DIY Emerald Physics CS2 Clones

0 Members and 1 Guest are viewing this topic. Read 77407 times.

BrassEar

  • Jr. Member
  • Posts: 248
Re: DIY Emerald Physics CS2 Clones
« Reply #20 on: 26 Sep 2008, 10:12 pm »
Position the speakers in the room as you would to listen, with the axis crossing slightly in front of the listening position.  Set the xover at 1000hz (48db per octave) and get/borrow/buy/rent a measurement tool......the Behringer equalizer with calibrated mic will do (I also have a "bouncing ball" Audio Control third octave thang that is very easy to see, but the Behringer equalizer/mic will show you the same thang).  Play pink noise and notice the deviation from flat.....now using the EQs in the Behringer xover set them until pink noise measures flat at listening position.  You might want to play with delay as well.  Of course, you might like a setting other than flat.  Have fun!

Thanks for the info Ric. Am I dense or are there only a few settings which could be involved?

1. Clearly the xover is set at 48 dB/octave around 1 kHz.
2. A shelving EQ filter must be applied to the low end to compensate for the OB design of the woofer setup.
3. Levels for the compression driver must be balanced with the woofers.
4. EQ must be applied to the compression driver/horn combo for flat on axis response.
5. Possible delay to the compression driver to compensate for  alignment.

Am I missing anything obvious here?

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #21 on: 27 Sep 2008, 03:11 pm »

1. Clearly the xover is set at 48 dB/octave around 1 kHz.
2. A shelving EQ filter must be applied to the low end to compensate for the OB design of the woofer setup.
3. Levels for the compression driver must be balanced with the woofers.
4. EQ must be applied to the compression driver/horn combo for flat on axis response.
5. Possible delay to the compression driver to compensate for  alignment.

Am I missing anything obvious here?

On #2, there should be two compensation: dipole rolloff 6db/oct and driver compensation. Dipole rolloff is fairly easy except we need to remember there is a hump before it which needs to be tamed using notch filter.

Driver compensation should replicate something like Linkwitz biquad so the 40Hz hump is tamed and the FR is flat towards 20Hz. I have been playing around with eminence 15 and the xmax is fairly easy to exceed.

6. There should be plenty of notches around. With my 45mm baffle I have dipole peak at around 300Hz and driver breakup at 2kHz per datasheet but this is fairly easy to correct if using DCX2496. I am using op-amps and it's a real pain to get measurement then build the circuits.

Btw. I have not heard the CS2 but not really sure that dipole + horn/waveguide combination would actually work? It would be dipole only up to 1kHz, and SL repeatedly mentioned that live reproduction can only be achieved if the radiation pattern is uniform across frequency ranges? Would love to hear how you go.

zybar

  • Volunteer
  • Posts: 12071
  • Dutch and Dutch 8C's…yes they are that good!
Re: DIY Emerald Physics CS2 Clones
« Reply #22 on: 27 Sep 2008, 03:35 pm »

1. Clearly the xover is set at 48 dB/octave around 1 kHz.
2. A shelving EQ filter must be applied to the low end to compensate for the OB design of the woofer setup.
3. Levels for the compression driver must be balanced with the woofers.
4. EQ must be applied to the compression driver/horn combo for flat on axis response.
5. Possible delay to the compression driver to compensate for  alignment.

Am I missing anything obvious here?

On #2, there should be two compensation: dipole rolloff 6db/oct and driver compensation. Dipole rolloff is fairly easy except we need to remember there is a hump before it which needs to be tamed using notch filter.

Driver compensation should replicate something like Linkwitz biquad so the 40Hz hump is tamed and the FR is flat towards 20Hz. I have been playing around with eminence 15 and the xmax is fairly easy to exceed.

6. There should be plenty of notches around. With my 45mm baffle I have dipole peak at around 300Hz and driver breakup at 2kHz per datasheet but this is fairly easy to correct if using DCX2496. I am using op-amps and it's a real pain to get measurement then build the circuits.

