[bassboy]

Richard had tried something with his set-up which was promptly dismissed. After thinking a little, I am not so sure it should have been dismissed so quickly. My intent was to set people thinking and looking at what was going on when the crossover was removed. It is not so obvious to me that it cannot work. Listening to the result of an audio experiment is great but I really like to understand why it sounds the way it does to increase my understanding which always leads to other interesting possibilities.

Actually, if you go back and look, I think every single person who responded to this thread speculated that fr might be improved without the filter (even florian in his own special way).  (Except for fergs, who did not submit a speculation)  I don't think anyone dismissed what he proposes, although it seems everyone except him is interested in why it might be better.

Richard, sorry for the lack of congeniality, but I wholeheartedly second MJK's last sentence.  To stumble upon something you like is a small victory.  To be able to recreate something you like (and all that that implies) at will is a huge victory.

[-Richard-]

Hi bassboy ~

Of course your intention to want to find out is an important part of any creative inquiry. I was merely sharing an insight with what I think I am hearing. Unfortunately my knowledge and understanding of speaker design is very limited. It was MJK/Martin who did the basic research and development and creative thinking and simulations and Scorpion/Erling passing it along in an AC post that lead me to try the Alpha-15A in my OB's... bless them both.

I understand my limitations and I try to work within them as best as I can. The one thing I can bring to this work is my willingness to engage in a DIY project and to try a few different simple variations on the theme.

Martin's graph... which he is very kind to share with us... is very very interesting to me. Thanks again, Martin... we all owe you an enormous amount of gratitude for your selfless willingness to share your important work with us.

If one was to biamp... so that the Alpha-15A's were controlled by their own amp... and if a simple inductor was used to cut off frequencies above 1000 to 1500 Hz... I think many people would find the combination of the often denigrated B200's... and Alpha-15A combination... to sound quite satisfying.

It was women who invented the spade and domesticated the herbs and grasses through careful study... observation... leading to the applications of cross-fertilization and hybridization. That is science... and that is why we are still here on this planet. Everything begins with observation... and the creative spark to want to know.

Warmest Regards ~ Richard

[panomaniac]

Very interesting....

The first thing that occured to me was polarity or phase (as bassboy mentions).  But not at only the top, but thru the lower midrange, too.  Could the 2 drivers be canceling somewhere in that range? That might be pulling out the mids, which is what it sounds like you're doing.

You don't mention what size baffle you use. Is it big?  My experience with small to mid size baffles is that the woofer has a terrible rising response because of the OB losses.  For me, it tends to make everything very upper midrange heavy - way too bright.

I have to cross my 15s very low, ~40Hz 2nd order and boost the hell out of them to fight the rising response.  It's only then that the rig starts to sound the way you mention, full and lush.  But different baffles, different woofers, different results, eh?

[-Richard-]

Further non-scientific insights into using a simple flat OB with the B200 and Alpha-15A drivers hooked-up in parallel.

I hope these insights do not frustrate those among us who like to measure... and use measurements to speculate or replicate audio experiences posted on AC.

I took out the Planet 10 phase plugs from the B200's... which in my set-up seems to interfer with the top end sense of "air" and potential for delicate resolution of the B200's... the addition of the phase plugs has made the B200's sound edgy and beamy... certainly non-musical.

I replaced them with a very thin foam-like packing material that I tried unsuccessfully to "mold" into a kind of dome shape (to replicate the original dust caps)... instead I wound up sticking this crude shape into the speakers pole piece (the "well" that the original dust cap covers) and instantly the B200'S sounded "resolved" on top... which is to say the instrumental timbre and textural life sounds entirely realistic and deeply etched. The top end no longer sounds edgy, beamy, "hot"... a remarkable transformation.

I also came to realize that allowing the Alpha-15's to play to their "natural" full-range resulted in a "bloated" presentation for at least some of my CD's... and tended to exaggerate the "spatial" dimension.

I took an older coiled inductor "crossover" that cut off frequencies around 300 Hz or so and uncoiled it to try and predict a 600 Hz or so frequency cut off (-6 db per octave above that). A nice chap at Parts Express explained to me that it is impossible to predict what the cut-off will be based on the amount of coil sniped off because inductors become electromagnetic devices when current passes through them and therefor do not conform to simple rules such as the length of the coiled metal material... hummmmmm.

In any case with a very high cross-over that I am assuming is around 600 Hz the sound is quite pleasing. I intend to order various inductors from Parts Express to confirm the best cross-over point.

