Notch filter at x'over freq. of woofer?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi, may I ask you learned folk another question please?
Could a parallel notch filter be placed at, or very near the x/over freq.
in a 2way system using a LR-2 network?

Misguidedly and naiivly I bought some 8" drivers cheap 'on special' a few years
ago. It was only recently that I discovered the response graph, and, aarrgghh,
(pls see attachment) they have a huge albeit 'balanced' peak at 2Khz.
So hence my question above. There is no way I would contemplate (another)
'cheap' 3way design and this is what they may only be acceptable for?

Or, is it just possible that they might be ok in a 2way?
Many thanks in anticipation, Grant (...newbie)
(unfortunately, no impedance graph or 'Le' available to Zobel; assume
'Le~~3mH?)
 

Attachments

  • cw-2136.jpg
    cw-2136.jpg
    95.2 KB · Views: 353
You may find that though there is a peak on-axis, when measured off-axis the response is pretty flat. If the driver starts beaming around there, then a peak can help keep the off-axis flat.

I see no reason why you could not use a notch filter in the Xover range though as long as you take into account the impedance (or put it before the Xover, maybe not a good idea).

Alternatively you could just crossover lower down. Have a look at the TangBand 871S, it could crossover at about 800Hz 2nd order and still give very decent SPL and its pretty damn good sounding too, though not ultra high-end.
 
Assuming the graph is correct it looks like about 6dB/8va. You could use a 1st order low pass at about 1kHz. That would make it flat out to 2K and give you a second order rolloff above 2K. Shouldn't be too hard to find a tweeter that can work down to 2K with a fairly sharp filter huh?

Just might be a simpler solution.

Alternately it looks like it could make a usable sub in a 2 to 3 cubic foot ported enclosure if the TS are right.

mike
 
Hi
Peaking most likely the result of cone material/geometry issues. Thinking if you can use crossovers developed for a metal cone type midbass along with a lowish Fs tweeter would be good for a 2 way. (RS180 crossover designs?)
3rd order elliptical lowpass would be an elegant solution to an otherwise 2nd order with a notch.

<edit>Peaking at this freq. for a 8 in woofer would indicate even much more wonky off axis responses, up to 20-30dB variation at 2-3 KHz. So a high order crossover is needed here.
 
Brilliant!!
Great ideas I never even contemplated. You gentlemen have
tremendous insight into response curve issues. Thank you
very much indeed!
Now how to Zobel them? Would, a guesstimate, of 3mH voicecoil
inductance be a 'ball-park' starting figure? I could plug the Re
and that into the Lalena calculator.
 
Hi Tenson, thank you, unfortunately no data on off-axis, but I understand! Crossing lower seems the way to go.

Thanks Mike. I like this idea, corroborates Tenson, so 1st order on woofer at 1Khz and second order(?) on tweeter at 2Khz (have tweets that should do this based on Fs).

Hi Infinia, I will have to 'look up' 3rd elliptical LP filters.

thanks to all for your valuable insights and best regards .. Grant
 
Hi,

Hmmm........ some good advice, some not so good.

The response curve of that driver may not be too bad, it is typical
of a certain type of driver specifically optimised for a simple c/o.

The idea is that a baffle width is chosen where baffle step is pretty
much all over by 1khz. The series inductor for the bass/mid unit is
chosen to implement baffle step. consequently above baffle step
the rising response is compensated and you end up with an ~
2nd order acoustic roll-off where the drivers response stops rising.

So the driver is designed to have a single series inductor.
Full zobelling is not the intention though partial zobelling could
be used to fine tune response, phase or impedance characteristics.

Also note fine tuning needs to take into account the diffraction
ripple, simply put ideally you need to measure and adjust.

The driver categorically does not need a metal cone type crossover.

The drivers parameters also indicate it is a poor choice for a reflexed
alignment. It shoud be used sealed 25 to 35 Litres. If bass extension
is the goal then TL or MLTL or MLTWQT is probably the best option.

:)/sreten.
 
If you choose a simple first/second order crossover for this woofer (not too great LF response here) then a 3-way is better suited. I can garantee at least 20dB off axis response for this driver at 2.5 KHz, so a high order LPF only as high as 1.8KHz, maybe lower would be the sensible aproach for a 2-way. (Hint look at Jon Marsh's designs for a RS180, the one he uses a 3rd order eliptical/Cauer at 1.8KHz) The response of the RS180 (metal cone) is similar to your driver.
A 1st order crossover approach is known for use on well behaved wideband drivers ala Dynaudio and Morel, unlike this driver.
 
infinia said:
...I can garantee at least 20dB off axis response for this driver at 2.5 KHz, so a high order LPF only as high as 1.8KHz, maybe lower would be the sensible aproach for a 2-way. (Hint look at Jon Marsh's designs for a RS180, the one he uses a 3rd order eliptical/Cauer at 1.8KHz) The response of the RS180 (metal cone) is similar to your driver.
A 1st order crossover approach is known for use on well behaved wideband drivers ala Dynaudio and Morel, unlike this driver.


Hmmm......

It would help if you actually knew what you are talking about.
This is not the first time that you given apparently sensible
advice based on a set of premises that are simply not correct.

I do wish you'd stop pretending you know more than you actually do.

:)/sreten.
 
Hi!
I'd like to thank ALL respondents for your very helpful replies!
Unfortunately, my knowledge, at this stage can not fully
comprehend all your replies. But I do thank ALL off you sincerely.

I have the '97 'cookbook' - can anyone recommend a more
comprehensive text or website, as I have spent eons searching
this wonderful resource for some basic definitions, like BSC,
which I presume is baffle step compensation.

