Use of large pro driver as OB-feasible??

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

I originally posted this question in the Ariel thread, which on hindsight was a totally silly thing to do as it understandably got totally overlooked so here it is all on it’s lonesome.

Tiny bit of background. I’ve only ever heard two dipole speakers, one was the Alon 4’s and the other the Orions. To be honest, they didn’t do it for me, BUT the dipole bass on the Orions I thought was very nice indeed.

So, it has got me thinking about converting my sealed bass to dipole. I use an 18 inch PHL sealed in about 110 litres. It doesn’t go deep naturally, with appropriate eq I can get it flat to thirty hz then it drops like a stone.

Now, I also happen to have four peerless 10’s a la the orions, but there is no point in going dipole with the tens as in the orions and then using the phl’s as a sub to go under them because they wont!! Ha ha.

I have a set of plans (from shinobiwan) that uses the four peerless as a sub to the high teens (I think) so I can use that for the bottom octave and a half, if I can turn the PHL’s into a decent dipole bass unit from 50-60 hz up.

7030 parameters

HIELE-SMALL PARAMETERS: TYPICAL (QC LIMITS)
Resonance frequency
Fs 34(±5)
DC Resistance 5.5(±0.5)
Mechanical quality factor Qms 4,8
Electrical quality factor Qes 0,41
Total quality factor Qts 0,38
Mechanical suspension compliance Cms 115
Effective piston area Sd 0,128
Equivalent Cas air load Vas m3 ,260
Max linear excursion Xmax ±mm 8,0
Linear displacement volume Vd 10-3.m3 1,024
Reference efficiency 0 % 2,5
Unity load volume Vas.Qts2 10-3.m3 41,4

I’m wondering if my real problems may lie in how high I want the bass unit to go. I don’t really want to take my mids lower than 300 hz, and so for all I know that may be the problematical area for running a dipole bass.

Does the fact that it will only be a single driver mean that vibration will be a huge issue??

All help will be appreciated, but I certainly don’t expect others to design it for me ha ha, your help will be more like in the way of links etc.

I’ve waded thru Seigfrieds site once or twice now, it gets easier each time but I dare say it will take a couple more times yet before my old Pooh Bear brain gets it all.

The actual xover implementations and eq that might be required are reasonably easy for me, it will be in a fully tri-amped deqx system. Although with the addition of the four driver sub, it will be in effect a four way setup, but will probably run the sub with a piece of behringer kit rather than an extra deqx!!!!

I will leave the mid and high as it currently exists, and just go for the dipole bass. The other option is too simply hack off the back of my existing sealed box and measure what happens, ie a version of suck it and see. The depth of the box is only about 300 mm, so in effect I suppose there wouldn’t really be much separation of the front and rear signal.

Hope I have given enough info that someone doesn’t have to ask for much more, and I hope I’ve done it without boring the hell out of you!!!

Thanks a lot

terry
 
Administrator
Joined 2004
Paid Member
Hey Terry - funny you didn't get any responses to this. So I'll bump it.

I haven't ever heard the PHL drivers, but they get good reviews.
No surprise that your 18" drops like a stone under 30Hz, at least that's what I would expect in OB, don't know about sealed. Still, that's pretty decent, you know. Do you really need that 20 to 30 Hz range? Or are you just not happy with the low end below 60?

You don't say what your mids are (or did I miss it?). A lot of good 18s will play pretty high, it's only beaming that becomes an issue. With that deqx you should be able to get out of the 18 real fast, if need be.

The Q of the driver is a little low for OB, but you may be able to make up for that with lots of power and EQ, a la Linkwitz

From what some OB users have told me, the single 18 can be a vibration problem. Sometimes they "walk" if you play them hard. I don't play my single 15s that hard, so mine stay put. :)

Let us know what progress you make.
 
OK, probably bored the hell out of you all ha ha. No, more likely that only those who helps themselves get help-and fair enough too I say.

To that end, just been to the local hardware and grabbed some roughed up gyprock (drywall I think you call it) for nix, and have decided to simply mount the woofers and then by measuring see what happens. Suck it and see indeed.

Anyway, will get started soon and post the measurements so others may learn something. It will probably confirm what some already know, but at least for me I will have learnt something.

will keep you informed.

isn't it always the way!!! post something and then find that someone has just posted ha ha.

