Programs and Spreadsheets

Home
Note:  Don't hold me responsible if anything goes wrong with your speaker design or your computer!!!  The program shouldn't do anything to your computer as it is just a "plug and chug" mathematical program.  It's not doing anything to any computer settings (like changing resolutions, or taking control of other processes) that could interfere with other programs.  I would think that if it crashed (say by dividing by zero or something similar), it would just quit with a warning and leave the rest of the system intact.  Of course, we are talking about microsoft operating systems ;-)

Also, I am very virus conscious and spyware conscious and have kept my computers and network virus-free and trojan-free.  I keep my virus scanner up to date, but that's not to say the web server may get one, or that my scanner may miss one.  ALWAYS check downloads for viruses, even from friends - ESPECIALLY from friends!  And if my web page and programs help you design a subwoofer that shakes your house off  its foundation - well, sorry, I'm not responsible for that either, though it might be nice to mention this site as providing the necessary tools for destroying your foundation ;-)

Note 2:  Spreadsheets are in Excel 97 format.

My program:
Leo's Simple Speaker v. 1.0.6 (LeoSimpl.zip ~1.6 MB zipped) - Now version 1.0.6!!!  

New Install Program:
On some WinXP machines, the program wasn't working standalone because some needed files weren't properly registering in the registry.  I've now written an install program (written?  Yeah, right - it's an installation wizard program included with VB) that properly registers the OCX file.

Enjoy it!  If you find any major bugs, email me at the address on my HomePage.  (Hopefully, I won't get 10 MB of bug reports!!!)  Please don't send requests for program changes, I may not even finish this one - including the crossover sections.  Notice: the program is now in a zip file called LeoSimpl.zip.  Sorry, the tiny original few 100 kb program grew to over 1.6 MB because of all the extra installation files required beyond my simple program.  If I figure out how to write an install routine rather than letting Visual Basic do it, maybe I can make it smaller.

You will notice tabs for crossover data and crossover graphics - NO they don't work.  That's still on the agenda if I ever get back to it.  Hey, it's taken too long just to get the main part into VB, much less the crossover stuff!  And it's hard enough keeping up with my updates.

Another important note:  If the graph axes labels don't line up with the lines, check your font dpi settings on your display settings ("Control Panel/Settings" or "Right Click" your desktop and select "Properties," then select "Advanced").  Most computers are set to small fonts, or normal 96 dpi.  I designed the graphs using 96 dpi.  Actually the graph is irrelevant, it's the text labels that rely on dpi settings.  If your font settings are set to large, 120 dpi, the axis labels for frequency won't line up with their respective lines.  Again, I may have to change the way I draw graphs.  Basically, I draw the graph, and just label the graph with Print (and Tab) statements.  This technique relies on the font dpi setting.  My machine (and probably 80% of machines out there) uses 96 dpi.  Basically, the dpi setting determines how widely spaced the characters are.  If you vary this setting from 96 dpi, the graph size stays constant, but the axes labels stretch or compress across the screen and won't line up with their respective lines.  If you don't want to change your font dpi settings, the graph is still accurately drawn, it's just that the labels don't line up with their respective lines.  Just keep in mind that the graph starts at 20 Hz, and goes up by 10 Hz at each line until you reach 100 Hz.  Then the lines are 100 Hz apart.

Installation and Removal:
As mentioned, just download it, unzip it, and run setup.exe.  The program shortcut installs in your start menu (by itself, not in a separate folder).  To remove the program, just go into "Control Panel/Add Remove Programs" and select it for removal like any other program.

Basic Instructions and Tips:

Try to hit enter after every input.  Previously, the TAB key did not pick up all values entered, but I think I fixed it.  If you change any volumes (or tuning) on the Box Model tab, you have to click the <Calculate> button again.   Oh, the <Print> buttons for the graphs print the form rather than the picture box - Sorry, I know it wastes precious ink.  However, I did change the Box and Vent Tab print routine so that it prints the data out on paper rather than printing the form, so it wastes less ink than it used to waste.  When you click <Save Image>, if you named the driver, it will save that image as "driver name"Frequency.bmp or "driver name"Limits.bmp.  If you don't enter a driver name, it will substitute "FrequencyResponse" as the driver name.

