[terry j]

hi all

did a search on thread titles, and there was nothing obvious (although I am sure the answer is in one of the threads somewhere, but which one?? ha ha)

In most cases I have seen, the discussion of parameters seems to centre mainly around the drivers used for the bass, in all likelihood to maximise the bass extension.

Am currently thinking about testing/converting my mids to OB, but am wondering if my driver is suitable (phl 1660).  Or, once we get past 'which driver to use for my woofer" does it become true that any driver can be used in OB??

Are there any possible pitfalls that might arise in going from sealed to OB eg danger of over excursion?

Thanks for your indulgence.

[jcake5]

Hi Terry,

I cannot provide any real technical information regarding specifications for OB mids.  My jump into OB mids was a just do it and learn from the results.  To date I have used various drivers that have the Qtc and Xmax all over the scale and have not been disappointed yet.  However, I admit that I volume limit my speakers to lower levels than required for most.  I have bottomed out a few voice coils on the low Xmax drivers and learn where the limits are for those drivers.  Also, using a woofer to mid crossover of 300-500 Hz really opens the door to many drivers since x-max is limited even at higher volume levels.

Like I said, I am no expert and I am willing to take a chance.  My wife even noticed a positive difference in sound quality when I went from a sealed mid to OB.

Have fun,
Eric

[terry j]

thanks eric, I appreciate the time you took to post your experiences and findings.

I guess you are right, the best way is always 'suck it and see', but I naively thought that there may have been some established guidelines in this area.

So I suppose the bulk of any OB theory is directed towards bass response??

Unfortunately, that would make me a PIONEER in ob theory!!! 

I will try my driver in OB alignment, and will report back here my subjective findings.  Unfortunately, as I'm an audio idiot (that might not be true but I certainly don't see myself on the cutting edge) the feedback may be as puerile as "yeah, sounds better" or " nah, sounds stupid", so I don't really see me as being able to contribute to the art.

But hey, on the face of it, that looks like it will be a worthwhile contribution as it seems that in the area of mids in OB there is, in fact, little or no empirical or hearsay data to go on!

My main concern is in the area of driver excursion, I aim to cross it at no less than 300 hz, I  may cross at 400 initially for hopefully a bit of safety and see if I can hear any stress, then lower it towards 300 if I can.I hope you all wish me luck cause it seems I am now PUSHING THE ENVELOPE whoo hhoooooo!!!!

Viva la revolucian!

[Rudolf]

My main concern is in the area of driver excursion, I aim to cross it at no less than 300 hz, I  may cross at 400 initially for hopefully a bit of safety and see if I can hear any stress, then lower it towards 300 if I can.

It is always a good idea to give some limiting condition, so people know where to start from.  :winkFirst you could simulate in EDGE, whether your planned (got a plan?) baffle width supports 300/400 Hz or if you need some equalisation (how many dBs) to counter baffle loss. Fs of the driver should be at least one octave below 300/400 Hz. From Qts and Fs of the driver you can see, how many dB the driver will be down at the x-over frequency. From there on everything which has been published about calculating woofer excursion limits does apply to your midrange driver too.

There is nothing like a "recommended set of parameters" for midranges in OB. Almost any midrange driver will work in one way or another - it all depends on what you want to achieve.

Just looked up the parameters of the PHL 1660. Should do well from 300 Hz up. Max.excursion before damage is given as ± 4 mm. I believe you can keep that under control by just checking with fingers.

[terry j]

thanks Rudolf,

you answered my question completely, but of course my ignorance has opened a few more! So, the actual T/S parameters do not play as much a role as for the woofers.  Makes sense, for the woofers it is how to estimate the bass response (I think).

Now that the parameters aren't really needed, I can see that I should have included a bit more data.  I intend to build a wide baffle (I have been inspired by the look of the Sonus Faber Stradivari), and as I use an 18" woofer I am by default forced to go wide baffle, so just accentuate it a little.

I've decided on using a baffle 750mm wide, and may make it about 1350mm high.  To be totally honest, I will not be worrying about the baffle step compensation.  I will be running it tri-amped (as they are now) by using the deqx.  It will of course make the measurements and create it's own correction filters, so a poor little 'air-head' like myself does not need to wrap his wits around one of the arcane parts of the speaker designers art.