Just to clarify, the 30 hz is the phl's in a sealed box, and eq'd to the max. I intend to see how low they go OB, and am not fussed if they only go as low as say 50 hz as I can put a sub under them.

The mids are also phl, will have to look up the number if someone is interested, but it is their dedicated - no compromise mid. If you can put any stock in marketing language!

The tweeter is the cabasse dom 40.

yes, will certainly keep any interested people up to date.

The bass is very tight and certainly visceral, just the way I like it!!
 
Disabled Account
Joined 2007
Hi. Two words: Linkwitz transforn. If you want bass from that driver below 30Hz, OB is not the way to go IMO. That PHL is perfect for LT, just needs lots of power. Your box maybe a little big for it, but easily corrected.
It's what I use, and I'd never go back.
 
If you are satisfied with 50Hz, with appropriate EQ, you should be okay.

I'd be more concerned about the 300Hz top end you want. Look for the drivers 1st breakup frequency...you may see the impedance curve "wiggle" at that point. Ideally, XO no higher than 1/5 that frequency, but 1/3 should be okay with a steep slope and a notch filter. This will minimize harmonics exciting the cone breakup.

7030 Mms is fairly low, so you may not have a big problem with vibration...though I can't imagine too much baffle mass for big cones.

Here is the BMS 18N850 in a 15" felt-damped U-baffle. Because of the baffle it doesn't need much EQ at 50Hz. However, Mms is high so even the 1.5" baffle is not heavy enough to prevent vibration. 1st breakup is just over 1k so I haven't tried it above 200Hz.

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

I notice you build the U frame woofer based on the 15" frame, you be able to advise me which sort of parameters one shuld be looking at in order to build a U or even H frame woofer.
example
Qts should be medium high around 0.4 or higher?
and what about other parameters?

thanks...
 
Wow, thanks guys. Nothing for a day or two, I wake up and all of a sudden there is heaps!!

mml2, currently they are in a sealed unit (not sure my first post was all that clear?) and as you say, eq'd heavily. But being big pro drivers they can certainly handle the power. From memory in the sealed box the Fb of the box is something like 70 hz, so an awful lot of eq is used to make it go lower.

Once they get put into an OB situation, they will I'm sure still require heavy eq. Those duties will be handled by the deqx unit. As I mentioned I'm not that fussed how low they end up going because I can throw a sub in under them.

BTW, if the sub only goes up to (say) 50-60 hz, will a single sub be sonically noticeable?? or should I split the single sub (four drivers) into two stereo subs of two drivers each and make it a true four way rather than three way plus sub.

Rudolf thanks for your suggestion, have come across ripole before. Will have to go and do a bit of looking up later (need to get some work done) but from vague,vague memory a ripole has two drivers??? As I say I need two go and have a look.

Paul, thanks. Interesting you should mention the upper frequency as the potential problem, I wondered if that might be the case as well. My measuring techniques at the moment are limited to frequency sweeps. Would the breakup frequency you mention be evident in that??

For my education, what is the effect of going higher than i/3 breakup frequency, does it simply leave a hole in the response for example? Bear in mind that the deqx can measure and correct the measurement, so any problems in the frequency domain I'm not to fussed with, but as you can tell the finer points of these interactions are a bit beyond me at the moment.

If I need to bring down the mids a bit I can I suppose, as the deqx would allow me to put say a 250 db/octave slope on them, which would cut out just a little of the LF energy methinks.

The potential vibration problem I guess would be minimized the higher I cut them off, but of course the first goal would be to have them go as low as possible.

Really appreciate everyone's responses, will post some measurements soon and we can all have a look see.
 
My measuring techniques at the moment are limited to frequency sweeps. Would the breakup frequency you mention be evident in that??

With adequate resolution and a low level of room reflections, yes. Breakup is usually visible as a narrow rise in SPL, or even sharp 5-15db peaks, just before the driver rolls off at HF.


what is the effect of going higher than i/3 breakup frequency, does it simply leave a hole in the response for example?

Harmonic distortion components generated within the driver are not attenuated by the crossover...so driver harmonics falling on the breakup rise/peaks are "amplified" by the rise or peak and can sound quite nasty. For example a 1k breakup can be excited by the 3rd harmonic of 333Hz, 4th of 250, 5th of 200 and so on. Really high order driver harmonics are usually very low level so they don't matter so much.

Also, many drivers exhibit energy storage just below the main breakup. So, even if you don't have the means to view ES, it is "good practice" to XO drivers well below breakup.

