Friday, November 27, 2015

Noes - Desesperation Crossover

What do you do when you have a long weekend, great new speaker drivers and your crossover parts are still a week away?  Improvise!

The Noes design calls for a 2.8kHz crossover point with a 3-5 db tweeter pad, among other things.  My plan was to tinker with the crossover during the long Thanksgiving weekend, and have them close to finished by Monday.  So I had two choices:

  • Do nothing and use the woofer full range
  • Find crossover settings that would work with the parts I had.
Desperate, I chose the latter.  I ended up with a crossover point at around 6.8kHz with the tweeter about 5 db lower than desired.  Why?  I only had 10 ohm power resistors left in my arsenal, and this was the lowest I could cross the woofer over with the parts I had.

How does it sound and look?  Surprisingly good!

Thanks to Omnimic I was able to quickly tell I had to reverse the tweeter.  Not my favorite situation, but desperate times call for desperate measures!

I've had the opportunity to listen to some fabulous recordings via our local classical internet station, KDFC.  The Mundorf AMT's are spectacular, I can't wait to finish the real crossovers.

 The crossover breadboards.  The top section is the tweeter, bottom is the woofer. 

Tuesday, November 24, 2015

Noes - First Mesurements

I still don't have crossover parts, but that's no reason why I can't do some interesting measurements.

Fee Air ScanSpeak Impedance

 Fresh out of the box in free air the speaker resonance is around 49 Hz at around 31 Ohms.  I believe I read somewhere that ScanSpeak specifies the Fs after break in, so I should probably measure this again in a week or so.

To begin with, a full-range frequency response chart for the woofer.  Incredibly clean at the upper end of the spectrum, even when averaged at 1/24th octaves.  This is with the speaker on a chair and microphone 3ft away on the woofer axis.

Next up is the impedance measurement in the cabinet.  The woofers have only been playing for about 4 days with low to medium jaz 24 hours a day.  I'm going to have to go back to Lee to ask for help interpreting both of these.  From the first chart I worry the box is a little large.  Below is the chart I really don't know well, the speaker + cabinet impedance. 

I can see the port resonance at 24 Hz, and a second resonance at 69 Hz.  The valley's nadir is around 35 Hz, somewhat lower than the driver's Fs when first purchased.  ScanSpeak specifies the Fs at 33 Hz after break in, in which case the tuning would be spot on.

I have a week before I can assemble the crossover, so I have time to plan ahead.  With Thanksgiving coming up parts will be slow in arriving.

Of course, there will be more later.

Saturday, November 21, 2015

My First Speakers

Well, at least the first speakers I have designed from the ground up, and I use the term "design" loosely.  Component selection was mine, cabinets are from Lee Taylor of Taylor Speakers, partial crossover design was from Madisound.  I say partial because they designed a crossover for a different tweeter than you see here.

The cabinet is much better looking than this, especially the veneer.  I'm afraid my phone doesn't do the quality of Lee's work justice, so please bear that in mind.  His web site has pictures much closer to reality.

In a sense, this is a fail.  I was trying to design some smaller satellites with these components, and Lee was trying to guide me to a floor stander.  These speakers are really very tall and deep so I'm not sure most satellite stands would work for them.  It's a bit like trying to have a baby and getting twins though, I am not at all unhappy!  I just got more than I was expecting.

Drivers

The original design called for a smaller Mundorf AMT tweeter but that was out of stock and weeks away from arriving in the U.S. so i went for it's bigger brother the Mundorf AMT 25CM1.1-R.  One weird thing about this driver is that it takes banana plugs, not spades or bare wires.

Thd mid-woofer remains the same though, the ScanSpeak Revelator 18W/4531-G00 7" driver.  The 4 Ohm version is more sensitive so mean less damping is needed in the tweeter section.   I chose the Revelator in part out of cost ($220 vs. $320 ea.) concerns but part was also that in the high end of the driver I liked the response better than it's Illuminator counterpart, then when I read that Zaph rated the Revelator better than the Illuminator in terms of distortion that cinched it for this design.  I admit the basket does not look as cool though.  

Despite the larger tweeter having a lower useful crossover point I'm keeping the crossover frequency the same for a number of reasons.
  1. I don't want to redesign the crossover from scratch
  2. Higher crossover frequency = higher power handling
 One big advantage here as well is controlling the vertical dispersion.  The taller the tweeter, the less floor and ceiling reflections and clearer the sound at the listening position.  Like ribbons and electrostatic speakers, we should get enhanced clarity and detail than with conventional round drivers. 

 

Test Wiring

Per Lee's suggestion I'm going to keep the crossover external until I have tweaked it perfectly.  For this to work of course I'll use a separate pair of banana jacks for each driver.  Internally I'm using plain old 14 gauge zip cord which happens to be made by Monster.  I had it lying around from the surround sound installations in previous apartments.  It's thick, flexible, finely stranded and best of all, free and taking up space under my bed in the wiring box.

The white cable with the bananas of course goes to the tweeter, the pinkish cable to the SS woofer.  This being a test harness I didn't want to spend too much effort, but after installation and looking at the kinks in the wires I decided that the heat shrink was going to be necessary for proper strain relief.

The Sound! 

Of course, anyone with a checkbook and drill can get speakers this good looking, but how do they sound?  Magnificently quiet.  You've never heard speakers that disappear like these..... My secret?  They don't work yet! :(  Crossover parts were delayed due to a series of errors on my part but I should be able to start putting that part together soon.  Of course, being unable to resist, I am currently breaking them in by running the woofer's full-range.  They sound surprisingly good by themselves!  I had to go back to the charts and look but they have useful output to 10 kHz so I can run these as my main speakers without going completely nuts due to the lack of treble.

Sunday, November 15, 2015

Focal Profile 918 - Final Charts

I'll be posting more info on this in the DIYaudio forums, but for now here are the pics.

The chart below shows the final measurements, taken from my seating position with 1/24 octave smoothing and a 2mSecond window.  The RED trace is with the grill (as recommended) and the BLUE trace is with the grill removed:


The image below is from the Sound and Vision review with the grill on.  Note in particular the 10K peak and ragedness around 5kHz which are gone, those are probably thanks to the liberal use of felt on the grill and tweeter housing:


And lastly the grill removed, which is the image most make fun of:
Not only is the grill less important, but the tweeter balance is different.  Instead of trying to push the tweeter up at the last octave, I've gone for a smoother response from 2-10k, that goes generally downward.  Crazy?  Well, by accident it actually measured much closer to the Focal Utopia's of 2004.  Check out this chart from Stereophile:

Of course, I'll never get the output, especially in the bass, of the Utopia, but it does sound much smoother and coherent than it did before.  Should you buy a Profile 918 and hack the crossover like I did to get a poor man's Utopai?  No way.  The value just isn't there overall.  The value IS there if you want to take this on as a learning experience though.