I see I can pull any dayton FRD files from parts express.
Outside of that I have to search the internet for every FRD/ZMA file to find a file for a given driver.
Where are you guys finding them?
I have a program that traces the file but it seems everyone is using graphs whose indices are not linear. This leads me to believe tracing is useless because it is not taking into account the differences in linearity of the reading of the graph.
I did some searching for databases and saw a lot of posts of people leading others to certain databases, however, those database links do not work anymore.
I don't know if ya'll protect your own FRD/ZMA file libraries like fishing spots but if someone is willing to share with me I would greatly appreciate it.
Thanks!
Outside of that I have to search the internet for every FRD/ZMA file to find a file for a given driver.
Where are you guys finding them?
I have a program that traces the file but it seems everyone is using graphs whose indices are not linear. This leads me to believe tracing is useless because it is not taking into account the differences in linearity of the reading of the graph.
I did some searching for databases and saw a lot of posts of people leading others to certain databases, however, those database links do not work anymore.
I don't know if ya'll protect your own FRD/ZMA file libraries like fishing spots but if someone is willing to share with me I would greatly appreciate it.
Thanks!
You can use VituixCAD to trace frequency response and impedance response images. Otherwise you need to measure the driver yourself, or find measurements posted online by others.
IMHO, the only way to accurately share that data is to establish a standard measurement scheme. Similar to the infinite baffle measurement provided for drivers. For designers, the drivers are mounted in an enclosure prior to creating the frd and zma ddata.The shape and size varies which means the data will also.
What I measure on my enclosure is not helpful to you on your enclosure, and neither are stock graphs.
Only measurements taken on your enclosure taking into account baffle width and edge treatment are valid (and internal volume/porting etc. for mid/woofers)
Trying to design speakers without actual measurements is not the way it should be done.
Only measurements taken on your enclosure taking into account baffle width and edge treatment are valid (and internal volume/porting etc. for mid/woofers)
Trying to design speakers without actual measurements is not the way it should be done.
That's exactly how it is. It's not just that the cabinet and its baffle have a significant influence on the measurement results. If you don't measure yourself, you also lack all information about the phase relationships between the various drivers of a multi-way loudspeaker, making it impossible for a simulation program such as VituixCAD to correctly "predict" the acoustic behaviour around the crossover frequencies.
Kind regards
Michael
Translated with help from DeepL
I could drop quite a few ZMA, FRD or PIR files here. But they are useless to you, unless you build the enclosure I built.
Instead, I'd say: buy a Focusrite Solo, a Sonarworks or Dayton measuring mike with individual correction file, a mike stand and some resistors and zener diodes. Build your own measuring gear for about €/$ 200. Invest time in learning how to use it all.
Or use Boxsim for Visaton drivers / VCad or other apps for Dayton drivers who publish the files on their web site. Did you see those?
The graph tracing applications take care of that. You set which axis is logarithmic and which is linear. They generate a text listing for the graph that is being traced. FPGraphTracer is the one that is usually used. On an earlier app it was possible to set the number of points set. This newer app tends to produce huge long list of values that are very closely spaced. OTT really. It depends how large the image is when it's traced.
A google will bring up instructions on using this app that is on this forum.
Boxsim is a bit different but only with Visaton chassis. The data files that come with it cover all including things like acoustic phase and offset. There is some evidence that it simulates pretty accurately but it might be wise once a chassis has been selected to knock up the box, fit the speaker and measure everything yourself including acoustic phase and it's offset. Data can be loaded on the basis that it was measured in a box - as designed as they put it. If a different brand of speaker is used the chassis has to be measured to get the data that is needed and there still may be a problem. The boxsim data includes polar data on their chassis. It's not in the usual form shown on spl traces. 😉 I don't see that as much of a problem as the usual traces show sufficient info about where xover points need to be. Sudden large departures in the general shape of some of the off axis traces can be bad news. Some may well argue with that view. Boxsim of course is intended to sell Visaton chassis. They provide more data than others do. Dayton for instance provide the acoustic phase but not the offset. Useless unless that is determined via measurements. Some do not provide any of this sort of info at all.
I use fptrace, works well although of course, manufacturer's' data are measured in whatever way they choose (e.g. infinite baffle, chamber) and by various methods and not measured in your cabinet.
FWIW, Dayton's measurements and those from SB Acoustics, to name a couple, are supposedly accurate and some others, which shall remain nameless, aren't. I build other peoples' designs and have found it really interesting to compare makers' data and the project crossover parts with the actual measurements - the differences can be amazing.
I've posted this here before, but have a look at the result for Paul Carmody's Classix II speaker: the top graph was from Dayton's files for the DC160 and my fptrace for the tweeter from Vifa BC25 data sheet. Bottom is the real speaker.
However, I find that FRD/ZMA graphs made from fptrace or the manufacturer to be useful in comparing different drivers' raw performance.
Geoff
FWIW, Dayton's measurements and those from SB Acoustics, to name a couple, are supposedly accurate and some others, which shall remain nameless, aren't. I build other peoples' designs and have found it really interesting to compare makers' data and the project crossover parts with the actual measurements - the differences can be amazing.
I've posted this here before, but have a look at the result for Paul Carmody's Classix II speaker: the top graph was from Dayton's files for the DC160 and my fptrace for the tweeter from Vifa BC25 data sheet. Bottom is the real speaker.
However, I find that FRD/ZMA graphs made from fptrace or the manufacturer to be useful in comparing different drivers' raw performance.
Geoff
An accurate file is done by taking your own measurements in your own enclosure. This is required to get baffle step losses and diffraction into the curves so you can best identify peaks or dips that may cause anomalies that may make the speaker harsh or dull or cause a problem "blending" the drivers in the crossover.
An "approximation" can be done using someone elses FRD data (whether manufacturer or another build) - but you need to what the driver was mounted in for measurements, so this can be "subtracted" then you "add in" the baffle response of your intended target enclosure. Although this can be done without taking measurements, it isn't as good and can cause 2dB+ errors if you aren't methodical about doing the subtraction / addition steps (or do not know the source data conditions). The most common mistake of using Manufacture data is forgetting to apply baffle step losses, meaning people design without enough compensation leading to a speaker that is overly bright or lacking bass / midrange.
An "approximation" can be done using someone elses FRD data (whether manufacturer or another build) - but you need to what the driver was mounted in for measurements, so this can be "subtracted" then you "add in" the baffle response of your intended target enclosure. Although this can be done without taking measurements, it isn't as good and can cause 2dB+ errors if you aren't methodical about doing the subtraction / addition steps (or do not know the source data conditions). The most common mistake of using Manufacture data is forgetting to apply baffle step losses, meaning people design without enough compensation leading to a speaker that is overly bright or lacking bass / midrange.