I am now using my Korneff 45 SET tube amplifier with a Passive Luminous "preamplifier". I recently had the resistor in the Luminous changed to accommodate the 2.5 watts of the Korneff (using the almost 2 volt signal from my cheapy Pioneer CD player). On most CD's the sound is extremely "present" and visceral... deeply satisfying... on some early recorded material it is not quite enough to establish a textural musical presence. In this particular set-up the Lloyd Peppard Magic 5 triode preamp, which supplies all the gain necessary, sounds hi-fi'ish compared to the passive non-tube circuit of the passive Luminous.

Insights: I do not recommend using Planet 10 phase plugs with the B200's (although Dave is admittedly a very nice guy). Triode SET amplifiers sound fantastic with OB's. "Passive" preamps or direct signal to SET amplifiers translates to a highly etched textural musical experience... deeply satisfying.

I just purchased a rebuilt 1950's tube integrated amplifier using EL84 tubes with tone controls to see what that brings to the equation... I should have that to experiment with in the early part of July. A full report to come.

Just keeping in touch and sharing my simple experiences with everyone.

Warmest Regards ~ Richard

[Dmason]

Two of the above insights I second, completely... among all the permutations my DIY home stereo has taken over the years, these two insights are overarching conclusions, to myself...

SETubes with OB is without a doubt in a whole other realm of sound; it is virtually holographic, everywhere. Entirely 3 dimensional with very well produced musical events. I am not talking about "imaging," or "soundstage," or any of these parlour tricks, but about a density of sound, a metaphysical equation of sorts, that sometimes can be almost like a "sound-cloud" surrounding the sound sources, whose texture changes with instrumentation, musical genre, recording technique, etc.

As direct input as possible, source to amplifier. Anything more, is less, it would seem. Anything seems to get in the way. Use hot sources of quality, with buffers, with low power amps. SET amps have amazing slewing capabilities, if done right. Use lossless contouring and volume, along with lossless filing of music. Do this, and even spent old drivers poached from old consoles found in the alley will sound wonderful. This I know for a fact.

[Graham Maynard]

Hi Richard.

I do not have Planet 10 phase plugs, and so cannot comment on them, but I did try many shapes, sizes and materials for centre phase plugs on the B200, some near flat at the centre, some long, some more bulbous like the Lowther ones, but I did not find any that sounded better than a simple circular section of soft (mini paint roller end) foam pushed into the voice coil former end, as here;-



Some shapes of phase plug could improve dispersion angle, but these then affected forward reproduction, and I must concur with your description of 'edgy'.  I concluded that the B200 did not suit a protruding phase plug of any kind.

I too glued alternate felt and foam layers to the pole piece, for this seems to damp both voice coil former cavity resonance and impede the direct path line of radiation from the tiny voice coil former ventilation holes which scissor through air motion with respect to the centre pole piece.

The best dust-cap mod I had with the B200 had the felt/foam layers glued to the pole piece, a new concave dustcap glued directly to the coil former end and then foam fingers externally adjusted to minimise beaming/ optimise dispersion.



An APT with physical position time alignment and used as an augmenting supertweeter was a final subtle help too, but eventually the arrangement shown in Post 19 here became my final B200 solution;-

http://www.audiocircle.com/index.php?topic=51673.0

As to what any 'measurements' might reveal I have no idea, but music listening takes another step forwards beyond plain B200 capabilities.


Cheers .......... Graham.

[-Richard-]

Update of my ongoing experiments with the B200's and Alpha-15A connected in parallel on a simple flat Open Baffle.

I received a pair of cheap simple 18 gauge coiled steel inductors from Parts Express that are valued at: .80... so that the cut off of the Alpha-15A's would be around 1200 Hz (with a progressive -6 db fade per octave above that).

Thanks to the important work done by MJK (see his graph above) of simulating the SPL response of the B200 and Alpha-15A we can clearly see that the Alpha-15A's begin to dip at around 1300 Hz. I wanted to hear what the combination of the B200 and Alpha-15A would sound like without the erratic resonances of the Alpha's after 1300 Hz.

With the new inductors inserted into the speaker cables going from the B200 to the Alpha-15A's which begin to cut-off the frequency of the Alpha's above 1200 Hz, it became immediately clear that a new smoothness and ease was present... the "bloat" that appeared on some Cd's was entirely gone. The more I listen to this the more I like it... s-m-o-o-t-h, liquid, clean, with a rich tonal pallet and deeply etched textural realism... but presented in a thoroughly relaxed atmosphere.

