, 10-28-2009 at 10:18 PM (6491 Views)
Now that you’ve read the tips…
I’ll be using files as I would name them, so you use whatever you want. It just makes things easy to identify.
Frequency Response (FR):
You either have to measure the FR and impedance (Z) curves, or trace/simulate them to be as accurate as possible. If you can find the manufacturers’ curves, or another DIYer’s curves; and they are reliably measured (Tang Band’s graphs leave a lot to be desired), you can proceed to tracing the raw-FR response. I have not used the new program SPL-Copy program, so I will utilize SPLtrace for this demonstration. I will not discuss the many ways to measure using a lot of $$ equipment this time, as this is the method I typically use most. FRD Consortium has most of these programs for download, but Google can find them for you too. Here is the site for the latest in Jeff Bagby's software:
MS-Paint is the easiest program to get the picture into SPLtrace.
-Open your drivers’ picture-file into Paint, and hit ‘Edit>Select All>Copy’ to get the image into the notepad.
-Open SPLtrace, ‘Import graph from Notepad’, and set your limits for SPL level, and frequency.
Be sure to use as much of the plot as possible, for a more accurate simulation. Tweeters can typically be traced down to about 250 Hz for maximum benefit, and if the woofer plot goes to 20 kHz, I will trace the entire thing.
When you ‘Start SPLtrace’ there will be a white line present on a gray field in the left. Line the gray field up with the lower limit for frequency, and the white line up with the first pixel of the FR curve at the lower freq limit. Click, click, click… following the curve as closely as possible. Try to anticipate where the gray-field edge of the next click will stop, and press the left mouse button. Repeat, repeat, repeat…
If you should meet the end of the trace limits, which sometimes I do, you can select ‘Trace>Options>Use Trace Limits’, and up your points value, etc to continue the process.
When you reach 20 kHz, got to ‘Trace>Stop Trace’ and ‘Trace>Save’ if it does not prompt you to Save. You can also go to ‘File>Save As’. Anyway you prefer, save the file as ‘(driver name)-Raw.frd’, and you are ready to go to the next step.
You can also tweak the equalization curves in the RM to simulate an FR curve instead of tracing it, or tweak your trace for better accuracy. Just hit ‘Transfer Modified to FRD Register’ for the tweaks to take effect on the curve in question.
I will typically measure the impedance and T/S and export the ‘.zma’ file from the WT2/3, since tracing them is not so accurate for me. Aside from this, you can simulate it in the Response Modeler 2.0. The area regarding the impedance in the RM is where we focus on this. You can import your traced data should go that route to modify it for the PCD simulation, or just enter the parameters for the driver. ‘Full T/S Parameters’ is for the value entry, and ‘Imported + MPE’ is for imported .zma files. If you have a plot visible to you for the driver in question, you can eyeball the curve and try to match it as closely as possible with the entered parameters. Once it looks like it should, save as ‘(driver-name)-Raw.zma’
Now- whether you traced, imported, or simulated it; you have the raw .zma file, and the .frd file from the previous procedures. Save these for future projects, or another attempt at the simulation at hand. This way you can start from scratch if needed.
The more accurate your data is, the better your simulation can be to the real-world.
Box design/altering the files:
Let’s go and set up the PCD, by altering the data through Unibox and the Response Modeler.
Use the same T/S values from the woofer in Unibox to get you box volume, a reasonable response plot, Xmax safety, and if vented- below the “OBS!!!” mach-factor warning. I usually optimize for the output being 100dB, as this is very loud for most people, and above that can really start to damage things and your hearing. Just vary the power spec in the upper left to get about 100dB. It usually is not very much power. (The govt. health recommendation is to wear hearing protection above 85dB.)
Save your Unibox file for your driver….and now we go to the RM again.
Response Modeler 2.0 (RM):
A lot of times when you reinitialize the RM, the past files are still present, and need to be removed and flattened. Hit ‘Clear Imported FRD Data’ and ‘Clear Imported ZMA Data’.
To get the altered files the way you want them, import both the .frd and .zma of the woofer into the RM. I do the woofer files first, and then the tweeter files, you’ll see why in a bit.
If the 4 following criteria are not zero’d out, please do so after importing:
-‘Clear EQ Register’
-‘Zero Out EQ Settings’
-‘Clear External EQ’
-‘Clear BDS Data’
Also- If you leave the RM open and decide to try another driver, you need to clear these out as before.
Now that the 2 files are imported, simulate the box you had in Unibox in the box-plot region. Even if you are using a dual-driver design, do them individually. This way when you import the files into PCD, you can use several different manners of design, and unless you have the response of the 2 drivers measured in unison in your design, you can’t use a single impedance file for 2 woofers.
Now, enter the T/S and Box Params, and select Sealed/Vented just above the ‘Box Response Model’, and simulate you Unibox box, or just use your simulated/traced/measured .zma to map your box as just prior. Once you get what you had, select ‘To Splice Box Plot To Response Click Here’, to relocate the box response onto the imported FR curve above. In ‘Box Plot Splicing’, adjust the level of the splice to match the FR curve, and select 100, 200 or 300 Hz for the best match. You don’t want a drop or step at the junction. It should be smooth. ‘Save Modified Result to FRD File’ as ‘(woofer-name)-RM-(project).frd’. The impedance will also be adapted to the Box Response completed, so you can also ‘Save Modified Result to ZMA File’, as ‘(woofer-name)-RM-(project).zma’. These will keep your files identifiable for future reference.
