Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Loren Jones]

I was gonna do biscuits or put cleats in the corners for assembly, but a friend of mine suggested going with the pocket screws.  I am really glad he did.  It made assembly easy and precise.  I think there are about 250 pocket screws between the pair of them.

[Phillip_Graham]

Loren Jones wrote on Mon, 01 November 2010 08:27

Quote: I like the design can you post that too. I would love to post the design, however it isn't mine to post.  I can put you in touch with the LABster who designed them if you would like to get the design from him.  He may have some slight modifications that he has made since my Fb came out a little lower than the design called for.  I don't know about that. Perhaps he can chime in here if he cares to regarding his comments on the design or other thoughts. Take care, Loren Jones

Since it appears I have been "outed" on this design, I will offer a few thoughts:

0. I, in NO WAY, put Loren up to this.  I didn't even see the thread until today, apparently a day old.  I am glad Loren is enjoying his new baby.  I should say Loren and I have never met, and that I am incredibly impressed with the cabinetry of the finished item!

1. This design came out of a previous LAB thread where I disagreed at length with Art Welter about the principles behind high performance vented box design.

2. In that thread I mentioned the RCF 21" because I realized that it was a very affordable driver for what it offers.  I became aware of this driver as a result of unrelated conversations with fellow LABster Jeff Knorr.  There are better 21s (e.g. B&C 21SW115), but I'd never seen a driver with better specifications for the money at retail.

3. Loren took a risk on building this design.  I say took a risk because this came straight from a model.  Even with the more advanced lumped parameter model, the port tuning is rarely absolutely correct.  There was no guarantee the tuning frequency would be correct for this tapered port geometry.  Loren went into the project with eyes wide open about the potential of the box Fb to be wrong.

4. I presented Evan with a modified version of the design to further test the accuracy of the tuning behavior of the model.  To the best of my knowledge he has not built the modified enclosure yet.  The modified design could end up only useful for firewood, model accuracy depending.

5. There are many more detailed measurements that need to be made of this cabinet before I would consider releasing a public design.  Measuring a speaker system with SMAART for system tuning at a gig is much different than the detailed measurements one should take of a professionally designed loudspeaker.

6. Any sort of plans would also include a completed collection of measurements far in excess of what one normally gets from datasheets.  Those measurements would provide unambiguous performance behavior, good/bad/ugly.

7. Should all those measurements ever get taken (by me or a vetted 3rd party), then I would figure out what to do with the design.  More than likely I would sell it, not so much to generate revenue, but to discourage "non-serious" builders from approaching the project.  I am not interested in becoming "tech support" for a vented box design

8. The box design requires 2 bands of in-band parametric eq and a highpass, so any builder would need access to DSP processing.  Also, the box really needs an amplifier capable of 180+ V_peak (360+ V_p_p) to get everything the driver is capable of. For those with small amps, there are probably better ways to go.

9. At the end of the day, people should be realistic that this is still a vented box.  While I feel confident that this box exhibits real extension to 35Hz, and output on par with a high-quality dual 18" (what a dual 18 will really produce output-wise, not what the datasheet says).  Those expecting to move mountains and cause earthquakes have read too many manufacturers spec sheets.

10. While the coupling efficiency of a 21" cone is superior to an 18" cone, one should not expect the efficiency of a single motor structure on a 21" cone to match that of two driven 18" motor structures for a given v_input.  I don't expect the "1w" at 1 meter specs to be as high as a dual 18 or tapped horn, even if maximum outputs would be comparable to the dual 18".

11. This new class of super-woofers stretches what we can do with vented boxes.  My modified port geometry seeks to address some of the limitations that now come into play with drivers that have so much Vd.  The goal was to reduce the limitations on box peformance that result as a consequence of turbulent effects in the port while still producing a design with acceptable ease of construction.

-Phil

[Phillip_Graham]

Loren Jones wrote on Mon, 01 November 2010 15:29

The model of this design predicted 128db max continuous output at 49hz.  The driver is rated for 1500 watts continuous based on thermal capability.  So it will take some power to exploit it fully, but I don't think it is any less efficient than other single driver reflex loaded cabinets.

The true low frequency sensitivity of this driver should compare very, very well with other single driver vented box systems with similar -3dB points.  The large cone and high BL product are helpful here.  The T/S model gives the driver sensitivity as 97dB.

Also, without knowing the port Q, the box excursion behavior is, for the moment, based on my experience with real-world port Q.  I'd have to have a measurement of the port behavior to check my estimates for port Q.

As for the long term output, a couple points of clarification:

1. 128dB continuous represents the output at V_max =200V_peak at the worst case frequency, excursion-wise, above the box Fb after considering BOTH voice coil heating and reduced magnet flux from a hot driver magnet structure.  This number should be considered a realistic, long-term value after a hard days use.  I have no interest in presenting inflated output numbers.  Peak output before driver damage would be a little more than 6dB above this point.

