Bossobass Mini GTG Thread

[MemX]

The hidden part (back plate, mag stack, top plate and basket) is superbly crafted as are the parts you do see (surround, cone and dust cap), and that goes for all of the SI drivers we've tested thus far.

 

Some of us can see the hidden bits all the time lol

[SME]

Some of us can see the hidden bits all the time lol

I wish there was a good way to see those bits without having to mount the driver inside-out.  I briefly considered trying to build an acrylic enclosure, but it looks like acrylic just isn't a good material to use as a sub enclosure.  I wonder if acrylic or glass inlays could be made to work and would reveal enough of the insides?

[maxmercy]

At least 3/4" thick plexi, small windows, well gasketed/sealed, lots of washered/flanged fasteners to avoid cracks.  Oh, and it scratches easily.  Glass would be more scratch resistant, heavier, need to be annealed to avoid stressing it too much, and drilling into glass (trepanning) is an adventure.

 

JSS

[nyt]

I'd have a glazier do that before attempting it

[maxmercy]

Grit, brass, water, patience.  I'm sure there are bits now that can do it faster, but it can go wrong.....and quickly.

 

Plexi/acrylic/polycarb is easier, but scratches easier too.  

 

 

JSS 

[SME]

Thanks for the ideas.  I talked with my wife who actually has experience with glass working.  She has a glass saw and the ability to cut small amounts into interesting shapes like discs.  She suggests using fish tank glass.  She also doesn't like the idea of bolting on the inlays.  Do you think an epoxy would be up to the task of bonding the glass to the wood?

 

I see a suggestion to use glass 3/4" thick.  That'll be tough to make work because I planned on the box itself being 3/4" thick.  I bet 3/4" thick glass is a lot more stiff than the same thickness of plywood.  Although, plywood has better damping.  I could glue reinforcement strips of plywood under the inlays, but that would both reduce volume and add weight.  Even without the inlays, these things are going to massive with 2 HST-18s each.

 

The wife suggested installing RGB LEDs inside to illuminate the drivers.  I love it.

[nyt]

You could build an inset into the enclosure for 3/4" glass.  I wouldn't use fish tank glass, enclosures are subject to lots of pressure oscillations, you don't want to put that on thinner glass.

[maxmercy]

If the speakers will not be moved often, and the windows will not see much of the ground/floor, just use acrylic, it is much more easily obtained and machined, and can be polished if scratched.  Leave the plastic on until you install to avoid scratching it while machining.  I would not trust an epoxy, and it will yellow over time anyway.  The LEDs are a good idea, but could be a little Auto-Zone/Car-Fi-ish unless done well.

 

I would not use fish tank glass, and if you don't use bolts, then use wood flanges above and below, think mortise/tenon support all the way around.  You can tenon the acrylic all the way around and 'mortise' the wood.  Use a plastic cutting router bit, and that way the epoxy/silicone will not be visible.  Install the plexi after finishing the box from the inside.

 

JSS

[Bossobass Dave]

Hollywood movies, wherein they use sugar-pressed-to-look-like-glass so that actors and fake gunshots easily shatter the glass substitute with no harm done to standers-by or stunt men being thrown through it, has given most people a false sense of the strength (weakness) of glass in general, IMO.

 

I use 1/2" glass for my glass top plate option. You can easily dance on the glass top and never approach breaking it.

 

Check out this YT showing the super-human strength of a sheet of 1/2" tempered glass:

 

https://www.youtube.com/watch?v=Q9oBZpwp6Ok

 

3/4" glass for inserts is waaaaay overkill. You can use PL S30 polyurethane roof sealant to secure the glass inserts with no additional fastening necessary. Paul turned me onto the stuff a while back. It takes a week to cure totally and when it does, the glass will not ever be easily removed and will certainly be sealed air-tight.

 

I'm looking forward to your build. LEDs inside to show off a great looking driver is a great idea!