For clarity, the notch filter I mentioned for a "3x crossover" won't help the driver harmonic situation either...the purpose of the notch is to keep acoustic response on target. (A fundamental signal will excite the breakup just as easily as a harmonic.)


a 250 db/octave slope on them, which would cut out just a little of the LF energy

Below crossover that is true, however the XO turnover knee goes slightly lower (closer to XO frequency) with steeper filters so peak driver excursion is actually higher with high order crossovers. (First order excursion continues to increase below crossover...a different can of worms.)
 
thanks Paul

my fault for asking questions that have answers WAY above me ha ha.

Will do the measurements and post them, and see what we can glean from that.

I'm not sure if i was clear, or indeed that you know what the deqx is, but I spose you do. The deqx of course is a digital xover, so do the phenomena you mentioned re driver excursion etc still occur with digital xovers?

Anyway, let's not waste your time answering questions that frankly at the moment I'm too dumb to understand, let's leave it till i have some graphs to post, and at least that way we can all put your answers (which are very much appreciated) into context with the help of the graphs.

Hope to do a bit of mucking around with it over the weekend, so please bear with me till then.

see ya
 
Hi all

Today quickly hacked an appropriate hole in the wallboard and took some quick measurements. Will quickly throw them up for comment, and then relying on the wide range of knowledge on these forums will get some suggestions of what to vary and measure, that is if it is of interest to others of course. To sum up, or recapitulate, I want my bass drivers in open baffle configuration to be able to go to at least 50 hz as I can supplement the bass with a sealed sub if need be, and I need the bass driver to be able to go up to around 300 hz or so, as I don’t really want to bring my mid down much lower than that.



I measured and corrected using the deqx unit, and then remeasured those results with REW which I assume a lot of you will be familiar with.

The driver is currently in a flat baffle 1.2 metres wide by 1.2 meters high, extending to the floor. The driver itself is centred approx 580mm off the floor, in the position in relation to the mids it will be in if the thing ever gets built. I have the bass unit as close to the mid physically as possible so that the two stitch up as best they can.

Quite obviously the finished product will not be that wide!!, although perhaps I could get used to it ha ha. I have more gyprock with which to add wings at angles etc if needed, but for now will post this and get some feedback perhaps.

I tried previewing my reply, but didn't see the attachment in the preview. According to the FAQ it should appear when I post so here's hoping that the image will indeed appear!

The blue is the raw response, the green after correction by deqx and with xover in action. The fuuny thing is that on the OB the response 9from memory) seems to be signinficantlt beeter and go lower than when measured in the sealed box!! Is that at all possible or has my memory totally left me?? Can measure the response in the box later if it would be of interest.

Paul W, there is no smoothing applied , does it indeed show the breakup frequencies you spoke of?? By the way, these measurements were done at a distance of around four inches from the dust cap, is that the right way to go about it and if not what suggestions might you guys have???
 

Attachments

  • phl 7030 raw and corrected in ob config.jpg
    phl 7030 raw and corrected in ob config.jpg
    93.2 KB · Views: 275
Hi Paul,
had a kinda brainwave, namely re-read your post! and sure enough, your statement of breakup looking like a series of 5-15 db swings just before the rolloff seems to be evident in the graph.

That first little dip after 1k is 1.3, then 1.72 and then bang on 2K.

So it looks like good news regarding our worries at the upper point of 300 hz, as it seems that the first sign of breakup starts at 1.3 khz which is well outside our rule of thumb of 3x the upper frequency cutoff point.

On the green graph, at 50 hz the measurement is 70.2 db, and so the frequency at which the measurement is 67.2 is 38 hz, which I am well pleased with.

Have not done any db measurements from listening position yet, and so until we get around to that have no way of estimating the excursion .

Just guesstimating the amount of boost to get 50 hz up in line with the rest of the graph, it looks to be only about 3 db which again is wonderful. That will probably change with a smaller baffle though!
 
Hi Terry,
Yup, you got it! The breakup region appears to be just over 1k.

A couple of suggestions:
15db per division makes good looking graphs, but it is also more difficult to see problems. Try increasing measurement resolution to ~5db per division.

Also, try increasing the measurement distance to about 1M. Measuring so close to the dust cap may mask other cone problems and might also underestimate the amount of EQ boost you'll need. The main issue with increased distance is room interference so don't be afraid to make multiple measurements...if you can take the speakers outside try that. Different perspectives can give you a better picture of what you'll hear in the end.
Paul
 
terry j said:
....By the way, these measurements were done at a distance of around four inches from the dust cap, is that the right way to go about it and if not what suggestions might you guys have???