2. With a cold fresh driver, and the same V_max= 200V_peak, the long term output would be about 1.5dB better.  X_lim is now a touch under 6dB above the improved output value.

If I were to make a spec sheet from the model, I would state the long term SPL at 129dB at 1m, and the SPL peak as 135dB.

Finally, just because the driver's AES power rating is 1500W, that doesn't make the driver happy, or sound good, after being hammered with that much integrated input.  A realistic, long term continuous input power level would be 300-400W in amplifiers that can provide such power limiting (e.g. iTech, powersoft, lab.gruppen PLM).  It should be mentioned that 400W continuous power input is quite high for the vast majority of music signals with a peak level of 200V_peak into to 6 ohms.


-Phil

[Loren Jones]

Quote:

0. I, in NO WAY, put Loren up to this. I didn't even see the thread until today, apparently a day old. I am glad Loren is enjoying his new baby.

This is absolutely true.  Phillip didn't ask me to post the build and certainly didn't ask me to mention or promote him as the designer.  He very kindly provided me with a design that clearly took a fair bit of thought and effort on his part.  He did ask me not to share the details publicly which is very reasonable given the effort and knowledge he put into coming up with it.  

Quote:

3. Loren took a risk on building this design. I say took a risk because this came straight from a model. Even with the more advanced lumped parameter model, the port tuning is rarely absolutely correct. There was no guarantee the tuning frequency would be correct for this tapered port geometry. Loren went into the project with eyes wide open about the potential of the box Fb to be wrong.

It was a smaller risk than I would have taken if I had just tried to make up a design myself using inadequate knowledge and poor modelling software.  While the Fb was a few hz lower than designed I think it speaks pretty well for Phillip's model given that it isn't just a straight rectangular or round port of a certain length.  I frankly am happy to have a little extra extension in exchange for possibly exceeding xmax a little sooner in the 45 to 50 hz range (this is where the box is apparently most excursion limited above Fb).  In our application we will never be pushing these as hard as they can go so the extra extension is a good tradeoff.

Quote:

At the end of the day, people should be realistic that this is still a vented box. While I feel confident that this box exhibits real extension to 35Hz, and output on par with a high-quality dual 18", those expecting to move mountains and cause earthquakes have read too many manufacturers spec sheets.

I currently have a borrowed SRX-728 and the aforementioned CE4000 amps both loaned to me by another exceedingly kind LABster, Mark Phillips.  Hopefully we will be able to get at least some meaningful comparisons to the 728 taken under the same conditions.  The tests won't be lab quality but hopefully it will give some idea of how this design compares to what is generally regarded as the default standard mid-level dual 18" vented sub.  Hopefully we can get those comparisons done this weekend perhaps.

Quote:

Should all those measurements ever get taken (by me or a vetted 3rd party), then I would figure out what to do with the design. More than likely I would sell it, not so much to generate revenue, but to discourage "non-serious" builders from approaching the project. I am not interested in becoming "tech support" for a vented box design

I will do all I can to get the box either up to Evan or down to Phillip for proper measurements.  I want to do this out of curiosity to see what the real performance is of "my baby".  Also I think that if good measurements help refine the design so that a really professionally designed and documented reflex loaded sub can be made available for the DIY types, then that is a good thing.

Thanks again for Phillip in kindly providing me with this design and lots of handholding along the way.  Thanks also to Jeff for the good deal on the drivers and to Mark Phillips for loaning me some gear to compare them to.

[Phillip_Graham]

Loren Jones wrote on Mon, 01 November 2010 19:52

I will do all I can to get the box either up to Evan or down to Phillip for proper measurements.

While Evan's offer is gracious and useful, I don't know that he knows how to perform all the measurements I would like to see made.  For instance, I don't know if he has the facilities and experience to do impedance testing at different input powers.

That said, Evan is certainly fully competent on general SMAART measurements, and on comparing performance with the 728.  Certainly he could measure the boxes, compare relative sensitivities, etc.

I look forward to seeing how that comparison turns out.

[Phillip_Graham]

Loren Jones wrote on Mon, 01 November 2010 19:52

I currently have a borrowed SRX-728 and the aforementioned CE4000 amps both loaned to me by another exceedingly kind LABster, Mark Phillips.  Hopefully we will be able to get at least some meaningful comparisons to the 728 taken under the same conditions.  The tests won't be lab quality but hopefully it will give some idea of how this design compares to what is generally regarded as the default standard mid-level dual 18" vented sub.

I found the JBL 2268H on the Altec Heritage website:
   http://www.audioheritage.org/vbulletin/showthread.php?11346- 2268H

On pp. 5 of the datasheet, JBL shows the Klippel Bl(x) curves.  The industry standard for defining Xmax for low frequency drivers from Klippel measurements is to the excursion where Bl is 70% of Bl_max.  By this standard definition, the JBL 2268H has a one-way excursion of 9mm.