Thankyou for looking up the driver, thought it may have cluttered here too much to list the parameters myself.  From your reply, (as you know) the Fs of the driver is 125 Hz, so is suitable for 300 hz crossover per your reply.  Using the deqx, I have the choice of very steep slopes on the x-over, for example up to 300 db/octave.  Out of curiosity, does being able to use such steep slopes help relax the rule of thumb about the Fs being an octave below the x-over point??ie does it offer much greater protection?

The next bit about using the Qts and Fs to determine needed boost etc is a little beyond me at this stage, certainly from just 'looking at the figures'.  I will however, when have a few spare moments, plug all those figures into Win ISD, and see if I can glean anything from that. 

I'm assuming it will give me useful data, but once again I have only  ever seen WinISD (and the others) used for bass drivers-can those programs also be used for midrange drivers?? (that is probably a very simple question, but hey!)

A tiny bit of background here, I have only ever heard two dipole designs, one of them being the Orion.  I was not that impressed, and from a couple of tracks that I was familiar with I concluded that quite a bit of detail was missing, and so kinda 'went off dipole' as not worth pursuing.

I realised that probably wasn't a fair test of different alignments, as I was using different drivers.  So for me it is a simple enough test to use my mids in a dipole type arrangement and then I will be comparing apples with apples.

As a result of a discussion elsewhere, I came to realise something I hadn't previously thought about.  And that is with my sealed mids, the sound off the back of the cone MUST, to some degree at least, come back out thru the cone itself which must muddy and colour the sound somewhat.

So from these two factors I thought it made sense to try my mids in an OB baffle, mainly with a view to compare the sound of the driver without the internal sounds coming back out thru the cone, rather than to 'take advantage of the dipole cancellation' at the sides.

I will probably try and absorb the rear wave thru copious amounts of stuffing (at least as an experiment) which will mean that it won't be acting as a dipole at all.  My thinking is that the rear wave bouncing off the wall behind the speakers is undesirable.  But hey, as I've already admitted I don't really know much, so that could all be rubbish.

[Rudolf]

terry,

this is a comparison for a 140 mm dia. cone placed at 1 m height in the middle of your 1350x750 mm baffle (green) and placed in the same location on a 1350x300 mm baffle (red).



Obviously it would be easier to equalize the red curve to the 6 dB level than the green one. For 300 Hz x-over the red SPL would be as good as the green one. You would have to reduce that green peak from 100-600 Hz anyway. So I would strongly recommend to taper your baffle accordingly.

Regarding Qts and Fs, have a look at this diagram:

]

With your qts=0,43 look at the qts=0,5 curve. At Fs=125 Hz (1,0 on the x-axis) the SPL is 6 dB down. At 250 Hz (2,0 on the x-axis) the PHL 1660 is about 2 dB down. So your driver doesn´t need any equalisation at the bottom end of the passband to compensate for a Qts-induced drop in SPL. Just use any x-over slope you find appropriate. Keep in mind that the 300 mm wide baffle would account for a 4 db/oct. slope on its own.
Anything up from 12 dB/oct. should be nice. 24 dB/oct. would be ample enough.

[terry j]

wow, those two graphs look different.  I've had the fixed idea for a while now that for OB the wider the baffle the better, I suppose that is a carry over from trying to maximize bass extension.

I don't expect you to do all the work for me, (but boy I do appreciate the time you have taken) so I have downloaded the edge and am in the process of nutting it out-by the scientific method of clicking buttons and seeing what happens, ie no knowledge at all!! ha ha.

I did manage to get a similar (but not exact) graph as the one you posted, so that is a promising start.

My main problem of course is using a large woofer, it is easier and more aesthetically pleasing to keep the large baffle all the way, and be totally lazy and rely on the deqx to make it all work OK.  But like everything, the less correction needed the better I would expect.

Will post some graphs later for 'review'! 

cheers.

So will muck about with various shapes and see what the response looks like.  In a general sense, do the real world results come close to the results predicted by the Edge??