Btw. I have not heard the CS2 but not really sure that dipole + horn/waveguide combination would actually work? It would be dipole only up to 1kHz, and SL repeatedly mentioned that live reproduction can only be achieved if the radiation pattern is uniform across frequency ranges? Would love to hear how you go.

The CS2's work extremely well.

George

BrassEar

  • Jr. Member
  • Posts: 248
Re: DIY Emerald Physics CS2 Clones
« Reply #23 on: 27 Sep 2008, 03:39 pm »

Btw. I have not heard the CS2 but not really sure that dipole + horn/waveguide combination would actually work? It would be dipole only up to 1kHz, and SL repeatedly mentioned that live reproduction can only be achieved if the radiation pattern is uniform across frequency ranges? Would love to hear how you go.

SL should go a hear the EP CS2. IMHO, they are better than the Orions in most respects. I just firmly believe that the weaknesses in the CS2 (harshness from the titanium driver and congestion from the WG?) can be easily eliminated. I am cloning them with respect to baffle, Xover, and Alpha 15's but I am using a 15-inch WG with a BMS compression driver. I will posts results and pix when I am done but don't hold yer breath.

tubamark

  • Jr. Member
  • Posts: 55
Re: DIY Emerald Physics CS2 Clones
« Reply #24 on: 27 Sep 2008, 11:52 pm »
. . . . . I am cloning them with respect to baffle, Xover, and Alpha 15's but I am using a 15-inch WG with a BMS compression driver. I will posts results and pix when I am done but don't hold yer breath.

Where did'ja find a 15" waveguide?
Price?
curious,

-- Mark

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #25 on: 29 Sep 2008, 05:29 am »

SL should go a hear the EP CS2. IMHO, they are better than the Orions in most respects. I just firmly believe that the weaknesses in the CS2 (harshness from the titanium driver and congestion from the WG?) can be easily eliminated. I am cloning them with respect to baffle, Xover, and Alpha 15's but I am using a 15-inch WG with a BMS compression driver. I will posts results and pix when I am done but don't hold yer breath.

You've heard both speakers? Could you please describe where each excels ? I can only dream of building an Orion, $2 grand drivers only and still needs 8 amplification :cry: ... so I was looking to build CS2 clone. In fact I have 4 Alpha 15 already which is very expensive to ship to Australia. But after playing around with the Alpha I conclude that this driver is not easily tamed to work up to 1kHz using analog filters. Let alone trying to equalise the horn. Maybe that's why the DCX was necessary. But then again just learning from my experience, without anaechoic chamber this could also be very trivial.

From DIY point of view Orion has more chance of a success but price prohibitive. CS2 is much simpler and cheaper to implement ... but there is no proven DIY settings out there (yet). The good thing is if someone is successful, implementation is "just an upload" away ....  :thumb:

dweekie

  • Jr. Member
  • Posts: 162
Re: DIY Emerald Physics CS2 Clones
« Reply #26 on: 29 Sep 2008, 11:06 am »

SL should go a hear the EP CS2. IMHO, they are better than the Orions in most respects. I just firmly believe that the weaknesses in the CS2 (harshness from the titanium driver and congestion from the WG?) can be easily eliminated. I am cloning them with respect to baffle, Xover, and Alpha 15's but I am using a 15-inch WG with a BMS compression driver. I will posts results and pix when I am done but don't hold yer breath.

You've heard both speakers? Could you please describe where each excels ? I can only dream of building an Orion, $2 grand drivers only and still needs 8 amplification :cry: ... so I was looking to build CS2 clone. In fact I have 4 Alpha 15 already which is very expensive to ship to Australia. But after playing around with the Alpha I conclude that this driver is not easily tamed to work up to 1kHz using analog filters. Let alone trying to equalise the horn. Maybe that's why the DCX was necessary. But then again just learning from my experience, without anaechoic chamber this could also be very trivial.

