Jump to content

Cornwall III journey....


aceinc

Recommended Posts

I became the proud owner of a pair of CW IIIs earlier last year at a great price as part of a larger purchase. Depending on how you value the various pieces I bought, the price could be anywhere from -$1,500 to +$350. They are B-stock and were originally purchased in 2013. I have had them in a 11'x15' den in my home for a number of months and occasionally listen to them mainly with the Mono Block Tube Amps and Tube Preamp they came with. Although I also listen to them with the tube pre & an Adcom GFA 555 Series II I use for testing.

Since I have so little invested in them I thought I would have fun with them. One of the bases/risers is water or rodent damaged, other than that they are cosmetically in 9/10 condition. Speaking of bases, Klipsch wants $150 for one base finished in Cherry. Bob Crites wants about half that for a pair unfinished. Unfortunately the Crites pair seems to be an inch deeper (14-12" vs 13-3/8") than what is currently on the cabinet. I have sent an email to Crites, and am waiting a response. Anybody have recommendations regarding a base(s)?

My initial plan for the speakers is to make the cabinets more rigid, add additional "insulation/sound absorption" and deaden any squawker horn ringing.

During my testing I played various bass/drum pieces to see if the cabinet panels would be excited/resonate. Using touch (not very scientific) I could feel the top sides & back all vibrating during different passages. So I feel some bracing could improve things. Also unscientifically I tapped the plastic squawker horn and the sounds it made seemed like they would be in the midrange (800hz-5khz), so I felt dampening that would be a good idea as well.

 

Here is my testing environment. I close the curtain on the door for testing to help limit reflections.
 

TestRoom1.jpg

Edited by aceinc
  • Like 1
Link to comment
Share on other sites

I have started the analysis phase of the project.

 

I moved one of the CW IIIs into the center of the room and using REW did a scan. I had the microphone about 39" from the baffle and about centered in the mid range horn.

 

From that I was able to get the following graphs. A few things surprised me;

 

  1. How smooth response was.
  2. The boost below 100hz.
  3. The dropout at ~5Khz.
  4. The waterfall plot seems too clean.

 

The boost below 100 hz is probably room gain/node. I have never worked with waterfall plots, but this one looks too consistent. Either everything is decaying too much at the same rate, everything is perfect, or I don't know what I am doing or looking at. I'm going with the latter.
 

CWBeforeSPLPhase.jpg

CWBeforeWaterfall.jpg

Edited by aceinc
  • Thanks 1
Link to comment
Share on other sites

I believe I answered the the reason for the cutout at 5 khz. It would appear that it is driver cancellation. The CW has two crossover points, 800 hz and 5 khz. Based on some of the knowledge gleaned from Danny's Tech Tuesday videos I decided to measure a few times from the same distance from the baffle and centered on the horizontal plane adjusting the height in 4" increments from the center of the mid range horn (31") to the center of the tweeter horn (35") to 4" above that. The results show the cutout frequency shifts. Based on Danny Ritchie's explanation I believe this is because the relative distance from the two drivers changing causes the cancellation frequencies to change. It can be seen in the 800 hz frequency range as well.

 

It's really cool to see with my own eyes stuff I learn. Thanks Danny.

 

 

CWBeforeSPLPhaseVarying heights.jpg

Link to comment
Share on other sites

14 hours ago, aceinc said:

I moved one of the CW IIIs into the center of the room and using REW did a scan. I had the microphone about 39" from the baffle and about centered in the mid range horn...From that I was able to get the following graphs. A few things surprised me;

 

  1. How smooth response was...
  2. The dropout at ~5Khz...

CWBeforeSPLPhase.jpg

 

 

That 5 kHz dropout is due mostly to the multiple-wavelength time misalignment at the crossover interference band around 5 kHz.  If you punch the "spectrogram" button, you will see something looking like this (if you set your preferences up to match):

 

Cornwa II Impulse Spectrogram.jpg

 

