[kfr01]

Sounds like calculator time for you. Sorry, this one should be simple math.

     Thats fine if you are good at math...

Me ?   Not so good  

Add the volumes of the rectangular areas inside the enclosure ... subtract any braces, etc.  

Work in inches.  

Volume = l * w * h

1 liter = 0.035 ft^3 ~= 61 in^3

Use target volume in in^3 to determine the number of inches to shorten up the rear...

Now pull that calculator out, the LCR plans, and a pencil.  You can do it.

[Timoxx4]

Ok well i had a go at it and found it needs to be about 6.5" deep ?

Dose that sound right or close?   That's an awful lot smaller than i was expecting. That's a good 10" smaller in depth from standard  

By the way, while i am here,  I have almost finished building the cabinets for the 2461's i have and am wondering if you are supposed to use a gasket or something where the speakers mount?  The tweeter has one already but all the other drivers don't?

Do you guys use gaskets or not?    

If so what do you use?   Whatever is used would have to be very thin as the flush mounting rebate doesn't allow for any extra and i have already cut the rebate so cant go any deeper now  

Thanks guys,
Tim.

[kfr01]

6.5 deep could be right.  Sealed enclosures can be relatively small.  Did you remember to subtract out the tweeter / midrange chamber?

I use this gasket tape for woofers:

http://www.partsexpress.com/pe/showdetl.cfm?&DID=7&Partnumber=260-542

[Kevin Haskins]

T/S parameters are remarkably accurate at capturing the small signal performance of a driver.

Krutke is not "some guy on the Internet"... he's a professional engineer, just as Dan Wiggins is. This is science, not a popularity contest. No need for personal attacks on either person; it adds nothing to the discussion.

The reality is that Adire's numbers are suspicious. Take a look at their own posted impedance plot:
http://www.adireaudio.com/Home/Images/ExtremisZ.gif
Notice how the impedance peak measured by Adire is around 37Hz, exactly the same as what John Krutke measured. Why, then, does Adire claim a Fs of 30Hz? Adire's claim, using their own measurements, doesn't make sense. Perhaps reality takes a back seat to marketing.

I'm sorry if I'm coming across as a personal attack on Mr. Krutke.   That is not my intention.   I have no association with him and I know little to nothing of his qualifications.   I've seen his web site and measurements and I'm impressed with his general knowledge and the work he has shown on his site.   I find little to disagree with after reading through his "Design Mantras".

I will point out a couple things and why I have an issue with what you are saying.

Fist off the nature of the beast is that drivers come off an assembly line.   The components that go together to assemble them vary in tolerance just like every other product in this world.   As an engineer it is important to understand the range of acceptable values and how they affect your final design.

If we look at Fs it is determined by two parameters, Mms & Cms.   The mathematical relationship is:  Fs =1/(2*PI()*SQRT(Mms*Cms))  I was incorrect earlier when I stated that Cms varies by up to 15%.   It can actually vary by up to 20% in production tolerances depending upon the design of the spider & materials used in its construction.    Mms varies very little... maybe a couple percent.    The result of this is that Fs can vary up to 10% based upon the tolerances of Cms alone.  

Cms also changes with amplitude so you can get another 10% variance depending upon the drive level you are taking the measurement at.   It is quite common to see Cms change another 15%-20% at very high drive levels.    

Now... you are basing your observations upon a test group of a single driver.  It would be more useful to compare several drivers off the assembly line.    A 37Hz Fs would be at the limits of production tolerances assuming a target 30Hz Fs.  If you measured it at low signal levels and especially if the driver wasn't broken in by a drive signal, it could even measure higher and still be within production tolerances.

If you model with the production variations of the small signal T/S parameters you will see a range of different suggestions for enclosures.   This isn't so much a problem with the tolerances in drivers, it’s a limitation of the small signal measurements.   The large signal level T/S measurements vary more than the production tolerances.   So the question becomes which set of T/S parameters should be used to design with?   We choose the large signal level measurements whenever possible because they are more representative of how the driver is actually going to be used in the real world.  

It is also important to understand how the production variation is reflected in actual use.   If the variation of T/S parameters is in the neighborhood of 20%, just to pick a number, how does that affect the final response in the enclosure?    