From DIY point of view Orion has more chance of a success but price prohibitive. CS2 is much simpler and cheaper to implement ... but there is no proven DIY settings out there (yet). The good thing is if someone is successful, implementation is "just an upload" away ....  :thumb:


It is really difficult to wade through all the opinions of various speakers.  I've read that the Geddes Summa was better than the Orion, which was better than the Hawthorne Sterling Silver Iris, which was better than the Emerald Physics CS2, which was better than the Orion, and we just cycle through over and over again. I've heard some highly touted speakers where I was the only one out of the whole room that couldn't stand to listen to them.  It's a tough world when everyone has a different pair of ears.....

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #27 on: 29 Sep 2008, 11:31 am »
Yes, that's why I did not ask which one is better but rather what qualities differentiate those design  aa. Of course if I was in the US I'd just go to RMAF and decide for myself  :drool:

Btw. I'm not sure what's with Gedes horn/waveguide. They still use box with all its inherent issues. But OB + horn like CS2, now that I really need to hear sometimes.

dweekie

  • Jr. Member
  • Posts: 162
Re: DIY Emerald Physics CS2 Clones
« Reply #28 on: 29 Sep 2008, 12:07 pm »
Yes, that's why I did not ask which one is better but rather what qualities differentiate those design  aa. Of course if I was in the US I'd just go to RMAF and decide for myself  :drool:

Btw. I'm not sure what's with Gedes horn/waveguide. They still use box with all its inherent issues. But OB + horn like CS2, now that I really need to hear sometimes.

Since you have 4 alphas already, have you considered some of the other projects posted in the past other than compression drivers?  The MJK project with Jordans and Alpha or the B200 wth Alphas come to mind.  If you feel there is weakness at 1K in the alphas, the full range driver mix might be worth your while. 

mightym

Re: DIY Emerald Physics CS2 Clones
« Reply #29 on: 29 Sep 2008, 11:19 pm »
If you haven't looked already, there's a long, long, but informative thread over at AudioKarma, called the econo wave guide thread.  If I were to attempt to clone the Emerald Physics CS2, I would look seriously at this for the top end.   The idea was * you supply the woofer, Here's the topend...*  From all accounts, it doesn't sound like a " horn ", but has verrry smoooth, and detailed sound.  I'm working on a rebuild of an old set of Realistic cabs that literally had the woofers kicked in, I plan to use the Econo-waveguide in them.

www.audiokarma.org/forums/showthread.php?t=150939


What kind of trouble are you having Re: the Eminence Alpha 15A's up to 1K?  since the Eminence Alpha's are on sale at PE for $52 freight paid ( Shh, don't tell anyone ), I had been thinking of a clone project of my own.

The info I was able to find looked to be 94Db +/- 2Db to (just) past 1K.

John

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #30 on: 30 Sep 2008, 12:16 am »
The design was Eminence 15 + a fullrange, as experiment only. My thinking was if I can't do it with easily implemented Full Range, there is no way I can make the horn to work.

frequency response is fine and can be measured and corrected. But it just does not sound right to cross at 1kHz. Perhaps due to time alignment etc. which is not easily done using analog filters. best result is to cross really low and then I did not need to correct the FR and transition between drivers is not too sensitive to the ear.

I believe these factors were the reason a DSP is necessary as they can correct them fairly aggresively. But spending AUD450 for DCX2496 without proven settings and an anaechoic chamber is not something I want to gamble on.

My 'final' prototype now runs a fullrange and the Eminence acts like a sub.

I still believe CS2 has the best and functional looking OB so far. Considering the price it looks like a really good value (in the US).

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #31 on: 30 Sep 2008, 12:55 am »

btw... I just found out the 12" waveguides has been discontinued from PartsExpress  :cry:

mightym

Re: DIY Emerald Physics CS2 Clones
« Reply #32 on: 30 Sep 2008, 01:35 am »
Too bad,

I looked the waveguide up, and they were recommending crossover at 1200 Hz.  I don't think you could make them work without the DSP.  But then what do I know...Emerald Physics says on their webpage that they crossed below 1000 Hz....