Notice the big time misalignment of the tweeter to the midrange at 4-6 kHz.  This is what is causing that dropout in your measurement.  If you wish to do something about this, you have two choices:

 

1) physical alignment of the acoustic centers of the midrange to the tweeter (moving the tweeter to the top of the cabinet by releasing it inside the Cornwall and moving it outside on top of box, then moving it towards the rear of the cabinet until it looks more like this:

 

Cornwall bass bin + AMT-1 6dB per oct spectrogram on-axis.jpg

 

or,

 

2) you can invest in a DSP crossover and tri-amp (or at least bi-amp the tweeter and midrange/woofer as the two channels) the Cornwall's drivers to digitally align the tweeter to the midrange using tweeter channel delay (about 0.75 milliseconds in the case of the first spectrogram above--but the CW III uses a shorter midrange horn and probably a higher order crossover filter that adds something like 90 degrees of phase misalignment (HF always leads the LF drivers using IIR/passive crossovers), so the delay will be different for your CW IIIs.  Digital delay will get the tweeter and midrange time aligned not only on-axis, but also off-axis, unlike the physical alignment case "1)" above.

 

Try physical alignment...for grins...and listen to the difference in the soundstage coalescing into a much bigger subjective image. I recommend listening to a mono Cornwall to hear it most clearly.

 

Chris

  • Like 1
Link to comment
Share on other sites

Chris,  

I would prefer not to either change the aesthetics/cosmetics of the cabinet (moving the tweeter outside & further back) nor bi/tri amp with a DSP. This limits the options.

One option might be to lower the crossover freq and steepen the curve, if the tweeter horn can handle that. This would help with the combing artifact caused by the distance of the center of the mid & tweeter horn being longer than the wave length of the crossover frequency introducing cancellation as well.

One half baked crazy idea I had is, What would happen if the compression driver was separated from the horn and an appropriate length of pipe were inserted between the driver and horn? Would the pipe act as a Hemholtz resonator or otherwise muck up the sound?

Link to comment
Share on other sites

3 hours ago, aceinc said:

One option might be to lower the crossover freq and steepen the curve...

This will make the problem worse in terms of time alignment, i.e., it will introduce more midrange phase lag--which is the source of the problem, and create a mismatch in the polar coverage below 5 kHz.  It will also create more SPL disturbances and subject the tweeter potentially to damaging transients.  It likely won't sound very good, either.

 

3 hours ago, aceinc said:

What would happen if the compression driver was separated from the horn and an appropriate length of pipe were inserted between the driver and horn? Would the pipe act as a Helmholtz resonator or otherwise muck up the sound?

I don't believe this is a very good idea. (But you can do anything you want--they're yours.)

 

3 hours ago, aceinc said:

I would prefer not to either change the aesthetics/cosmetics of the cabinet (moving the tweeter outside & further back) nor bi/tri amp with a DSP. This limits the options.

This actually eliminates your options to solve your 5 kHz disturbance. Klipsch has already minimized the issue in mono-amping with passive only crossover filters and no time alignment. I wouldn't change it if you don't intend to eliminate the source of the problem.

 

I would think that just listening to them will likely be the best option. Measurements and machinations about them can thus be avoided. 

 

Chris

Link to comment
Share on other sites

The fun has begun. So far I have;

  1. Dampened the mid horns with Butyl Automotive Sound Deadener.
  2. Dampened one woofer with the same material.
  3. Added braces front to rear under the port shelf on one cabinet. They support the shelf and rest on the bottom of the cabinet.

Things to be done;

  1. Stiffen the sides, top and back.
  2. Replace foam rubber stapled sparsely to the interior of the cabinet.
  3. Remove and determine if the tweeter can be "lengthened."

What are folks thoughts on foam rubber vs rock wool for the interior of the cabinet?

 

I have pieces of a mattress topper which is quite thick I could line the cabinet with. I also have most of a bale of rock wool.  The rock wool looks more challenging, as I think I need to trap it behind cheese cloth and staple that to the inside of the cabinet walls. The foam I can either use adhesive, or see whether my staple gun will penetrate the thickness and bite into the wood.

