[BrunoB]

I bought my speakers 626R as a kit about two years ago and I just can't stop modifying them. This became a new hobby.

Here is a summary of the mods:

 -three way first order series crossover (all three drivers!) using  TRT Dynamicaps and  Mundorf Zero Ohm coils (for the woofer)
- all drivers are in phase
 - time aligned and convergent drivers angled for a listening distance of 5.5 feet
 
 -home made ribbon foils for the tweeter, "Spaghetti ribbons", cleaned and silcleared contacts between ribbon and aluminium support

-midpanel contacts  treated with Siclear (don't try this - it is very easy to damage the midpanels)

- main enclosure open  in the back integrated with the room treatment
- main enclosure lined with Dynamat, durafoam, memory foam and filled with Fiberfill (high resiliency polyester fiber)
 - small braces that push the panels appart with a high pressure
-midrange enclosure lined with modelling clay and river stones stuck in the clay, filled with miraflex
- enclosure covered with a soft microfiber carpet for diffraction reduction
- tweeter front plate removed, ribbon closely surrounded by cotton balls for  diffraction reduction
- 100 w variable L-pads for the tweeter
- variable L-pads treated with Caig Pro Gold (thanks to Jim Romeyn for the L pads tweaks)
 -Megawoofer magnet covered with 1 pound of alternate layers of modelling clay and aluminium foil

Measurment:

The only measurment I have is the impedance curve (thanks to Rick from Selah Audio for this measurment):




Sound: very sweet treble and midrange, fast bass. I am very happy with the result. The nicest comment I received from another AC member is "Your speakers are the most detailed non-fatiguing speakers I've ever heard in a long, long time. (can't remember any actually) "




Bruno

[mac]

That impedance curve isn't from a ported box speaker.  Did you plug your port?

[BrunoB]

That impedance curve isn't from a ported box speaker.  Did you plug your port?

No. I removed it. I also have a large opening in the back of the speaker, behind the woofer,  where the crossover was placed. The idea is to have a "finite-baffle" system. If I could I would have removed the whole back of the speakers. Cancellation of the front and back wave is not problem in my setup.

 This large opening in the back gives an impedance curve that looks like a sealed enclosure.

Bruno

[BrunoB]

I feel that I should describe why I decided to go with a new crossover. The main is reason is that I did not like the treble coming out of the midrange panel. I tried to improve the midpanel treble by using different sound adsorbing materials as I described previously. http://www.audiocircle.com/index.php?topic=9535

So the next step was to decrease the crossover frequency to 5500 Hz. I have no experience in crossover design. All the information I have is coming from the book “The loudspeaker design cookbook” and from the Web.  I decided to use a 1st order series crossover because (1) 1st order series crossover can have a steeper slope than 1st order  parallel crossover,  (2) they are forgiving to  - a difference in driver impedance, network capacitance or inductance will change the crossover frequency and slopes but still maintain a flat summed frequency response (based on a document from John Kreskovsky, “First Order Series Crossovers, What’s all the fuss?”). This is important, because I don’t have any measurement tools; I can not check the actual crossover frequencies and slopes.
Another factor that helped to improve treble is the TRT Dynamicap capacitor.  When I upgraded my network with the TRT’s, (I used also Sonicaps and Audiocap Theta), there was a very significant improvement in treble quality.

Note that 5500 Hz might be too low for the tweeter when using first order slopes. I suspect that the actual crossover frequency is even lower. But since I don’t play very loud (nearfield listening), this hasn’t be a problem.


Bruno

[BrunoB]

I have continued working on my crossover. I purchased  SoundEasy and a microphone to make measurements. At least now I can check what the crossover is doing:

Gallery with measurements and simulation

The transition between midpanel and tweeter is less than 4KHz. I haven't heard any buzz or had any problem with the tweeter ribbon yet. The xover frequency between the mid and the woofer is similar to what I simulated with SE for the 626R original xover but I have much steeper slopes (zeta = 0.5).

