@ziggurcat
Forum Replies Created
-
Posts
-
I have not even made any attempt to build this yet – it was always going to be a project to start sometime this year (I started thinking about it late last year) once the weather got warmer. I also have a tendency to plan these things, and abandon them due to laziness. It’s been a long time since I’ve been on this forum, but I believe I was sort of targeting a Q of 0.7, and the math would have been based on whatever WinISD spit out at me.
Oh, please don’t feel as though you’re obligated to respond. 😳
Unless there’s a trick with filters to extend the low end without exploding the cone excursion, I think I might be able to move on to the details surrounding the box itself?
And if there are any other users here that can add their voice to this, please do!
I know it’s been some time since my last reply, but I did want to clarify (in case it came across this way) that citing the specs from Dayton/PE was not meant in any way as a any kind of “Nuh uh you’re wrong” kind of thing, I was only looking to find out whether the Dayton/PE specs Nominal impedance spec should be entered or if the Znom was to be set to 2ohm. I did load in the driver with Znom set to 2 and 4ohm, and quite literally the only difference was something like -/+ 0.3 to 0.5dB in SPL, which is audibly not going to make any difference.
At this point, I think I have the driver selected as I haven’t come across any other driver that has naturally extended as low as the UMII18-22. I am sure that some of the low end will get a boost from room gain, so I’m not too worried abotu that. I’d rather keep things as simple as possible for my very first attempt.
So given all of that, I might be at the stage of figuring out the cabinet size/damping materials, etc… While I personally wouldn’t be adverse to build an 11 cuft box, I do enjoy having my head on my shoulders, so I think I would ideally like for this to be somewhere around 7 cuft as a max box size. I’m hoping that it’s as easy as what I’ve seen across a few other forums (1lb of polyfill = 0.75 to 1 cuft), but I’ve also seen posts that seem to suggest that it’s not that simple (in my case, though, I’d use carted wool instead of a synthetic polyfill). I’ll continue to search around the internet, but I think it might be a little difficult to sift through all of the proverbial chatter, which has the potential for analysis paralysis.
Anyway, I can’t thank you enough for all of the help in this. I certainly hope that I haven’t been too annoying or caused any amount of frustration 😳
Oh dear… 🤣 I must have had that ROSSO driver in my head because 1256.6 is the Sd for that. I’ve corrected those variables, and added the Le (it wasn’t auto-calculating, and it’s not a number that I recall Toid entering in his video so I haven’t been including that). It (obviously) now matches what you’ve shown in your screenshots. The only question I have would be about the Znom. You mention to change it to 2ohm, but on the PE site (and even the Dayton Audio site), it does say that the impedance is 4ohm. So wouldn’t that make the Znom entry 4ohm? And is everything else was okay?
Under Signal, I do set the distance to 7ft (2.134 meters) because that’s technically how far the sub would be from my MLP. That doesn’t appear to affect the cone excursion when I change it to that value, thankfully. I also nudged the system input power to 880W since that’s 80% of what a Crown 1002 would do in its 4ohm bridged mode.
I would like to build the box a little smaller than the volume generated in WinISD, so what’s the “best” approach to shrinking the box without compromising things too much? If I shrink the box by 4-5 cuft, is there a known mathematical formula to calculate how much “stuffing” to put inside the box for the driver to see it as 11.398 cuft? I get that it probably depends on the material used since there are differences in density, but even a general “guideline” would be helpful. Ben here has addressed some of this, but I welcome as much information as possible in order to gain a better understanding.
Dayton Audio suggests a 4 cuft closed box for this driver (Q target of 0.8). Reducing the box in WinISD to 6.51 cuft gets a Q of 0.8, which tells me that there’s ~2.5 cuft of space (less the space taken up by the bracing, and woofer displacement) that is filled with some kind of “stuffing”, yes?
Understanding that room gain will play a part, is there a way to bring the low end extension up a little bit without causing the cone excursion to skyrocket? When would one use a Linkwitz transform filter (I know its application is only for sealed/closed boxes)? What do those 4 values mean (f0 Q0, fp and Qp)? It seems that this UMII18-22 driver is good on its own without any real tweaking like that, but say if a different driver that didn’t extend as low as the UMII18-22 was chosen… though I haven’t come across a driver that “beats” the UMII18-22.
I do apologize for all of the dumb questions. Once again, thank you for your time, and patience (I get how frustrating it can be when dealing with people who are completely new to something like this – I regularly teach silkscreen printing to people who have little-to-no experience with the medium).