If you're wondering why I didn't start with the cursor in the "Driver Name" textbox , it's because I often find myself just wanting to see how a group of drivers model, and it's real quick to enter Fs, Vas (and units) and Qts immediately, then hit <Enter> (if <Calculate> is highlighted) or click <Calculate>.  You should be able to immediately hit <Enter> again, since <Frequency Response> should be highlighted, and you will see your graphs.  Basically, five "Enters" and you have a graph - Pretty quick, huh?!  Since I often want to change box size or tuning, I put the <Calculate> button on the Box Model tab, too.

For Newbies, to calculate and model boxes for a driver, you need three parameters at a minimum:  Fs, Vas, and Qts.  Then you can click <Calculate> and get recommended box sizes and tunings.  If you want to see Maximum SPL graphs, you need:  Power Handling of the Speaker (for thermal limits), and Vd.  If you don't know Vd, you need Sd and Xmax.  If you don't know Sd, you can input the driver's diameter and Xmax.

Sealed boxes should model fine for any Qts.  Remember, you can't input a Qtc that's less than Qts.  The program adds 0.1 to Qts to get Qtc if the Qts is above 0.707 (The program's normal Qtc default) which will give you a big box.  Don't fret if it's a huge box, just change the volume to something you can tolerate, or shoot for a higher Qtc and re-calculate.  Vented boxes model best when Qts is between about 0.2 and 0.8 (and using Qloss = 7).  I used some curve fitting programs to get very accurate formulas for this range of Qts (and Qloss = 7) using The Loudspeaker Design Cookbook and Bullock on Boxes.  Beyond that, the routine kicks out to a traditional power function vented box formula which may not be the best or flattest model.  At least it doesn't give a negative volume which is what my new formula would calculate if Qts gets very low.  But, again, don't fret, just input various box volumes and tunings and click <Calculate> and then click <Frequency Response> to see the graph.  Keep changing values until you find a response that suits you.  Note - don't just click the graph tab, you have to click the <Frequency Response> or the <Displacement Limited Output> button to initially graph the response.  After that, if you don't change any inputs, you can click the tab to see the graphs.  Actually, there is another way to make a graph.  If you've already clicked the <Calculate> button and then a tab for a graph, just hit the save button, and it will draw the graph.  That's just a side affect of how I save the image.  Of course this does save an image, but it will be blank if the page started out blank.  ;-)

On the Box and Vent page you can input extra volume for bracing, etc... as well as add a predetermined percentage if desired.  For sealed boxes, you may want to take into account stuffing by making this percentage, say 20% or 30% less (in other words, make this box 80% or 70%).  With vented boxes, you may want to overdesign by 10% or 20% so input 110% or 120%.  I thought about making a separate percentage volume change box for each box type (sealed and vented), but figure you're already interested in one or the other at this point, so I didn't.  I may later.  You can also input various port diameters.  Hit return after inputting each of these values, and the values change instantly.  One other thing I ought to explain, the Displacement Limit Graph isn't based on overall excursion limit formulas per say. What I used was the maximum displacement limited output formula in the LDC which is truly only valid above F3. This formula uses a K constant that for sealed boxes, depends on Qtc, so that it basically changes for different box volumes as you would expect (bigger box, lower Qtc, less power handling). Try different volumes and the sealed box curve changes. However, for vented boxes, it just uses a constant value of 3 for K. Therefore, the vented graph calculations don't change as you would expect due to different box volumes and tunings. (I still haven't found a reasonable formula as of version 1.04.)  Granted, these formulas in the LDC are only truly valid at F3 and above. What I have done is to use the formulas, but varied the "F3" across the graph from 20 Hz to 500 Hz - a somewhat bastardized process, and then I flatline it at maximum power. The upshot of all this? Comparing the sealed box displacement curves is probably pretty reasonable to get a relative idea of power handling, but the vented curves should be taken with a grain of salt. :-)  I would think that the "true" vented box power curve would follow the sealed box for most of the way as you go down in frequency, but would drop like a rock below Fb.  Don't be mislead and think that the vented box always has more power handling.  It doesn't.  As a rule, vented boxes have similar power handling down to the sealed box F3.  Vented box power handling is probably better than (or equal to) the sealed box from that point down to the vented box F3 at which point the vented box will have much better power handling.  Finally, below the vented box F3 (or more accurately, the tuning - Fb), the sealed box will have better power handling since the woofer is unloaded below this point in the vented box.