If you take a group of Extremis drivers, put them in our suggested enclosures and measure the ACTUAL response of the system you will find they vary very little.   Why?  Because in the acoustic system the suspension is dominated by the enclosure, not the Cms of the driver.   If you look at vented alignments the shift in parameters occurs in such a way that they offset in complimentary ways.   In other words, the actual response in the system varies very little with the shift in parameters.  

Your claim that reality takes a back seat to marketing is nonsense.   We make the best suggestions we can for enclosure design.   It is in our best interest to give people as accurate of information as we can so they get optimal results.     We have nothing to gain by doing anything else.

As a final note on measurements I'll take a quote directly from Mr. Krutke's web site.   I agree with his statement because he understands the variability that comes from using different measurement equipment and techniques.   We use Praxis with higher drive levels to approximate real drive signals for modeling our enclosure recommendations.   We use LSPCad to model them.  

All parts of this test are done with Soundeasy 10.0. This is a modeling package with a lot of extra testing applications included. It is not a full acoustic measurement system such as CLIO or Praxis however, so take these results with a grain of salt. What is important however is that all of these drivers were tested in the same way. Because of this, my test results should not be compared with tests done elsewhere by other people, but comparisons within this group of drivers are perfectly valid. That's why tests like this are best done in groups.

There are a few types of tests that are important, but not shown here in the interests of brevity. Keep this in mind and know that these results need careful interpretation.

Make note of the "careful interpretation".

[Kevin Haskins]

Ok well i had a go at it and found it needs to be about 6.5" deep ?

Dose that sound right or close?   That's an awful lot smaller than i was expecting. That's a good 10" smaller in depth from standard  

You can use anywhere from about 20L - 33L without a problem.   That ranges from a Q of about 0.70 - 0.57.    That is much smaller than the ported alignment that should be around 43L.     Use around 12oz-16oz of polyfill because 100% fill is 16oz per cubic foot (28.32L).  

You can be off in your volume for the sealed cabinet by a fair amount and still get great results.   Just get the cabinet in that range and call it good.

Do you guys use gaskets or not?      

Typically I do in small sealed enclosures so you don't get any whistles or noise.   Some weather stripping sold at hardware stores often will do the trick.

[Watson]

Now... you are basing your observations upon a test group of a single driver. It would be more useful to compare several drivers off the assembly line. A 37Hz Fs would be at the limits of production tolerances assuming a target 30Hz Fs. If you measured it at low signal levels and especially if the driver wasn't broken in by a drive signal, it could even measure higher and still be within production tolerances.

Your argument about production tolerances doesn't make sense if you think about it in any depth.

If the target Fs is 30Hz and somehow the two measured drivers having an Fs of 37Hz represent exceptional examples on the outer limits of manufacturing tolerance, then given a typical manufacturing distribution, we'd expect some drivers on the opposite side of the distribution, i.e. having a Fs of 23 Hz.  The problem is that a Fs of 23 Hz for a 6 and a half inch driver is borderline ridiculous.  It's just not believable.

If you take a group of Extremis drivers, put them in our suggested enclosures and measure the ACTUAL response of the system you will find they vary very little. Why? Because in the acoustic system the suspension is dominated by the enclosure, not the Cms of the driver.

Of course that's true, but what does that have to do with the quoted specs of the driver?  What you're saying is true for any driver.

Make no mistake, in a sealed enclosure with a low Qts driver like the Extremis, a difference in Qts of 45% (0.482 measured versus 0.332 in the Adire specs) will cause a major difference in the response of the loudspeaker.  Using the Adire numbers, the tendency will be to make the enclosure too small, leading to a speaker with a Qtc of about 1.02 and a +1.5dB boost around 90Hz.

Again, it's easy enough for anyone with a computer (for playing sine tones), a single resistor, and a multimeter, to measure Qts of whatever driver s/he has in hand.  You'll then be able to see exactly what size of sealed cabinet to use and you won't have to trust anyone's word for it.

[Kevin Haskins]

Of course that's true, but what does that have to do with the quoted specs of the driver?  What you're saying is true for any driver.
...

Sure... the tolerance issue is only good for about -+ 10%.   So if we took 30Hz as the reference it would be 27-33Hz, not 23Hz.    The other part of the variation comes from measurement technique.    If you measure a non broken-in driver it is stiffer than one that has been jiggled at 20Hz for 24hrs.   The offset is always a stiffer Cms, not a looser one.    If you measure at low drive levels the Cms show up as stiffer also.   Once again the offset is always towards a higher reading rather than a lower one.  

