radulescu_paul_mircea

After adding two more braces from one side to the other, I did another round d of bursts.

4 meters, added 6 dB for 2 meters equivalent.

25-126.7 dB

30-136.7 dB

31,5-  137.3 db

40-141,4 dB?

50-144.2 dB?

80-146.4dB?

90-144.9 dB?

100-144,2dB ?

The sensitivity at 30 Hz is 99.8 dB at 4 meters at 2.83 V. Just a dB higher than the original. 

His 3 brothers are on their way, I'll have 2 with 21Ipal and 2 with 21ID but with slightly modified bracings. This one didn't behave as I wanted it to and there was a strong vibration on the large panels so I need to add 3 more pieces of wood from one side to the other. Another problem is the weight, at 165 kg, it is pretty hard to manage.

After all , I will not build the back ones for cardioid configuration, I'll use them as they are.

These are the test enclosures, to learn and to try things. They will be for comparing my new designs and to see what kind of improvements I can manage to squeeze , if I go bigger but I'll keep the dimensions constraint to 140*140*70/75 cm...

Each driver on one FP14000 clone chanel, CEA2010 burst, calibrated mic, 4 meters, added 6dB for 2 m equivalent. Where you see a question mark, the amp turned itself off for protection. the closest buildig was at about 10 meters behind the sub.

Bph221

25-124.1dB

31,5-135,1 dB

40-141,1 dB

50-144 dB?

80-145.3 dB?

90-144.9 dB?

100-144,2dB ?

4 hours ago, Ricci said:

External dimensions?

140*100*70 cm .

These are the modifications i made to this design, it is based on my sketches. I changed all the bracing, it was easier to make all of them with constant width because of the bigger width. The much bigger back chamber will surely increase the efficiency and power under 45 hz, I didn't feel the need to improve the horn on this one, it is good enough with just 20 cm extra length and an S1 correction. the material is already being cut

Already ordered some wood, in fact a lot, I needed to order more to get the best price and I also need it for my other projects.

 I thought a lot of what you said and the sub would weight too much to be practical. So I decided to go modular. The idea is to make two SKHorns , but the depth will be 40" and the width 28", and port tuning will be a little higher , at 34 hz approximately, Just to get the best deep end I can, but also to match the top end a bit.

The back enclosure will be an Ss15 based design, where I try to match the phase as well as I can, but I need the best efficiency  to be centered around 50 Hz  if I can, because I will make an end-fire configuration with that center frequency and the response of the front enclosure will have a high shelving in the back of it, so the frequency response for not need to match . 

The two enclosures will be installed back to back with screws buried on the sides . I wanted to use an electromagnet like the ones for doors , but that would complicate things a bit and unfortunately I'm a bit short on time

I ended up with end-fire instead of gradient configuration after I did a very cool experiment. I put the RTA I'm REW on infinite averaging for a  5 hours set only with the music they are playing and the conclusion was that the energy is centered at 52 hz with a difference of 6 dB compared to the 40-110 hz band. So if I can get the attenuation of -15-20 dB at 52 hz in the very far field, over 100 meters distance in the back of the speakers, compared to a normal setup, it will be extremely helpful.

I've already got my drivers, the difference between the B&C and the 21ID is night and day, power handling, excursion, shear mass, thermal inertia etc. But they are both extremely well.made drivers, with different applications for the same approach.

The plan is to build these speakers for use in this application and to use them for comparing my own designs which will be build in the next month.

I'll post some pictures when I'll start building

The drivers are already here! The 21ID is so small and light compared to the 21Ipal.the cone is much softer,  also, the coil is way smaller, with a much thinner wire and it gets hot more quickly. Probably twice as fast. 

The driver gets very noisy after about 3.5 cm excursion pk-pk , in comparison the B&C can travel about 5 cm before any strong noise . We clearly have a winner here, but that was not what I wanted to do.

@Ricci you are right, the setup is a bit stupid proof, it's easier to get it right and to avoid any tempering with the setup. 