I’ve attached an update .wpr for the UMII18-22 (the data for this driver should be correct, now?). Please let me know if there’s anything horribly wrong with this. I could probably benefit from some help regarding the Linkwitz transform filter. But it might not even be necessary since a simple 2nd order Butterworth Highpass gets me to about the same end result, but with more system input power.
As always, thank you for your incredible input so far.
Can you take a screenshot and show me because I am certain I had the correct values (i.e. mm, cm2, Tm, mm/N, etc…)
I’ve also just installed WinISD on the PC I use at work, and entered the values of the ROSSO 18SW800 (the driver cited in those screenshots), and got the same results as what I see at home.
If you’re referring to the .wpr file I attached above with the UMII18-22 driver – again… I am certain the values are correct. Also, the UMII18-22 is different from the previous UM18-22, and I was referencing Dayton’s own datasheet.
I even just downloaded the .wpr file I attached, loaded it into WinISD at work (where I’m at right now), and all of the values (i.e. mm, cm2, Tm, mm/N, etc…) are all correct vs the UMII18-22’s datasheet
Edit: Sorry – just looked a little closer. For some reason the Re and Bl values weren’t correct, so I have adjusted those… don’t know what happened there, but I definitely didn’t enter 16.8 ohm or 38 Bl for that driver (and they were set to the correct ohm and Tm settings…)
Edit 2: I figured out why it set those specs higher – when I changed the number of voice coils to 2 (which is has), then the ohm and Bl numbers jumped up. I also set it to Series because the data sheet says, “*Dual 2-ohm voice coils. All specifications measured with voice coils wired in series.”
-
This reply was modified 1 year, 6 months ago by
.
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
What do people typically do in that case? Do they retain the SPL or do they sacrifice SPL in favour of low end extension? Or is it one of those “it depends on…” answers? This project is meant to add a second sub into my system (and if it goes well, build it again, and replace my HSU -ULS-15 altogether), so that is another thing to consider with regards to low end extension.
The HPF option really seems to be the easiest option, and I’m not too distraught about the results with the UMII18 driver – again, 102dB at 18Hz is still very loud (and yes, I am aware of the equal loudness curve, etc… so I know for it to be “audible” it requires significantly higher SPL).
The stuffing I would use wouldn’t necessarily be the typical polyfill. I’d probably use actual carted wool (which I think is denser, and more effective) – my wife is a knitter, and spins her own yarn sometimes, so I have a good, cheap source for that – or some of that recycled denim insulation (2″ thick) or a combination of both.
This is an extreme example, but this is what I’m talking about (I think it will appear below). This is the same ROSSO driver as above (post #17, I think), but I set the system power input at the full 800 watts (the driver’s RMS value) to exaggerate the effect. It’s there when I set the system input power for that driver even at the 80% value (560 watts) of the amp I would’ve probably used in this particular case (Crown 1002 – its 8ohm bridged value is 700 watts, and the ROSSO is an 8ohm nominal driver). This shows up when the “SPL is Xmax limited” box is checked in the Advanced section under “Project” in the bottom left corner of the screenshot. In Toid’s video about creating the “perfect box” he says to check that option. You can see in the graph that it starts to roll off, and then at about 55Hz, there’s a sharp elbow, and the graph goes down in a straight line instead of a gradual curve. It seems to be related to cone excursion, but even if I apply a Highpass filter to drop the excursion below the line, it’s still there.
That’s sort of what I am looking into finding out with the Linkwitz transform filter – i.e. what can be done to manipulate it to squeeze a bit more low end without causing the cone excursion to go off the charts. The only thing that seems to do that is lowering the system input power, but then I’m sacrificing a lot of SPL, which I may or may not want (ultimately this all gets level matched down to 75dB anyway once it’s integrated into the system). I also know that none of this accounts for room gain, so I am sure some of that low end will get a bit of a boost from that.
What I think I might end up doing if that Linkwitz thing is a bust – should I stick with the UMII18-22 – is leave it as is, and apply a simple 2nd order Butterworth Highpass filter to get that cone excursion down a little bit since it does creep up a bit past the red line. Because really, 18Hz at 102dB is still pretty okay in my books, which – again – would be a little bit more due to room gain.
Now, would I be safe to assume that the Dayton Audio guy in the video is correct when he says that 1lb of stuffing = 1 cuft, which would get the driver (UMII18-22, in this case) to see a 6.5 cuft box as a 10.5 cuft box if I stuffed it with 4lbs of damping material? If so, and I just leave things as is with a simple Highpass filter then I think I might have my build, but I am going to look at a few more drivers in the meantime to see if there’s anything “better.” There are a lot of options out there, and I’m still not sure which one is the perfect fit. A lot of these drivers only seem to have Fs values around 30-40Hz, so there’s gotta be a way to get more extension via DSP without causing them to explode, no?