It is interesting that I read quite by accident today that the classic BBC speaker design purposely created a midrange "hump" in order to create a more visceral presence in that region. Quite honestly... what I am hearing now with this particular inductor is so smooth that the "overlap" of the 2 speakers in the midrange area is entirely undetectable... I could not at this moment be happier with the result!!!

Suggestion: If you have a similar set-up to mine do try this crossover point for your inductors: .80 / cheap 18 gauge steel coil.

Also: based on Graham's suggestion (and photo) I pushed my crude thin foam packing material deeper into the voice-coil-former and added another thin foam piece on top of that for a "denser" treatment... sounds quite nice. I can see why some people feel the need for a tweeter on top... not the last word in "air"... but it is so satisfying otherwise that I have no interest right now in exploring that option... perhaps one day if I understand what I am trying to do.

Hi Dmason ~

To say it is a pleasure to read your always scintillating and honest prose is an understatement... I look forward to reading your insights more than any print or emagazine writer on audio... I trust your take on what you are hearing and it influences my approach to exploring the possibilities of achieving a more vital experience of music through audio. I hope you continue to share with us your penetrating insights.

Warmest Regards ~ Richard

[panomaniac]

Richard, we still don't know the dimensions of your baffle and the driver placement.  Well, at least I couldn't find it.
Also, no side wings?

Knowing the size of the baffle you're using would be a big help toward getting a feel for what you're doing. Thanks!

[D OB G]

Hi Richard,

I could be wrong (I'm either totally wrong or right!!), but the BBC used a 2-3 kHz dip?  They certainly used a hump around 120 Hz, to give the impression of bass when none existed.

(Linkwitz also uses a dip around 3 kHz- as an option).

David

[dweekie]

Richard, we still don't know the dimensions of your baffle and the driver placement.  Well, at least I couldn't find it.
Also, no side wings?

Knowing the size of the baffle you're using would be a big help toward getting a feel for what you're doing. Thanks!

Hey Pano, just a few posts down from here.

http://www.audiocircle.com/index.php?topic=47314.40

I'm going to have to get my B200's back in service with the Alphas.  After a few failed sales, I think I'm destined to keep these forever 

[nullspace]

Thanks to the important work done by MJK (see his graph above) of simulating the SPL response of the B200 and Alpha-15A we can clearly see that the Alpha-15A's begin to dip at around 1300 Hz. I wanted to hear what the combination of the B200 and Alpha-15A would sound like without the erratic resonances of the Alpha's after 1300 Hz.

Hi Richard --

Happy to hear that your trials continue to go well...

I would encourage you to go onto Eminence's website and grab the pdf for the Alpha-15As: Alpha 15A datasheet. While I suspect that in general they are prone to the usual manufacturer smoothing, it shows a freq response quite different than that graphed in Martin's simulation. Like many 15" pro drivers, the freq reponse begins rising ~1khz before reaching cone breakup at 2khz.

The Theile/Small-based simulations in the MathCad worksheets are invaluable, but after a certain point the T/S #s won't accurately predict a driver's frequency response -- I would suggest that >1000hz is easily into that territory.

Regards,
John

[fergs1]

Greetings , well my alphas are heading across the the pacific as we speak Martin I know you get bombarded with dumb arses like my self pestering you for crossover details but would you consider looking a little closer into using the b200 and Alpha together with regard to rolling off that large peak at 2khz on the alpha and getting the two to mate together smoothly. I'm afraid I have no knowledge base regarding simulations and crossover design. If it sounds too much like hard work I fully understand as your contribution to the diy scene has already been enormous. Building speaker baffles is no proble as I have already built 3 pairs of horns and an OB system using the b200, but when it comes to the electrical side of things I'm like when ever they show homer simpson doing the handyman number. The hammer comes down, the nail bends, the saw is pushed forward, it jams and bends. You get the picture
Antway I'm looking forward to making some more sawdust, a domain I know much better
                                          peace and goodwill  fergs

[Graham Maynard]

Hi Fergs,

http://www.prodance.cz/protokoly/alpha_15a.pdf

It appears that an additional medium Q notch filter could be used to take out that 2kHz peak.  This would give an additional undercut of the Alpha response at frequencies already covered by the B200.

Cheers ......... Graham.