Besides that, the LDC says you can use Xmax + 15% to calculate Vd, but I just used Xmax (no extra 15%), so my values should already be somewhat conservative. Basically, I would just use the curves to compare driver outputs, but not read into them that they're at the absolute maximum output. You still have a little headroom.  Don't go around saying, "Leo says this is the maximum, blah, blah, blah...," or that, "Vented boxes handle the most power."  You just might be wrong.  ;-)

This program doesn't cost anything, it's not even shareware.  Well, on second thought, let's call it "Lunch-ware."  If you're ever in town (say, seeing McJerry from Eminence or something) stop by and buy me lunch.  McJerry knows where I live.  Or perhaps we may meet at a speaker building meet, say in Dayton or Lima or where ever.  Besides that, we can chat about speaker building.  :-)

Have fun, and I hope you find it useful!!!

Bug Fixes for version 1.06:
  1. First major fix was to make an install routine to properly register the TAB ocx file as well as the install data.  When trying to use the standalone version of the program and the OCX file, machines that didn't have Visual Basic or the OCX file installed may or may not run the program properly.  (It always worked on my machine because I had Visual Basic of course.)  Corrected.
  2. I realized there is a possibility that axis labels might not line up with their respective axes on graphs.  See "Another Important Note" above.  Corrected  Oops, not corrected, I should say that now you're aware of it.
  3. Printouts from the "Box and Vent" tab showed the last ratio label as "2.60 : 1.60 : 1" instead of "1.59 : 1.26 : 1".  The actual dimensions under height, width, and Depth were correct, just the label showing the ratio was wrong.  Corrected.
  4. On the Frequency Response Graph, "Cu m" units printed a closed parenthesis ")" after the "m".  Corrected.
Bug Fixes for version 1.05:
  1. I use a "while" routine when doing some calculations, and certain combinations of low Fs and Qts=0.37, would cause it to "resonate" or bounce between two values instead of closing in on a difference approaching "0."  That programming routine was in all program versions including my old DOS GWBasic version.  Apparently it didn't affect the old DOS version, though.  It probably has something to do with roundoff of double data types in the two different programming languages.  I've inserted a catch for when the routine bounces indefinitely like this, and rerun the routine with different start points.  Corrected.
  2. I added a "GotFocus" routine that now highlights the text as you tab through textboxes.  Previously, it just placed the cursor in the box but didn't highlight any text already there in order to overwrite it.  It was annoying having to double click the box to highlight the previously entered text instead of tabbing or hitting enter, and being ready to overwrite the existing data.  Corrected.
  3. I also updated the Compliance (VAS) units options so that when you change the units, You have to hit calculate again.  It also clears the text boxes on the "Box Model" tab "reminding" you to hit <Calculate> again.  Previously, a bug would change some calculations.  Corrected.