For the last set of four drivers I measured the Fs varied between 33-35Hz.   That was with about an hour or two of break-in.  The variation is much tighter than 10% on this particular driver.   If you crank up the drive level those parameters will shift lower yet.    The Qts varied from 0.35-0.38 so it is somewhere in between where the single sample that you are pointing out, measured by someone else and our random sample of four.  

Stick an Extremis in a small sealed box and measure the low frequency response the numbers line up pretty good with modeling the Adire T/S parameters.    How do I know?   I've been doing a LOT of small box measurements with the Extremis lately.

[skrivis]

T/S parameters are remarkably accurate at capturing the small signal performance of a driver.

Krutke is not "some guy on the Internet"... he's a professional engineer, just as Dan Wiggins is.  This is science, not a popularity contest.  No need for personal attacks on either person; it adds nothing to the discussion.

The reality is that Adire's numbers are suspicious.  Take a look at their own posted impedance plot:
http://www.adireaudio.com/Home/Images/ExtremisZ.gif
Notice how the impedance peak [b ...

Looks like the Fs is higher than 30 Hz to me too.

[DanWiggins]

Watson,

I've seen a few Extremis measure out with an Fs at 26 Hz.  Yes, it does happen, just not as often.  In fact, with the way drivers are built, I reject those that measure low to start, because they will come down over time.

The trace we posted was a small signal measurement, like most people will see when they run a Woofer Tester, or do a standard T/S measurement.  However, I do not believe that is really applicable; I believe that you should measure the T/S at a level consistent with typical use of the driver.  With the Extremis, that would be with 2-3V driving it.  And you'll find that in addition to Qes creeping up a bit, Cms will increase too, meaning Vas rises and Fs drops.

Drivers are mechanical systems - they will change with amplitude and frequency input.  They will not behave identically at all frequencies and drive levels - yet.  We're working towards that.  But with existing suspensions that are currently available, I felt this was the best compromise I can come up with.

This isn't coming in terms of marketing-speak; this is coming in terms of an engineer who is hired by companies to design drivers for them.  Right off the production line, at at small levels, I like my Cms to be a bit stiffer than spec.  That way after some use at at higher levels - where the end client will tend to operate the driver - the driver will be closer to the nominal levels.

Dan Wiggins
Adire Audio®

[Timoxx4]

Sup guys,

I have a question about the cross over in the LCR now.

I see the tweeter and the mid are wired backwards.  The + goes to the - of the speaker and so on.    And the woofers are + to +.


My question is i have the 2641 kits aswell but there tweeter and mids are wired the other way.  So this would mean that the LCR would be out of phase with the 2641 mains yes?

So should i wire the woofers backwards in the LCR and the tweeter and mid + to +    - to -   ?  Just like the 2641's are ?

Thanks,
Tim.

[kfr01]

Sup guys,

I have a question about the cross over in the LCR now.

I see the tweeter and the mid are wired backwards.  The + goes to the - of the speaker and so on.    And the woofers are + to +.


My question is i have the 2641 kits aswell but there tweeter and mids are wired the other way.  So this would mean that the LCR would be out of phase with the 2641 mains yes?

So should i wire the woofers backwards in the LCR and the tweeter and mid + to +    - to -   ?  Just like the 2641's are ?

Thanks,
Tim.

No.  Follow the schematic.

[Timoxx4]

Hang on, let me put it another way.


Should i just connect the finished speaker backwards at the binding posts and leave the cross overs as per the schematic? ( which is exactly the same as what i asked before just doing it an easier way)  So they would be in phase with the 2641's then?    If i don't the LCR will be out of phase with my 2641 speakers?    


Or is someone not telling me something and the centre LCR is supposed to be out of phase with the 2641 speakers?   Some how i don't think so.

If i where building 3 LCR kits as left, centre, right.  They would all be wired exactly the same and in phase with each other.

Have i just answered my own question        

[kfr01]

Hang on, let me put it another way.


Should i just connect the finished speaker backwards at the binding posts and leave the cross overs as per the schematic? ( which is exactly the same as what i asked before just doing it an easier way)  So they would be in phase with the 2641's then?    If i don't the LCR will be out of phase with my 2641 speakers?    


Or is someone not telling me something and the centre LCR is supposed to be out of phase with the 2641 speakers?   Some how i don't ...