[MJK]

Greetings , well my alphas are heading across the the pacific as we speak Martin I know you get bombarded with dumb arses like my self pestering you for crossover details but would you consider looking a little closer into using the b200 and Alpha together with regard to rolling off that large peak at 2khz on the alpha and getting the two to mate together smoothly. I'm afraid I have no knowledge base regarding simulations and crossover design. If it sounds too much like hard work I fully understand as your contribution to the diy scene has already been enormous. Building speaker baffles is no proble as I have already built 3 pairs of horns and an OB system using the b200, but when it comes to the electrical side of things I'm like when ever they show homer simpson doing the handyman number. The hammer comes down, the nail bends, the saw is pushed forward, it jams and bends. You get the picture

I have seen the response plot for the Alpha 15A a few times on the Internet and two things jump out that one should keep in mind. First the low frequency roll-off starts near 100 Hz and is definitely not representative of the driver on a baffle, it would suggest that it is the Alpha 15A measured in some form of box. The Alpha 15A has a fs of about 40 Hz and a Qts of about 1.2 which would not yield that response curve if measured in a baffle. Second the 10 dB break up peak at about 2000 Hz immediately catches your attention. It definitely looks ugly but I don't think it is a factor. If you follow my methods then you will low pass the Alpha 15A no higher than 200 Hz with a 2nd order filter, this means the driver is down 30 to 50 dB by the time it gets up to 2000 Hz. I have turned off the amp to my Lowthers and listened only to the pair of Alpha 15A's and I don't hear any nasty artifacts due to this break-up peak.

I don't own the Visaton B200 drivers so the simulations I have run in the past were using T/S properties I pulled off of the Internet. I looked back at these results and found the following information. When I combined the B200 with a single Alpha 15A per side the best calculated SPL response I got used a 2nd order low pass filter at 200 Hz and a 2nd order high pass filter at 500Hz. That would be my starting point based on the information I have for these drivers.

Hope that helps,

Martin

[scorpion]

I incidentally ran into a frequencyresponse curve of the Alpha15A on the net: http://www.usspeaker.com/ALPHA15-1.htm . This looks like an IB-measurement though you would have expected a bit of a hump down at fs. I mailed US Speaker and asked about their measurements and got the answer 'The manufacturers provide them' (Sic !). However this measurement seems much more in line with the Alpha's bassresponse than the PDF on the Eminence site and the breakup response is also there.

I have tested the Alphas in my 'Volks-OB' (straight switch with the A&D R1524 crossed at about 200 Hz 12 db/octave Linkwitz-Riley) and do like what I hear !

/Erling

[fergs1]

Greetings, thankyou for you time Martin. Greatly appreciated. Still haven't arrived(the alphas) so I still have a bit of time to prepare.  Now how do I go about actually designing these said crossover points and are they appropiate for using the completed speakers with the one amp(not biamped, or was that how you figured it)Next step the timber yard, or as it is known now the cardboard and particle yard
  Will keep everyone posted.             cheers  fergs

[scorpion]

Martin,

Going back to your simulation of Alpha15 and B200 above I also ran it with identical result in your 2-OB model. What is puzzling me is the sharp dip in the overall response between 1 and 2 kHz what do you think can explain that ?

/Erling

[scorpion]

I measured the Alpha15 fullrange in my 'Volks-OB' mic at 1 m distance and 90 cm (36") above floor in my livingroom. This is the result:



Quite astonishing, nothing of the cone breakup. In fact it is quite pleasant to listen to fullrange. The fundamentals are well reproduced. So revision is called for !
I think Richard might have a very good case when complementing the B200 with the Alpha going higher. As Martin said, you should never rule anything out beforehand. 

/Erling

[MJK]

Going back to your simulation of Alpha15 and B200 above I also ran it with identical result in your 2-OB model. What is puzzling me is the sharp dip in the overall response between 1 and 2 kHz what do you think can explain that ?

Hi Erling,

I believe the dip is just two drivers adding destructively. The locations and output magnitude and phase just tend to cancel each other at the listening position.

Martin

[MJK]

Now how do I go about actually designing these said crossover points and are they appropiate for using the completed speakers with the one amp(not biamped, or was that how you figured it)

For a passive crossover for this type of OB design, I treat each driver as if it were a pure resistor. The crossovers are high enough in frequency to be above the driver's resonance impedance peak and not so high that you need to worry about the voice coil inductance.

For the Alpha 15 A I would start with 9 mH in series and 68 uF across the driver.

For the B200 I would start with 35 uF in series and 2.75 mH across the driver.

Remember that I am guessing based on driver properties pulled off the Internet so you may need to adjust things a little by ear. But these values should be a good starting point.

Martin