I have dozens of questions, but I need to study before I could
make a sensible reply.
Thanks to all of you, most graciously, and respectfully, Grant
 
Hello again sreten and thank you kindly for your considered reply.

I am embarrassed to ask such a probably stupid question but,
(as I haven't found the definitive thread yet) is baffle step defined by
the horizontal displacement of drivers on the baffle, i.e. the tweeter
should be mounted several inches *behind* the woofer,(hence slanted baffles)
or does it mean the distance between the drivers 'frame' and the edge of the
enclosure? and...
Full/partial Zobelling? I would have assumed full was required to get a
flat impedance curve (after resonance)?
Diffraction ripple - will also have to put that on my 'study list' but I assume
similar to 'fluid (water) waves diffracting around an obstruction'!
Yes! I am planning to put these in a sealed box.(but MLTWQT? I'm lost, Lol!)
Thanks again, very best wishes, Grant
(sorry for delay in posting)
 
sreten said:



Hmmm......

It would help if you actually knew what you are talking about.
This is not the first time that you given apparently sensible
advice based on a set of premises that are simply not correct.

I do wish you'd stop pretending you know more than you actually do.

:)/sreten.


Hmmmm...Some weird dynamic going on here. Sreten why do I deserve a tirad like this? You should look to your own posts, and put in more quality of instead of quantity. I am prepared to defend myself from any technical argument you can muster. Sreten with a response like this, you have just proven you are not a gentleman by any means.
 
grantnsw said:
Hello again sreten and thank you kindly for your considered reply.

I am embarrassed to ask such a probably stupid question but,
(as I haven't found the definitive thread yet) is baffle step defined by
the horizontal displacement of drivers on the baffle, i.e. the tweeter
should be mounted several inches *behind* the woofer,(hence slanted baffles)
or does it mean the distance between the drivers 'frame' and the edge of the
enclosure? and...
Full/partial Zobelling? I would have assumed full was required to get a
flat impedance curve (after resonance)?
Diffraction ripple - will also have to put that on my 'study list' but I assume
similar to 'fluid (water) waves diffracting around an obstruction'!
Yes! I am planning to put these in a sealed box.(but MLTWQT? I'm lost, Lol!)
Thanks again, very best wishes, Grant
(sorry for delay in posting)


Hi,

BSC you'll have to search on - but its not horizontal alignment.
Related to BSC it has ripples it is not a smooth progression.
BSC is caused by diffraction, the difraction has ripples.
Start here : http://sound.westhost.com/bafflestep.htm

Many speaker designs do not zobel the bass driver. Full is required
for flat impedance, but partial can be used for flatter impedance
(but not flat) and to manipulate response, phase + impedance.
There is a whole range of possibilities for the series L and parallel
C (with / without an R in series with C) that can be manipulated
between 1st order with Zobel and a second order c/o.

MLTQWT (not mltwqt) is mass loaded tapered quarter wave tube,
a variation of TL's and TQWTs,
see : http://www.quarter-wave.com/
also see http://www.t-linespeakers.org/

Sealed box loading is fairly simple but as Fs is not particularly
low you would need to do something for genuine deep bass.

At this point I must say you cannot base a design on a published
response graph unless you can absolutely trust it. Basing the c/o
design on interacting with the response of the driver is asking
for trouble if the drivers response is not measured and confirmed.

You really should measure the driver, in fact, you need to.
If you can't you have to ask yourself what is the point of
what your trying to do ?

Perhaps you'd be better off simply defining what it is is you
ideally want and finding an existing design to suit, then at
least it wouild be worth all the building effort.

Checkout http://www.zaphaudio.com/ for the speaker
design process , deals with BSC and models diffraction.

:)/sreten.
 
Grant,

I've used a couple similar models to this (jaycar with Carbon fibre cone, Altronics Poly cone), They are at present working well as woofers in a 3 way system, in 30L vented enclosures.

I've used what's called an 'EBS' alignment, which is when you tune the box lower than recommended - this extends the response lower, but you lose a bit of level. I make this up by turning up the bass tone control a tad - works very well in my system.

You can see how this works by playing with one of the software progs that let you simulate a box design.

But I'm only using them up to around 160Hz.

I'd try the 'xover with a hole in it' approach to see how they sound as a 2 way system, then think about getting a couple of nice mid-woofers to make them 3 way (I'm using the old Peerless 850488s).

Cheers,
Pete McK
 
Thanks very much again sreten!
I'm reading your helpful links and taking it all in (but it will take a while, I guess).
Thank you for this (and I quote):

"At this point I must say you cannot base a design on a published
response graph unless you can absolutely trust it. Basing the c/o
design on interacting with the response of the driver is asking
for trouble if the drivers response is not measured and confirmed.

You really should measure the driver, in fact, you need to.
If you can't you have to ask yourself what is the point of
what your trying to do ?

Perhaps you'd be better off simply defining what it is is you
ideally want and finding an existing design to suit, then at
least it wouild be worth all the building effort."

I agree entirely! I was initially wondering if these woofers would be even remotely feasible in a 2-way, maybe I didn't state that specifically enough? And if so, asking for some general help which to now has well exceeded any expectations.
The off-axis response, impedance graph and Le for these
drivers is not shown, so it really does make me question the validity of the published data. I think I may have indeed wasted my money on them and that 'all the building effort' would certainly not be warranted especially considering my current scant knowledge and lack of test equipment. It seems that I should 'junk' them and cut my losses, and as you suggest find a suitable (and well received) 'existing design'. Many, many thanks, Grant

(to the other people who have replied in the last few days, also many thanks
I am trying to reply to your posts in turn)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.