Bug Fixes for version 1.04:
  1. Some boxes, most notably Fs, Qms, Fb and the "Overvolume percentage" boxes, would tell you to input a number >0, then say "Are you sure about this number?" if you hit enter with the box left blank.  Corrected.
  2. I finally wrote a print routine for the "Box and Vent" tab instead of printing the form.  As an added bonus, if you choose metric units on the "Box Model" tab, it now prints out in metric units.  Note that the screen tab still shows english units, but the printout will show metric.  I didn't feel like writing much more on the tab to force it to show metric units.  Sorry non-USA people, I'm just lazy.  Be thankful I'm at least printing it out.  Corrected.
  3. I noticed that on the "Box Model" tab that if you changed units, the vented box routine wasn't changing appropriately.  I now fixed this, and have made switching units on the fly much easier by incorporating that unit option buttons in a "Change" routine.  Just pick different units on the "Box Model" tab (between sealed box volume and Qtc) to see your speaker modeled in either english or metric units.
  4. I also changed the graph "notes" at the bottom of the graphs.  If you change units on the "Box Model" tab, then hit <Frequency Response> You will see the pertinent data in those units.  I just noticed that the "Cu m" units prints a closed parenthesis ")" at the end.  Oh well, I'll fix that later.
  5. On the "Input" tab, I added a "keyboard tab" routine to include the units on Vas.  Now you will start out inputting Fs, then the program will tab to Vas, then it will tab to the Vas "units" where you can input "l", "f", or "m" for liters, cubic feet, and cubic meters respectively.  Of course it starts out on "liters."  Sorry Parts Express, I love you , but get with the program and use liters like 99.9% of other speaker companies instead of cubic feet.  :-)
  6. I know there were a few other changes, but silly me, I didn't keep too good of notes on what I fixed.  Oh well, it's better.  That's all I can say.  :-)  Eventually I will finish those crossover routines.
Bug Fixes for version 1.03:
  1. An error could occur if the <Save Image> button was clicked before calculating data.  I had various "catches" to keep you from going here, but there were still a few ways around it, so I made the <Print> button and the <Save Image> buttons unavailable until <Calculate> is clicked.  Corrected.
  2. If the <Save Image> button was pressed on the Displacement Limits Tab, then the <Save Image> button was pressed on the Frequency Response Tab, a bug changed the Qtc to 1.1 without the users knowledge and it would output a graph based on that.  If <Calculate> was pressed, the Frequency Response would return to normal when the <Frequency Response> button was clicked.  Corrected.
Bug Fixes for version 1.02:
  1. I Changed the color of the vented box to the appropriate blue on the displacement graph.  Corrected.
  2. Also, when the graph is less than -30 on the displacement graph, it now plots a red line.  (The curve continues to draw along the bottom edge of the chart, it just should have drawn in red so it didn't overwrite the chart outline which is drawn in red.)  Corrected.
  3. If you made Qtc =0 or "" on the Box Model tab, even if box volume was chosen, it would say, "input a number > 0" and switch you to the front tab.  Corrected - mostly :-)  (If you make Qtc=0 or " ", it adds 0.1 to the Qts, then calculates box volume.)
  4. On the Boxmodel tab, if you totally deleted the Vented Box Volume, and clicked <Calculate>, it would recalculate normally.  But if you changed the volume to 0, it calculated to a different volume.  It was jumping to the "old formula routine."  Corrected.
  5. Also, It didn't format the Box Volume text box when jumping routines to the old formula.    Corrected.
  6. I forced the X/Y coordinate readouts (dB, Hz)  to only reflect mouse position within the graph limits.  Corrected.
  7. Previously, if you hit <Save> or <Print> on the graphs, it kicked you back to the second tab and cleared the graphs.  Now it leaves you in the graph and clears it and redraws it.  Corrected.
  8. If you changed original inputs, the box sizes and tuning didn't reset on the next tab.  Now they do if Qts, Vas or Fs change.  Corrected.
  9. I put a copy of the Calculate button on the front page, and now have let it switch to the second tab.  I'm just trying to keep button pushes to a minimum.  Now you can enter Fs, Vas, Qts, and hit <Enter> and it moves you to the second tab, and it automatically highlights the <Frequency Response> button, so you just have to hit <Enter> to see the graph.  Corrected.
  10. It now prints the Peak Frequency and the Peak dB at the bottom of the Frequency graph.  Corrected.
  11. I turned off the ability to click into the mousemove X-Y boxes.  Corrected.
  12. I limit Fs to less than 500 Hz, so the program only calculates values where Fs is < 500 Hz since that is the limit of my graphs  Plus, it took longer to calculate as I have every frequency go through a "while" statement when calculating F3 for vented boxes.  Corrected.
  13. I added a question about input if it is way outside a "normal" range (say if Q>2).  It says, "Are you sure about this number, it is outside the normal range for this value".  Corrected.
  14. Oh, when the mousemove events are inside the Text boxes which show cursor location, I leave the pointer as a cross.  Corrected.
  15. I turned off the ability to change the cursor location boxes on the graphs.  It didn't hurt anything, but it implied you could type in your own values.  Corrected.
Bug Fixes for version 1.01:
  1. On the Box and Vent tab, the sealed box original volume and ratio volumes doesn't appear to change when changing to metric on the Box Model tab. If you expected your answer in cu ft, you were fine ;-) Vented box is fine.  Corrected.
  2. Also, on the same tab, the Additional Volume percentage box implied (in the tool tip) that it accounts for stuffing which should allow it to decrease box volume. Well, it didn't. That's what I get for trying to combine a box for additional stuffing based on extra percent and stuffing correction to allow a volume reduction due to stuffing. Re-written such that you can put in a value using 100% as meaning the same original volume and numbers like 80% mean 20% less volume due to stuffing.  Corrected.
  3. Frequency Graph - If you had designs which caused the graph value around 500 Hz (normally fairly flat) to move up or down from the 0 dB line, the text below the graph will float up and down, and sometimes overlapped the graph. The program used that last point graphed to determine where to start printing text.  Corrected.
  4. Changed a routine which added 0.01 to Qts to get Qtc so that it now adds 0.1 to Qts.  When using 0.01, it was giving huge box volumes, heck, that's just about an infinite baffle design!  Corrected.
  5. Added an icon to the "About" box.  Okay, extremely trivial, but it looks better.  ;-)
Well, all in all, not too bad so far. I would just hate to see a really major error pop up.