Plugging the JBL driver into a box model,using the box volume calculated from the JBL 728 datasheet, and assuming the JBL engineers would pick an intelligent alignment, I made a quick comparison of the single 21 vs the dual 18".

It appears that the JBL dual 18 will have similar output at 100Hz, +2dB output at 50Hz, and -2dB at 35Hz relative to the single 21" when both systems reach their respective Xmax.  This is very similar performance--certainly room response error and driver batch variability would have more measurable influence.  The 21 requires additional input voltage to reach the same output level, obviously.

Since we often push loudpeakers beyond Xmax, it will be interesting to see which of these designs holds their subjective composure better under additional stress.  My gut is that the dual 18 will fare a little better here.

The single is 21 approx. 70% the size of an SRX 728, so if there is any take home story message here, it is that 21" based designs are probably here to stay.

[Phillip_Graham]

Doh, hit the wrong button, double posted.

[Tim McCulloch]

Phillip Graham wrote on Wed, 03 November 2010 09:43

Loren Jones wrote on Mon, 01 November 2010 19:52 I currently have a borrowed SRX-728 and the aforementioned CE4000 amps both loaned to me by another exceedingly kind LABster, Mark Phillips.  Hopefully we will be able to get at least some meaningful comparisons to the 728 taken under the same conditions.  The tests won't be lab quality but hopefully it will give some idea of how this design compares to what is generally regarded as the default standard mid-level dual 18" vented sub. I found the JBL 2268H on the Altec Heritage website:     http://www.audioheritage.org/vbulletin/showthread.php?11346- 2268H On pp. 5 of the datasheet, JBL shows the Klippel Bl(x) curves.  The industry standard for defining Xmax for low frequency drivers from Klippel measurements is to the excursion where Bl is 70% of Bl_max.  By this standard definition, the JBL 2268H has a one-way excursion of 9mm. Plugging the JBL driver into a box model,using the box volume calculated from the JBL 728 datasheet, and assuming the JBL engineers would pick an intelligent alignment, I made a quick comparison of the single 21 vs the dual 18". It appears that the JBL dual 18 will have similar output at 100Hz, +2dB output at 50Hz, and -2dB at 35Hz relative to the single 21" when both systems reach their respective Xmax.  This is very similar performance--certainly room response error and driver batch variability would have more measurable influence.  The 21 requires additional input voltage to reach the same output level, obviously. Since we often push loudpeakers beyond Xmax, it will be interesting to see which of these designs holds their subjective composure better under additional stress.  My gut is that the dual 18 will fare a little better here. The single is 21 approx. 70% the size of an SRX 728, so if there is any take home story message here, it is that 21" based designs are probably here to stay.

Hi Phill-

You nicely illustrate the trade-offs.  It takes more voltage swing for the 21", but the box is smaller and probably lighter.  Amp output is cheaper than it's ever been (the original Peavey CS800 was the same 'cost per Watt' as an ITech 8000).  An amp has much less cubic volume that the 30% sub space.

Thanks for the additional info.

Have fun, good luck.

Tim Mc

[Josh Ricci]

Nice looking cabs Loren. The build quality especially. I think the rcf 21 is one of the better bargains out there.

Philip a lot of what you posted above regarding the performance and limitations of a vented design are spot on with my way of thinking. Good insight there. I completed a pair of vented cabs housing b&c 21sw152's earlier this spring. Unfortunately my tune came in much lower than I had anticipated. I had aimed for 30hz (yes I know. The application is a little different from normal) and hit 25hz instead due to loading of the port against the rear wall and proximity of some internal bracing. I underestimated the amount of shift this would cause. I thought it would be marginal at best. Learned a lesson there. They are still very nice theater or music system subs. Anyway I do have a rather large data set for them taken outdoors that should be viewable in a week or two fwiw. Unfortunately I've never seen any comparable data from a respectable double 18 to compare it to. Next spring I'll be building diy horn cabs for the 21's to test out also.

[Phillip_Graham]

Josh Ricci wrote on Thu, 04 November 2010 10:10

Unfortunately my tune came in much lower than I had anticipated. I had aimed for 30hz (yes I know. The application is a little different from normal) and hit 25hz instead due to loading of the port against the rear wall and proximity of some internal bracing.

Did you calculate your port lengths using the "classical" port formulae, or by some other method?  I use models of ducted flow of the style one would see in a program like Akabak.

Even still the values don't come out on the money, mostly because of the end effects you allude to.

Cabinet impedance testing at different values of large signal V_in is the most straightforward way, in my mind, to see the port compression effects, and well as shifts in port tuning frequency due to air behavior in the port with increasing Reynolds number.

I'd suggest switching to the B&C 21SW115, it has substantially more Xmax than the SW152.