[maxmercy]

Ensuring the proper anneal and temper is the problem unless buying quality stuff from a known good source.  While the video is impressive, it is not indicative of what you can find for purchase in most places, and strain is difficult to evaluate unless you have large polarizing sheets and a decent light source.  Poorly annealed glass is a disaster waiting to happen, with high internal strains waiting for just the right amount of force to let it break; I have seen poor 1/4" glass break into many pieces with a due to a quarter hitting it.  Properly annealed and tempered glass is tough stuff.  Personally, I would use acrylic/plexi.  Much easier to machine for little lost in other areas, and you can blacken the edges to make them disappear easily.  If it gets scratched, you can re-polish it without too much fuss, as there are many foggy-headlight-lens polish kits out there.

 

JSS

[SME]

Thanks for all the great ideas!  Sadly, it's going to be a while before I do this.  I'm still unemployed, despite having many leads.  But instead of working on fun projects, I've been getting a property ready to sell.  Which is very good because we're going to make a stupid amount of money off of it, some of which will likely help finance the eventually planned house addition with dedicated room. 

 

The suggestion of making boxes before buying woofers is a nice one, but I am still inclined to wait until I have the actual woofers.  While I doubt Nick would suddenly leave the industry or remove the HST-18 woofers from the market any time soon, it is still his option to do so.  I feel it's prudent to assure parts availability before investing much time.  I'm actually prioritizing building a more powerful PC-based DSP solution because I stand to learn something in the process that might help me in my job search.  It also opens up the path to upgraded mains speakers.  I believe I have a lot more to gain per dollar by pursuing these upgrades first.

 

Now I just need to get hired for a job with a start date some time next year, and I'll actually be able to get some cool stuff built!

[MemX]

Best of luck for the job search   Gaining employment can be a soul-destroying process!

[Electrodynamic]

Thanks for all the great ideas!  Sadly, it's going to be a while before I do this.  I'm still unemployed, despite having many leads.  But instead of working on fun projects, I've been getting a property ready to sell.  Which is very good because we're going to make a stupid amount of money off of it, some of which will likely help finance the eventually planned house addition with dedicated room. 

 

The suggestion of making boxes before buying woofers is a nice one, but I am still inclined to wait until I have the actual woofers.  While I doubt Nick would suddenly leave the industry or remove the HST-18 woofers from the market any time soon, it is still his option to do so.  I feel it's prudent to assure parts availability before investing much time.  I'm actually prioritizing building a more powerful PC-based DSP solution because I stand to learn something in the process that might help me in my job search.  It also opens up the path to upgraded mains speakers.  I believe I have a lot more to gain per dollar by pursuing these upgrades first.

 

Now I just need to get hired for a job with a start date some time next year, and I'll actually be able to get some cool stuff built!

 

No need to worry about the part in bold. I've invested too much in DHL Express'ing parts to keep up with the recent surge in HST-18 purchases to suddenly flee the market. I'll be making the HST series of drivers for quite some time.

[Bossobass Dave]

For practical output comparisons of systems I prefer to use the maximum long term output sweep myself. These are much more representative of what can be actually realized with an appropriate "safe" amplifier capability.

 

About the output comparison of the 24 and LMS. Yep...Can't buy LMS's anymore. Kinda moot at this point. Yes you get 3dB in sensitivity for each doubling of drivers. You are also halving the power to each driver and dropping the excursion requirements. If you appropriately power each driver the same as the first you get another 3dB for the full 6dB. The LMS is right about 89-90dB sensitivity in a 4 cube box over the 40-100Hz area. The 24 is about 92-95dB over the same. If you quadruple up the LMS you reach about 95-96dB over that range which is higher than the 24. The 24 still wins out decisively below 30hz. The CEA-2010 burst numbers for both drivers below 40Hz were nowhere near using the full K10. If we look at the maximum long term sweeps for both the SI was only requiring 100v and the LMS 120v. Considering both as a simple 4ohm minimum load we get values of 2500w maximum for the 24 and 3600 for the LMS. You don't need 4 K10's to power the 4 LMS drivers to realize their full displacement potential. You only need one K10, A14K, Sanway, SP12000, etc...I don't find that to be an unrealistic scenario. The 24 is limited by the excursion below 16Hz during the sweep with 100 volts going to it in that box. It could probably use a bit smaller box to limit excursion while the LMS in 4ft is managed pretty well. Personally I'd probably use something around 10ft or maybe even 8ft for the 24 and bring a bigger amp. When the 24 is driven to it's limit and the LMS is driven to it's limit the 24 has about a 9dB lead in deep bass output. So it takes about 3 LMS's to keep up with it. If that somehow indicates the 24 is anything other than a displacement monster I don't know what to think.