My Old MS-DOS Speaker Building Program (~64 kB zipped) - Recently updated - not as bad as "MacroSucks Gatesware" Corporation :-)  Written in QBasic.  Runs full screen now - to print graph, hit <cntrl>-<PrntScrn>, then open Paint and hit paste - may not print this way in WinXP or Win2000, or at least it won't in my WinXP).  Yes, it's old and I know there are still some bugs, but I don't even want to hear about them.  I don't intend to do any more to this little program, but it gives quick and fairly accurate designs.  Just pretend I work for Microsoft and that there will be a major patch out soon.  Yeah, right!!!!! :-).  The vented box routine starts going crazy if the Qts is not between about 0.15 and 0.6, and may give negative numbers for box volume (especially for a low Qts), but even in this case, you can still do a custom design using any volume you want and the curves and numbers should change to then be okay.  There is still a loop in the vent subroutine (or maybe it's the box volume?) that says value must be greater than zero even if it is greater than zero, but usually it works.  The box simulation using design tables is for use with the charts in the Loudspeaker Design Cookbook.

Miscellaneous Crossover Formulas (49 kB) - This spreadsheet contains a parallel trap routine (notch filter), sloped attenuation routine and tweeter attenuation routine.  Input data into the blue boxes to find the proper filter values from the Loudspeaker Design CookBook.  Note that the formula for the parallel trap routine is incorrect in older LDC's.  I informed their (a?) web site of this and they said the error still went out as late as the 6th edition.  I don't know if it's been corrected in the 7th edition.  At the bottom of page 121, 5th edition (p. 119, 4th ed.), the formula for Z should be:

Z=1/SQRT[1/R2+( 6.2832*f*C - 1/(6.2832*f*L))2 ]

(if I did my math correctly).  In the LDC, in the highlighted section, the fractions were "inverted" or 'flipped" to what is shown here.

Frequency Response plot (74 kB) - Finally, here I have a spreadsheet that plots the frequency response of  your meter readings so you don't have to plot them by hand.  It's already setup for 1/3 octave warbled sine waves (1/3 octave pink noise should work also, though the overlap between frequencies may make it not as accurate).  If you want 1/6 octave tones, just change the 1/3 to 1/6, but then you will have to lengthen the table (just copy the last lines on down), and readjust the limits of the graphs and move the correction factors up or down or add accordingly.  Not difficult, just tedious.  However, most people use 1/3 octave tones - pink noise or warbled sine waves.  Don't use steady sine waves (except perhaps at the lower frequency extremes say below 80 Hz), or your response curve will look like the Rocky Mountains.  There are two worksheets you can click through at the bottom.  Just input into the first table on the first page, all other tables will fill in accordingly.  Aren't spreadsheets great?!!!   The difference between the two pages (worksheets) is that the first page utilizes MY correction factors for the RatShack SPL meter, and the second page utilizes HTF's factors.

I hope they are clear enough.  Have fun.  I hope you enjoy these simple little things.

TOP OF PAGE

Last updated 6/23/07