 

Link to comment
Share on other sites

They say it isn't real unless you post pictures. There is a 2 MB total limit to pictures per post so this may take a few posts.

Things you'll see in the pictures; (may not be the same order)

 

  1. Squawker & woofer treated to stop ringing.
  2. Braces used for one cabinet.
  3. Inside of cabinet before & after bracing (after removing foam, and before putting new foam in).

 

20210103_150253.jpg

20210110_121741.jpg

20210117_113530.jpg

20210117_113813.jpg

20210117_140528.jpg

Link to comment
Share on other sites

The squawker damping was applied to the panels which seemed to resonate in the bandpass. I intended to do the entire woofer in deadening material, but I did every other section first and the sound of the steel frame when struck with my fingernail went from almost bell to somewhere between a tick and a thunk.

 

The angle braces I used were all from various pieces of  1.5" scrap lumber (yellow pine) which I had laying about. They were cut to 1.5"x1.5"x9.5" long and had a 45 degree angle cut in each end. I used 3 up each side and 2 on the top. There was already two 2x4 (nominal 1.5"x3.5") yellow pine braces connecting the baffle (motor board) to the back. The braces in the ports were fabricated out of some primed 3/4" pine I had laying about. These were the most painful to install, I would show a picture those, but I guess the 2MB limit is per user, not per post.

I have not listened to the results yet, because the cabinets did have one issue I needed to deal with, the bases. When I purchased them the previous owner showed me the bases had been water damaged, they looked good unless you tuned the cabinets over.. I have fabricated one base and replaced it. The other base is built, but is waiting on paint to dry and holes to be drilled. The new bases are made from 1x3 (nominal 3/4"x2.5") finger jointed and primed pine painted black Duratex which covers my inadequacies as a cabinet maker. This increases the cabinet height by ~1/2" taller.

I ran REW on the same cabinet I had run previously (its base was the first one). There appear to be subtle differences in the waterfall plot at frequencies above 120 hz. Some of the slow decay has been reduced in the time domain which I believe will translate into a positive change during listening.

  • Like 2
Link to comment
Share on other sites

@aceinc

 

You can try to fill this hole at 5Khz. I did the simulation on VituixCAD and at home it sounds good. So to do this put a 6 or 8µF capacitor instead of the 2.5µF or 1.75µF capacitor depending on the version of your filter. Measure the curve and tell us if it's ok, but many prefer to listen to the steep slope with 1.75uF-0.2mH-2.5uF than the softer one with 6uF, even if the latter is better: in the end your ear will be the judge.

 

hajx.png

 

nhgr.png

 

nb Thank CANT for scheme

 

 

 

It's here

xrzr.jpg

 

If you don't want to take off the capacitor in place, you can put in // 1 or capacitors to make the desired value. For example you add 4uF to 2.5uF already on the network and you get 6.5uF. to have 8uF add 5.5uF to the 2.5uF in place. Same reasoning if at home it is 1.75uF on your network.

 

edit : On my CWIII with the DE120 I put 3.5UF (2.5 + 1uF) and I inverted the phase + on - and - on +, and with Lpad.

 

 

😎

Link to comment
Share on other sites

mustang, this looks like an interesting mod and simple enough I could do it myself. I will pull the plate off the back and see what I can see.

I am not an even an amateur regarding electronics so I will talk here in generalities.

Does the current curve at the squawker, tweeter crossover leave a gap, which the suggested modification of reducing the steepness of the slope addresses?

One test I did and documented above was to move my testing microphone vertically in three 4" steps. This caused the cutout frequency to shift downward. From what I think I understand, this would seem to show that the the two drivers are cancelling each other out around the crossover frequency and when you change the relative distance between the two drivers it would cause the cancellation frequency to change. If this is true. wouldn't making the slope gentler increase the width of the cancellation and make the cutout wider?

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...