I think you're thinking too hard about this.  Wiring the tweeters and midrange with reversed polarity at the crossover level doesn't put the entire speaker 180 degrees out of phase with the 2641s.  

Why?  The polarities of the tweeter and midrange were flip-flopped to account for the phase shift (delay) that the crossover filter created.  i.e., the polarity change on those two drivers was done to bring those drivers -back- into phase in the crossover region and in the non-attenuated region.

In other words, simply because two drivers are wired out of phase in a crossover design does not mean that the entire speaker should be wired 180 degrees out of phase with other speakers in the system.

Sure, all filters add delay which can cause some final phase problems in a multi-speaker system.  However, you'd need some more sophisticated measurement equipment to determine the summed delay of the crossover networks in the  speakers.  You'd need even more sophisticated equipment to predict or determine what final result these delays would have on the frequency response at your listening position in your room.  When you remember that differences in distance between the drivers, speakers placement in a multi-channel system, reflections in the room, shifting your head a few inches, etc. can also create phase related cancellations, any small amount of filter delay becomes a moot issue.

To summarize:
1) Kevin's speakers sum flat if wired to spec.  Wire them to spec.  
2) Just because two drivers were wired out of phase to compensate for the phase shift caused by the crossover filter operating on them does not mean that the entire speaker is 180 degrees out of phase with other speakers in the system.  Observe polarity at the binding posts.
3) All filter networks cause some delay, but this surely can't be fixed by changing the phase 180 degrees on the binding posts.  If you're very concerned about your final multi-channel frequency response being very flat at your listening position you need to do some active equalization with a device like the deqx, tact, or a computer.

[Kevin Haskins]

Uhh... what Karl is saying is right on.   Follow the directions and everything will come out peachy.

[Timoxx4]

Ahh now i see....


I did not realise the cross over adds some delay. That explains a lot.

Thanks.        

Now, on with the building  

Later,
Tim.

[Timoxx4]

Update....


Well i just finished building my sealed LCR today      Turned out pretty good i think.  Just a nice size too.

Anyway.   I wanted to test my theory as mentioned 5 or 6 posts back about the phase thing.

I think i was right.  It WAS out of phase with my 2641's     I disconnected one of my 2641's and put the LCR in its place and played some music.   I could clearly tell it was out of phase with the other speaker. So i reversed the speaker wires at the binding posts on the LCR and yep, that fixed it  

So can anyone explain just exactly what the deal is?   You say it should be in phase with the 2641's due to cross over delay but this is clearly not the case as i just did the experiment and found out.     And yes i did follow the schematic to the letter and quad drupel checked all the wiring to make sure it is all good.

[Kevin Haskins]

You may simply flip the polarity of each driver in the LCR.   Flip the woofers, midrange and tweeter and you will still get the same frequency response.   The woofer polarity of the LCR will then be in phase with the woofer polarity of the 2641s.   Easier yet just wire the LCR with reversed polarity and you will get the LCR woofers in phase with the 2641 woofers.

[ezeg]

Timoxx4,

...or anyone

How does the sound compare between a 2641 and an LCR?

Thanks,

[kfr01]

Update....


Well i just finished building my sealed LCR today      Turned out pretty good i think.  Just a nice size too.

Anyway.   I wanted to test my theory as mentioned 5 or 6 posts back about the phase thing.

I think i was right.  It WAS out of phase with my 2641's     I disconnected one of my 2641's and put the LCR in its place and played some music.   I could clearly tell it was out of phase with the other speaker. So i reversed the speaker wires at the binding posts on the ...

Well, there's something to be said for practical experimentation v.s. theory.

In any event, I'm glad you found what sounds best.

Another way to mess with phase problems in a home theater, and easier than actually switching wires, is to play with the time delay settings of your center channel speaker relative to your left and rights.

[Timoxx4]

Timoxx4,

...or anyone

How does the sound compare between a 2641 and an LCR?

Thanks,

I think they sound much the same. I think the 2641 might have a little more low end but you would have to be rather picky to be able to pick it  

I cant really give you an accurate comparison as i have built my LCR as a sealed speaker instead of the ported type it is supposed to be.  So a ported one will go lower than mine.  But the sealed one is actually quite good down low.   If you are using subwoofers ( as i am )  and you cross over the sealed LCR to a sub then you don't miss anything and the sealed LCR has plenty down low to cross at 70Hz, 80Hz.