Speaker design, X-Over, and MTM Questions - Please help a noob!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
frugal-phile™
Joined 2001
Paid Member
Jennice said:
Why should less energy be transferred to the cabinet with this push-push design?

In push-push the two drivers are tightly coupled mechanically. They have equal but opposite motion. They therefore activly cancel each others newtonian forces so that this (mechanical) energy is not transferred to the cabinet.

The sane amount of acoustical energy is impaxting the enclosure - one of the reasons i like TLs is that they tend to be low pressure.

dave
 
Dave,

Do you mean that the push-push drivers should somehow be physically connected (glued baqck-to-back), or do you refer to theair in the enclosure when you say that they are mechanically connected?

If glued back-to-back, I can't see how it'll ever make it to the rear of the speaker, unless you have very funny looking cabinets. ;)

Jennice
 
frugal-phile™
Joined 2001
Paid Member
Jennice said:
Do you mean that the push-push drivers should somehow be physically connected (glued baqck-to-back), or do you refer to theair in the enclosure when you say that they are mechanically connected?

Some people go as far as gluing them -- i prefer to be able to disassemble them.

There are some few drivers designed so that a bolt can go thru the polepiece to attach the driver to the cabinet... a pair of those would allow very good coupling.

An externally hosted image should be here but it was not working when we last tested it.


More common is the use of a brace such as a dowl to connect the 2 drivers.

An externally hosted image should be here but it was not working when we last tested it.


If you add the ready rod to attach the fronts and pull the drivers together you get reaaly close to ultimate.

Just a 2x4 with hot glue used here:

An externally hosted image should be here but it was not working when we last tested it.


dave
 
Dave,

Thank you very much for the pic's. They say more than 10^3 words. :)

One question, though: How would you glue the second driver (on the speaker that is shown with the driver removed)?
If the glue hardens slowly enough, how do you ever get the driver out again, or is the spacing rod only glues to the one driver?
Is it made a tiny bit loner than the distance between the drivers, so that there is tension on it to begin with?

Jennice
 
frugal-phile™
Joined 2001
Paid Member
Jennice said:
One question, though: How would you glue the second driver (on the speaker that is shown with the driver removed)?
If the glue hardens slowly enough, how do you ever get the driver out again, or is the spacing rod only glues to the one driver?
Is it made a tiny bit loner than the distance between the drivers, so that there is tension on it to begin with?

The trickiest thing is getting the brace the right length... if it is shorter, then you can add a spacer (on 1 sub we did, a single sheet of paper was all that was needed.

On the pictured speaker i used hot glue... it basically just holds stuff in place and doesn't really glue it. The driver easily comes out with a bit of a tug. (matter of fact the stuff is so wimpy you sometimes find yourself "gluing" it a couple times because it doesn't even hold the brace in place.

dave
 
anyway....

I do recommend anyone to search and choose the drivers before to start any crossover choice or optimization.

This is specially true if you are newbie on DIY speakers matter. I do this because different brands (and models also) really have different response too.

If a midbass unit starts to roll-off at 3.5K and respond -20dB at 8K, you can not cross it near 3K or higher unless you want to sum driver's response to crossover one and get +20 dB/oct slope acoustic filter at low pass section.

I think MTM D'Apolito configuration is a good one since you know the features and use them. Uniform vertical response, combined with the fase relationship of second or 4th order crossovers (acoustic ones, of course) does very good results and the industry knows it. I'm trying to pick up the URL's where you can read about this and will post here soon.

I hope do not got you even more confused. Anyway I did learn it reading, reading and reading and reading :D
Internet is a great source of good information, since you are at right location.
;)

My 2 cents...

Euclides.
 
links

here some links to read about
:D ;)

http://www.rane.com/pdf/note107.pdf

http://www.rane.com/library.html

http://sound.westhost.com/articles.htm

this well known but...

http://www.linkwitzlab.com/crossovers.htm

this also:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=41240

:cool:

hope to help.

P.S. on above I meant uniform response as simetrical response of the polar lobes, as they are equal above and below the tweeter axis, and stucked by false cognate... "fase" as phase...
sorry my bad english...
 
Hisatugo, thanks for all the links, very helpful...in making more confused! :D
The ESP article says that more than a 2nd order is useless, and impedience compensation is good. The rane article extolled the virtues of a 4th order LR crossover (active, though). Everyone has their own opionion, and usually they have some merit. As a beginner, it is difficult for me to gauge which is reasonable and which isn't. Good stuff...
 
oh... sorry... (long answer)

As I would know, several opinions make this indeed.

:angel:

I was trying to show you how important is to choose your drivers first. If you make this choice and have the data on hand, you will make all the decisions more confident.

