TH-18 Flat to 35hz! (Xoc1's design)

Hmm...ok. The question is: did you ever proof your cabinet design to fit to your faital driver???
I took the 18hp1060 because i got a very attractive price for that. That driver was a fixed point for me. At the beginning i tried to use the xoc1's plans or nearly similar plans by building the first prototype. I quickly found out that the driver needs a differnet design to work best.

That's curious - I sim'd the 18HP1060 in the TH 18 and the results don't seem to indicate a driver/cabinet mismatch. And HornResp suggests that it will take almost 1.2kW to drive the driver to linear Xmax. See sim'd results attached. And a 30Hz 24dB/oct. filter should be enough to keep excursion below the TH's resonance frequency in check.

Of course real results will trump sim'd ones any time, but if I was testing this build for problems, I'd start at limiting the power input to 300W and slowly adjusting it upwards while pulsing it with 50 Hz 6-cycle tones (in-band maximum excursion occurs at that point). I'll bet that the build should start to give problems somewhere north of 1.2kW (the driver is going non-linear at this point of course, but should not be driven to Xmech). If however it starts to show problems well below that, then it's time to take a closer look at the build itself IMO.
 

Attachments

  • 2017-07-10 (3).png
    2017-07-10 (3).png
    25.3 KB · Views: 628
  • 2017-07-10 (4).png
    2017-07-10 (4).png
    19.9 KB · Views: 619
For starters this cabinet wasn't designed for only one speaker. There are drivers that perform better than others but most of the newer premium drivers with parameters that lend themselves to tapped horns should do pretty well in it. I am not saying that mine are doing poorly, they get loud and deep, it's just that I thought that I would be able to drive more power into them before reaching a mechanical limit. They are a pretty solid build with enough bracing, I doubt build quality is hindering performance. I can't imagine I have any air leaks either after all the glue I added after the fact.

@BertlSound. I have a huge respect for all the time you spent running multiple hornresp sims and actually building many prototypes before deciding on the modified versions you built. I also did some sims of a slightly longer horn path and while they did show some additional bass extension they also seemed to run out of xmax earlier a little above tuning (50-60hz range). Looking at your pic at post 2449, unless the total box dimensions were changed, the horn path looks only slightly longer than the standard TH18. If you are getting more output than I am, which I assume you are, then I doubt the extra output is due to the horn length but more due to the fact that S2 is so small. You built "cone correction" into your horn path. I opted to add mine to the original design. I am using cone correction as a generic term. I know some people say they concept is fake and it may be.

@Djim. I am not sure what the popping sound is. But if your theory is right then the driver is too weak to be hitting xmech. So what is the noise then? The cone being so non-linear that it is scraping or something? Also why does BertlSounds 18HP1060s work so well when his design puts the hornpath almost into the cone causing S2 to be real small?

I did incorporate the cone correction as separate pieces so I could remove/modify it if needed. I suppose I could remove it from 1 cabinet and compare the 2.
 
Hi SonicXtasy,

I’ll try to explain, with a 'weak' motor I meant it loses control over the cone in a TH18 before its published Xmax is reached. Losing control = losing the electromagnetic damping over a moving mass, a force needed to stop it.

The Rme of 123.0 N*s/m, as calculated in post#2600, only counts as long the VC stays in the gap. This can be extended a little outside the gap for instance by using a split-VC but I don't know if that is the case with the 18HP1060. Lets assume FaitaPro did something and their published Xmax represents more or less 70% of BL. At this point the motor still has a BL of 17.36 N/A but only an electromagnetic damping of 60.27 N*s/m.

As far the 'popping' sound, it is difficult to tell what you experienced. If you used proper filtering hitting Xlim doesn't seem likely, but not impossible. The VC hitting or scratching something is also possible. In post#2602 you wrote: "I did incorporate the cone correction as separate pieces so I could remove/modify it if needed."... maybe it was one of the separate pieces?!

Regards,
Djim
 
Last edited:
For starters this cabinet wasn't designed for only one speaker. There are drivers that perform better than others but most of the newer premium drivers with parameters that lend themselves to tapped horns should do pretty well in it. I am not saying that mine are doing poorly, they get loud and deep, it's just that I thought that I would be able to drive more power into them before reaching a mechanical limit. They are a pretty solid build with enough bracing, I doubt build quality is hindering performance. I can't imagine I have any air leaks either after all the glue I added after the fact.

@BertlSound. I have a huge respect for all the time you spent running multiple hornresp sims and actually building many prototypes before deciding on the modified versions you built. I also did some sims of a slightly longer horn path and while they did show some additional bass extension they also seemed to run out of xmax earlier a little above tuning (50-60hz range). Looking at your pic at post 2449, unless the total box dimensions were changed, the horn path looks only slightly longer than the standard TH18. If you are getting more output than I am, which I assume you are, then I doubt the extra output is due to the horn length but more due to the fact that S2 is so small. You built "cone correction" into your horn path. I opted to add mine to the original design. I am using cone correction as a generic term. I know some people say they concept is fake and it may be.

