 2 but is half the height and much narrower. I've deliberately traded off bass extension for sensitivity but still have a rising SPL problem above 700 Hz:

An externally hosted image should be here but it was not working when we last tested it.

(I assumed 3/4" stock).

May be the problem above 700 Hz could be solved with a notch filter? Hopefully Martin or others will be following this thread and comment.

 

[holdent]

I looked again at Martin's "Recent Design Improvements..." section for Project 2 to double check and see if he'd implemented a notch filter. The actual far field freq response that he obtained for this design does not exhibit the "double hump" response my model predicted... He found it necessary to add a BSC but not to do anything else.

Any guesses why my model went wrong?

[Incidentally I can't overstate how much fun I've had with the MathCad models and how valuable I've found them to be! They are highly recommended].

 

[MJK]

Any guesses why my model went wrong?

I don't think your model went wrong. With respect to the rolling SPL response above 150 Hz that is seen in the ML TQWT calculated response, I don't think this is unique to the ML TQWT enclosure. These are cause by floor reflections, the path length difference between the 1 m distance from the driver to the listening position and a path that reflects off the floor before arriving at the listening position creates a series of humps and valleys in the SPL response. If you change the listening position to 2 m or 3 m the rolling response will start at a different frequency. Any full range driver mounted in an enclosure at approximately ear level will have a similar calculated response. A notch filter is not the answer because the frequency of the humps and dips will change with distance to the listening position. If you look carefully at the measured responses in my project 2 documentation you can see hints of this effect but the scale is not big enough to clearly be able to identify the peaks and nulls. The default set-up in the ML TQWT worksheet is for this design and shows the corrrected SPL response achieved with the BSC circuit.

With respect to your alternate ML TL design. The driver appears to be much closer to the floor so the rolling response will start higher in frequency. If you place your design on a stand to put the driver at ear level I think you will see the same type of response as the ML TQWT design.