It may be more ecological than running a farm of physical PCs, but the poster does not need a farm of physical PCs to run two copies of HornResp.
FWIW, what you described is exactly what I'm using now - an i7 box with 16GB of RAM that's running Windows 8 Pro as the host platform and using Hyper-V to host three guest OS - Windows 8 Enterprise (my "office PC"), Windows XP (for the stuff that does not like Windows 8) and Ubuntu 13.1. I don't use the guest OS however to each run different copies of HornResp 🙂.
HP filters
Going back to the talk of high pass to protect from over excursion, I have noticed that in many instances a BW12 will allow a little higher power, and sometimes as much as 2 dB higher power before hitting Xmax vs. running a BW4 electrical high pass. This seems to be true with all filter alignments..ie BW, BE and LR.
The actual XO frequency can not be lowered much at all, but the power can be raised.
.
This goes against my intuition, but when looking at the excursion plots on the filter wizard in HR I can see how the shape of the excursion plots allows this to happen.
Just an observation,
Dave
Going back to the talk of high pass to protect from over excursion, I have noticed that in many instances a BW12 will allow a little higher power, and sometimes as much as 2 dB higher power before hitting Xmax vs. running a BW4 electrical high pass. This seems to be true with all filter alignments..ie BW, BE and LR.
The actual XO frequency can not be lowered much at all, but the power can be raised.
.
This goes against my intuition, but when looking at the excursion plots on the filter wizard in HR I can see how the shape of the excursion plots allows this to happen.
Just an observation,
Dave
Why DIY?
Hi Y'all,
I second DHAA's Post #233, for me it's all about the learning, and I appreciate the generous efforts of so many people on these forums.
Regards,
Hi Y'all,
I second DHAA's Post #233, for me it's all about the learning, and I appreciate the generous efforts of so many people on these forums.
Regards,
Marc - I installed Ubuntu once but didn't do much with it. I will have to experiment with that once I get this tapped horn thing under control. I might have to start and "Ubuntu for Old Men" post though. Both of my son's are computer professionals, so for an old man I am not a total Luddite. Thanks for the tip.
Hey Brian, I think i may actually qualify as having a PC Farm (or a PC burial ground, depending on how you look at it). In my house I have two i7's, one i3, MacBook Pro, MacBook, two pentiums, two atoms, five 27" monitors, one 22" monitor, and more. All have been hacked to run OSX. My sons dump their "old" computers off, which are often only a year old. The grand kids spend a lot of time here, so it's nice for them to be able to to do homework and stuff without having to wait line to use the computer.
Question of the Day - High Pass Filters for Subwoofers
O.K., I actually wasn't going to have a "Question of the Day" today as, I need to concentrate and make some progress on my horn fold. But Dave brought up a good question here. I have just started playing around with the HR Filter Wizard, and am trying to make sense of what I am seeing. I fully understand the difference between Butterworth and LR filters, the purpose of high pass filters, and all the basics - so I am up to speed there. But how about a discussion on the what/why/where of the types of filters you use? Thanks.
O.K., I actually wasn't going to have a "Question of the Day" today as, I need to concentrate and make some progress on my horn fold. But Dave brought up a good question here. I have just started playing around with the HR Filter Wizard, and am trying to make sense of what I am seeing. I fully understand the difference between Butterworth and LR filters, the purpose of high pass filters, and all the basics - so I am up to speed there. But how about a discussion on the what/why/where of the types of filters you use? Thanks.
I usually only use active 4th order BW. If you don't have standalone dsp this is probably what you are going to be stuck with in real life.
Technically it really doesn't matter what type or slope of filters you use to sim, as long as you understand that you SHOULD sim with filters that you will actually be able to apply in real life, especially if you are going to build the sub and expect it to perform as predicted by the sim. All the types of filter will have their own differences and you can use the one that complements your design best in the sim, but you have to be able to provide that filter in real life too. A lot of newer amps are coming out of the box with full dsp capabilities so that makes everything a lot easier, so if you have one of those just pick the filter type and slope that works best with your design.
Oliver, I have a quick question for you. The above quotes comes from this post #127 http://www.diyaudio.com/forums/subwoofers/248026-hornresp-brainiacs-help-old-man-13.html#post3763469 which included a drawing of a horn flare.