 

(Bolded part by me)

 

I intend to show the results of part of the tests Paul and I have conducted that show a result contrary to what Josh posted here and that others have agreed with in comparing results of the standard CEA tone burst and sine sweep tests. The missing piece is that max excursion plus multi-octave spreads of additional simultaneous playback requirement, without additional excursion extremes, requires additional amplification for an accurate result at the seats.

 

I've mentioned, ever since Josh began testing with a K-10 connected to mains of a dedicated 220V-50A line, that the results, especially the max short term burst and max short term burst averages, should not be expected by DIYers who connect similar systems to a Behringer iNuke connected to shared 120V-20A mains, or similar popular choices +/- a couple of dB.

 

Taking Josh's example of the LMS-Ultra-18 only requiring 120V @ 10 Hz to achieve maximum excursion, working out to 3600W, and reasoning that a K-10 amplifier is not needed, we take a soundtrack effect and input that into a system and find that it will require a good bit more than 3600W because the input signal is a full spread IN ADDITION TO the requirement at 10 Hz.

 

The GTG vids of the pair of HST-15s is a perfect example of what really should have been the obvious over the years in this regard.

 

The input signal in this case was the cop car bashing HULK unleashes on Abomination in chapter 17 of The Hulk.

 

 

So guys, remember the cop car scene from TIH we cranked on Nick's 2 HST's?  I just went back and replayed it with everything set exactly the way it was when we shot the video for it and took some measurements...

 

Turns out we were pushing 8912.32 peak Watts (give or take some equipment measurement error)  to the 2 HST's in that video.   The drivers were moving 42mm one way.  Hahahahahaha! !

 

Disclaimer: We do not recommend that you feed that much power to any of Stereo Integrity's products.  It was done only in the name of GTG excess and fun. 

 

 

That's about 4500PW for each HST-15.

 

Here's a SL cap of that last punch, followed by a CEA 2010 shaped tone burst at 10 Hz.

 

Here it is again with the level of the soundtrack effect unchanged from the first SL cap but with the CEA burst increased by more than 10dB to show the difference in requirement to reproduce the shaped tone burst vs the soundtrack effect:

 

The point is that, although it may take 'X' watts to drive the HST-15 to X-max, or maximum excursion, at <10 Hz, it will always take 'X' +'Y' watts to reach that point at <10 Hz when adding the wide spread of additional frequencies all the way up to crossover, or as much as 5 additional octaves of content that typically comprises a world class ULF effect.

 

We also conducted tests of 4 drivers (8 voice coils) connected to an amplifier in 3 different wiring schemes to show that all driver receive equal voltage and therefore react with equal excursion regardless of the wiring scheme. This is because the subject has come up before and recently Mojave has posed the problem at AVS regarding noticeably different excursion in his mains arrays.

 

I will also post the SHOOTOUT results in which we loaded a Raptor enclosure with A, a pair of BHT-15s, B, a pair of LMS-R-15s and C, a pair of SI-HST-15s and conducted a series of in-room tests for comparison. Shootouts are always fun stuff and this one pits the lowly BHT-15 against industry heavy hitters. Stay tuned...

[BeastAudio]

Looking forward to that for sure

[Infrasonic]

Good post, Dave. Look forward to more data you want to bring to the table. Always interesting.

[tuxedocivic]

Taking Josh's example of the LMS-Ultra-18 only requiring 120V @ 10 Hz to achieve maximum excursion, working out to 3600W, and reasoning that a K-10 amplifier is not needed, we take a soundtrack effect and input that into a system and find that it will require a good bit more than 3600W because the input signal is a full spread IN ADDITION TO the requirement at 10 Hz.

 

So, if we run a 10hz sine wave and measure 120V, then run pink noise... we will measure what? Not 120V?

[3ll3d00d]

So, if we run a 10hz sine wave and measure 120V, then run pink noise... we will measure what? Not 120V?