There are drivers that fit well for 2nd order Xovers and others 4th order (acoustic I mean, if the driver roll off 1st order and you sum first order xover you get 2nd order acoustic response if things are right designed)

But I think more than 2nd passive is something difficult to realize indeed. Have you tried to simulate the circuit in some computer software? To get phase and response fit the requirements for linear and coherent is a real pain. To simulate response of the drivers is another pain. :xeye: So you can imagine this in real life soldering, sticking, testing, hearing, soldering again and again and again...

So, If you don't mind I say it one more time: choose your drivers first and after that, have it on your hands. After that again, make some measures, if possible in the room you intend to hear them. The companies can NOT do this beacause they never know where their loudspeakers will play. BUT YOU CAN. This is an advantage, and if you move, you can modify it to new location.
;) Is this not good??


So with the measures on hands, (impedance, Qts, Le, RTA frequency response) you can sit down and think on how to cross it. ;)

Unless you are at least an eletrical eng. student, I do sugest also to search for drivers and projects well known and already succesfully mounted. They include xover circuits, already tweaked to last note and you will not get in error. ( Mr. Linkwitz has very good ones). You will learn A LOT OF DIY audio with them.

The initial projects choosed and realized by this way will not discourage and you can start a new one with progressive challenges.

I also suggest you do not loose the focus of the project: is it to be a construction project OR is it a loudspeaker for music and pleasure project? You can do both but it will NOT be the first one that let you get both. Because you (and me too) will always think something could be better and the next project will be better and so on

;) If you intend to hear music do remember to consider you taste for music when choosing drivers also. Hear some of them before to buy... if possible of course. Or at least, follow recomendations of somebody with similar music preferences but with more experience with different drivers.

To construct good loudspeakers IS A REAL possiblity. They should satisfy nobody but ourselves (of course some friends are part of us :D ). I would consider me as a newbie here, as our fellows have years more of experience and critical reading this forum was a great source of good info. I constructed 8 pairs of loudspeakers since the first one, 6 years ago. They ended trash or sold... I have only the last and the new one.

This 8th pair intend to be the last as I think is time to spend energy on amplifiers and maybe the digital source now.

Good learning!
:cool:

Best Regards. Sorry for long answer, took too much of your time.

Euclides.:smash:
 
Hisatugo-
Thanks for the advise. I appreciate all the help I can get, I am new at this, but like to learn. Reading many opinions is good, it just seems that many are contradictory, and without any experience it is hard for me to judge the good ones from the bad.

Thanks again!
 
frugal-phile™
Joined 2001
Paid Member
I have to agree with Hisatugo. You need to choose drivers 1st (unless it is a case of what you are gifted with). I like to choose drivers that are very extended, i usually start with a "full-range" to do the middle since i prefer not to have any XO points between 300 & 5 K and further out is better (80-150 at the bottom, and 6-15k at the top (if needed)). If the "mid" speaker is listenable without any help at either end, i consider that a good start.

I also like to get as high efficiency as i can, unfortunately with my favored 4" in the middle, this usually limits me to 90-93 dB.

dave
 
uhm...hum...

Sbolin,
oh, you are welcome...

So let's show you what I did, as Dave did. So you can learn collecting experiences...

Drivers are Dynaudio 21W54 / 15W75 / D-260 (out of sales)

The last (8th) speaker have LP at 360 Hz, 3rd order electrical (4th acoustic) BP at 360 Hz/ 2600Hz 2nd electrical (4th acoustic) and HP at 2600 3rd electrical. I'm serious thinking on active as soon as my gain clones are ready. Passive in this case are very difficult to realize as I have 6 drivers each channel (two of each one) in WMTTMW config. Response has several small dips that I think comes from the room response and reflections I am dealing with.

Unfortunately I can not stay out of crossing in the 300 to 5K as Dave recommends (good advice). The response curves of the drivers do not permit or would make circuit a real challenge.

Regards..

Hisatugo
 

Attachments

  • hime4 1.jpg
    hime4 1.jpg
    39 KB · Views: 135
Dave,

I'm really starting to like your concept of the push-push woofers for "active" BSC.

There's still one important thing (probably mre things later :rolleyes: ) which I haven't figured out.

Obviously I need to figure out the formula for calculating the BS frequency, and based on that I need to find the appripriate frequency for the LP filter for the compensating Mid/woofer.
BUT: Is it a simple 1st order phenomenon, requiring just an inductor in series with the compensating driver, or what lind of filter would it need?

The units I had in mind are:
Tweeter: Scan-Speak D2905/9300000
Woofers: Peerless HDS 6¾" 850439 (I'll need two for each side then).

My plan was to make an active XO for a 4 ch.amp, maybe with 2nd order, crossed at 3.5kHz. One channel for each tweeter, and one for each mis/woofer section in each side. Then I'd still need the passive filter for the BS compensating woofer.

Would this be good, or can you see any "obvious" mistakes you'd make fifferent?

My reference (for comparison)sound at the moment is the B&W 309 floorstander, and I'd like to make it better than that.

