Driver suggestions for high output open baffle speaker

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

Thanks for your comments.

In order to make SPL measurements you really need a calibrated meter of some kind. Otherwise it's just guesswork.

If our numbers don't agree, it's about the numbers and how they are worked out.

It's not so much that our numbers don't agree, but rather that your measurements don't agree with your own calculations. I am simply trying to help you work out where the error comes from.

At the moderate average power levels you are using, power compression will not be significant effect. With a 50W amp, the average power levels are probably of the order of 5W or so, not enough to cause power compression even with puny drivers.

I doubt that two pairs of speakers in stereo is going to add 6db to the output

I assume you mean one pair. I agree that a stereo pair will not sum to give exactly a 6dB increase over a single speaker. At lower frequencies program material is usually presented equally left & right, but at higher frequencies the effect will be less. It will be program dependant. The point is that the second speaker will add significant extra SPL to the room.

I'm also not sure the room is going to contribute 3db to the output.

Again it depends on the listening environment. In an anechoic chamber there is no reverberant field. In a reverberation chamber the reverberant field will effectively double the average acoustic power in the room. A normal listening room falls between these two extremes and without knowing the specifics of your listening environment, 3dB is a good estimate based upon my own measurements. You can easily measure the effect for yourself.

In any case, it seems clear that you want more SPL. The first place to start is with more amplifier power. I recommend at least 200W per side for mids & highs as being a good amount that will allow both higher average & peak levels. Again, with the high crest factors of music, the thermal capability of electro-dynamic drivers and the unyielding clipping point of most amplifiers, amplifiers need to be considerably over-rated.

Cheers, Ralph
 
Here's the power spectra of pop/rock music that I compiled some time ago.

My aim was to work out the SPL required for each driver in a multi-driver speaker system.

Conventional wisdom suggested pink noise (-3dB/oct) was a suitable substitute.

I actuallity I found that there was more MF & HF energy than pink noise. The drop was only 1.5dB/oct.

The program material was various pop/rock/electronic music. 14 different types in all averaged togther into a composite spectra. The filtering was done in Cooledit in octave wide frequency bands.

Cheers, Ralph
 

Attachments

  • average power spectra for pop rock music.gif
    average power spectra for pop rock music.gif
    9.5 KB · Views: 446
My SPL readings accuracy relies upon:

1. accurate specification of the sensitivity of the electret from the manufacter

2. correctly entering in the sensitivity into ultracurve

I have no reason to doubt that the published sensitivity is wrong. I will have to check that I have calibrated the level correctly.

Would a $50 radioshack SPL meter be any more accurate? They may not agree, but then how do I know which is more accurate?

I think there is a relatively simple explanation for the numbers. The actual input power is less than I might choose to simulate.

The SPL for 10w is 89 db. Add 3db for reverberant field and we have 92 db. This is probably closer to what is really going on in the midrange. If I try to push more, the sound is strained.

To get a better idea of what my speakers really can do, I should hook up the behringer. I have done this before, but not to find out how much juice I can get. I recall the bass was more than I thought they were capable of, and better dynamics. But still, this won't get near my target.

How do you know there is little or no power compression. If a 250w pro driver designed to deal with thermal compression has 2db of compression at 40w and 5 db @ its rated power, surely a hifi driver will have at least the same if not inferior performance. I'd expect the hifi driver would range from 2db at low power levels going up to about 5db at full rated power. The hifi driver will normally have more excursion related compression I'd guess, since it relies more on excursion which is not linear, while a greater portion of the pro driver's compression is related to thermal effects.

Ralph, interesting site. Is it a commercial site in the making?
 
Hi Paul,

Would a $50 radioshack SPL meter be any more accurate?
I can't say how accurate a Radio Shack meter might be, but there are reasonably priced Class 2 meters that are calibrated to within +/-1dB and that's plenty good enough.
How do you know there is little or no power compression
I don’t believe that power compression is an issue with your setup. You simply don’t have enough amplifier power to cause power compression in your midrange driver. For example, consider that the JBL 2226H incurs 2dB of power compression when driven by band limited pink noise at 50% of its rating. For the 2226 with a rating of 600W that’s 300W average. Considering the crest factor of pink noise is 11dB, that means the driving amplifier is putting out undistorted peaks of 1kW! Admittedly the 2226 is a very well designed driver with a very high power handling ability.

At these power levels special techniques are needed to dissipate the heat. For a Hi-Fi driver the problem of heat dissipation is much less severe as the power levels are much lower.
For a 50W amplifier the average power level of music will only a few watts. That amount of heat is easily dissipated by any modern driver.

