Unibox sealed box excursion question

[Greybeard191]

I'm playing around with a Tumult 15D2 in Unibox and I've run into what seems to be a conflict.

If you build a 2 cu ft box and set input power to 1000 watts, Unibox says on the sealed box that that excursion of 33mm is exceeded at about 30 hz or slightly under (on one of the graphs). Yet on the actual design tab (where you enter the numbers), it plainly shows that progressively making the box smaller drives the max power up... this is what I would expect to be correct. And 2 cu ft means on the design tab that the box has a max power of about 1300 watts.

So is the excursion graph to be ignored? And excursion limits based on whether your box size allows more max power than your amp can put out (making box power-limited rather than excursion limited)?

GB

[cjd]

Excursion graph is simply a map of excursion over frequency at a given input power level. It is accurate.

C

[ThomasW]

Drivers in small boxes are output limited by a combination of the excursion and the thermal heating of the VC.

If I put a 15D2, 1000 watts into a 2 cu ft box I get the following.

[bossobass]

GB,

Though I haven't yet loaded the new Tumults, the old model is similar and actually requires a slightly smaller box for equivalent system Q compared to the new one.

You will be fine as far as excursion headroom at any frequency using net 2 cubes (actually, up to 2.5 cubes) and 1KW.

I have several iterations of sealed and sealed L/T Tumults at around net 2.25 and 2.5 cubes. I run 1200-1500 watts into each Tumult and have never exceeded Xmax, even when playing BHD and WOTW at levels that drive the amps into clip with 12 dB of L/T boost.

With music program (down to 30 Hz) you're perfectly safe.

http://jeff-n-jenn.com/Dump/123-2326_MVI.AVI

Bosso

[Mark Seaton]

One factor not often considered is what exactly is going on at the "linear limits" of a driver. At this point we usually see either or both the BL dropping to ~70% and the suspension greatly stiffening. At the lowest frequencies you can expect to see less excursion when approaching rated Xmax in the real world due to this effect. The loss in BL can sometimes cause a slight rise in Xmax from predicted at the in-box Fb (simultaneously less excursion below Fb), which will also shift slightly higher.

Finally, also keep in mind how the power from your amplifier will get used in the real world. As soon as the amplifier has to deliver wide band spectral content (more than just one frequency) it can no longer put out full power at any one frequency.

More food for thought. Of course the easy way out is to just over-design the system so you stay away from these limits.

[Jack Gilvey]

Of course the easy way out is to just over-design the system so you stay away from these limits.

That major advantage the DIY-er has, in my estimation.

[Greybeard191]

OK... I've sorted out Unibox and figured out why my graphs were whacked out.

Here's the next question:

The box volume is the biggest thing controlling excursion. Box volume, driver, and the amount of stuffing varies Qtc of the box in Unibox.

I've also got this FRD consortium LT spreadsheet which is a handy tool. The problem is trying to transfer data from Unibox to the LT spreadsheet. The LT spreadsheet seems to treat the box volume as if it was unstuffed... for example you enter a given volume and it gives you the unmodifed (by the LT) Qtc of the box/system. This reasonably closely matches what Unibox will give you for either an unstuffed or lined-wall box with same driver/same volume (the final number is probably mid-way between the unstuffed/lined box numbers for Qtc)

Anyway, If hypothetically I wanted to LT a heavily stuffed 60 liter box with a Qtc of .677 (as modelled in Unibox), do I enter a sufficent box volume in the LT spreadsheet to get an unmodified Qtc of .677? This would be a volume closer to 85-90 liters. I guess ultimately my question is does the LT spreadsheet , or even the LT circuit simply modify Qtc and Fb of a system (which might be dervived from a variety of box volumes and stuffing levels), and volume plays no part in the equations.

Hope this is understandable.

GB

[bossobass]

It would be incorrect to say that volume plays no part in the L/T's transormation of a sealed box.

You can use the L/T to change Fb and system Q, but the 2 are tied to each other by a formula. You can only alter system Q (Qtc) proportionate to the amount of boost (change in Fb).

IOW, if there was no boost applied, there could be no altering of Qtc. When a change in Qtc is attempted that is beyond the capability of the L/T due to too little boost, it simply affects the entire FR instead of Qtc.

Model the L/T with the apparent box size due to stuffing, but use the excursion graph of the model that has the actual size of the box plugged in. IOW, stuffing affects system Q but not excursion resistance. Don't attempt a radical change in system Q if you only intend a small L/T boost of, say, 3-6 dB.

Hope this helps.

Bosso

[Greybeard191]

Model the L/T with the apparent box size due to stuffing, but use the excursion graph of the model that has the actual size of the box plugged in. IOW, stuffing affects system Q but not excursion resistance. Don't attempt a radical change in system Q if you only intend a small L/T boost of, say, 3-6 dB.