Jennice
 
frugal-phile™
Joined 2001
Paid Member
Jennice said:
Obviously I need to figure out the formula for calculating the BS frequency, and based on that I need to find the appripriate frequency for the LP filter for the compensating Mid/woofer.
BUT: Is it a simple 1st order phenomenon, requiring just an inductor in series with the compensating driver, or what lind of filter would it need?

luckily the frequency you need to roll the back one off at is right near were a woofers impedance curve is usually its flattest so an inductor is often all you need. Sometimes a zobel to flatten the impedance rise might be needed.

The f=4560/w (in inches) gets you in the ballpark. "The Edge" simulator or the FRD BDS will get you even closer.

The beauty of the 0.5 woofer on the back is that as long as you err on the side of too high, because it is in the shadow of the cabinet you still get perfect baffle-step compensation, you are only worried about extra total energy in the room coming from the back speaker. You could even start out by just letting the back speaker run ragged and see how it interacts with the woom/

My plan was to make an active XO for a 4 ch.amp, maybe with 2nd order, crossed at 3.5kHz. One channel for each tweeter, and one for each mis/woofer section in each side. Then I'd still need the passive filter for the BS compensating woofer.

Thst works -- you could also use 3 amps/side (an advanatge if you are trying to get max level out of 2-8W tube amps)

dave
 
Ho Dave,

Thank you for the description and and the links.

I have 150W x 4ch at my disposal, so I don't think I'll run out os juice. :D
My only thought is, if it's overkill to make this bi-amp thing? If I could use a reasoably simple passive filter, maybe it would be just as good?
Then again, tweaking filter values and volume levels is a lot cheaper when talking about small signal components.
Speaking of which...
I have been thinking, whether an LC-based low pass filter or two RC filters after each other gets me the best result. Theory says I should get 2.nd order from both (IIRC), but what does real life say?

Jennice
 
frugal-phile™
Joined 2001
Paid Member
Jennice said:
if it's overkill to make this bi-amp thing? If I could use a reasoably simple passive filter, maybe it would be just as good?

Anything more complex than a single cap on the tweeter, go active

Speaking of which...
I have been thinking, whether an LC-based low pass filter or two RC filters after each other gets me the best result. Theory says I should get 2.nd order from both (IIRC), but what does real life say?

Not many LC active XOs get made since RC is easier.

With that 4 k peak on the HDS, 3.5 kHz is going to be interesting.

dave
 
planet10 said:

With that 4 k peak on the HDS, 3.5 kHz is going to be interesting.

dave

Hmm... Somehow I don't like that word... "Interesting". ;)

Where would you cross it? The tweeter data sheet says that I should go no lower than 2.5kHz with a 2nd order filter, but I was thinking of using the drivers roll-off.
I don't like going too close to the 1-2kHz either.

The beauty of the 0.5 woofer on the back is that as long as you err on the side of too high, because it is in the shadow of the cabinet you still get perfect baffle-step compensation, you are only worried about extra total energy in the room coming from the back speaker. You could even start out by just letting the back speaker run ragged and see how it interacts with the woom/

I don't think I understand this part... or the consequences.
What do you mean by "letting the back speaker run ragged"?

Jennice
 
Hmm.. interesting thought.
If there's a rise at 4kHz, then why can't the summed (with the tweeter) SLP be corrected reasonably well by letting the tweeter HP start a little "too high". If the woofer had a flat response, it would cause a fall in the total SPL, but maybe the peak could compensate?
If this approach was to be used, shouldn't I try to create a filter for the tweeter, which has a similar steep curve, to match the woofer roll-off, giving a constant SPL in total?

I always thought it was a sin to let a driver be pushed to it's natural roll-off, but is that only for tweeters then (dur to damage with LF signals)?

Jennice
 
frugal-phile™
Joined 2001
Paid Member
Jennice said:
If there's a rise at 4kHz, then why can't the summed (with the tweeter) SLP be corrected reasonably well by letting the tweeter HP start a little "too high". If the woofer had a flat response, it would cause a fall in the total SPL, but maybe the peak could compensate?
If this approach was to be used, shouldn't I try to create a filter for the tweeter, which has a similar steep curve, to match the woofer roll-off, giving a constant SPL in total?

An externally hosted image should be here but it was not working when we last tested it.


The peak is probably from ringing in the cone which has more problems than just the FR

I always thought it was a sin to let a driver be pushed to it's natural roll-off, but is that only for tweeters then (due to damage with LF signals)?

A midbass that rolls off smoothly on its own is a good thing IMO. This is a natural 2nd order rolloff. This is the kind of driver i'm always looking for.

Arguments against just using the natural rolloff is that the off-axis response gets pinched just below where the tweeter comes in... choice of tweeter can minimize this, and the rest is your choice of compromizes. I like as few reative components as possible between amp & speaker.

Going active gives you a lot more choices.

dave
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.