What you have just completed is the simulation of near-field and far-field measurements, and combined them for the simulation of in-room response, minus Diffraction and Baffle-Step losses. The reason these 2 have to be spliced together is you can’t get good bass FR measurements without doing it outside and with a long gated time period. The typical shorter timing gates used catty-corner in your room at 1 meter are to prevent reflections in the measurements, and are only good to about 400/500 Hz on a regular basis. The secondary close-mic’d measurement at about Ľ” from the driver is your near-field plot. The main difference between the sim and measurement at this junction is that the measurements would have already factor in the BSC/Diffraction issues.
Now scroll down to the last section in the Response Modeler to add in the box dimensions and edge treatments. Enter the box dimensions, edge radius, and the type of directivity to expect from the woofer being used. You can tweak position for the flattest Diffraction signature if you so choose, or leave it square in the middle and adapt to this compromise. It’s your choice. Click ‘Save Baffle Diffraction Curve to BDS Register…’ to redirect and adapt this information to the FR plot up above. What is visible after this change is the full-BSC/6dB FR plot. If you want full-comp, resave the .frd file to ‘(woofer-name)-RM6-(project).frd’, or ‘(woofer-name)-RM-full-(project).frd’. If you want less than full compensation, you can select ‘Inverted’, and remove a percentage of the baffle step in the case of boundary-proximity to the design’s intended position. Closer to a wall you get, the less BSC you require. Save accordingly as ‘(woofer-name)-RM(dB-comp)-(project).frd’. Now that the Woofer is done, clear its files and the 4 criteria listed above. Import the tweeter’s .frd file. Select your position on the previously entered baffle dimensions, and the directivity for the tweeter. ‘Save Baffle Diffraction Curve to BDS Register…’, and select the same compensation percentage as before, and resave as ‘(tweeter-name)-RM(dB-comp)-(project).frd’. You do not have to alter the tweeter’s impedance curve unless it is unchambered. This is the same for dome-midranges that are chambered/unchambered. For a cone-mid, go through the same procedure as the woofer, unless it is chambered. Use the separate volume intended for the Midrange in the simulation. Just be sure to treat all of the files for the same project the same way with respect to the intended box.
Now that you are finished altering all of the files, you have to ‘Auto Extract Phase From FRD File/From ZMA File’, to the *altered* RM files, not the RAW files. These buttons are to the right of the box-model graph. Resave the extracted results to the same file name: ‘(driver-name)-RM(dB-comp)-(project).frd/.zma’. These are the files you import into PCD! Now you are ready to import.
Without actually having the gear to do the measurements yourself, the simulations just completed will get you ready for the PCD. If you successfully traced, simulated, and altered the files to the best accuracy you can, you should have a fairly accurate simulation coming your way.
Passive Crossover Designer 6.2 (PCD):
Open PCD, name the drivers, Load/Load/Load the 4/6/8 graphs required for your design, and enter all your offsets for the drivers in question. Use the tweeter as your 0,0,0 reference, and match offsets of the drivers to the tweeter position. The tweeter is typically the listening axis. Remember Z is not a negative value in PCD6, however, it IS negative in PCD7. Enter the piston diameters for off-axis response, and toggle how many drivers of each you are using with their offsets as well. I covered this in the ‘tips blog’, with pictures for an explanation. Select Design Type and Load the wiring configuration, and finally SAVE all of this setup as a .csp session file for future reference as initial setup. This gives you a backfall in case you want to start over. Save each progressive session as a new simulation for the ‘project(number).csp’, in case one works better than another.
On to the woofer-
Select target, level, and slope you want to hit, and click ‘Initialize Textbook Values’, tweaking to match the target from there.
Repeat for tweeter, midrange.
You will become accustomed to the values associated with the right frequencies, and the textbook emulator won’t be of much use any longer. For notch filters, conjugates, and zobels, you can use the 6dB reference equations to get a relative idea of what value to use in that range, with its impedance. They are as follows:
C = 10^6*[1/ (2*pi*R*F)], where C is capacitance in uF, R is resistance, and F is Frequency.
L = 10^3*[R/ (2*pi*F)], where L is inductance in mH, and same as above.
Once you have met the targets, you can hit ‘sum’ in the graphs, and tweak to a better result.
So- there’s the setup! I hope that was thorough enough to understand…
Your #1 Source for Audio, Video & Speaker Building Components
Deal of the Day
View Our latest
Order our FREE 336 Page Full Color Catalog
Speaker Component Categories
Home Audio Speakers
Professional Audio & Guitar Speakers
Car Audio Speakers
Measurement & Design Tools
Subwoofer Plate Amplifiers
Full-Range Plate Amplifiers
Cabinet Hardware & Speaker
Subwoofer System Kits
Speaker Repair Parts