I’ve also looked at the Dayton RSS460HO-4, which should be the closest thing to the UMII18-22, but it rolls off a lot sooner than what Dayton’s specs seems to indicate when I pop it into WinISD, so I don’t know what’s going on there.
And someone has to know what those kinks in the graphs mean… it’s only there when the SPL Xmax limited box is selected, but I would like to know what it means.
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
-
This reply was modified 1 year, 6 months ago by .
I’ve tinkered around with the newer Dayton Ultimax II driver – can someone have a look at this file, and let me know if there’s something wrong (I’m still not sure about the system input power, for example)? This is all just testing the waters, so nothing is set in stone or anything like that. Bear in mind that I do not quite know what I’m doing, here, so I apologize to those who are far more experienced. The volume of this is 10.5 cuft to hit a Q of 0.707. The Linkwitz Transform filter is set at the default of what it spit out when I added the filter. The system power input was reduced to 150w in order to get the cone excursion below the red line.
If I remove the Linkwitz transform, bump the system power input to 1240 watts (80% of the 1550w available with a Crown 1502, bridged 4ohm), then simply adding a 2nd order Butterworth Highpass filter cutoff at 10Hz looks like it might be better/simpler? Only “issue” there is the F3 ends up being about 30Hz, F10 would be about 18Hz (though that’s still at about 102dB).
I watched a video from Dayton where the guy goes through the construction of the previous Ultimax 18″ kit they had, and when he talks about stuffing, he says that 1lb of stuffing = 1 cuft. My understanding from that is if I were to build a 6.5 cuft box for this driver, I would add 4lbs of stuffing to get the driver to see it as a 10.5 cuft box, yes?
It doesn’t seem to matter what driver I select, in my tests so far. Even the Dayton RSS460HO-4, which has a spec’d Fs of 18.8Hz or the new iteration of the Ultimax 18″ (Fs of 22Hz) kind of does the same thing (though the UMII18-22 does seem to have an easier time). I don’t know what the specs are for the driver in my HSU ULS-15, but it hits 20Hz in my space (though that’s with room gain), and it’s only a 15″ driver in an 18″ cube. What I would like to know is why that kink in the graph appears. I do have the Xmax limited box checked in the Advanced tab (which is what Toid says to do in his tutorial video).
I’m absolutely not ignoring the cone excursion, which is why I want to figure out how to correctly set the Linkwitz transform filter. Each time I apply that filter at whatever defaults it shows when i add the filter (this differs depending on the driver), the cone excursion skyrockets, so there has to be a way to do it where the excursion doesn’t go off the charts. Increasing the box size doesn’t appear to lower the excursion, either, it seems to make it worse, in fact. Shrinking the box does lower the excursion, but the box Q goes up past where I want it to be.
Backing off the system input power does drop the excursion, but it doesn’t drop below the red line until I set the system input power to 5 watts (at 7ft, which is where the sub would be located from my MLP) with that ROSSO 18SW800. That drops the SPL down to 95dB with the Linkwitz transform is disabled (down from 114dB) before it starts to roll off. My own personal listening levels aren’t that high – about 60 to 65dB on average – so I don’t know how much of that I have to factor into this whole thing. Also when taking measurements in REW, I bump the MV up so that the measurement SPL level is around 40-50dB above the noise floor. I just don’t want to build something only to have it explode if I try to take a measurement.
Ok, so I think I’m going to need a bit of help understanding some of the settings in WinISD. I’m looking at an SB Audience ROSSO 18SW800 driver in this example. I’ve entered all of the necessary driver information, and it’s spit out whatever WinISD spits out. I will list my questions below:
1. Under Signal > System input power, what value do I enter? It’s an 8ohm nominal driver, and if I, say, use a Crown XLS 1002, it gives a Bridged 8ohm rating of 700W. I’ve also seen in a few videos that they enter 80% of the total value of the amp they are planning to use… if my idiot brain is understanding this right, would be 560W? Now, what I see when I select SPL, the graph starts rolling off, and when it hits about 45Hz, there’s a kink in the slope and the roll off isn’t as gradual. I’ve attached a photo of what that looks like on my end. When I, say, lower that to 400W, the kink goes away. The overall SPL drops a bit before it starts to roll off, but only by about 1-2dB.