I understand you use a spreadsheet to lay our your horn, but where exactly does this script command come into play. I don't see any way to do that in HR. So does the script command happen within the spreadsheet, which then is imported into AutoCAD?
Spreadsheets are not my strongest point, but maybe I should figure that out too, it seems like it offers a high lever of perfection and saves a lot on mind numbing number crunching. Thanks.
Posts #127/248
Hi DHAA,
Oh ja, just a quick question 🙂 I've up since 5AM, and am still way behind, oh well:
The script is an ASCII file (text file) w/ commands that can be imported into AutoCAD, and is then automatically executed.
After you get your horn data from Hornresp into a spreadsheet you manipulate the data until you have only two columns left over, the first one is the length column, and the second one represents the duct heights. For a 3D representation you'll need to add a thrid column for Z. This gives you the XY(Z) coordinates for each data point.
Then you export that as a CSV (comma separated values) file. You can open this file in a text editor (Notepad). There you add the respective commands, e.g.: for AutoCAD it is very easy: at the top you enter the command "pline" in its own line, and at the bottom of the file, after the last set of values you hit enter twice (this is for AutoCAD). You can then save the file, and rename it, e.g. from: horn_data.txt into horn_data.scr (or save it as .scr). This file can be imported into AutoCAD, and will automatically draw a pline (AutoCAD for polyline) representing the duct height information. If you used 1/2 duct height in you spreadsheet, you can now mirror the duct height line around a center line (which you draw from 0,0 to total duct length), and you have your horn flare as a drawing. In this drawing you can pick any position you want, and use the CAD dimensioning utility to tell you how high the duct is at that distance.
Take a look at SketchUp, it has to have some form of CSV import facility. Maybe one of the SketchUp experts can help here. It will probably need three (XYZ) values.
Thanks for referencing the old post, had all but forgotten about that one. As diyaudio will not allow the uploading of a .scr file I had to change it to .txt.
Regards,
Hi DHAA,
Oh ja, just a quick question 🙂 I've up since 5AM, and am still way behind, oh well:
The script is an ASCII file (text file) w/ commands that can be imported into AutoCAD, and is then automatically executed.
After you get your horn data from Hornresp into a spreadsheet you manipulate the data until you have only two columns left over, the first one is the length column, and the second one represents the duct heights. For a 3D representation you'll need to add a thrid column for Z. This gives you the XY(Z) coordinates for each data point.
Then you export that as a CSV (comma separated values) file. You can open this file in a text editor (Notepad). There you add the respective commands, e.g.: for AutoCAD it is very easy: at the top you enter the command "pline" in its own line, and at the bottom of the file, after the last set of values you hit enter twice (this is for AutoCAD). You can then save the file, and rename it, e.g. from: horn_data.txt into horn_data.scr (or save it as .scr). This file can be imported into AutoCAD, and will automatically draw a pline (AutoCAD for polyline) representing the duct height information. If you used 1/2 duct height in you spreadsheet, you can now mirror the duct height line around a center line (which you draw from 0,0 to total duct length), and you have your horn flare as a drawing. In this drawing you can pick any position you want, and use the CAD dimensioning utility to tell you how high the duct is at that distance.
Take a look at SketchUp, it has to have some form of CSV import facility. Maybe one of the SketchUp experts can help here. It will probably need three (XYZ) values.
Thanks for referencing the old post, had all but forgotten about that one. As diyaudio will not allow the uploading of a .scr file I had to change it to .txt.
Regards,
Attachments
Last edited:
Question of the Day - Close Enough for Rock n' Roll?
I am still struggling to get horn model into an actual cabinet, but that brings up a very important question.
HOW PERFECT DOES IT NEED TO BE?
I am trying to do this old school style without a spread sheet, and if I can get within 10% variance overall, is that close enough for a prototype? Should I strive for 5% variance, or even less?
I have trashed quite a few attempts so far, and I am slowly getting closer to where it needs to be, but at what point should I just accept it as being good enough?
I am still struggling to get horn model into an actual cabinet, but that brings up a very important question.
HOW PERFECT DOES IT NEED TO BE?
I am trying to do this old school style without a spread sheet, and if I can get within 10% variance overall, is that close enough for a prototype? Should I strive for 5% variance, or even less?
I have trashed quite a few attempts so far, and I am slowly getting closer to where it needs to be, but at what point should I just accept it as being good enough?