The decision is not final and it's pretty hard to get the fold good enough to match the phase and response of the front part. I also wanted to put the port and the horn in the center to get the acoustic path as close as possible vertically and horizontally

I've just ordered 2 21Id drivers, so I'll probably build these. They'll have 2 21Ipal on the front in series and one 21ID on the back . I'll use a Powersoft K10 on the front drivers and a FP14K clone on the ones on the back, one on each channel . Processing will be a DBX driverack. 

It will be very helpful to get some direct comparisons between the 18Sound and B&C drivers, I'll also get to play with some mammoths, because they'll get very large

What do you guys think of the idea of making this a cardioid sub. With 2 21" IPal on the front and one 21ID on the back to be able to use a 2 channel 2 ohms stable (ish) amp, like a clone , on each sub ? It would need to be 50% lager, practically to just go deeper and make a fold that works exactly like the front . 

I know the enclosure will vibrate more and it would be a bit hard to match the tuning for the port, but I think the hardest part is to get that horn right and to be able to find a good folding. It would probably need different hatches also

But I need that configuration and I bet it would be awesome!

Off course, it could be easier to build 3 SKHorns and stack them with one backwards but the place I would be using these is a bit wide and I need to spread them to have a better coverage.

On 3/9/2018 at 12:45 PM, eggchen said:

Did you end up comparing the 18ipal to the 18SW115? I have 2 TH18's with the 18SW115 (4 ohm) powered by a Crown Itech 6000 and am very happy with the output and sound quality! 

I didn't test it after all , because I changed the 18"s with 21"s to use them in my own designs. I don't want to use that design anymore, it's better to use my own The difference though is huge, the efficiency and thermal stability are way higher than those of the SW115. I'll use the 18 IPal in one of my other designs and after that I'll use the 18Sound where normal drivers are needed.

And the 21" Ipal is so powerful compared to anything else, the sound is so much better!

1 hour ago, Ricci said:

Odd because there does appear to be a lot of noise in the 10m measures. Thanks for clarifying

If you leave the response as it is , you'll see that the impulse response is truncated at +500 Ms from ref time. If you use the tools to change it at 125 Ms and to increase the zeros before ref to keep the resolution, you'll find that the response gets clearer. This is because I was surrounded by solid buildings and their reflexions were very powerful

3 hours ago, Ricci said:

. If you ever do this type of testing again try using a 100w nominal input voltage for all of the measurements. 20 volts for a 4 ohm nominal cab for instance. This will be a lot louder but it is necessary when doing measurements like this in a noisy environment especially at 10m.

I used about 28 volts for these measurements , but the voltmeter wasn't very precise so until I get a better weather to be able to take another reading with a new ,much better meter, I'll settle for what I have now.

I have used one FP14000 clone for them , and adding another dual 18, 3 drivers per channel, I managed to pop a 16A breaker at 230 V. It is quite power hungry

2 hours ago, Ricci said:

looked at this set of measurements. It looks like you were taking measurements at distances of 1/2/4 and 10 meters with the following orientations: Directly on axis, turned 90deg, 135deg and 180deg? This set of measurements was conducted on a single cab, dual cabs, and quad set of cabs. Is this correct?

One thing that immediately jumped out at me is that the background noise floor is quite high. Looks like about 75-80dB

On 2/21/2018 at 6:18 PM, Ricci said:

 

They 1, 2, 4 in a 2 by 2 stack and 4 in a vertical line stack.

The measurements were taken at 0 deg, 90 deg ,135 deg and 180 deg from axis on ground. The noise floor wasn't that high. It was lower than 60 dB for most of the time and when the wind was harder, I paused the measurements until it  settled. 

7 hours ago, Ricci said:

I finally DL'd this. I'll have a look at it when I get a chance. What model of subs were used? 

I used a set of 4 single 18" custom subs loaded with RCF LF18G401, made for a client, which are used as low kick.

For anyone who wants to see the measurements for the directivity of 1,2 and 4 single 18" subwoofers in different configurations, on and off axys, you can access my Drive account and download the zip with REW measurements. Where there is a V, it means vertical possition

You would need to change the Impulse response windowing to Left Window Tukey 0.25 400 ms, Right Window Blackman-Harris 7- 150 ms and Ref. T is left in place for each sweep.

https://drive.google.com/open?id=15MmO5qj9v1RLZfO1_BYkUIAf6MDLi8M2

I hope it works.