2. You can see in the attached photo that by the time it hits 20Hz, the SPL has dropped by a fair amount. I understand that none of this accounts for room gain, etc… but I would like it to be higher at the 20Hz point. My understanding is that I would probably need to add a Linkwitz transform filter to achieve that, but I haven’t the foggiest clue what numbers to enter in order to create the “correct” filter. For example, if I am aiming to hit an F3 of around 16Hz, what do I enter for the f0, fp, Q0 and Qp values? In my example, WinISD seems to automatically set the f0 to the Fsc value under the Box tab and the Q0 automatically gets set to 0.449. Fp and Qp were set by default to 20Hz and 0.707. I’ve attached a photo to show what that looks like on my end (you can see the kink more clearly – I left the system input power at 560W, and dropping it to 400 removes that kink). I have also attached a photo of what the cone excursion graph looks like with this filter applied.
3. Is all of this fruitless with the driver I am modelling in this example? Would a different driver be a better fit?
Thanks for that – I did go and watch Hexibase’s videos on T/S parameters. Because I have no previous history with this process, a lot of it admittedly went over my head, but it was all very informative.
I think what I am getting hung up on right now (and might need a nudge to get out of) is the box size. I still have to look at a few other drivers to see what that looks like in WinISD, but the one driver I played around with the one driver came back with a volume of 0.905 cubic feet (around 25.6 liters). What my idiot brain is struggling with is how does that work? Clearly I’m not building a box that isn’t even be deep enough to fit the depth of the driver, so how do you determine the actual size of box to build when the math gives you values that are so small?
The other thing I’m going to need to understand is what filters (i.e. Linkwitz transform or 2nd order HP, etc…) do you apply in WinISD to get more out of the driver. I know the models don’t account for room gain, but again… no experience to know what to do at this time.
Once again, thanks for the input. I do subscribe to Toid’s YT, so I can go watch some of those videos shortly.
My mentioning of the cabinet volumes was more to verify that I was doing things “correctly.” I understand that there are things like bracing, and damping materials that would effectively increase the box size in order to hit the numbers (particularly those drivers that yield a much smaller net volume). I was plunking the results into the online SubBox pro calculator (which does account for the woofer displacement, but not any of the bracing, etc…), and in some cases I would get boxes that were, say, 22″ x 22″ x 5″ or something silly. I’m not going to worry too much about that until I look at things in WinISD, and move on to the next steps.
I’ll check out some drivers that have lower Qts values – are there any particular drivers I should absolutely avoid (i.e. brand)? And should I be worrying about whether the driver has a neodymium or ferrite magnet?
And agreed – if there’s anyone else who can offer their advice/opinions, please chime in!
EDIT:
So I am watching an intro WinISD video from Toid and when entering the driver data, one of the specs he mentions – Cms – isn’t even listed in the driver specs (example the B&C 18DS100-4 driver). What do you do in that case?
EDIT 2:
I was able to find the Cms info from loudspeakerdatabase.com, so if I run into that issue again, I can always check there. I did download the actual file from that site, but it wasn’t showing up when I tried to import in WinISD. I think it’s just due to the way I’m running the software on my Mac. If I decide that I want to pursue this more beyond this project, then I might pony up the dough for a Parallels subscription, and run it through something more “official.”
I am a Mac user, but I was just able to get WinISD running via Wineskin, so I can start to tinker around with that. Which Toid video do you recommend I watch?
I’ve also seen that there’s a specific order you have to enter the information. Is that still true?
Cool, thank you! So that means my math was (more or less) correct? I did get 285 liters plus change, too, if I rounded the Z value up to 0.87 instead of leaving it with the 6 or so decimal points. If I went with the Ultimax II 18″, is there a way to reduce the size of the box a little bit without negatively affecting the results too much or… ? Like, raising the Qcab to 0.8 reduces it to a 6.31 cubic ft cabinet volume, but I don’t know what effect that has (yet).
And what about a driver like the SB Audience NERO-18SW1900D? When I did the same math, I get an Inside Cabinet volume of 2.82 cubic ft (70.75 cubic liters). That seems smaller than it should be (I got a Z value of 1.98 and box resonance of 57.23Hz)? Would you lower the Qcab value in that case (say, 0.6 or 0.5)?
Fs of that driver is 34Hz, Qts is 0.42 and Vas is 157.9 liters
And with regards to driver selection, is there a particular type of driver that is better-suited for a sealed cabinet vs. vented?
-
This reply was modified 1 year, 6 months ago by