[Rudolf]

wow, those two graphs look different.  I've had the fixed idea for a while now that for OB the wider the baffle the better, I suppose that is a carry over from trying to maximize bass extension.

Below 200 Hz the graph shows very well, what you want a wide baffle for.

... so I have downloaded the edge and am in the process of nutting it out-by the scientific method of clicking buttons and seeing what happens, ie no knowledge at all!! ha ha.

Very good decision! In the end it is always better to acquire knowledge for oneself than having to ask it from others.

My main problem of course is using a large woofer, it is easier and more aesthetically pleasing to keep the large baffle all the way, and be totally lazy and rely on the deqx to make it all work OK.  But like everything, the less correction needed the better I would expect.

You can get linear results on axis by moving the driver near to the edge of a large baffle. But by the same time this will change the polar response significantly, making response to the right very different from that to the left. That can, but must not be an issue. But in that case you better exactly know what you are aiming for.

So will muck about with various shapes and see what the response looks like.  In a general sense, do the real world results come close to the results predicted by the Edge?

Edge presumes an ideal driver with linear frequency response from 1-100.000 Hz. Edge does not account for any reflections from floor or walls. Within these limitations the Edge simulation is as good as it gets. I believe almost all simulation progs rely on the same math algorithm for edge diffraction.
Especially for the midrange, where most drivers behave quite linear and almost no floor boost exists, the results of edge should be quite reliable.

[terry j]

hi rudolf, and others!

been playing with edge, and have managed to get a graph that doesn't have that insane dip in it (peaks don't bother me at all, the deqx will easily handle them).

The simplest thing is to quickly knock up a test baffle and measure it, as the main point at the moment is the nature/character of the sound vs sealed.  Besides that, a numbing paralysis can set in if I just fiddle with simulations, if you get my drift.

So the quick question, if someone has experience with it, how far from the baffle should i do the measurements and then correction??  With the sealed setup, I normally measure from a meter, not sure if that is sufficient distance for a dipole.

Very soon I will be building a pair of subs, using four peerless xls 10's a side, but arranged vertically (when that height box is placed next to the simulated wide baffle, ie 'extending' the width and the mid is now asymmetrical in the extended baffle really helps knock out the huge dip in the simulations) which I could run up to say 80 hz.  The reason I mention this is that MAY allow me to run the (now) upper bass unit open baffle as well, as I'm idly wondering whether the transition from sealed bass to OB midrange (given the x-over occurs at 300-320 hz) may be discontinuous sonically at those frequencies.

Does anyone have experience with that, or am I making a problem which in all likelihood doesn't exist. But, the thing that I did love with the Orions was the OB bass, so it has it's own intrinsic appeal.

[Rudolf]

So the quick question, if someone has experience with it, how far from the baffle should i do the measurements and then correction??  With the sealed setup, I normally measure from a meter, not sure if that is sufficient distance for a dipole.

From my own experience: If the mic distance is at least the same as the baffle width you should be quite safe. 1 m will be fine. But that is only for the direct response. If you got that linear enough, you need measurements from the listening position to fine tune baffle distances from front and side walls.

Very soon I will be building a pair of subs, using four peerless xls 10's a side, but arranged vertically (when that height box is placed next to the simulated wide baffle, ie 'extending' the width and the mid is now asymmetrical in the extended baffle really helps knock out the huge dip in the simulations) which I could run up to say 80 hz.  The reason I mention this is that MAY allow me to run the (now) upper bass unit open baffle as well, as I'm idly wondering whether the transition from sealed bass to OB midrange (given the x-over occurs at 300-320 hz) may be discontinuous sonically at those frequencies.

If you "extend" the baffle by placing a box next to it, you will add depth. This additional asymmetry does not show up in Edge, but will have significant influence on the FR.
I would be happier if you could place the transition from box to open baffle one octave lower (40 Hz). This seems to be more natural since dipoles tend to "loose" below 40 Hz anyway.
Please try that bass box without the back (mimicking a U-frame) before you decide to seal it. May be you will prefer "open" bass all the way down as you heard it with the Orions.

Rudolf

[terry j]

good idea on trying the four xls 10's in a sealable box, listen sealed/unsealed, any required equalisation is easily achievable with the deqx.

