ROAR18

Personally I don't like it at all, for all the reasons previously discussed in the previous thread. That huge 20 db dip right above the passband is going to cause all kinds of problems with proper crossover implementation. The terminology they use is made up, the whole thing is a 1/4 wave resonator, not just the big bit at the mouth. The frequency response is terrible and will just get worse with power compression. They claim this huge dip right outside the passband is a benefit - but it certainly is not.

Circlomannen and Martinson latched onto this idea when trying to figure out how the Danley sub works. Needless to say their ideas about a velocity wave creating an extra segment in front of Danley's horn out of thin air was ridiculous, as was the notion that Danley's frequency response looks anything like this thing, or even that the Danley sub was a tapped horn. (It's not.) Their analysis of Danley's horn was remarkably off base and their analysis of this type of creation is equally off base.

You can get the same frequency response with a much more traditional looking flare, although I have no idea why you would want to because it's terrible.

I'm sure it makes a lot of noise though.
 
ROAR12_and_linesource.jpg

My Roar12 is comparable to a THAM15 in both max spl and usable frequency response.

It has a more physical character. Where the Tham15 has a slightly better attack, the Roar12 will make stuff rattle and jangle with lots of energy in the lower part of its passband. A B&C 12PS100 seems to have roughly the same max spl in a Roar12 as a B&C 15PS100 in a Tham15.

https://youtu.be/xZ227k8Gf9U?t=25

My oldest son has a lot of fun with my Roar12 and a 10 watt cheap class D amp.
When standing close to the mouth of the Roar12 the bass pulses feels quite unpleasant in your chest (with only 10 watts).

Cheers,
Johannes
 
Last edited:
S1,2,3 / Sd (compression) for the ROAR series :

ROAR10 - 0,73
ROAR12 - 0,67
ROAR15 - 0,64
ROAR18 - 0,65

In the ROAR designs the driver is additionally loaded by a resonator segment, and it has a double sided tapping, this combined with the rather low compression should make for decent operating conditions for the driver.
 
Last edited:
I think I'd better leave that one to Johannes seeing as he has one, I have only heard the ROAR12 one time, with a brand new driver so the T/S parameters where most likely a bit away from the spec sheet, but it still made an impression, I remember thinking that it it sounded massive, not as "light" as the THAM designs.

I'd like to hear two RORAR12 (or any of the ROAR designs) with drivers that has softened up a bit, matched with a decent top system in an out door setting before attempting to give a more detailed description.
 
I strongly believe in using as low compression (S1,2,3/Sd) as possible. I have seen cones buckle and crack with compression as low as S2= 0,5XSd in tapped horns and tapped pipes.

The ROAR-series is designed to treat the driver cones as gently as possible when pushing the limits with sustained high power close to the drivers maximum ratings.

The ragged response in the Hornresp-sims is the low level cold voice-coil theoretical response. The response is quite different with real life frictional losses and from thermal compression and an increase in Qes due to a hot voice coil.

ROAR is a "heavy duty" horn meant to be used with lots of power in real life events. It is not meant to simulate nice in the very theoretical world of Hornresp.

I tried to describe the general character in my first post in this thread. It is quite physical and "heavy" but with a nice "punch" in the midbass. Quite similar to the THAM15, but with more "body".

It does not change character much with increased power levels. It just keeps getting louder. There is no audible compression or noise from turbulence.

The ROAR18 should sound quite similar to the ROAR12 but louder, even more physical and powerful with 10 Hz deeper low end response.

Cheers,
Johannes
 
I strongly believe in using as low compression (S1,2,3/Sd) as possible. I have seen cones buckle and crack with compression as low as S2= 0,5XSd in tapped horns and tapped pipes.

I seriously doubt that. If there was a cone failure with compression ratio of 2:1 then there was a serious problem somewhere. Either it was an absolutely terrible design or it was used without a high pass filter or there were air leaks or something along those lines. Or maybe a completely unsuitable driver was chosen, one with a very weak cone. Either way it's a problem with the design or build, not the compression ratio.

I've seen reports of up to 8:1 compression ratio used with no consequence. There is no reason whatsoever for cones to buckle under a 2:1 compression ratio in a competent design.

The ragged response in the Hornresp-sims is the low level cold voice-coil theoretical response. The response is quite different with real life frictional losses and from thermal compression and an increase in Qes due to a hot voice coil.

Yes, as I mentioned a long time ago, the response is quite different with heavy power compression. It's much MUCH worse as any reasonable person would expect.
Bottom row left - response at small signal level
Bottom row right - roughly estimated response with heavy power compression, Re doubled

An externally hosted image should be here but it was not working when we last tested it.


I didn't bother to estimate the effect of friction losses, but as any reasonable person would expect, it would make things worse, not better.

ROAR is a "heavy duty" horn meant to be used with lots of power in real life events. It is not meant to simulate nice in the very theoretical world of Hornresp.

Well this much is clear because the simulated response is terrible. The sim itself is not even an accurate representation of the enclosure. Look at the schematic in the pic I posted. See that flared section near the end in the sim? That flared section does not exist in the actual enclosure. So not only did you not strive for decent frequency response, you didn't bother to sim the enclosure properly either. If Hornresp can't do it use Akabak, don't throw out garbage sims that don't accurately reflect the plans.

I also take GREAT exception to your wording and the implied attack on the accuracy of Hornresp. "...the very theoretical world of Hornresp ..." ?!? Give me a break. If you accurately sim what you build the sims WILL match the measurements at small signal levels. That is beyond argument at this point, there's plenty of proof. You want to slam Hornresp for being "too theoretical" and then post a garbage sim that doesn't even match the plans? This kind of behavior will not stand, I will call you out every time.

