[dewar]

O.K knowledgeable folks, this time I'm really in over my head and need your help.

In my quest for perfection I've decided to add some tweeters to my B200 OB's. Have decided on the B&G Neo3 PDR.

In talking with a highly regarded speaker designer I have been advised to cross them over at 2k as the tweeters are of higher quality than the B200's from 2k upwards.

I feel a bit heretical putting a passive crossover in the signal path, largely due to having often read on these pages of benefits of a crossoverless design that full range drivers can provide. (PC based active crossover beyond my means at the moment, and Behringer DCX puts me off as signal manipulation is not done in the digital domain from what I understand)

I was hoping some of you could weigh in opinions here. 

I'm also wondering if I'll need DEQ to smooth out phase related peaks and dips that I'll get above and bellow the crossover point? I was told that it would be more of a problem with my original idea of highpassing the tweeter at 10K(the B200'snatural roll off ) but I cant work out why it would be less of a problem at 2k than at 10k.

Thanks for any input,

Bevan

[Dougl]

In my quest for perfection I've decided to add some tweeters to my B200 OB's. Have decided on the B&G Neo3 PDR.

Nice tweeter.

and Behringer DCX puts me off as signal manipulation is not done in the digital domain from what I understand)

The Behringer is a digital crossover and is useful and versatile.  There are several filter algorithms, and some people are religious about it.  Other people are put off by its op-amp's, etc.
My personal advice:  if you need an active crossover, buy the DCX2496.  Prototype with it.
It has enough parametric equalization to be useful, its easy to use, and has a good re-sale value.  Then make up your own mind.

I'm also wondering if I'll need DEQ to smooth out phase related peaks and dips that I'll get above and bellow the crossover point? I was told that it would be more of a problem with my original idea of highpassing the tweeter at 10K(the B200'snatural roll off ) but I cant work out why it would be less of a problem at 2k than at 10k.

  I tend to use as little equalization as possible.  Definitely try it without eq first.  Also, try several crossover points.  Set it up and play tunes for a week.  Then change it and see what differences you can hear.

[konut]

I've thought about doing just what you're proposing. My thoughts in reverse order.

1. Its been pointed out that just high passing the PDR and running the B200 full range will cause cancellations. Not good.

2. Using a symetrical XO, at 2000hz, will take away the large appeal of midrange coherance.

3. Using a DEQ, while being versitile, causes its own problems as far as intergation.

4. Ditto for DCX.

If I had the avaliable funds, this is what I'd do. I would buy the DCX. I could then model an asymetrical XO. I'm not 100% sure the DCX could do this. This would be 6db per octave slope starting at 2000hz. Remember, because of the B200s' rising response, this would effectively flatten the curve above this range. I would then play with the slope and frequency at which the PDR integrates well. What you would have then would be a set of values you could replicate, on a passive, either speaker level if you wanted to use only one 2 channel amp, or line level, if you had two 2 channel amps. Optimally an either 6db per octave or 24db per octave slope would be used to high pass the PDR to maintain uniform phase. It would be relatively easy to measure the results using an SPL meter and test tones. Does this make sense?

[mcgsxr]

Can I ask what cancellations would occur using the b200 full range, and bringing in the PDR at upper frequencies?

What if you used a 12-15K, 12db/octave for the PDR?  Would that not fill in the upper frequencies, and lose efficiency as it crossed over with the b200?

Never done it, happy to explore with y'all!

[dewar]

thanks all.

mcgsxr, you answer lies here. basically you'd need a 120db/octave slope. surprised me too. Can this be done passively?

Doudl, I should have said that as far as I know the DCX does not have digital out(which would require multiple external dacs). It's their internal dac which concerns me. You reckon it's up to scratch?

"Using a DEQ, while being versitile, causes its own problems as far as intergation." konut, could you say more?