If you’re convinced that power compression is an issue for you, you can measure by how much. Using pink noise, measure the RMS power fed to the driver and the SPL at a low power level, say 1.5WRMS; and then repeat the measurement at 15WRMS ( the maximum for a 50W amplifier to reproduce undistorted peaks in pink noise) and measure the deviation from the expected SPL. I doubt you will see a measurable difference.
If I try to push more, the sound is strained.
I can understand if your midrange driver is sounding strained as it is probably under powered. Even with adequate amplifier power you may still feel it sounds strained. In that case driver non-linearities causing distortion are likely to be responsible. A better quality driver might be called for.

Large diameter midrange drivers that I’ve had good experiences with include the JBL 2123 and PHL 2460. Both of these drivers are very linear at high SPL’s. Interestingly, due to their high efficiency, only moderate drive levels are needed to produce huge listening volume.

I find it's often a good exercise to see what commercial products are offering. For example the SCM150 ATC studio monitor uses a 1" dome tweeter to acheive SPL's of 117dB/m. That's loud! The speakers are tri-amplified with a dedicated 50W amp just for the tweeter. That's about the limit of what's achievable with an unaugmented 1" dome in a 3-way system. To go louder both ATC and Genelec move to a 34mm tweeter powered by a larger amplifier.

A waveguide can add up to another 6dB to the direct sound field. But keep in mind that a waveguide achieves the extra on-axis level by reducing the off-axis energy. For a semi-reverberant listening environment, some of the apparent gain in SPL is therefore lost. Something like a 3dB gain in loudness would seem realistic with waveguide loading compared to an unaugmented dome tweeter.
Ralph, interesting site. Is it a commercial site in the making?
Thanks for the feedback. I’d love to be able to make it a commercial site one day, but at the moment I’m just trying to make it a useful resource for hobbyists and constructors. I believe that speaker design is mostly an engineering challenge, and I want to show the steps and thought processes needed to take a design through the system design process.

Cheers, Ralph
 

Attachments

  • power compression.gif
    power compression.gif
    35.1 KB · Views: 398
paulspencer said:


AJ:
I will happily have drivers shipped from overseas if I think I can get better drivers, or more suitable, or better value. I'm quite fussy with drivers - not happy with something local if I think I can get something I like from overseas. I bought my subs that way.

I don't know if it will meet your driver "look" factor (something irrelevant to me) but I'll give you my 2c on what you might consider. I'm building something similar with more pistonic materials, but you want a softer resolution picture, so here goes:
350/450hz on up: http://www.bmspro.info/index.php?show=item&usbid=10284&id=5059992
If no BMS dealer there try http://www.assistanceaudio.com/02_BMScones.html stateside.
If you have a Coda dealer down under you could hopefully audition something like http://www.codaaudio.net/index.php?show=item&usbid=18612&id=5060773 to get an idea of what type of sound. Not the same driver, but a taste of the sound.
XO the tweeter 1.8-1.9k as recommended by manufacturer.
70-350/450hz http://www.eminence.com/pdf/kappalite-3015lf.pdf
2/ch, available http://www.bltsound.com/index.php?s...162052833&update_item=0&edit_item=0&x=63&y=15
or possibly locally. Mounted WCXW. @400hz 15" CTC (all drivers)places them within 1/2 wavelength for good summation. EPD in the 500mm range will be a enough (450w x 1700h perhaps?) and still keep the (flat) baffle size reasonable. Make it 1 1/2- 2" thick, no (or minimal) wings.
All drivers are very lightweight so shipping (if necessary) won't be astronomical.
Good luck.

cheers,

AJ

p.s. keep in mind the the above piston drivers only approximate point sources at low frequency
 
To anyone- have been listening to planar 2 way sat dipole with mono sub. Read somewhere best results from sealed treble and mids instead. First attempt at open baffle ( hybrids? )as follows:mids 108 db phenolic 1",treble 105 db titanium, both feeding 12" round waveguides, fmod highpassed tube amp at 500 hz. Cap filtered tweeter at 8000 hz. 98db 15" woofers fmod low passed Nait 5 at 200 hz, which ought to flatten out a "knee" in the drivers' response . Baffle 19mm void free birch, 24"x 57". Can get decent sound, but Naim is wound up to 11:30 even at lower listening levels. Is this normal to keep up with a 6 or 7 watt tube amp? Am I in danger of damaging the Naim. There's no way I'm getting 98db on axis even summing two 15"ers. Any help appreciated.Btw, is it normal to struggle to blend a monopole sub with a dipole woofered speaker. Just trying to fill in @ 35 hz seems to ruin pitch.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.