The Econo waveguide is designed to cross in the region of 1400 Hz., the Eminence Alpha 15A's have a nasty big peak above 1000 Hz.  even with a 24 Db crossover ( passive ), you'd have a bit of trouble NOT having a fair peak, or would spend lots of time and $ perfecting the woofer filter.


John

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #33 on: 30 Sep 2008, 02:06 am »
I did 24/db LR filter (active) and added notch filter (also active) at that 2k peak. It actually worked well.

I did encounter problems with bloated 300Hz and 500Hz -- again without anaechoic chamber this might be just room effect.

I wonder if doing measurements outside would be a valid one ?

If we can replicate the DSP using analog filters I think it would be great (and better than AD -DA perhaps).

This is nearfield measurement of an eq-ed Eminence 15 to compensate dipole roll-off. I forgot whether it had the 1kHz xo or not. That 2kHz peak can be removed fairly easily. Does anyone have a measurement of Selenium + 12" waveguide ?


BrassEar

  • Jr. Member
  • Posts: 248
Re: DIY Emerald Physics CS2 Clones
« Reply #34 on: 30 Sep 2008, 03:13 am »

Where did'ja find a 15" waveguide?
Price?
curious,

-- Mark

http://www.gedlee.com/Summa.htm

BrassEar

  • Jr. Member
  • Posts: 248
Re: DIY Emerald Physics CS2 Clones
« Reply #35 on: 30 Sep 2008, 03:16 am »
frequency response is fine and can be measured and corrected. But it just does not sound right to cross at 1kHz. Perhaps due to time alignment etc. which is not easily done using analog filters.

With the Behringer DCX it is very easy to set the xover, slope, gains, and eq. But how does one measure and correct the time alignment issues?

BrassEar

  • Jr. Member
  • Posts: 248
Re: DIY Emerald Physics CS2 Clones
« Reply #36 on: 30 Sep 2008, 03:18 am »
FWIW, a pair of Alphas on an OB the size of the CS2 were very easy to measure and EQ. The only reason that you can't run it full range is the 2K spike. It is actually very smooth right up to 1.4K. If you xover there at 48 dB/octave I think it sounds excellent.

BrassEar

  • Jr. Member
  • Posts: 248
Re: DIY Emerald Physics CS2 Clones
« Reply #37 on: 30 Sep 2008, 03:25 am »
Does anyone have a measurement of Selenium + 12" waveguide ?



I assume for informational purposes only? I find that Selenium to be bright, hard, and congested near the xover region and is what really holds the CS2 back IMO.

gainphile2

  • Jr. Member
  • Posts: 65
    • Gainphile
Re: DIY Emerald Physics CS2 Clones
« Reply #38 on: 30 Sep 2008, 04:46 am »
frequency response is fine and can be measured and corrected. But it just does not sound right to cross at 1kHz. Perhaps due to time alignment etc. which is not easily done using analog filters.

With the Behringer DCX it is very easy to set the xover, slope, gains, and eq. But how does one measure and correct the time alignment issues?

I saw DCX screen before where you can introduce delays. In theory one just measure the distance between woofer and tweeter coils, and calculate in ms. the time difference.  My understanding DCX has two types: Long and short delays. I don't know more details.

With analog allpass, the phase correction is much more difficult to design, and even then the results can be argued/debated.

This is what I meant:

Quote
Adjustable delay for all inputs and outputs allow manual or automatic correction for room temperature, phase and arrival time differences

mgalusha

Re: DIY Emerald Physics CS2 Clones
« Reply #39 on: 30 Sep 2008, 01:31 pm »
Quote
I saw DCX screen before where you can introduce delays. In theory one just measure the distance between woofer and tweeter coils, and calculate in ms. the time difference.  My understanding DCX has two types: Long and short delays. I don't know more details.

The DCX has the capability to do set this automatically if a measurement microphone is connected to it's "C" input. I haven't tried it but the DSP can do it.

If you are setting delay times manually you can set them by distance or time.