Apparently I'd need to send drivers and baffle to the speaker designer to have a 2 way crossover designed. Might be cost prohibitive. Think I might go back to the safer plan of rear mounting the Neo3, which wont result in any phase cancellation issues.


cheers

Bevan

[konut]

There are several 'DEQ' units avaliable. Which one were you thinking of? I'm sure I could find something wrong with each of them.      By all accounts the internal DAC of the DCX is just fine. There is a looong thread at the DIYaudio forums dealing with integrating it into a home stereo rangeing  from simple attenuation to all out mods that claim state of the art sound. If you like I could post a link. Bottom line, for me, is that its a pretty complicated piece of gear to begin with. Why do I need to add something that 'out of the box' is not going to be useable long term without a further investment in time and more equipment. All I want is a Xover for gosh sakes!

The problem with rolling in a tweeter to the full range B200 is that there is still significant HF content in the B200 causing interactions  depending on distance between the 2 drivers and seating positions, resulting in vastly different peaks and nulls. While you could find a spot that measures resonably flat, a measurement taken even an inch or 2 away might look like a roller coaster.  There was a thread where Danny Ritchie pointed this out. I could hunt it down for you if you like.

The approach I outlined would require the temporary use of the DCX only to find the proper xo points and slopes and then, perhaps, get a pair of Marchand passive line level XOs at the reqired values. While this might not get you perfect linearity, it would come darn close. You could always add the DEQ later to squeeze the last bit later.

[JohnR]

Think I might go back to the safer plan of rear mounting the Neo3, which wont result in any phase cancellation issues.

You could set it up as a dipole and then experiment by covering the front or back. Just a thought...!

[Dougl]

Doudl, I should have said that as far as I know the DCX does not have digital out(which would require multiple external dacs). It's their internal dac which concerns me. You reckon it's up to scratch?

No piece of equipment is perfect.  However, the DCX2496 has been in my system for a year and a half, and I am not dis-satisfied with it in stock format.  It is true that several people have modified it, especially the power supply and analog outputs.  If you have questions on the DCX2496 capabilities, download the current configuration utility.  It runs on a PC, is graphical, and you can download your work to any DCX2496.  The DCX2496 does not have digital out.  you are using the onboard DAC.

mcgsxr, you answer lies here. basically you'd need a 120db/octave slope. surprised me too. Can this be done passively?

 To me, this is a design trade-off.  The person who specified the 120 dB crossover strongly believes in crossing over at a half wavelength.  He has a valid point, especially in the mid-range.  I have heard "helper" tweeters work with full range drivers @ 10K before with no issues I could hear.  You need to decide for yourself.  Crossover slopes of 120 db are only available digitally.  Since the tweeter is less efficient than the woofer, this would be a challenging passive crossover anyway.

"Using a DEQ, while being versitile, causes its own problems as far as intergation."

The Begrenger DEQ2496 has the same DAC as the DCX2496, and some of the same issues with modifications, etc.

2. Using a symetrical XO, at 2000hz, will take away the large appeal of midrange coherance.

I Agree.

Think I might go back to the safer plan of rear mounting the Neo3, which wont result in any phase cancellation issues.

 Its your call.  it would certainly help.  

[dewar]

when I said "your answer lies here" this link was meant to follow

http://www.audiocircle.com/index.php?topic=31585.0

it's Danny Ritchie giving a very good account of the problems we're dealing with.

Will report back in a few weeks when the tweeters are in place.

cheers,

Bevan

[markC]

Keep this in mind too. As good as that B200 sounds, from 2kHz and up most any soft dome tweeter will clobber it in performance. If you are going to have to add one any way, and with that driver you will, then cross it off no higher than about 3kHz. I'd try it as low as 2kHz because the off axis response of the 8" woofer will already be falling off by then.

Above is a quote from Danny and I took his advise.
I introduced a 3rd order x-over @ 2100 hz and retained the baffle step circuit. The results are Very satisfying. The open and effortless mid range is there in all it's palpable glory and the tweeter quite simply does a much better job with the upper mids and highs.
Efficiency drops substantially to about 88-90 db, but well worth it to my ears. Be aware that you can't just drop a text book x-over in and expect good results. Much tweaking is required and a spice simulation program took a lot of trial and error out of the exercise.
The tweeters are staying!