Last edited:
Only testing will 'tell the tale', though for a two octave pass-band [max] POC, what mainly matters is that there's no major impedance mismatch through the bend to hamper 1/4 WL pipe action otherwise it will be hard to figure out what else may need changing since it's become two 1/2 WL resonators coupled together rather than a single 1/4WL one, unless one of course intentionally designs it that way. 😉
GM
GM
Last edited:
How perfect does it need to be? That's up to you. For a simple single fold tapped horn it should be possible to get it absolutely perfect.
The horn folding link I provided near the beginning shows a fairly complex tapped horn design by Littlemike. I reverse engineered the plans back to simulation inputs, and IIRC we ended up with almost exactly the same Hornresp inputs, no more than 2 percent different than each other in the worst case and much much closer to that through most of the flare.
Unfortunately, this level of accuracy is not common and people are not even putting their intermediate checkpoints (S3, S4 if applicable) in the right spots, as tb46 pointed out (linked to) in post 185 and I mentioned in the horn fold link.
How much little errors matter depends on a lot of things. If you want to know how much it will matter you can incorporate your folding errors into the Hornresp simulation and see what happens to your graphs. Or you can just work on the fold until it's right.
I prefer to get the fold as accurate as possible.
I am no Sketchup expert but I believe the first step in Soho54's horn folding with Sketchup tutorial is getting the flare into Sketchup, so there is some type of import feature. (I don't use that method so I don't know much about the import feature but I know it's there.)
The problem I see is that unless people have folded horns in a different way (like on paper) folding in Sketchup seems to be an almost insurmountable task for some people, for some reason.
I'm not sure why that is, folding is a simple job even without a computer. It takes a lot of time and work, but the work is simple. Lots and lots of simple steps and iterations.
Last edited:
Thank your Oliver for taking the time to explain that. You folding instructions from previous posts have been very helpful, that is primarily what I am relying on to get this thing done. I am going to try my preliminary fold posted later today to give you guys something to laugh about.
I just tried to import your attached .txt file into SketchUp - No Go, it seem to only allow importing "image" files. I will take a look at my book tonight to see if it shows any way to do it. Maybe I would have to upgrade to SketchUp Pro?
Thanks though, once I get thing thing darn fold thing happening I will revisit this spreadsheet idea of your further. I like playing around with HR and SketchUp so much, if a spreadsheet is the most efficient tool to combine those two pieces of software, I guess that is what I need to do.
Ignore me the rest of the day if you busy. I am just the opposite. The company I work for lost a major client recently, and for the past few months things have been getting progressively slower. I am able to spend a few hours every day here at work doing HR/SketchUp/tapped horn research, so I am not complaining, but I guess I should be starting to worry about it. Thanks.
Last edited:
just a guy, I will have to check that out again over the weekend. It seems I have to keep going back to things again as I learn more - eventually beginning to fully comprehend what they are talking about.
Well, that is definitely true for me. What my major hinderance is seems to be I am learning SketchUp as I am doing this. I think if I had already developed a working knowledge of SketchUp things would be going a lot smoother. But until I memorize all the steps involved and all the short-cuts, it is very frustrating.
I had experimented with folding using graph paper, but I think I was making my models too small. Do you model the horn on a larger scale, cut it up and re-arrange it, then re-draw it again? If this is the approach you use, what scale do you do it in? Thanks.
Good Enough?
Shouty, aren't we 🙂?
No built cabinet will ever measure exactly like a simulation, period, full stop.
You need to decide what is perfect, your idea of perfection is different from mine, which ten years ago was closer to yours, until I actually heard low bass from a car an octave below what my PA was producing...
Common practice is to EQ and cross over a PA with a huge 10-15 dB (or more) "haystack" below 100 Hz, which I don't go for, but is "perfect" for those that do.
Even though the haystack is common, the shape of the haystack has wildly differing preferences.
Around the throat, a 5 or 10% variance may have more effect than at the mouth, simpler fold patterns reduce that problem.
At the mouth, some rather large differences can be tolerated, and it would be up to you to decide which is "perfect".
Below you can see the result of changing the Keystone mouth area from it's "normal" (determined after dozens of different area and shape tests) 25" height to 15" and 8", over a 100% change in area.
