La Scala Mod Project Completed

[mwiener]

I just completed this project doing a number of mods to my 77 La Scalas.  During the last two years I built the La Scala risers for bass extension, added the ALK CSW crossovers to replace the aging AA crossovers, and I replaced the K77 horns with the Eliptrac DE120 also from ALK.  Thanks Al K. for the recommendations, it was spot on.

 

While doing this, I measured the incremental improvements in each mod to make sure that there was actually a measurable improvement.  Chart showing frequency response improvements below.    I can say now after the fact that I am really happy with all of the mods.   I still have a little work to get to my perfect 2 channel setup, but the work I did hear in 2 years eliminated about 90% of the gap between where I was - to where I want to be.

 

For the frequency testing I used a Studio Six Digital iTestMic which is a calibrated test mic equivalent to having a pro Type 2-class test & measurement mic.  The mic is $200 from Audio Control.  It plugs into the iPad directly and bypasses the mic on the iPad.  Also on the iPad you run an app from the same company called Audio Tools.  It is a $20 purchase on iTunes.  The Audio Tools app has a number of applications, the ones I use the most are FFT, Real Time Analyzer, SPL Meter, and Signal Generator.  The SPL Meter for home theater setup beats the radio shack analog meters hands down.  I got much better results with that application and the iTestMic.

 

When you buy an app in iTunes, you get the use of that application on up to 5 physical devices that are tied to the iTunes account.  When I bought Audio Tools app, I have access to it on all 5 Apple devices I own.  That allows me to run pink noise on the Signal Generator on 1 iPad and feed that through the stereo by using a standard RCA input into an unused preamp input from the iPad headphone output.  I run a second iPad with the iTestMic running the Audio Tools FFT application to measure the results.

 

With Signal Generator you can do sweeps, specific frequencies in a sine wave or square wave, white noise, and pink noise.  I used pink noise for my tests which is equal intensity across the entire spectrum 20-20,000 hz on the signal generator application.  Ideally the graph resulting from the Audio Tools FFT application would be a straight line representing exactly the same frequency response generated from the La Scalas  across all frequencies 20-20,0000hz as heard by the iTestMic and the FFT application.

 

My testing was as simple as setting up the volume on the amp, running pink noise on the signal generator at a pre set volume, putting test microphone on a tripod connected to second iPad running FFT, and testing each configuration one after the other without changing anything.  I used the screen capture capability of the iPad to capture the frequency response in a jpeg image.  I did get some coaching from Andrew at Digital Six Studio, I can share the details for FFT setup if you want them.

 

Keep in mind these results are specific to my room.  My stereo in your room could be perfect, or it could sound horrible.  Also keep in mind low frequency measurement is dependent upon where the microphone is physically positioned.  Move the microphone around the room, you change the measurement.

 

Here is a picture of the testing setup:

 

I built my risers a little different.  I shortened the riser by 3/8” so I can place a grill using the same speaker cloth material as the material covering the tweeter and horns.  BTW, I am looking for some black Heritage #17 cloth if anyone has any.  Eventually I will make the grills and hold that in place with small neodymium magnets.  I also used the Precision Port flared 4” port tubes from Parts Express.  I started with the port tubes at 8” instead of the recommended 7”.  I can always shorten them, but to my ears the 8” sounds like a good compromise between the lower frequencies and the fast bass response I am used to in my La Scalas.  Interestingly with the flared ports, if you want an 8” port, you actually make the tube structure 9” because they do not count the second flare on the inside as part of the tube length.  I also used ¾ dowels to add support around the doghouse.

 

You can see the Precision Port 4” flared  port I used at Parts Express, it is part number 268-352.  You can also Google 268-352 and see it that way.

 

I have switched back and forth several times. To make it easy to switch back to stock, I chose to put the sealing foam that goes between La Scala and riser on the bottom of the La Scala speaker itself.  Then I can place the stock speaker on the original plywood base, or I place the speaker on my riser.  I can clearly hear a loss in the low end without the risers, although with the test information you can see that the differences are not huge, they are measurable, and to my ear the difference is greater than indicated in the graph.  Keep in mind that I listen 10 ft behind where the test microphone is placed, this could account for hearing a bigger improvement than is depicted on the graph because the bass response may be greater at that listening location.

 

Here is a picture of the riser build.  This shows the inside of the riser before I put top on with the doghouse cutout.  The top is glued to those dowel sticks as well.  It is pretty solid when all put together.  The epoxy I used is really good stuff, we race wood boats made with it and it lasts for 10s of years running at or over 100 mph on the water.

