how does one tame fostex drivers ?

[zobsky]

i thought of posting this question since many fostex drivers can have a piercing sound that is awful on some music (but strangely doesn't become offensive on others)

for exampls, damar and a phase plug (and the whizzer foam mod to some extent) evened out the frequency response on my FE166E in backhorns but i can't help feeling that there is a peak somewhere in the response perhaps in the midrange or low end of the treble, ... (had the same feeling with the whizzerless FF165K when i had it) . reading various bits of info on the web, i'd like to think that some other fostex drivers may have the same issues.



can any experts comment, Mark McK, .. I'd love to hear what you have to add.

One thing I can think of (but haven't tried) is a notch filter, .... any comments (curiously, no posted results out there for the popular 166e driver)

but are there other alternatives ??

[MarkMcK]

zobsky,

Information will be coming on several Fostex models. I have, however, tried to include others in underwriting (to some small extent, one driver) the work.

That puts me at their mercy for when I will be able to test the driver. As soon as I receive the promised driver, I will be able to say more about the performance of a very small sample of drivers (one per model), and about consistencies and differences across what is also a small sample (three) of all the drivers Fostex markets.

Best,

Mark

[Sjef]

A parametric equalizer wilp help a lot. I know that most of the people here are instantly shocked by hearing the word equalizer but it can cure so much when properly used. It sure does cure the shouting sound of all the Fostexes, Lowthers and others out there.

[TC]

i thought of posting this question since many fostex drivers can have a piercing sound that is awful on some music (but strangely doesn't become offensive on others)

Are you sure the sound is not simply a more faithfull rendering of your system?

***What is the rest of the system?****

Fostex can sound sharp, tilted up.

I find amplification that works with Fostex, rarely works with anything else, and vice versa. Cabling can also have sharp brightness (exposed brilliantly with Fostex). Ditch the silver if there is any in the system, that rarely works.

TC

[doggy]

Just got the ff165k's in the mail today. I have the fe206e, fe126e.
This driver does not have as much upward tilt as the other fostex drivers I have and at this point these are brand new and don't have the same harshness. They may not have as much treble but at least they can be listened to without wincing. female voice sounds nice. These are set up in a existing tqwp 47" high cabs I had from before. The cone has a foam surrond which looks more more delicate than the fe series .

cheers
doggie

[coredump]

zobsky,

suggested read: Current Source Amplifiers and Sensitive / Full-Range Drivers

The graph in there shows a peak at ~4000-5000Hz (when driven by an SS amp) for your driver. Could this be your problem?

The document helped me quite a lot designing a BSC fpr my FE108EZ-driven Buschhorn. I was pragmatic about it: I have an amp that I like and given room acoustics, so adjust the speakers via positioning and BSC.

And yes, the FE108EZ has definitely worse issues if not corrected. :-)

Best regards,

Oliver

[zobsky]

thanks all for the replies,

Terry, I know what you mean about amp / component matching with fostex drivers. I've had wildly different results with different amps (and other system components too). Some of the better matches i've tried that take the edge of the sound withough making it dull or boring are the little sonic impact amps, panasonic sa-xr10 reciever, my pilot 903 6v6 amps (although those can be a little strident) . Also, I agree with you to an extent that these drivers can be picky about music that they render well,, partly a function of the recording.


however, i feel the issue with my drivers at this point is more peak related rather than frequency related, as coredump points out (the reason for this thread). definately, your point is completely valid with the stock drivers, which mine are not.
if i can grab a measurement mic, i'll try and run a sweep to see if i can identify the annoyance.


coredump, how do you go about designing a BSC filter?

doggy, i had the ff165k and found them a little treble shy and not as dynamic as the 166e (one look at the curves will tell you that), .. i wonder if a tweeter crossed in at around 8K will help things a lot (they shouldn't beam as much as the stock 166e), .. even then, i found that these drivers sound as a little hot somwhere in the spectrum. this can be cured to an extent by using some felt or other absorband material on the inside of the basket legs facing the rear wave (be careful not to use too much or it will dull the sound)


thanks for all the other replies as well

regards

[TC]

however, i feel the issue with my drivers at this point is more peak related rather than frequency related, as coredump points out (the reason for this thread).

Agreed. Measurment will show a small peak. What your hearing I believe is a larger peak that exists (in a whole system) when music injects more energy into the freq region where the peak is hard to control (midrange). What you need is as close to a current source type situation as possible. Low power triodes do this well, seemingly the lower power the better. Conversely, an old SAE 50 watter works great that I have used in testing several hundred of the 166e since they were introduced here in 02. Not a realistic sounding amp but a good test rig. Also I cannot emphasise enough about -some- amps needing better cabling. SOme amps do fine with any kind of wire+Fostex while others cannot control midrange without creamy wire, Cardas or Nirvana. Damping is needed, electrically, and some wire provides this as does RLC circuits as a last resort.

A simple noise scan or sine wave shows the peak, but a music system can excite the peak potentially far beyond what the test will show. Especially if you listen to Rene' Flemming. Vocals will expose system shortcomings faster than any electrical/acoustical measurement could possibly.

TC

[coredump]

zobsky,

zobsky wrote:
coredump, how do you go about designing a BSC filter?

I am in no way an expert on this, but I can try to describe my way a bit. Firstly, you have to be certain about the annoying freq range; this can be achieved - as you already mentioned - via a freq sweep either via computer/soundcard->amp or as .wav output, burned to a CD (something like 200Hz rise per second should give you a rough idea where to start). Once you know F1 and F2 (the -3dB points of your filter) you could use this and this calculator; they gave me similar values, which I would regard as a good thing. ;-)

One of the most interesting effects was adding resistance between the driver and the amp. Mr. Pass mentioned in a post that he prefers a Gainclone with 1R series resistance. The DC resistance of the coil and the speaker cable I use add up to around 0.9R, which feels 'just right' in terms of control and 'grip'.

I also agree that particular SS amps to not match Fostex speakers; for an interesting experience, try an old Naim Nait1; you would be amazed. :-)

Best regards,

Oliver

Edit: What would actually happen if I calculated a notch filter like a 'normal' crossover? Say, F1 with a coil (calculator) and F2 with a cap (calculator)?

[coredump]

I did some further reading; apparently both ways are OK. The calculator links I gave above seem to work best at attenuating the drivers' resonance peaks. For a high-Q notch filter (meaning: attenuating a broader freq range) you can use the following way:

* calculate the inductor's value using F1 with this calculator
* calculate the cap's value using F2 with this calculator
* attenuate to your likening trying various resistors

Perhaps one of the local speaker gods could comment.

Best regards,

Oliver