Electro voice used to call the reduction of port area a "step down" mode, extending LF at the expense of output, the TH analogy is the same.
If you look at 30 Hz, the 8" step down is +7 dB over normal, which sounds nearly twice as loud, and would require two normal cabinets with twice the power to achieve. However, with the low frequency extension there is a penalty in all the rest of the bass range, the compromise not worth it for most genres.
But each of the three curves might be "perfect" for a person's particular desires. And with four T-nuts and a plate, each could be employed when desired.
"Rubenesque", or "fat"- all depends on taste .
Art
WHAT IS PERFECT?
Shouty, aren't we 🙂?
No built cabinet will ever measure exactly like a simulation, period, full stop.
You need to decide what is perfect, your idea of perfection is different from mine, which ten years ago was closer to yours, until I actually heard low bass from a car an octave below what my PA was producing...
Common practice is to EQ and cross over a PA with a huge 10-15 dB (or more) "haystack" below 100 Hz, which I don't go for, but is "perfect" for those that do.
Even though the haystack is common, the shape of the haystack has wildly differing preferences.
Around the throat, a 5 or 10% variance may have more effect than at the mouth, simpler fold patterns reduce that problem.
At the mouth, some rather large differences can be tolerated, and it would be up to you to decide which is "perfect".
Below you can see the result of changing the Keystone mouth area from it's "normal" (determined after dozens of different area and shape tests) 25" height to 15" and 8", over a 100% change in area.
Electro voice used to call the reduction of port area a "step down" mode, extending LF at the expense of output, the TH analogy is the same.
If you look at 30 Hz, the 8" step down is +7 dB over normal, which sounds nearly twice as loud, and would require two normal cabinets with twice the power to achieve. However, with the low frequency extension there is a penalty in all the rest of the bass range, the compromise not worth it for most genres.
But each of the three curves might be "perfect" for a person's particular desires. And with four T-nuts and a plate, each could be employed when desired.
"Rubenesque", or "fat"- all depends on taste .
Art
Attachments
Last edited:
You're welcome! Neither, acoustic impedance mismatch that causes reflections back to the 'throat', i.e. sets up eigenmodes in the sections and why the line needs to always be expanding.
GM
GM
Thanks for the info Art. The interchangeable "plates" look interesting - thanks for sharing that graphic. That idea may prove very useful for my needs, as I will be building these for three different generations of people, and it seems each generation want's to go lower. But with the interchangeable "plates" we could use them in different combinations or combine them for some blow-out event. Thanks for the idea.
Last edited:
The TallSkinnyFlatFifteen Has Lost Her Girlish Figure
Here is an example of my fold for the TSF15 I have been slaving over. Although I made a valiant effort, I finally came to the realization that there was no way I was going to ram my horn into her slim 17" x 17" x 48" figure.
So, I fattened her up a bit (the interior width is about 2" wider) and things between her and I are going much better. So here is where I am at:
- From the end of the baffle UP, that dimension is almost perfect.
- From the end of the baffle DOWN, that dimension is almost perfect.
- From the end of the baffle to the right side of the cabinet interior, that dimension is a little bigger than it should be. So, If I just make the baffle a bit longer to fix that, it looks like this will work. The lengthening of the baffle will have very little effect on the previous UP/DOWN measurement, so I should still be good there.
Does anyone see any other problems here? I believe that I am incorporating the "Advanced Centerline Method" properly, aren't I? Am I finally doing this right? Thanks.
Here is an example of my fold for the TSF15 I have been slaving over. Although I made a valiant effort, I finally came to the realization that there was no way I was going to ram my horn into her slim 17" x 17" x 48" figure.
So, I fattened her up a bit (the interior width is about 2" wider) and things between her and I are going much better. So here is where I am at:
- From the end of the baffle UP, that dimension is almost perfect.
- From the end of the baffle DOWN, that dimension is almost perfect.
- From the end of the baffle to the right side of the cabinet interior, that dimension is a little bigger than it should be. So, If I just make the baffle a bit longer to fix that, it looks like this will work. The lengthening of the baffle will have very little effect on the previous UP/DOWN measurement, so I should still be good there.
Does anyone see any other problems here? I believe that I am incorporating the "Advanced Centerline Method" properly, aren't I? Am I finally doing this right? Thanks.
Attachments
- Status
- Not open for further replies.