11 hours ago, Ricci said:

I've never been on social media. I need another time sink like a hole in the head. 

Haha. Yeah, that group can be a time black hole. But lots of good stuff overall, even though like any other forum or group there is some background noise. 

https://www.facebook.com/groups/733449990043427/?ref=group_header

This is an awesome group on Facebook and I've been posting there for a while. Some of my measurements can be found over there and many other discussions mostly for pro audio.

You can find manufacturers there , like Brandon Heinz (from Renkus Heinz) Dave Millard, representative from full far audio, Sean Hennessey and some others.

Please, feel free to join!

Testing the Directivity index of different configurations of subs and summation. Wow!! With only 4 single 18 in a square stack, one gets a huge front to back directivity.
I've tried a single, double vertical, double horizontal, 4x 2 by 2 stack, 4 tall and I forgot about 4 horizontal but I was already frozen by then. 
On axys , 90 deg, 135 and 180 degrees off axis, at 1, 2, 4 and 10 meters.
Well calibrated mic , ECM8000 with absolute SPL calibrated from a B&K 2270 to 0.1 dB ( hard work to do that unfortunately, I'll get a new mic very soon), Focusrite Scarlett 2i2, REW with loop, 20 meters from any wall, sweep measurements and RTA. I also did CEA2010 bursts , maximum SPL and loads of fun!
Plus some playing with a closed enclosure loaded with dual 21Ipal isobaric series.
Amplifiers were a FP14000 clone, one Crown XLS 2002 and one Martin MA2.8.

I'll need to find a way to post pictures here, I have some graphs, I have 130 measurements , I'll make some plots , some excel work on them.

@Ricci yes I am using ABEC 3 but barely. It is way to complicated for now and I don't pretend I am getting any easy progress. I have an Akabak script on my older laptop (it doesn't run on Windows 10)  that I could copy and post or send to PM. It is not very detailed but with the additions I managed to get a higher Directivity and a closer response.

Welcome Moar!! Thanks for the link, it is a very useful tool. I needed that in my life

Playing with it I've been getting very interesting results, some of them really incredible. Even from a single enclosure, the Directivity starts to go forward even at very low frequencies like 40 Hz.

This made me go out in a parking lot with 4 single 18 subs and to take directivity measurements.

I've managed to get this sensitivity curve when I simulated the acoustic load of the air around the enclosure, as if it would be another waveguide with a very rapid expansion, but acting on the enclosure . In fact that, coupled with some acoustic resistance in the enclosure, higher losses through friction, increased viscosity in the port and enclosure gave me the most similar response. 

Is there a way of explaining how is it possible that the sensitivity of the speaker to be 107 dB at 100 watts/ 10 meters at 100 hz? I've tried everything to get that sensitivity in Hornresp and Akabak but with no results.i usually get a play response from 40 to 120 hz at about 102 dB. The rising response is not present in any simulations, except if I take into account the directivity increase at higher frequencies , but even with that, I can't seem to duplicate the results. 

More good news things will come from this. B&C takes the wheel at Eighteensound

http://www.proavl-mea.com/#/ondemand/id/3s2NvmQMAC/

Here they are!

4 hours ago, chrapladm said:

Will be interesting to see what your results are. I do like the IPAL but havent been able to make a design that warrants the extra cost yet.

Oh, but they warrant the cost on their own. In anything you put them, that has increased performances. But only with the right amp. There's in fact the problem. Here we only have Powersoft amps that handle very low impedance and high reactivity subwoofers. SpeakerPower amps would be the others, but I can't really get my hands on them easily. Probably some Crests are also good and stable and also Crown amps. But not as good.

No, I missed a punctuation mark. They are using the 18SW115-4 driver in the original TH118. I will build 2 identical enclosures to the TH118 , that are now easily found on the web.in those enclosures I will try the ID and IPal drivers to see what are the differencies and if it worths the cost difference. In my sims, the IPal version should go flat to 33 Hz instead of 37 and it should keep the same max SPL because it has more excursion and long term power handling than the SW115.