HT 18 DVC subwoofer build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thanks for the horn tutaledge (pun intended). I was never into horns too much. I think for simplicity sake and size consideration the standard port ought to to. If I can get 110 dB in the living romm I'm sure that will be more than adequate. Anything louder than that and the wife's plate collection would be at risk. A pair of MTMs shouldnt have too much trouble hitting that. Also, if a single driver will do that, which ibsuspect that it will, then the DVC woofer can be wired in series for 4 ohms as well, so that will work out well as a fair number of the MTMs are also 4 ohms. Not a big deal but I'm all for keeping it simple.

I'm also heavier on the sim side of the fence. There is no reason why you can get close with a sim and then tweak a bit if necessary. I'm really not sure why some, who will remain nameless, have such a hard time wrapping themselves around that!
 
Hey JAG -- is there a particular software that you can recommend to model the ported box. I remember reading that you had to make some adjustments for the SI HT18 in order for it to sim well. Also, I guess the SI provided properties were a bit off. Seem like Mark had put some values out there:

Re = 3.8905 ohms
Fs = 17.8978 Hz
Zmax = 37.0127 ohms
Qes = 0.5616
Qms = 4.7809
Qts = 0.5025
Le = 3.6920 mH (at 1 kHz)
Diam = 387.3500 mm ( 15.2500 in )
Sd = 117841.1543 mm^2(182.6542 in^2)
Vas = 184.9278 L ( 6.5307 ft^3)
BL = 25.6356 N/A
Mms = 843.5236 g
Cms = 93.7440 uM/N
Kms = 10667.3457 N/M
Rms = 19.8412 R mechanical
Efficiency = 0.1774 %*
Sensitivity= 84.5076 dB @1W/1m
Sensitivity= 87.6385 dB @2.83Vrms/1m
Krm = 52.781E-03 ohms Freq dependent resistance
Erm = 637.558E-03 Rem=Krm*(2*pi*f)^Erm
Kxm = 48.447E-03 Henries Freq dependent reactance*
Exm = 699.829E-03 Xem=Kxm*(2*pi*f)^Exm
;------------------------------------------------------------------
Ftest = 15.867 Hz
Ftest/Fms = 0.8866
Test Mass used = 229.7000 g (Equal to 45.9 nickels)
Test Mass (Ft=Fms*0.90) = 197.864 g (Add -31.836g for Ft=16.108)
Test Mass (Ft=Fms*0.75) = 656.074 g (Add 426.374g for Ft=13.423)
---------------------------------------------------------------------------------------
* The base TS model will be affected by Cms variation caused by break-in, heating, or *
* changes in suspension components. BL, Re, Rms, Mms etc are assumed to be constant *
---------------------------------------------------------------------------------------
pct Fs-Hz Cms-um/N Vas-L Qms Qes Qts
25.0% 8.9489 374.9761 739.7112 2.3905 0.2808 0.2513
31.5% 10.0448 297.6187 587.1091 2.6832 0.3152 0.2820
39.7% 11.2749 236.2202 465.9889 3.0118 0.3538 0.3166
50.0% 12.6557 187.4881 369.8556 3.3806 0.3971 0.3553
63.0% 14.2055 148.8094 293.5546 3.7946 0.4457 0.3989
79.4% 15.9451 118.1101 232.9944 4.2593 0.5003 0.4477
100.0% 17.8978 93.7440 184.9278 4.7809 0.5616 0.5025 <- base TS
126.0% 20.0896 74.4047 146.7773 5.3664 0.6303 0.5641
158.7% 22.5498 59.0550 116.4972 6.0236 0.7075 0.6331
200.0% 25.3113 46.8720 92.4639 6.7612 0.7942 0.7107
252.0% 28.4110 37.2023 73.3886 7.5892 0.8914 0.7977
317.5% 31.8903 29.5275 58.2486 8.5186 1.0006 0.8954
400.0% 35.7956 23.4360 46.2319 9.5618 1.1231 1.0051

AFAYK, are those about as good as it gets without testing?