Am I to conclude from your comments that only having the mids OB might conflict with the sealed box below it??

I'm a little concerned that my current woofer is not suitable for running OB down to as ;low a frequency as 40 hz (PHL 7030), indeed rigged up a test baffle a month or two ago and it didn't look like going very low at all, but I 'may' be able to get it running Ok as OB to maybe 80 hz-measurements will tell.

Is it a realistic proposition that four of the peerless will get me lower than (say) 30 hz in OB?  Not sure which program I can use to simulate that, but as you say, if I build it with a removal back then I can measure it. The reason I use 30 hz as the benchmark is that as it stands now I can get flat in room to 29 hz as it is now, don't see the sense in not going lower with an extra 8 woofers!!!(except maybe it would sound that much better).

Come to think of it Rudolf, I think you replied in my thread on running my 18's as OB over on DIY, you suggested that they may have been suitable for a ripole??? 

Anyway, thanks for your help, and do appreciate your time. Best bet is to just start and see how we go.

[kyrill]

hi terry
i discovered your post and i am very interested in yr findings. I use a focal 7W audiom as mid which has similar specs as your phl 1660 (7W: Fs 117 hz, QTS 0.49) I also use a DEQX but i "hate" wide baffles. The smaller and taller, the more i like them 
I prefer solid spruce or maple material for the baffle as the vibration characteristics (pitch) of the wood will sound (much) more musical if i may believe the theories of C37 of Ennemoser ( http://www.ennemoser.com/c37theory.html or http://www.mother_of_tone.com ) Not that I want the baffle to be a "secret" 3rd loudspeaker by moving air thru its vibrations,  but because it WILL pass thru its resonances to the magnet of the speaker, by the way of mechanical coupling.*
no way this can be avoided from happening as long there is contact between speaker unit and baffle.  The magnet as i understand it should ideally be inert and motionless. theory of designing speakers assumes this motionlessness to be there. This "motionlessness" is the reference for the moving magnetic field that drives the cone. All vibrations of the magnet however are distortions and this distortion will ratio 1:1 be in the moving magnetic field and immediately colour the sound of the driver. This "colouring" is distortion and probably impossible to avoid with speakers attached to baffles. As Aluminum has the worst sounding characteristic a baffle of aluminum would be last choice. mdf  sounds dead and closed in compared to spruce but more alive and dynamic than marble or stone. So in choosing baffle material you choose in a way for pleasant or unpleasant unavoidable colorations/distortions.
However the existence of vibrating panels is much more serious in box speakers then in OB speakers.
But how is your project going? Can you post photos?
*
the existing theory to make the baffle  stiff and heavy or very light but stiff ( see thread about baffle material) is still very valid. but even a 1000 kilo baffle will vibrate with the music and so influences the speaker magnet. If you hold an empty (!)  wine glass with the open end to the wall, you can listen to a conversation in the other room! If you listen to yr. stereo even the walls sing and dance along with the music

[el`Ol]

I discovered something that could be a good OB midranger, the Beyma Pro 6MI
http://car.beyma.com/ENGLISH/listaprod.php?id_familia=13
Maybe combined with a large Mundorf AMT (also 4Ohm and good sensitivity), plus active bass,...

[Dmason]

Oliver,

that does look mighty interesting. I always was impressed with the pro audio versions of this drivers, and you have found one that is Q-perfect for the OB gig. Nice. US Speaker has them as well. With alittle active contouring, these drivers would likely be all you need if you have older ears. Most people over 50 are typically down 6db @ 8KHz, something like that.

[D OB G]

Hi kyrill,

re baffle materials and magnet motion, for an alternative approach you might like to look at "An OB design" of July 19.

David

[kyrill]

thx David

[terry j]

hi kyrill

sorry to be so rude, but have been busy and not logged on for a while.

Funny I  popped on just now, cause in about an hour I will finally! start to move on all this.

You probably have way more 'speaker building smarts' than me, but I for one am glad I have the deqx as it will make my life so much more easy heh heh.