The literal answer is a waveform with a different (lower) crest factor (which depends on the type of pink noise used as they are not all created equally) so a true rms meter would then measure a lower value. I am not sure this is what you are asking though. My interpretation of the quoted post was that it implies real content would have a higher crest factor, at least for short periods, hence the sustained power demanded goes up.

[Bossobass Dave]

The short answer is, yes... not 120V.

 

First, there's no 10 Hz content in pink noise, unless you create it yourself. The point of Josh's post is power requirement to push whatever driver to "full displacement".

 

Drive the sub to maximum excursion using a 10 Hz sine wave. Drive the same sub to maximum excursion using pink noise from 10 Hz to 80 Hz and yes, of course, you will measure higher voltage from the amplifier.

 

It's really just common sense that it will require more power to drive a sub to full displacement plus a full spread of content above that point than it would to drive the sub to full displacement with only a single frequency sine wave.

[3ll3d00d]

 

It's really just common sense that it will require more power to drive a sub to full displacement plus a full spread of content above that point than it would to drive the sub to full displacement with only a single frequency sine wave.

I'm not sure common sense applies here tbh.

 

My thinking was

 

moar frequencies at "full scale" == increasingly square wave like behaviour == reduced crest factor == more sustained power demand 

 

i.e, the peak power required doesn't change but the duration you need that power for goes up 

 

is this what you mean by common sense? if not, what do you mean exactly?

[tuxedocivic]

It does sound like common sense, but as mentioned, I don't think it is that straight forward. It's a bit counter-intuitive. Now, I'm the first to admit this isn't my strong suit, but my understanding is that if it takes 120V to drive the subwoofer to xmax at a VLF, then that is the max power you need.

 

You yourself Bosso have said that we only care about VLF/ULF output when comparing systems as the MF and LF are easily accomplished when you have enough discplacement to achieve ULF. This has come up when people tout the advantages of sensitive pro subs.

 

So then the question is about whether more power is required as you add bandwidth to the signal. We've already found out that it takes 120V at 10hz. Now if it takes more power in the higher frequencies to achieve the required bandwidth, I believe you've changed your position about the usefullness of sensitivity in the upper bands. It's inferred. (of course, high sensitive subs usually come with a xmax limitation, so there's another problem thrown in the mix).

 

Furthermore, I don't think the reasoning that there's more peak power demand due to increased bandwidth is correct. Whether it's a single frequency sine wave or a complex sound effect full of hundreds of sine waves, the voltage requirement stays the same. The current demands are also not increased. Counter-intuitive; yes. Could I be wrong; yes. I would really have to dig out some books to formulate the proper response if I were to hang my hat on this.

 

Finally, there's some problems with your example of the HST15 requiring 4500watts to reproduce the Hulk scene. The first problem is that you're comparing apples to oranges. The LMS 18 is larger and probably does require less power to reproduce the scene. Second, the 10hz tone burst you put on the graph,,, at what level did you chose for this? It seems arbitrary. To arbitrarily increase it 10db doesn't show that you need 10x the power to reproduce the Hulk scene, unless the original level came from something??... Third and last, is the spec lab capture the digital output, or is that in room? Either way, the output requirements of your sub would be a lot less (perhaps down around the 10hz tone burst you displayed) thanks to room gain. Another phenomenom you've correctly argued in favor of previously. Perhaps I didn't catch something, but it all seems unrelated.

[tuxedocivic]

I'd also be interested in knowing how power can be calculated from a single frequency to multiple frequencies if X+Y really is the required power.

[Bossobass Dave]

It does sound like common sense, but as mentioned, I don't think it is that straight forward. It's a bit counter-intuitive. Now, I'm the first to admit this isn't my strong suit, but my understanding is that if it takes 120V to drive the subwoofer to xmax at a VLF, then that is the max power you need.

 

You yourself Bosso have said that we only care about VLF/ULF output when comparing systems as the MF and LF are easily accomplished when you have enough discplacement to achieve ULF. This has come up when people tout the advantages of sensitive pro subs.