I was looking through the 17 hz organ thread, and I saw some sims for the OPs box, but I couldn't seem to find any for an idealized box.

Also, you think 110 dB seems a little optimistic? It really might not even need to be that loud, but it would be nice to have a little headroom. I got 4 drivers if I need them!
 
No those t/s parameters Mark measured are garbage, you need to keep reading that thread. 843g MMS is absolutely ridiculous.

After I complained that it was ridiculous he said his scale was not calibrated correctly so he corrected for the difference and came up with a new set of parameters with 543g MMS, which is still too high. Both times he has tried to defend these results but they are wrong.

Mark measured the t/s parameters in a car which is unbelievably naive. I would use ANY of the probably dozens of t/s measurements of that driver that have been posted instead of Mark's. I've looked over quite a few and they are all pretty close but Mark's is way out in left field. If you search avs forum a bunch of people have measured this driver.

I would personally just use the t/s parameters measured by data-bass. It was a prototype that he measured but probably no significant changes have occurred since then because all subsequent measurements I have seen except Mark's are pretty close.

As far as a sim program, I use Hornresp whenever I can but I use MJK's worksheets for mltl and sometimes tl design, TL.app for stuff with multiple ports or parallel elements and Akabak for the more obscure features that other programs don't have.

Hornresp has my large coil inductance compensation built right in. You absolutely will not get an accurate sim if you don't account for this issue. My adjustment method isn't perfect but it's proven to be a lot more accurate than a regular simple sim.
 
I was looking through the 17 hz organ thread, and I saw some sims for the OPs box, but I couldn't seem to find any for an idealized box.

Also, you think 110 dB seems a little optimistic? It really might not even need to be that loud, but it would be nice to have a little headroom. I got 4 drivers if I need them!

I can run a sim if you want but I need to know stuff like the max size you can live with.
 
If you want to know more about the large coil inductance issue my paper is here - https://sites.google.com/site/amate...coil-simulation-accuracy-issue-and-adjustment
If you use Hornresp the adjustment is built in to the loudspeaker wizard (if you select it) but if you use any other simulator there's an adjustment formula you can use.

... and this image is a preview of what to expect (overlaid but not adjusted for sensitivity). I was shocked at first when LTD02 sent me this image, I knew my adjustment worked but I had no idea the results would look like this.

An externally hosted image should be here but it was not working when we last tested it.
 
I can run a sim if you want but I need to know stuff like the max size you can live with.

I think if it is a ported box it's not going to be too big. I could live with 5-6+ cu ft per driver. Some of the horns just get rediculously large. I'll also try and download hornresp tonight as well and would be interested to know how you come up with the corrections. I haven't had a chance to read through your post on it yet, but I will.
 
Yeah, 5 or 6 cu ft isn't really going to cut it, especially if you are talking about external dimensions. Here's an average size box for this driver - Cycl?ps 18 - AVS Forum | Home Theater Discussions And Reviews
This design is 12 cubic feet external and 9 cubic feet internal.

Here's a sim using data-bass measured t/s parameters in this design.
First screen - Hornresp inputs (Bl adjusted for large coil inductance issue)
Second - schematic
Third - frequency response
Fourth - port velocity
Fifth - displacement
Sixth - frequency response NOT adjusted for large coil inductance issue

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


Notes:

1. Shown at 600 watts which is the rated power handling for this driver. It can take more if you play music with high dynamic range and low duty cycle. It will be destroyed at this power level if you play low dynamic range high duty cycle like a sine wave for any amount of time.

2. Port velocity if higher than I'd like to see but it's still in the acceptable range. BUT if you use a high pass filter you will cut it down significantly.

3. This design is tuned a bit higher than you'd like, you can see in the displacement graph that the minima is at about 18 hz. You would want to tune a couple hz lower if you use a high pass filter and more than a couple hz lower if you don't use one.