I am not showing the frequency response of my speakers yet because it is not easy to get correct measurements for frequencies below 400 Hz (= I have no experience doing this - I am still learning).

Bruno

[Skynyrd]

Great Post regarding your 626 mods!
Skynyrd

[BrunoB]

Great Post regarding your 626 mods!
Skynyrd

Thanks.  FYI, here is a thread with a few (old) audiocircle posts  about the sound of my modded  speakers and system.

Bruno

[BrunoB]

My crossover is first order and the drivers are time aligned. My goal is  to have all drivers in phase and to be able to reproduce a square wave. My problem is that  when I connect the midrange driver in phase, the frequency response does not look nice and the speaker does not play as loud:



The red line is the FR with the midrange connected out of phase whereas the green line shows the FR with the midrange connected in phase. The red out of phase line looks much smoother than the in phase green line.

How I should connect the midrange panel: in phase (poor FR) or  out of phase (nice FR)? I would have expected that with a first order crossover,   all driver are in phase provides the best FR. But this is my first crossover design, so sorry for such a trivial question.

FYI, here is the impulse and the FR of the midrange driver alone measured by myself with the help of SoundEasy:



I also plotted the phase, but honestly, I don't know how to interpret it.

Thanks in advance for any help or advice.


Bruno

[Jose R.]

Hi Bruno

Interesting graphs.  I have the RM30C and have no experience with your speakers but I have had similar results using digital amps and crossovers.  In my room, there is a dip in the 300-700Hz range with the drivers in positive polarity.  Changing the woofer polarity to negative gives a better measurement curve with a flatter response between 300-700 but I felt that the sound was not as coherent.
Changing to Auricaps and megawoofers did not change the measurement curve much.
The rising tweeter response above 20KHz may be a microphone effect - do you have a calibration file for your microphone?
My solution has been to cross the panels at 720Hz (6th order Linkwitz) but the megawoofers do not easily go this high.
One solution is to change the woofers to something that will be able to cross higher than the meagwoofers.

Regards

Jose

[BrunoB]

Hi Bruno

Interesting graphs.  I have the RM30C and have no experience with your speakers but I have had similar results using digital amps and crossovers.  In my room, there is a dip in the 300-700Hz range with the drivers in positive polarity.  Changing the woofer polarity to negative gives a better measurement curve with a flatter response between 300-700 but I felt that the sound was not as coherent.
Changing to Auricaps and megawoofers did not change the measurement curve much.
The rising tweeter response above 20KHz may be a microphone effect - do you have a calibration file for your microphone?

Jose, thanks for your input.


No I don't have a calibration file for my microphone (Behringer ECM 8000). I think that the rising treble is due to the horn removal of the tweeter and the fact that I like it hot. I measured another speaker (Ascend Acoustics CBM170), and it did not show the rising treble.

Bruno

[BrunoB]

I am in the process of  OXOing my speakers - removing the xover and install it in a separate external enclosure. So, I took everything out of one of my speakers and since my box is empty, I decided started to line the walls of the cabinet (Note that I don't have MLS cabinets). I choose 1 cm thick heavy granite tiles for lining material. The finger knocking sound on the external panels changed completely: from a hollow guitar sound to a  dull sound. Curiously, after lining only one wall, the other walls were quieter as well. I reinstalled and reconnected the woofer when I had only one wall lined with granite. I  played some music, put my hands on the walls and could feel vibrations on the untreated wall but barely on the granite one. I am almost done lining all the walls and finger knocking almost hurts. Knocking near the edge or on the middle of a side panel gives the same sound. And I still have to add braces and melted modelling clay. The estimated cost for two 626's is about 26 euros (11 for the tiles, 5 for the glue and 10 for the tile cutter).

Credits: last issue of the German magazine Hobby Hifi (they used thinner ceramic tiles for a small speaker). http://www.hobbyhifi.de/

PS: the tile cutter does not work well for granite tiles - but then I don't have have any experience with tiles

[fredgarvin]

Bruno, hadn't you lined your cabinets previously with some dampening sheets? Mine have been lined with something similar to the Blackhole5 Brian uses. Are you saying the granite is better? Thanks.