 

 

Here is a picture of the finished speaker and riser:

 

 

If you want the information on the finish used on the risers, you can find that at:

https://community.klipsch.com/index.php?/topic/148545-any-consumer-black-lacquer-spray-paint-that-matches-lascala/

 

I took measurements after each mod, but probably the best one is the comparison between the stock speaker and all the mods.  To do that I had to do a photo shop modification of the stock measurement in one layer and the last measurement in a second layer with both graphs visible.  This is where the differences are really apparent.  The pink line is the original stock measurement, and the white line is the measurement after all the mods.   The CWS cross over gives you the capability to bump both the tweeter and the squawker up by 1db increments.  I currently have the squawker set to 1db above the factory preset of -6db I can see a spike at above 650-700 hz that I might be able to fix by pulling the squawker down to -6db.  I will get to that eventually, but I suspect that while I might fix that anomaly, I might create other problems.  I set the Eliptrac DE120 tweeter as high as it would go.  You can see its better than it was, but I still have an anomaly I would like to fix at 7k where it steeply drops off then comes back at 8k.  I like a hot tweeter, and this is getting close to perfect.  I used to not like any recordings I had without adjusting the treble, but now depending on the recording I am pretty happy with it totally flat with the tone controls out of the music path.  If I could get a little more out of the tweeter, I would probably like all my recordings without treble, that is why I am thinking if I can fix that 7k drop off, it might be perfect for me.  Looking for any suggestions on flattening out between 7k and 8k.

 

 

 

Let me know if you have any questions. I certainly have a lot more data, but I tried to keep this as brief as possible, otherwise it would be a pretty long and potentially boring read.

Edited June 12, 2015 by mwiener

[PrestonTom]

It is not boring by any means. I am not familiar with your software package but do you have a means to run an impedance measure? This can help you fine tune the port length or stuffing.

 

Also, in the original recommendation by DJK on the ported design, he advocated using a filter at the lowest frequencies. IIRC, this was to limit the driver excursion and to also boost the bass a bit. Have you implemented that or are you planning to? Most folks forget that part of the design.

 

Anyhow, congratulations and good luck, I admire folks who actually are willing to measure the results of their mods,

-Tom

-

[mwiener]

Thanks for the reply Tom.

 

The audio tools software does have a speaker impedance app and an impedance plotting app for $29.  I would give it a try, but it currently only works with the  iAudioInterface module which is the big brother to the iTestMic.  It is roughly twice the price but adds a lot of additional functionality.  An adapter to use the iTestMic with the impedance test is in the works, but no idea when it will be available.

 

I do remember reading about the filter, but I could never find any details.  I have no idea what would be involved, but I would be willing to give it a try to see what kind of improvements I can make.  Keep in mind the last set of mods took 2 years with the time I had available, so it might be a while before I report back.

 

Mike

[djk]

"but I could never find any details."

 

[mwiener]

Thanks djk.  As soon as I saw that picture I started thinking about my pre amp.  Its an old Adcom GTP-500II.  It has a low pass filter built into it.  According to a review done in 2011 on that unit that I found doing a Google search, the low pass filter rolls the frequencies below 28 hz.  I think it was meant for feedback or imperfections in records.  I was really surprised that a full analysis on a 20 year old pre amp was done in 2011.  I think it is against forum rules to put the link in the response.  If you Google adcom gtp 500 II it is one of the first hits.  Would you agree that the 28 hz low pass pre amp filter would work?

 

The other thing I think I want to do is use the normal output out of the preamp versus the lab output.  I am currently using the lab output to an Adcom GFA-545II.  I am going to research that a little more before I switch it.

 

Thanks for the suggestions.

 

Mike.

[Full Range]

Very nice project that is similar to my La Scalas

Difference from yours would be in the port length as I currently have 11" with the Crites 15" cast

The bass bin makes a huge impact in the lower register

[Horn Fanatic]

I have seen this mod being implemented for quite some time. What I find odd, and it's only a guess, is the chap who first came up with the mod obviously has not read any papers written by Paul Klipsch on horn design. By turning the La Scala into a vented system, the reactance annulling provide by the chamber Klipsch wrote about is essentially null and void. There is a formula Klipsch cited that he borrowed from Kellogg that specifies the chamber volume based off the cut-off frequency, ergo, the length of the horn where the area from the throat approximately doubles. I use this formula as a redundent back up along with volume calculations based off the work of Keele and Beranek which are integral to a horn design program I wrote. The chamber volume and throat area are inter-related. Changing one or the other throws the horn design parameters out the window. A larger chamber volume translates to an increase in throat area.