If you want to say the measurements won't match the sims when the box is pushed hard into heavy compression just say that, don't take a passive aggressive swipe implying Hornresp is "too theoretical" because it can't do something it was never meant to do.

It does not change character much with increased power levels. It just keeps getting louder. There is no audible compression or noise from turbulence.

This is either completely untrue or you did not listen to it at the levels you indicate it was created for. EVERY speaker in the world changes character when pushed into heavy power compression. Yours is no exception as I just showed in the picture.



Everything about this is bad. Terrible frequency response, the suggestion that frequency response is somehow ok at high power (character doesn't change), the sim not matching the plans, the implied attack on Hornresp, the fact that this design is being promoted without ever having been built or measured AND with all the other problems previously noted.

You have a lot of studying to do. This was clear years ago when you came up with this concept in the Danley thread where you presented so many wildly ridiculous ideas (like an extra box segment at the mouth created out of thin air by vortexes, implying the BC horn was a tapped horn and that it had a frequency response even remotely similar to this thing). I mentioned the vast amount of problems with this design way back then, it seems you have learned anything at all since then.
 
The Power Compression tool can also be used, rather than having to manually adjust Re.

Yes but I can't use it to good effect. For example I do know that when pushed hard a driver's Re can measure 2x higher than it's original cold value. BUT I have no idea what VC temperature is equivalent to this phenomenon. Either way, predicting power compression is always going to be a rough estimation (as far as I know) so I like to keep the variable in the sim as realistic as possible - I do know Re can double but I'm not as confident specifying that in a temperature related way.
 
Just for fun, let's look at what good design looks like as far as power compression is concerned. This is a tapped horn I built back in 2009. I designed it but it's basically just a scaled down TH_SPUD with a single much smaller driver. So in effect you could say this is how Danley designs, and if you sim his designs you will see this behavior over ad over.

Some things to note are that under heavy power compression, ideally the low knee will shift just a tiny bit lower. This doesn't happen in your design. And ideally under heavy power compression the frequency response should actually either get better or at least stay moderately viable. Again, this is not what happens in your design.

The only way for the frequency response to stay viable with heavy power compression is to start with a box that seems a bit too large in the first place so the initial low power response looks a bit undamped like so -

Bottom row left - response at low power level
Bottom row right - estimated response under heavy power compression, Re doubled

Note how the frequency response actually gets FLATTER as power is increased and losses incur. What would Danley do? THIS is what Danley would do. Note that this sub is no exception, it also changes character under heavy compression, but at least the frequency response doesn't fall apart into a hot mess.

An externally hosted image should be here but it was not working when we last tested it.
 
@Mrscy

I'm responsible for the published design documentation of the ROAR series, let me know what measurements are you missing or need to refine and I'll se if I can help you, the ROAR series was designed using CATIA v.5. and I'm not sure about 3d model file format compatibility with the Autodesk suite.

CATIA offers good tools for geometry analysis, such as areas, path lengths and volumes, it is also most useful for analyzing driver mounting/demounting envelopes and tool space, let me know if you have any questions.
 
@Mrscy

Just build it as per the plans. The measured respons is much flatter the the theoretical simulated hornresp respons indicates. Hornresp does not account for frictional losses in the tapped pipe sections.

Cheers
Johannes

Frictional losses in the path have nothing to do with anything. It's been proven over and over that if you accurately sim what you built the sims will match the measurements at small signal levels.

You don't know how this particular sub is going to measure because no one has built it and you didn't bother to simulate it properly. All you can say for sure is that the frequency response is going to be terrible, nothing is going to get rid of that terrible massive dip right outside the passband, which will ensure there's no way to properly cross it over to the mains.

HiFiForum.nu - ROAR-12 på G

There is some measurements a few posts down in the link above. The dark blue line with the word "Mynning" is a near field measurement of my roar12 with a new B&C 12PS100.

Where's the sim to go along with that? Where's the measurement details? Measured inside or outside? At what distance? How much smoothing in the measurement?

You can't say this new design is going to measure flatter than Hornresp predicts based on a measurement of a completely different design with no sim comparison and no measurement details.

This whole thread is ridiculous.
 
M_tningar-_ROAR12.jpg

The blue line is with the microphone in the mouth of the ROAR12.
The red line is with the microphone at the exit of the tapped pipe section at the bottom of the front-resonator.

We did not think that a very low compression was enough to prolong the working life of the driver. We tried to use an as symmetric loading of the driver as possible in order to avoid warping the cone and the suspensions of the driver at high excursions. The ROAR-series has two quarter-wave resonators tuned to two different frequencies loading the two sides of the cone in an as symmetric way as possible. In a "normal" tapped horn like the THAM-series the driver couples into an expanding unsymmetrical horn throat. This will warp the cone and the suspension when the driver is being pushed hard. With the magnet protruding into the mouth of the horn you get turbulence and a discontinuity of the horn-expansion curve.
Even though driver-failures seems very rare in the THAM-series, we wanted to maximize the working life of the drivers and minimize the risk of driver-failures as much as possible in our new ROAR-series. As linear excursion, Bl and power-handling keeps on increasing in every new generation of drivers we though we needed a new approach to take advantage of these factors.

Based on my first ROAR12, we feel we have achieved our goals. Due to my illness I have not had the health and strength needed to properly test and evaluate it. The posted measurements above is with a raw driver and indoors. I use an Umik and REW.

I hope you build a ROAR18. If you like the THAM-series then I am certain you will like the ROAR-series.

Cheers,
Johannes