I've got nothing better to do right now so I have time to present some more info to explore this issue in a bit more detail. Hopefully this will greatly clarify my position. The last paragraph is a simplified summary for the tldr folks.
So how to decide if a flh needs a hpf or not?The first thing that needs to be addressed is that front loaded horns are different than resonant enclosures that do not use a sealed rear chamber in some important ways. This is extremely important, especially when comparing performance aspects of different alignments, but this issue is never mentioned and different alignments are almost never compared fairly.
Here's a picture of three different alignments all using the same driver. Each alignment has the same 300 watts input. The top two pics are frequency response and excursion of a front loaded horn, the middle two pics are a ported box and the bottom two pics are a sealed box.
I didn't include examples of transmission lines, tapped horns and other examples of chamberless resonant alignments, but those alignments are fairly accuately represented by the ported box shown above as far as the points I am going to discuss.
Tuning (as defined by the lowest frequency sharp notch in excursion) happens to be ABOVE the low frequency knee in front loaded horns but BELOW the knee in resonant enclosures which do not employ a sealed rear chamber. This is the trade off that the chamber provides - for the same LF knee in both alignments the flh excursion control is traded for a higher actual tuning. This is a very important consideration because the ported box (or tapped horn or tl) can use a hpf to knock down and control the excursion peak below tuning without affecting the frequency response curve shape or magnitude of SPL at the low knee significantly. Front loaded horns will see a reduction in SPL at the low knee of about 3 db if they use a hpf to reduce the excursion peak below tuning to the level of the next highest excursion peak, as shown in post 61 - http://www.avsforum.com/t/1484963/any-reason-not-to-do-a-tht/60#post_23659531
This means that you CANNOT accurately compare front loaded horns without a hpf to resonant alignments that do not have sealed rear chambers unless you consider the front loaded horn's highest magnitude excursion peak as it's xmax limited spl. Otherwise the front loaded horn will appear to have a huge advantage that doesn't actually exist.
To explain this further in case it's not clear, when we present our excursion limited max spl graphs the common convention is to show the SECOND highest magnitude excursion peak as the xmax limited max spl. It's generally accepted that the highest magnitude peak (the one below tuning) will be controlled with a hpf. This means that if you show a flh in this manner without intending to use a hpf in real life the highest excursion peak will NOT be filtered and will push the driver considerably past xmax at the simulated power level - usually by a factor of 2x or more. If you don't intend to use a hpf with a front loaded horn, you must limit the highest magnitude excursion peak (not the second highest) to the value of the driver's xmax in simulations and that makes front loaded horns look a lot less appealing when compared fairly to resonant enclosures that do not use a sealed rear chamber. Since ALL chamberless resonant devices should use a hpf to protect the driver below tuning it's fair to show them with the second highest magnitude excursion peak as the xmax limited max spl but it's NOT fair to show front loaded horns in the same way if they won't be using a hpf.When compared fairly, front loaded horns almost never outperform tapped horns (especially in terms of LF output vs size in acoustically small single subs or stacks) except in very specific situations.
But of course it depends on the exact goals for the system, the design and the driver used, and whether or not the flh uses a hpf. This is a whole topic in itself but the point here is that they are almost never compared fairly. This difference between chambered and chamberless alignments has enormous implications when deciding whether to choose a flh vs any other alignment in the first place, and in fairly evaluating their respective performance envelopes. This info is also extremely important in determining WHERE to put the hpf if one is used with a flh.
On to the next point of interest. In the above graphs it's easy to see that the alignments with sealed rear chambers do in fact control excursion to a large degree. But with front loaded horns there's always going to be a big excursion peak right below tuning. Excursion does level out to a lower level at frequencies below the peak but the peak itself cannot be ignored because the driver will be destroyed if it is driven to xlim forcefully at the frequency of the peak. Even if not driven to xlim the driver will flop around like a fish out of water at that frequency and cause IM distortion. This will simply not happen with ported boxes or tapped horns since the required
hpf for these latter alignments will limit the excursion peak below tuning to the same level (or at least close to the same level) as the much smaller next highest excursion peak.
Next, in order to decide if a hpf is required or not for any particular flh it's important to evaluate how bad this flh excursion peak below tuning really is. Below is a quick example to show that it can be somewhere between negligible and catastrophic at high power. This shows two very different horns using the same driver. In the first example the excursion peak below tuning is only 1.5x higher than the next highest excursion peak but in the second example it's 3x higher.
And herein lies the problem with buying an IP protected design with no objective excursion data and that has not been honestly and clearly assessed and described (wrt to performance, limitations and potentially necessary protection filters) in detail by the designer. There's no way to know how bad the peak will be unless you simulate it. Most of the time with competently designed and realistically sized bass horns the highest magnitude excursion peak is going to be around 2x (or a bit more) than the next highest peak (and my tht guess simulation is right in that range) but you really should have hard data in order to make an informed decision.