 

If Klipsch saw fit to turn his folded horns into vented system, it would stand to rerason he would have done so.

 

Just sayin'...

 

H.F.

Edited June 13, 2015 by Horn Fanatic

[Full Range]

Ask my ears what they like best ? And they will answer

For the room the speakers are installed - we like the La Scalas better with the bass box mod

Some things are what they are - they work or they don't

A modification can only be proved or debunked if it's built and implemented to test - simple as that

Question - Do you think the bass box would still be fitted to my speakers if the sound was worse

[Horn Fanatic]

I never suggest the mod was a bad idea, just stating some facts about horn design. I see the mod as compensating for the shortcomings of the La Scala, which by horn design parameters is not so much a horn, but a discontinuity. I've heard plenty of La Scalas, and I think the stock design would benefit more by being mounted in a wall to reduce the solid angle. I also think the La Scala would benifit more by including angled panels in the rear corners. Doing so can inhibit back reflections off the side panels, and perhaps get rid of that 250-500 Hz dip in the responce. The 22.5" internal width of the La Scala translates to the wavelenght at approximately 602 Hz. Half that wavelenght translated to approximately 300 Hz, where that frequency bouncing from sidewall to side wall can cause resonance. Then again, the mouth area may be the culprit.

Edited June 13, 2015 by Horn Fanatic

[Full Range]

Fair enough

Good of you to clarify

[djk]

"the reactance annulling provide by the chamber Klipsch wrote about is essentially null and void."

 

Obviously someone that has not examined the LaScala in Hornresp (it is not annulled properly).

 

If you reduce the back volume of the LaScala it will be properly annulled, and it measures better too.

 

" I also think the La Scala would benifit more by including angled panels in the rear corners."

 

Pretty good speculation, but it doesn't measure that way.

 

" I think the stock design would benefit more by being mounted in a wall to reduce the solid angle"

 

Again, Hornresp doesn't agree.

 

Stock LaScala, Crites woofer (data from Carl Huff).

 

 

LaScala with back volume reduced by a third.

 

 

Hornresp has been vindicated! Note that the amplitude response of the short horn flattens out, much like it did in the 32Hz ported version. Clearly the Klipsch La Scala can be improved by reducing the volume of the sealed enclosure.

 

Reduced back volume with padding behind woofer (2 foot by 2 foot sheet of 1 inch foam to the legacy sealed cabinet. The notch at 200Hz is all but gone and the one at 575Hz is substantially reduced. And notice the new gain above 700Hz in the red circle).

 

 

 

The notch at 575Hz pulls up but we lose amplitude at 1kHz which may or may not be important dependent upon the upper horn or waveguide that you have in that cabinet.

 

These are not in 'living room' measurements. These are open space 'in garage' measurements. The cabinet is positioned near an open garage door so as to avoid room effects.

 

************************************************************************

My buddy Thermio and I both scratch built a set of LS bass cabs and he did the djk mod to them. He took measurements on one cabinet as the mods progressed. Here is the data:

 

200 160 125 100 80 63 50 40 31 25 20 HZ

77 80 79 73 72 78 68 60 61 52 51 dB Stock, mouth braced
76 78 78 73 69 75 65 60 66 56 52 4.5 cuft, no ports
76 78 77 75 74 80 70 60 68 54 54 Two 4"dia 10" ports
76 79 78 76 75 80 71 60 68 54 53 Ports cut to 7"

 

The interesting thing with 10" ports is the boxes went deep, but the snappy, fast horn
sound was lost! Cutting the port length to 7" brought the snap and speed back to same
as stock, plus the lows remained, esp. at the 31Hz box tuning point.

 

This data is withK33E driver and light stuffing. We think a K43E in this box would be superb, but haven't tried it yet.I'm going to try the mods using two 4" ports with both ends flaired (Thermio used 4" PVC). Flaring
is supposed to reduce distortion and possibly increase the gain. Parts Express sells the kits.

 

Neither of us is going back to stock.

 

Pete

***************************************

Note: the 7" port length tunes to about 35hz, the EV Interface A box is also tuned to 35hz with a Q=2. That will pick ut the bottom octave by about 6dB.

 

Sentry III boxes may be used by changing two resistors per channel.

 

The Adcom (or any other piece with a low-cut filter) may also be changed with only two resistors per channel.

 

Measurements by Thermio and Pete show about 7dB more gain over stock (sealed) at 31hz using the dual 4" dia by 7" ports.