Bosso

To restate this the way I think I understand it, and you can yea or nay me?

1) The volume that gets plugged into the LT spreadsheet is whatever volume is required to achieve the system Q prior to application of LT... this volume may be larger than the volume from unibox if unibox is modelling the box as heavily stuffed

2) The excursion graph to be used is the one from unibox. A smaller box can sufficently limit excursion for a given power. e.g. it is possible to build a box that is power-limited, not excursion limited.

IOW, box size and stuffing do effect system Q, but box size overwhelmingly effects excursion (stuffing actually does have minor effects to excursion, at least on the graph). The LT spreadsheet deals with system Q, Fb, and modifications to them, as allowed by the amount of boost/reduction in system Q.

GB

[Mark Seaton]

Hi GB,

While it should be fairly obvious, I think we should remind everyone that when looking at any sealed system, SPL is excursion. In other words, you can't change the excursion you need to get a specific SPL. The box size and stuffing only change how much power you need to put in at a given frequency to get some SPL level. This is sensitivity vs. frequency.

Whenever I used to use the LT spreadsheets, I would generally use the separate tools for what they do best. IOW, set the box size while working with the power I would be using. Now use the LT spreadsheet to see what components & resulting EQ you need to get the desired response. All you care about in the LT spreadsheet is for the un-EQ'd driver response to match that of your model. As you play with the resulting response, you can see how much gain is required for various responses. Since your driver/box model defined the maximum power you have on hand and/or the maximum excursion of the driver, the gain in the LT curve ends up being subtracted from the high frequency response of your driver/box model.

All this tells you is the maximum level you could sweep a sine wave through the system and not clip the amp or overdrive the sub. That's not real program material. Remember that if the soundtrack or music has a huge peak in the upper bass range, enough signal can still drive your amp to maximum output, even if the LT is cutting that range by 6-12dB.

To say it again from a little different perspective, the raw, driver/box/amp model let's you know what your amp can potentially produce from your driver in its box. From there you have to decide how you can modify/EQ the input signal to give you the response you need, and the output you want without clipping the amp or bottoming the driver. If you can't seem to get what you want in loudness while maintaining the low frequency output, you need more subwoofer. Before you go buying 8 15" drivers to hit your target at 10Hz, build or measure one sealed subwoofer, and measure what it delivers to your seats. The toughest part is getting some feel for how much is "enough" at a given frequency.

[Dennis H]

The problem is trying to transfer data from Unibox to the LT spreadsheet.

You might find SL's spreadsheet easier. After all, he's the L in LT. All it asks are the F and Q from unibox along with your desired F and Q.

http://www.linkwitzlab.com/filters.htm#9

http://www.linkwitzlab.com/pz-eql.xls

[bossobass]

To restate this the way I think I understand it, and you can yea or nay me?

1) The volume that gets plugged into the LT spreadsheet is whatever volume is required to achieve the system Q prior to application of LT... this volume may be larger than the volume from unibox if unibox is modelling the box as heavily stuffed

2) The excursion graph to be used is the one from unibox. A smaller box can sufficently limit excursion for a given power. e.g. it is possible to build a box that is power-limited, not excursion limited.

IOW, box size and stuffing do effect system Q, but box size overwhelmingly effects excursion (stuffing actually does have minor effects to excursion, at least on the graph). The LT spreadsheet deals with system Q, Fb, and modifications to them, as allowed by the amount of boost/reduction in system Q.

GB

1. Correct
2. Correct
3. What Mark and Dennis said

As Mark said, in so many words, the tradeoff for excursion protection (without using a HP filter to limit FR) is efficiency.

You really only do need the Fc and Qtc of the actual natural sub. Measuring is better than a model, but it's not so critical if your software is good. I'm not familiar with Unibox but, in the end, you'll build the sub with a driver, a box, some stuffing and an amp so measuring it's actual end result is easy enough to do before you build the L/T circuit.

Bosso

[Dennis H]

1) The volume that gets plugged into the LT spreadsheet is whatever volume is required to achieve the system Q prior to application of LT... this volume may be larger than the volume from unibox if unibox is modelling the box as heavily stuffed

That will be close but not quite right. Stuffing changes Qt but it doesn't change Fb. Making the box bigger changes both. As Bosso said, ideally you'd measure the real Qt and Fb of your box with an impedance sweep and plug the numbers into SL's spreadsheet.

[Greybeard191]

so adding/removing stuff from a box of a given volume simply changes Qtc? And what? The impedance spike at Fb changes in amplitude, not position (WRT frequency)?

I'll have to figure out how to measure Qtc then... or calculate it somehow from what Fb actually is.

GB

[Brian Bunge]

Stuffing a box changes both the impedance spike at Fb, as well as Fb itself. If you haven't played around with Unibox, you should download it and start using it. You can answer a lot of "what if" questions with it.