4. At 600 watts you are only using about 18 mm excursion. So you could increase the power a bit (if the music signal is not too demanding) OR you could size the box larger which would increase the low end sensitivity while eating up a bit more excursion.

5. The 6th screenshot shows a regular sim - NOT adjusted for the large coil inductance issue. This is how most people will sim this driver but it will not be accurate. The third screenshot shows the frequency response after it has been adjusted for this issue. It might not be perfectly accurate but it will be quite a bit more accurate than a simple sim. If you read the paper I linked to regarding this issue you will see that this type of driver does not simulate accurately at all and you can get much closer if you make a small adjustment as I have done in the third screenshot.

So you can download Hornresp and play around with this a bit. Lower the tuning, play with the box size. You can also look at whatever tb46 designed for the organ sub in the other thread.
 
Man, it deleted my response!!! I got most of the curves to match, but don't know how you got the port velocity to plot. A couple of the variable I didn't find:

Eg?
Rms?
Vrc (box volume?)
Lrc?

Is there a tutorial on hornresp somewhere?

I'd trade a little efficiency for lower flatter response, but am having a little trouble altering the variables to get it (maybe it doesn't do that?). It doesn't seem like the acoustic output is very sensitive to box volume changes...

Also, when I tune the port down I get about 20 mm disp before the port tuning kicks in. since the xmax is 23.5 so maybe that is not a problem...
 
What do you mean you didn't find the variables? You don't know what they are?

Eg - power input in volts
Rms - this is a driver spec, if you double click the rms box you can enter qms instead (you can double click all the boxes and see what happens)
Vrc - volume of rear chamber
Lrc - length of rear chamber

There are a few Hornresp tutorials, just google Hornresp tutorial, there's also instructions included with the program. And if you hover the mouse over any box it will tell you what it is in the bottom of the program.

You won't get lower flatter response unless you make the box exponentially larger. That's what happens with these large coil drivers with inductance issues. The apparent motor strength is a problem, it's much lower than the specs would indicate. if you use the measured spec for Bl (21.3 or something) you can get flat response as low as you want really easily as I showed in the X'ed out 6th screenshot, but that is not reality. The driver will not measure the way a simple sim indicates.

20 mm excursion isn't a problem.
 
Last edited:
OK, I've been playing with hornresp and getting the hang of it. I ran through some scenarios and made some observations..

First the 'small' box. I wanted something that would layout in a 4X8 sheet of plywood. First the 'small' The biggest volume you can get for one box out of a 4X8 sheet is 2X2X3. Internally that is 22.5X22.5X36 internal = 10.5 cu ft- port = 300 - 16 = 284 liters (say 275 with a little strategic bracing).

Next if you get the 4 sides out of a single sheet (4 boxes 5 sheets) you end up with a box 2X2X4 or 22.5X22.5X48 internal - 14.1 cu ft - port = 400 - 10.9 = 389 liters (say 375 with a little strategic bracing)

results:
 

Attachments

  • sm_box_input.png
    sm_box_input.png
    9.8 KB · Views: 57
  • sm_box_acoust_pow.png
    sm_box_acoust_pow.png
    16 KB · Views: 57
  • sm_box_drive_disp.png
    sm_box_drive_disp.png
    15.5 KB · Views: 55
  • lg_box_input.png
    lg_box_input.png
    9.8 KB · Views: 54
  • lg_box_acoust_pow.png
    lg_box_acoust_pow.png
    15.7 KB · Views: 51
  • lg_box_drive_disp.png
    lg_box_drive_disp.png
    15.2 KB · Views: 14
  • ex_lg_box_flat_input.png
    ex_lg_box_flat_input.png
    9.9 KB · Views: 16
  • ex_lg_box_flat_acoust_pow.png
    ex_lg_box_flat_acoust_pow.png
    15.9 KB · Views: 17
  • ex_lg_box_drive_disp.png
    ex_lg_box_drive_disp.png
    14.3 KB · Views: 18
I seems like the larger the box the flatter you can egt the response curve, but the driver displacement goes up in the larger box (or less power for the same displacement). Is there something on the DIY site here that I can use [model] to come up with the high-pass filter for the sub to keep the infrasonic out?