@Djim. I am not sure what the popping sound is. But if your theory is right then the driver is too weak to be hitting xmech. So what is the noise then? The cone being so non-linear that it is scraping or something? Also why does BertlSounds 18HP1060s work so well when his design puts the hornpath almost into the cone causing S2 to be real small?

I did incorporate the cone correction as separate pieces so I could remove/modify it if needed. I suppose I could remove it from 1 cabinet and compare the 2.

Hi SonicXtasy

it was just a guess that you made no sims because i couldn't read out that from your posts. sorry for that. you are right that i made also "cone correction" into the hornpath. i think it's a waste of hornlengt doing it in other ways. compression ratio is high that's also true. what i've seen is that S2 is the most sensitive part of this design. with the origin design it works also but not really good. i had to adjust the cab not only slightly to get a better response.
but anyway you should not get this popping problems when you drive the speaker in a normal way.
i'm driving my self-powered models with 1500W(8ohms). the driver seems to be always save also at maximum power. i never had such problems you described. next thing i will spend some time for is to measure the subs completley for better comparison (FR, Sensitivity, Splmax, Power compression - my amp modules can measure the load-impedance while playing).
all the best bertl
 
Thanks Art for the info. There seems to be no difference in the gap between the 4 and 8 ohm version which means some distributors :nownow: need to correct their Thiele-Small parameters.

I made both HornResp to compare the difference between the 8 and 4 Ohm version. Based on 1 watt calculated for each driver.
Eg for the 4Ohm version = 1,8V
Eg for the 8Ohm version = 2,3V

As you can see there is no difference between sensitivity and a very minor difference in response. As suspected the minor differences you only will see back in Group Delay and impedance.

18SW115_8ohm_vs_4ohm_01.jpg

Sorry to bring up a very old post but I am trying to compare the Xoc1 Th-18's sensitivity and max continuous output (at xmax) to the Danley TH-118.

The Danley TH-118 has a sensitivity of 108dB (2.83V at 4ohms) = 105dB at 1w. DJim's sim above (using the same 4ohm B&C 18SW115) shows the sensitivity to be about 98dB (averaged from about 50-80hz like on Danley's measurement graph).

Is the TH-18 really about 7dB less sensitive than the Danley TH-118?? I haven't tried measuring the 1w sensitivity myself yet (I don't know how to) but 7dB less sensitive can't be correct!
 
Last edited:
Fenis, first of all the Xoc1-Th18 is a TH in his own right. According to Xoc1 the design was coincidentally almost the same design as the DSL TH118. The DSL TH118 has a much higher compression ratio in the first section of the path. (S1/S2)

Later versions of the Xoc1 TH18; MK1, MK2 etc where more changes to 'mimic' the TH118 performance with cone compensation pieces added, or slight changes in the path construction.

The sensitivity story is most likely (like all/most commercial pro builds do) the highest response pointer in the graph.
 
Last edited:
Hi

yesterday i did some sensitivity measurements with my TH18s. I borrowed a mesuring equipment from a hifi workshop which is very easy to use. its a closed hardware based system from a german developer called Kirchner ATP. The system can calibrate the amplifier voltage directly for sensitivity measurings. Impedance measurings, TS parameters and many other can be done directly.
It's an older system running on XP and it's made for hifi applications. it can't handle heavy PA levels above 130dB so i have to check splmax separatley.
The 1W1m measuring was made indoors and looks good (driver = Faital 18HP1060)
 

Attachments

  • IMG_8327.JPG
    IMG_8327.JPG
    349.7 KB · Views: 521
Is the TH-18 really about 7dB less sensitive than the Danley TH-118??
No, definitely not. Using the same driver (18SW115), you might see up to something on the order of 1.5db less sensitivity or so due to the build quality and refined design of the TH118 design.

As mentioned, the TH18 design was pseudo clean-room designed, but the reality is its performance is very comparable to the TH118.
 
Hi

yesterday i did some sensitivity measurements with my TH18s. I borrowed a mesuring equipment from a hifi workshop which is very easy to use. its a closed hardware based system from a german developer called Kirchner ATP. The system can calibrate the amplifier voltage directly for sensitivity measurings. Impedance measurings, TS parameters and many other can be done directly.
It's an older system running on XP and it's made for hifi applications. it can't handle heavy PA levels above 130dB so i have to check splmax separatley.
The 1W1m measuring was made indoors and looks good (driver = Faital 18HP1060)

Looks good BertlSound! Is that 18HP1060 4 or 8ohms? Mine are loaded with the B&C 18SW115 4ohm with cone compensation so I'd imagine mine would be about 104dB as well. 2 of them next to each other being 110dB 1w is very nice efficiency!
 