[warnerwh]

Never thought of lining speakers with Granite. That may be a good idea for my subwoofer project especially being as I have some 3/4" granite left over from a kitchen remodel. That's a great idea, thanks.

[BrunoB]

Bruno, hadn't you lined your cabinets previously with some dampening sheets? Mine have been lined with something similar to the Blackhole5 Brian uses. Are you saying the granite is better? Thanks.

Today, it would be for me  impossible to compare directly the results  of BH5 with granite.  Having said that, I believe that damping structural vibrations of the box, granite is much better than BH5 (I tried BH5 a long time ago). On the other hand, BH5 does  reduce air borne vibrations as well (granite can't). Thus, ideally, one layer of granite plus BH5 would be perfect (the glue I used for the granite is extremely strong, it binds to the wood way better than BH5 could;  BH5 would stick perfectly to the mirror like surface of the granite). Note that for MLS cabinets, granite would not be necessary (I have seen one and it is quite dead).

Note that with granite lining, the speaker becomes very heavy.

Hope this helps,

Bruno

[fredgarvin]

Thanks Bruno, my cabinets sound pretty inert but if I end up going in to change the crossover for cdwg I might throw on the granite while I'm in there.

[John Casler]

Today, it would be for me  impossible to compare directly the results  of BH5 with granite.  Having said that, I believe that damping structural vibrations of the box, granite is much better than BH5 (I tried BH5 a long time ago). On the other hand, BH5 does  reduce air borne vibrations as well (granite can't). Thus, ideally, one layer of granite plus BH5 would be perfect (the glue I used for the granite is extremely strong, it binds to the wood way better than BH5 could;  BH5 would stick perfectly to the mirror like surface of the granite). Note that for MLS cabinets, granite would not be necessary (I have seen one and it is quite dead).

Correct on all counts.

They do different things.  The granite is stiffening the actual cabinet, but does no "airborne" resonance damping.

The BH5 reduces the airborne vibrations and standing waves, but offer "minimal" cabinet stiffening/damping.

So the thing is to use them together, if you have the non-MLS cabinet.

I have also had clients use ashphalt based bathroom floor tiles too, along with BH5 or other type of stick on damping to good results.

They are pretty easy to cut and shape, very "non"resonant, and can be applied like the granite with lots of glue.

[John Casler]

Bruno,

I forgot to mention, GREAT TAG LINE

A good enclosure is a dead enclosure.  A good room is a dead room.

[BrunoB]

I have also had clients use ashphalt based bathroom floor tiles too, along with BH5 or other type of stick on damping to good results.

They are pretty easy to cut and shape, very "non"resonant, and can be applied like the granite with lots of glue.

Hi John,

I tried  something that is probably very similar to asphalt based bathroom floor tiles: "Waschmachineunterlage" : a  1.5 cm thick heavy rubber mat that is used to isolate the washer from the floor. It helps but it was not as good as granite to reduce the panel vibrations (I had both at the same time in my speaker, granite on the left side panel, rubber on the right side)). I might glue the rubber mat on top of the granite ... hmmm

What about lining one of your subwoofers with granite?


Bruno

[pjchappy]

Have any pics, Bruno?

[BrunoB]

Have any pics, Bruno?

Here are my pictures:
My album

I have been thinking about opening the back of my 626 for quite some time. I started the work a few weeks ago.



Up to now I have connected only the woofer and the midrange. Stuffing  is done with Miraflex and seems to be very efficient (density: 4 lbs/cf). I haven't installed the crossover yet. But when playing music using only the midpanel, I can barely hear something coming from the back of the speaker.

I measured the impedance curve of the stuffed speaker with the woofer only:


 The upper blue line is the impedance multiplied by 10.

Looks strange to me.  The max impedance at resonance is only 11 ohms!

To be continued ... next year

Bruno