I think I poke around a little more with the box sizing -- there might be a boz just a little larger that I can live with!
 
Is there something on the DIY site here that I can use [model] to come up with the high-pass filter for the sub to keep the infrasonic out?!

Hornresp will simulate a high pass. After you Calculate go to Acoustical Power window, then Tools menu and select Filter Wizard.

Also if you haven't figured it out yet you can double click the Eg box and enter whatever power you like in watts and reference to either the driver Re or the nominal impedance (Re will show actual power at 1 watt, nominal impedance will not).

And to get the port velocity go to Acoustical Power window, Tools menu, Output - Port and then Tools - Particle Velocity. You might want to check on that, your port is really small, just over half the size of the one in the Cyclops design.

And why 1 pi? Unless you are putting the sub in a corner 1 pi will prove to be a bit overstated.

Are you masking resonances? If so you probably don't want to do that.

600 liters is a bit big and if you prefer the sound of sealed boxes you probably won't like that max flat sim you created that is flat down to 12 hz.

Looks like you are getting the hang of it. Carry on.
 
Last edited:
The boxes will have a common floor/wall intersect, not a corner. The pick shows a yellow zone that looks like that, so I figured...

I'll checkbthe filter thing tonight when I get home as well as the port velocity. I assumed a 6 in diameter pvc.

Let's say I like the clised box sound or at least something not as ringy as a typical old-school ported box -- what should ibbe looking for in the alignment?
 
The boxes will have a common floor/wall intersect, not a corner. The pick shows a yellow zone that looks like that, so I figured...

If your boundaries (wall and floor) are infinitely rigid and create a perfect reflection with no losses then 1 pi will be appropriate. An underground reinforced concrete bunker would be close. If your listening room is not in a concrete bunker it's better to sim 2 pi and just be happy if you get a couple or even a few more db than expected.

I'll checkbthe filter thing tonight when I get home as well as the port velocity. I assumed a 6 in diameter pvc.

Port velocity was already pretty high in the cyclops design (assuming no high pass filter) and that port was almost twice the size. With a 6 inch port velocity will be higher and you may or may not notice audible chuffing, but port losses due to port compression start to happen way before you hear audible chuffing. The conservative guideline is to keep velocity at or under 10 m/s to control port compression losses.

Unfortunately you can't sim a high pass filter and port velocity at the same time in Hornresp but you can in Akabak. It only takes about 5 minutes to import a Hornresp design into Akabak and check it out if you know what you are doing. I might be able to help with that but probably not, it's pretty busy around here. Maybe TL.app will let you sim port velocity and a high pass filter, not sure, I've never checked.

Let's say I like the clised box sound or at least something not as ringy as a typical old-school ported box -- what should ibbe looking for in the alignment?

Reread post 5, I was pretty clear about this. If you look at the frequency response of a sealed box and recreate that with a ported box it will sound like a sealed box.

With this driver in particular, even if you do this it won't sound like any sealed box you have ever heard with a low or mid qts driver because this driver will produce a pronounced "inductance hump" at 30 or 40 hz instead of a purely rising response that rises all the way up into the midbass. Only high q drivers produce this hump, and with the large coil issue factored in this driver is effectively a high q driver. You can see the hump in the measured response that I showed. It was measured by data-bass.com, that is a very reliable source, and it's an actual measurement of an actual driver, so you know it's true.

But to be very clear, if you like the sound of typical low/mid q drivers in sealed boxes the thing you DO NOT want to do is make a 600 liter max flat ported box that's flat down to low teen digit frequencies.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.