I don't mind the wide baffle look (if you haven't already done so, have a look at the sonus faber stradivari and THEN tell me it's not gorgeous!!), but I'm sorta forced to go in that direction as no matter what I do any baffle I use has to accomodate an 18" woofer, so I may as well make that a design feature.

As Rudolf pointed out earlier, from the edge simulations of 'our' mid driver on a wide baffle, wide may not be the optimum way to go, so that aspect seems to be on your side. Me???  Think I'm first off just gonna throw it on a wide baffle, measure it and see what (if anything) the deqx can do about it.

I also have 'dreams' of being able to run my PHL 18 as OB, already did some rough and ready testing on it in OB format a while agoout.  It doesn't really want to do it, but I'm hoping that I can get it to do it successfully from say 80hz up (even that may require a bit of eq!) because what I'm about to start today is build some sealed subs that hopefully will do nicely up to 80 hz when the PHL can take over in OB.

The trick I need to work out is that I already am running the deqx as tri-amp, so hoovew do I get my subs into the system (ie essentially a four way now which a single deqx does not support, and I'm sure I don't want the expense of another) yet still retain the capabilities of the deqx as a preamp with the fantastic remote real time eq capabilities.

I've just started experimenting with running a splitter from the bass output of the deqx, with one end going to the analog input of a behringer dcx 2496, which I can then use it's x-over facilities and eq for the sub. 
\
As the subs will eventually be under the 18 on the baffle, my hope is that if I can 'marry' the sub (via appropriate xover values and eq) to a close 'knit' with the properties of the 18 above it, then when I eventually measure the bass response the mic-it not knowing any differently will hopefully think the signal is coming from one driver rather than two seperate ones-will pick up the '''one''' signal and then do a correction which will give me a flat seemless transition from the low sub up to the OB upper bass unit.

I think I know what I mean, however I am very much less sure that you or anyone else for that matter could follow that tortured description!!!!

Will keep you informed

[kyrill]

hi Terry

Low bass is not too direction sensitive. So if you use (1 or 2) ) sidepanels, you can do the 18 inch on a side panel.

DEQX only have 3 analogue out per channel/.But a very low sub ( possible "under" the freq. of the 18 inch woofer ) can easily be done by ear. Just neglect  the sub in all DEQX measurements. Your sub amp needs a volume knob and a variable cut off  freq knob. Most class D subamps have them incorporated . I have a sub as well and it is only allowed under 60 hz. I put my finger on the subwoofer and feel it comes "alive" before i hear it ( the rest of the speakers play the music)
i do a "smidgen" extra and sit on my listening position, recalibrate until i am satisfied. with many more different music labels you get the feeling of the "music" setting. I use the sub for the HT as well for movies and put it a bit higher to feel the thunder when appropriate

PS  The Sonus faber stradivari pfff an intimidating speaker in most rooms. You must have a room of 24 feet+ wide and much more than that deep, and the Stradivary will become an obedient member on equal footing  with the rest of the setup, albeit never humble

[terry j]

hi kyrill

little setback tonite, looks like on of my sub drivers not a'workin, will have to have a look tomorrow.

I can't really put the 18's on the side, (you and many others may laugh) because I run them up to 300 hz and meet the mids there, the mids can't really go any lower but luckily for me the 18's handle that job superbly.

Funny you mention those room dimensions! my room is about 8m wide and 6m deep and 5.5m high, but with a large 2m deep bay right behind me, so in effect 8m deep.  Bang on your estimates ha ha. So all you have done I'm afraid is encourage me

I hope my driver is ok

[kyrill]

 A room 5.5 mtrs high? You live in a church, oh my
 must be,
 you are a holy man seeking to find redemption by listening to God thru thy speakers
 i bow to yr project 

ok i understand your baffle falls back to good looking dimensions again compared to the room you have.

Hmm i did not know that 100-300 hz need to be directed to you,  so it could not be in a side panel. But you dont need it as your baffle width fits your room easily.

300 hz sounds very good for a mid to begin. I play around 300-600 hz. Not made a choice yes as my open baffle midrange is not ready yet. Beautiful easy with the DEQX to play around with X-0ver freqs.

I hope yr driver recovers

show us yr photos