These first two points combined with my guessed tht simulation in post 61 (assuming it's close to accurate, and I'm pretty sure it is) indicate that a tht with 300 watts and no hpf has an excursion peak that is signifcantly past xmax when used with a 16 mm xmax driver
. That excursion peak will be below xlim (assuming the classic Tempest is used with 300 watts) but definitely past xmax. Power should be reduced to limit the excursion peak below tuning to 16 mm to evaluate it fairly or compare it to other alignments like ported boxes or tapped horns OR a hpf needs to be added to keep that excursion peak down. In that light the tht without a hpf is not nearly as impressive as it might seem when compared to a similarly tuned tapped horn that does use a hpf.
The next thing to look for when deciding if a hpf is appropriate for a flh is a bunch of different end user system goals and unique room considerations that I'll lump together here - the expected frequency range of media content, the acoustic environment and the desired max SPL level. It's trivially easy to use Audacity (or similar software) to find out the frequency range of media. If the media does not go as low as the frequency of the excursion peak below tuning you do not need a hpf at all. Next, the acoustic environment must be measured to find specifics about room gain, but in general a small sealed concrete bunker will have more acoustic gain (up to 12 db/oct) than an open concept house with thin single sheet drywall and vinyl siding exterior walls (which may actually have closer to 0 db than 12 db of room gain at 20 hz and below). Corner loading the sub will produce more spl vs excursion at all applicable frequencies than placing the sub against a single wall or in the middle of the room or outside with no boundary loading. And desired spl is self explanatory but goes hand in hand with the acoustic environment. So to put this in perspective, a guy with a corner loaded sub in a small sealed concrete bunker that doesn't need more than 90 db peaks isn't going to need a hpf (but there will still be the consequence of higher than necessary IM distortion even at low levels). On the other hand, a guy with a single sub for an outdoor theater with the subwoofer placed 20 meters away from an audience that expects reference level spl is probably going to benefit from a hpf, especially if he's got more than 300 watts on tap. Add to this the fact that the tht webpage advocates eq'ing below tuning and Bill says a hpf is not necessary and you have potential for disaster.
Next let's look at the driver itself. Drivers with medium xmax and medium power handling (like the classic Tempest) are hard to assess for the necessity of a hpf when used in a flh without specific objective data on the driver AND horn combination and other factors as outlined above. Drivers with low xmax like the Dayton Classic which used to be recommended simply don't have enough xmax to provide much SPL without a hpf, making the size of a horn a wasteful proposition. Drivers with even more xmax than the classic Tempest can be useful to provide more LF SPL output potential but at the same time higher xmax drivers (especially those that also have higher power handling) usually have problems with high inductance which can severely limit the useful HF corner in a way that simulations cannot accurately predict. In any case it's incredibly useful to be able to simulate the horn with the driver to be used (with measured t/s parameters, not manufacturer published specs) and also to know how the driver reacts wrt power compression and inductance related HF rolloff.
So in summary the choice to use a hpf or not with a flh is not a simple single sentence answer even when talking specifically about a single particular design. There are advantages and disadvantages. A hpf (if used with a flh) will take away significant spl at the LF knee, unlike a ported box or tapped horn since the actual tuning frequency does not correlate with the LF knee in these different alignments in the same way. This means that front loaded horns are almost never compared fairly to other alignments that do not use a sealed rear chamber because of the way xmax limited max spl is usually shown (especially here in the forums it's almost universally shown as the second highest magnitude excursion peak being the max spl limit which gives the flh a fictional advantage unless it's used with a hpf) and this is very important because it has implications for frequency response curve shape, xmax limited power handling, IM distortion, etc. For a flh, it's important to know the magnitude of the excursion peak below tuning in order to decide if a hpf is required. Additionally, the frequency range of the media, the acoustic environment and the desired spl level is very important in deciding if a hpf is required for a flh, but ideally no design (other than simple sealed boxes) should be boosted below tuning. If a hpf is used with a flh there will be less chance of hitting xlim, less excursion in general, less IM distortion, and overall higher excursion limited spl if the driver can thermally handle it. But probably most important in this discussion is the fact that if my sim is anywhere near accurate the tht with a 16 mm xmax driver and 300 watts applied with no hpf is considerably past xmax at 22 hz by a factor of about 3 db. In fact it can only handle 170 watts within xmax. The situation is exponentially worse with an 8 mm xmax driver as per the plans circa 2007.
I didn't bother inputting the Dayton Classic specs into my tht guess simulation but using the Tempest specs and adjusting excursion to 8 mm xmax in my sim, it took only 45 watts for the excursion peak below tuning to reach and exceed 8 mm as shown in the pic below (12.55 volts into 3.5 ohms is 45 watts). In fact there's a good chance that drivers with 8 mm xmax may even be destroyed with 300 watts in the tht if used without a hpf regardless of the wiggle room that may be provided by decreasing Bl and increasing compliance considerations that are seen when pushing the driver to the ragged edge of mechanical destruction. And while I realize the current plans may say the Dayton Classic is not acceptable it was a recommended driver for this same design just a few years ago. I would love to be able to present a 100 percent accurate simulation of the tht but I can't and regardless, I am confident that my simple guess simulation is way closer than it needs to be to demonstrate the broad strokes of the issue. To wrap this up, while using a hpf or not will always come down to an end user personal choice, that choice should be informed by objective info, and once that info is processed it will be found that the correct answer is almost never to just run a flh without protection unless the system is WAY more capable than it needs to be. So is it me that doesn't understand this issue or is there something else going on here?