[JANDG]

Nice approach..what tweeter did you use for the 2100hz XO 3rd...? I have went with rear fire ribbon only @ 10k1st order, with a line level correction circuit for the B200 & a 150 HP also included in the line level circuit. When I A/B front & rear ribbons @ 10k the rear one easily, I played with both & padded accordingly but it is harder than it looks running di-pole ribbons. So try that tweeter rear mounted when you get it. worth a shot for sure. also have you looked @ Geoff's H. line level correction circuit.. ? cheap , easy & effective .

[corloc]

What is the point of using a full range driver?  Why not use a pro mid driver? Or a woofer with higher qts.  I'm not a purist, but aren't you loosing a lot of the full range driver benefits by crossing so low?

Chris

[markC]

What is the point of using a full range driver?  Why not use a pro mid driver? Or a woofer with higher qts.  I'm not a purist, but aren't you loosing a lot of the full range driver benefits by crossing so low?

Chris

The point is not so much using a full range driver as using the driver that I have and knowing how it sounds.
If I were starting from scratch, there's no way the b200 would be my choice. But, since I'm familiar with the driver and it's characteristics, and I already have them, it was an obvious choice to try and see,( or hear.)
The mid range is the magic of this driver for me, most notably the vocals. Especially female voice. That magic remains.

[markC]

Nice approach..what tweeter did you use for the 2100hz XO 3rd...? I have went with rear fire ribbon only @ 10k1st order, with a line level correction circuit for the B200 & a 150 HP also included in the line level circuit. When I A/B front & rear ribbons @ 10k the rear one easily, I played with both & padded accordingly but it is harder than it looks running di-pole ribbons. So try that tweeter rear mounted when you get it. worth a shot for sure. also have you looked @ Geoff's H. line level correction circuit.. ? cheap , easy & effective .

I'm using a Visiton silk dome. I've tried, (for months), to cross over @ the b200's natural roll off with less than good results. With this set up,  the highs no longer "shoot" at you, but instead "wash over you".

[dewar]

Still waiting for the tweeter...

I'm reminded of that old story about the guy with his favorite axe, the one he's replaced the handle of 5 times and the blade twice.

You get to the point where if not the magic is lost, then at least one of the main reasons for owning a B200 - simplicity and great performance for the price. Yet every tweek might be an improvement and worth it's marginal cost. maybe I would also have started with a different driver if i"d know I'd be looking at all these mods, yet I'm still very happy with what my speakers produce, and importantly, enjoying all the fiddling.

I'm pretty convinced that a 2k crossover is beyond the talents of a beginner like myself, so 10k rear firing the tweeter will be. If only I had the flexibility if JohninCR's pc based system...

Got one more question. Is the reason a 10k passed rear facing tweeter will not effect measured FR because signal is so attenuated , or is is because the signal is so broken up by bouncing off the front walls that it has no one faze that could cancel out the B200's output?

I'm wondering what would happen if you put in a rear firing 10K HP'd tweeter and the use a DEQ to boost the measured signal to be flat for 10 to 15 or 20K. Would there be too much rear firing energy even though we measure flat?

cheers

Bevan

[nodiak]

FWIW, I read Danny's statements too and had always felt the B200 and Hemp 8's lacked satisfying treble. I didn't go to any great trouble to design a nice xo as MarkC did, but I did devise a simple 1st order electrical xo for the Hemps and a Scanspeak 9300 tweeter (cap on Scans, coil on Hemp). The cutoff reduced the effeciency of the Hemps midband to make a perfect blend to the 90db Scans with no attenuation needed.  Quite a bit better sounding - retained mids of Hemps I like and treble of Scans I've always liked. The Hemps have whizzers and dustcaps removed and phase plugs added (when I did this the treble extension was decreased so I treated them like a midwoofer).
This was all done by ear, so it could be much improved, but the idea rang true.
Don