 

Note: Carl Huff had the mic in front of the speaker, while the ports are on the rear, the speaker was also one foot off the floor, both of these will affect the measurements below 100hz. Thermio and Pete were measuring in room, and it showed much better low end. Their room had a null at 40hz that showed up in all four variations they measured.

 

Edited June 14, 2015 by djk

[Full Range]

In the near future I will devise a sliding port length system

I want to see how different port lengths can change the bass signature

Because all the drivers used are not Klipsch. Then a little experimentation is what's needed

[mwiener]

Well, I got a little time this afternoon to try the lo pass filter.  I definitely would not use it for listening.  The bass response drop is readily apparent as soon as you put the low pass filter in.  I spent a few minutes to use the tools to plot out the difference.  The pink line is with the lo pass filter and the white line is without.  There is a full 1db drop in SPL as well with the lo pass filter engaged.

 

 

This test was done with the normal pre amp outputs.  Instead of using the audio tools signal generator to generate the pink noise, I cheated and used a pink noise recording on a test CD.  It was a little quicker, and I did not have to fight my wife for her iPad.

 

I did find out that the normal pre amp outputs roll off frequencies below 10hz, while the lab pre amp outputs go all the way to zero.  Not sure that I see any reason not to use the lab pre amp outputs.

 

Based on this test, I don't think I would want to use a filter.  I am open to other suggestions if you don't think this was a valid test. 

 

Mike

[wvu80]

These are not in 'living room' measurements. These are open space 'in garage' measurements. The cabinet is positioned near an open garage door so as to avoid room effects.

 

Thread Drift:  Hey DJK, did you put a bunch of hyperlinks in your posts?  Most seem to point at ebay listings.  I've never seen that before on posts within the Klipsch forums.  The word "open garage door" had a link to a completed ebay listing, but the link did not follow my quote from djk.

 

(see post 11 for hyperlink reference of "open garage door")

 

I just hope something weird is not infecting the forums.

Edited June 14, 2015 by wvu80

[Full Range]

These are not in 'living room' measurements. These are open space 'in garage' measurements. The cabinet is positioned near an open garage door so as to avoid room effects.

 

Thread Drift:  Hey DJK, did you put a bunch of hyperlinks in your posts?  Most seem to point at ebay listings.  I've never seen that before on posts within the Klipsch forums.  The word "open garage door" had a link to a completed ebay listing, but the link did not follow my quote from djk.

 

(see post 11 for hyperlink reference of "open garage door")

 

I just hope something weird is not infecting the forums.

I see nothing

However it sounds like malware infecting your browser with key word recognition

[djk]

"However it sounds like malware infecting your browser with key word recognition"

 

Must be. I'll have to get someone from IT to look at my computer (I'm at work).

 

Edit:

 

Got rid of the hotlinks by copying into Notepad (which stripped off the links), and copied back to here.

Edited June 14, 2015 by djk

[djk]

"Based on this test, I don't think I would want to use a filter. "

 

Looks like you have enough room gain that you don't really need the boost.

 

I still recommend a filter if using records and a high powered amplifier.

 

As previously mentioned, filters may be changed to have any shape required. A Q=1 filter would be 0dB down at it's cut point, with a fraction of a dB lift in the 40hz~50hz region (assuming a 28hz~31hz cut point).

Edited June 14, 2015 by djk

[zobsky]

In the near future I will devise a sliding port length system

I want to see how different port lengths can change the bass signature

Because all the drivers used are not Klipsch. Then a little experimentation is what's needed

These are what I will be using http://www.parts-express.com/parts-express-speaker-cabinet-port-tube-4-5-16-id-adjustable--260-329?utm_source=google&utm_medium=cpc&utm_campaign=pla

[mwiener]

Those adjustable port tubes would probably be easier to install than the flared ones I used.  I needed a 6" hole saw to cut the port tube opening, then I still needed to use a router with a rounding bit to open it up enough to fit the flare through the opening.

 

One positive thing about the ports I used, you can buy each piece individually or you can buy the port tube as a complete unit. So you can cut them to 6", then buy the pieces to patch them to 8" by buying a joiner piece and another section of 4" tube.  When you find the right length, you use PVC plumbers pipe glue to put together permanently.

 

Admittedly, I glued all but 1 joint.  The idea being I could trim down to 7" at that non glued joint and put them back together if I did not like the results at 8".  I don't intend on making any changes to the length at this point.  I think I just got really lucky hitting a good length the first time.

 

Mike.

Edited June 17, 2015 by mwiener

[Full Range]

Thanks for the tip and link Zobsky

I will look into it