 

So then the question is about whether more power is required as you add bandwidth to the signal. We've already found out that it takes 120V at 10hz. Now if it takes more power in the higher frequencies to achieve the required bandwidth, I believe you've changed your position about the usefullness of sensitivity in the upper bands. It's inferred. (of course, high sensitive subs usually come with a xmax limitation, so there's another problem thrown in the mix).

 

Furthermore, I don't think the reasoning that there's more peak power demand due to increased bandwidth is correct. Whether it's a single frequency sine wave or a complex sound effect full of hundreds of sine waves, the voltage requirement stays the same. The current demands are also not increased. Counter-intuitive; yes. Could I be wrong; yes. I would really have to dig out some books to formulate the proper response if I were to hang my hat on this.

 

Finally, there's some problems with your example of the HST15 requiring 4500watts to reproduce the Hulk scene. The first problem is that you're comparing apples to oranges. The LMS 18 is larger and probably does require less power to reproduce the scene. Second, the 10hz tone burst you put on the graph,,, at what level did you chose for this? It seems arbitrary. To arbitrarily increase it 10db doesn't show that you need 10x the power to reproduce the Hulk scene, unless the original level came from something??... Third and last, is the spec lab capture the digital output, or is that in room? Either way, the output requirements of your sub would be a lot less (perhaps down around the 10hz tone burst you displayed) thanks to room gain. Another phenomenom you've correctly argued in favor of previously. Perhaps I didn't catch something, but it all seems unrelated.

 

You're crossing into other ares of sub design here^^^, one of which at least you've addressed yourself, that being the displacement disadvantage of a higher sensitive driver.

 

Most subwoofers have to employ limiters to inadequately reproduce the total wide bandwidth effect because the sum of power is always higher for the frequency spread than it is for the single frequency that will overdrive the excursion limits of the driver or drivers in the system because they are displacement limited, where the headroom certainly always is (and as you've mentioned that I have said more than once in the past).

 

That's not the crux of the discussion.

 

Ilkka's short-lived multi-tone test shows that several sine waves of 90dB sum to 110dB. There's no way you get higher output from same input power. Josh did a few himself, so you can ask him how he calculated total power in dBSPL.

 

 

In the graph above of the spectral contamination test of the LMS-U-18,  just imagine for the sake of discussion that the first tone at 20 Hz drives the sub to max excursion. If so, the reading with a single tone at 20 Hz would be 92dB. Adding the additional tones doesn't drive the sub beyond max excursion, but the total is 110dB and certainly more power is required to achieve that than was required to generate the 20 Hz tone alone.

 

I'm curious to know why you believe it would not require more power to reproduce the multi-tone spread than it would to reproduce only the single tone at 20 Hz?

 

 

I'd also be interested in knowing how power can be calculated from a single frequency to multiple frequencies if X+Y really is the required power.

 

The REW SPL meter in Z weight mode can auto-calculate the total from a spread of infinite separate simultaneous frequencies in dBSPL. To measure volts and amps and watts is a simple matter of having a high quality DMM and a calculator.

[tuxedocivic]

Yay, the site is back online. Was it dead for everyone else?

 

 

 

I'm curious to know why you believe it would not require more power to reproduce the multi-tone spread than it would to reproduce only the single tone at 20 Hz?

 

Well, I'd don't actually believe either yet. I'm asking for intellectual sake. A few things struck me to ask the question. For one, if it works as you say (the more spectral content, the more power required), then it would seem to me we need gagillions of watts to even run a tweeter. Also I find your examples inadequate, as mentioned in my previous post.

 

But maybe I'm just not putting the dots together. I would like to see what you come up with. This really isn't my strong suit. I just don't understand amplifiers very well. If I'm wrong, hopefully I'll get it eventually. If I'm right, it'll take someone who knows this stuff better than me to explain it, cause I'm not gonna try and go through my text books to figure this stuff out

[3ll3d00d]

It still seems to me this is predominantly a question of duration (crest factor) rather than peak amplitude. Think about what the waveform looks like in the spectral contamination case, say they are all full scale sine waves then sum them. No doubt the peak will be larger in some parts of the waveform (hence in real content it either clips or the mixer has to attenuate) but it will certainly be a more sustained load. Just referring to volts alone is therefore insufficient info as it has no temporal info.