Hy Bertl,
thank you. Did you compensate for groundplane in your graph??
Regards

Hi Sabbelbacke

i was a bit confused about the steep roll-off in the mesuring. i also thougt about your advice for the groundplane compensation. your are right in that point, i didn't any compensation.
so i repeadtet the measuring yesterday with same conditions and spend some time playing around with the ATB PRecision. doing so found out a curious thing:
as i said i was wondering about the steep roll off, because i had already done measurings with other systems like Arta, Carma4 and with my Neutrik Analizer before. All that measuring didn't show such a steep roll off. so i checked the ATB in it's self by measuring power-amp out to ATB input. Than i got a little surprise.The ATB PRecision System came with an preamp and a power Amplifier (Mackie m1300i). a magnaditude measuring showed the the shitty preamp causes a roll-off by 6dB starting at 200Hz!! have a look at the graph attached. unbelievable. the flat line is ATB connected directly to the mackie amp. measured with adjusted 2,83V at CH1 out.
then i checked mic sensitivity. this was adjusted (8mV/Pa) correctly. after that i made new measurement outgoing from the first graph.
blue: first measuring, 1W1m, with shitty preamp, mic on the floor, indoors
brown: new measuring, 1w1m without preamp, mic oin the floor, indoors
red: new measuring, 1W1m, without preamp, mic 0.5m obove the floor, indoors

didn't think it would be tricky like that measuring sensitivity but at the end - i hope, now it's done - the result is ok. thank you for your helpful input. please tell me again if there is something wrong with my approach.
 

Attachments

  • IMG_8360.jpg
    IMG_8360.jpg
    454.8 KB · Views: 536
  • IMG_8359.JPG
    IMG_8359.JPG
    426 KB · Views: 555
  • IMG_8358.JPG
    IMG_8358.JPG
    280 KB · Views: 569
Looks good BertlSound! Is that 18HP1060 4 or 8ohms? Mine are loaded with the B&C 18SW115 4ohm with cone compensation so I'd imagine mine would be about 104dB as well. 2 of them next to each other being 110dB 1w is very nice efficiency!

Hi

i'm using 8Ohm drivers. so i can drive also 2 with my Pascal ampmodules (1550W@8Ohm, 2400W@4Ohm). big advantage: just one powercord 230V/16A for 4Subs+2Tops. because of the high efficency most of my jobs can be done with this setup;)
 
Hi
...so i can drive also 2 with my Pascal ampmodules (1550W@8Ohm, 2400W@4Ohm.)


Have you recognized any problems with power reduction on the Pascal XPro modules.
They deliver long term around 1kW.
Sinus bursts on the Sub-Channel around 2,4kW and the other two channels 800W (XPro-3).


I´m planning to use the XPro-3 with two 8Ohm Subs (3" VC) and two small tops.
I´ve already mountet the modul on a 45x35cm Aluminium plate , added a Xcillica DSP and a seperate heat sink with forced air colling for the module.
I´dont think I have to worry about the XPro-3 modul.
 
Have you recognized any problems with power reduction on the Pascal XPro modules.
They deliver long term around 1kW.
Sinus bursts on the Sub-Channel around 2,4kW and the other two channels 800W (XPro-3).


I´m planning to use the XPro-3 with two 8Ohm Subs (3" VC) and two small tops.
I´ve already mountet the modul on a 45x35cm Aluminium plate , added a Xcillica DSP and a seperate heat sink with forced air colling for the module.
I´dont think I have to worry about the XPro-3 modul.

Hi Dominik

no i didn't have any problems. the power ratings of the xPro modules are differing from 2kW up to 2.4kW depending on the manufacturer using the modules (for instance PL-Audio says 2.4kW). The power supply delivers 2kW. I think they go along with todays class D amps from other manufacturers. it's nearly the same technology Camco/L'Acoustics is using. i recommend reading the test from Production Partner release 2/2017 regarding L' Acoustics LA12X. find some pics attached - maybe you can read it. on the last page you can see an advertising from four-audio which is founded by Anselm Görz, the german audio godfather (he does the testings for ProductionPartner, he developed also monkey forrest, maybe you know that). They also use the pascals..... in daily use i did not see any kind of weakness by playing different styles of music (live or playback). sure it will suck the capacitors by playing long term bursts, but have a look at the pcb of the LA12A. you won't find a big number of capacitors. i regarding power rating of class-D amps it's all about the definition.
 

Attachments

  • IMG_8361.jpg
    IMG_8361.jpg
    881.5 KB · Views: 524
  • IMG_8362.jpg
    IMG_8362.jpg
    745 KB · Views: 502
  • IMG_8363.jpg
    IMG_8363.jpg
    820.6 KB · Views: 156