I hope you're not thinking that the measurement represents any consideration given to dipole cancellation, because at that distance you've taken the rear wave out of the equation, and left yourself with the driver's free air response (essentially infinite baffle).

Measure at the listening position. Then you'll see a sloped response.
 
Thanks all.

I had absolutely no real idea where to measure dipoles from, so just to start somewhere ZI did a near field measurement. It at least was useful in seeing the breakup nodes (I can pretend I know what I’m talking about now), but I always suspected that it wouldn’t be a true picture as no summing of front and back wave had occurred yet, thanks for the clarification John.

From that viewpoint, I guess you were right John, and as good a position as any to take the measurements may as well be the listening position, which is the graph below. Both measurements the same, one with 1/6 octave smoothing for a bit of clarity. The mic is now 3.3 metres from the driver.

Of course, straight away I see a few things, the most obvious being the (now apparent) rapid fall off in the frequency response, which by the look of it will require quite a bit of boost. That’s OK as it stands I suppose, it’s to be expected and these drivers already cop a huge amount of boost in the sealed boxes ( and much lower than 50 hz by the way), so the amount of eq required in and of itself is not too daunting (perhaps).

But there must be quite a bit of room interaction in these measurements, (yes?), and so any correction of that signal using the deqx I would be a bit wary of. But, boys will be boys and so I couldn’t resist a bit of rough, quick and dirty mucking about with eq just to get a feel of what can be done with the eq to get the response a bit flatter over the band of interest, and that little attempt is in the second graph, again one raw and overlaid the same with 1/6 smoothing.

The overall trend is kinda level, but with those huge peaks and troughs evident. Would they be the additions and cancellations due to baffle size?? Or just simple room interactions and nulls??

The next step is of course to remeasure these from the same distance but this time outdoors and see what clarifications we can glean from that. Bit of the proverbial pain, but probably needed to make some sort of final decision.

My gut response from looking at all this is that at the end of the day this plan won’t be a goer, but we may learn more after the next measurements. As I know very little really in all this, what would your gut feelings be at the moment regarding the possible success of this enterprise??/

Did a tiny bit of research regarding ripoles as suggested a bit earlier, and I at least found out that they don’t require two drivers ha ha. So as some sort of possible back up will do a little more looking into that. Is anyone able to provide some sort of summary regarding the similarities or otherwise of the sound between dipoles and ripoles??, first reaction would seem to be that the ripoles are more closed in, does that mean we lose the essential characteristics that we are after in the dipole??

Any other suggestions on how or what to measure with these? Hope that was of a little bit of interest, of course we have yet to vary the size of the baffle, but as the results already seem borderline I somehow don’t think going to a smaller baffle will help!!!
 

Attachments

  • before eq, raw phl 7030 in ob.jpg
    before eq, raw phl 7030 in ob.jpg
    88.3 KB · Views: 203
terry j said:
Did a tiny bit of research regarding ripoles as suggested a bit earlier, and I at least found out that they don’t require two drivers ha ha. So as some sort of possible back up will do a little more looking into that. Is anyone able to provide some sort of summary regarding the similarities or otherwise of the sound between dipoles and ripoles??, first reaction would seem to be that the ripoles are more closed in, does that mean we lose the essential characteristics that we are after in the dipole??


This comparison of dipole and ripole radiation shows that ripoles are indeed "compromised", but you don´t loose the essential characteristics. Left diagramm is CB, middle open baffle dipole and to the right ripole.
 
Terry,
The 3.3M measurements are good for examining room response, but now we are so far away we can no longer see what's going on with the driver itself because of the baffle and room. The dust cap itself may dominate your 4" measurements, and the baffle/room dominates 3.3M, so I'll again suggest distances around one meter (or maybe .5M) to see if you can get a good shot of the overall driver response with minimal room interaction. (Will the DEQX do time gated measurements?)

Low frequency driver measurements indoors are very tricky. Going outdoors to a large flat surface simulating the floor, but not room walls, should work. Ground plane measurements may work even better.

Personally, I wouldn't use a ripole at 300Hz. Direct frontal radiation should present fewer problems at that frequency.

I wouldn't give up just yet.
Paul
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.