Because you can not be sure what chassis and crossover, build 40 years ago, you found in the garbage bin, you have to measure them. Impedance curve and response. Together these two will tell a lot about the state the chassis are in. If you get serious, measure distortion, two.
Do you have REW? It is free, so you don't have to buy it, easy, well documented, many tutorials and capable of more things you will ever need.
All you need are some old wires and a resistor for speaker measuring. The most expensive thing will be a microphone. In my world the entry level mikes are around 35US$, but they need an XLR phantom power interface for around 25$.
Even cheaper and simpler, is a Dayton Audio thingy called iMM-6 you plug it into your phone. It even is calibrated. Got one for 25$ new lately.
If you want to make a good product, you got to invest a little. Some learning (the hardest part), some time, a little money for tools. Without all of that you can erase "good" or "decent" from what you produce. Sad but true...
Do you have REW? It is free, so you don't have to buy it, easy, well documented, many tutorials and capable of more things you will ever need.
All you need are some old wires and a resistor for speaker measuring. The most expensive thing will be a microphone. In my world the entry level mikes are around 35US$, but they need an XLR phantom power interface for around 25$.
Even cheaper and simpler, is a Dayton Audio thingy called iMM-6 you plug it into your phone. It even is calibrated. Got one for 25$ new lately.
If you want to make a good product, you got to invest a little. Some learning (the hardest part), some time, a little money for tools. Without all of that you can erase "good" or "decent" from what you produce. Sad but true...
While everything stated in this thread is indeed true, the nature of the xover is the Q of the filter. When you can measure the nominal impedance of the drivers, and calculate the Q, you can approximate the type and slope of the transfer function. Using L and C values with the impedance attached calculates the Q of the filter. 0.5 Q will be LR, 0.707 will be Butterworth; and these are the most common. If you calculate Q that is twice or half these values, then the intended impedance attached should be half or double the impedance used to calculate it.
However, this is a big misleading rule of thumb, as drivers never measure ideally, and the textbook filters usually are not designed to be applied to real drivers.
However, this is a big misleading rule of thumb, as drivers never measure ideally, and the textbook filters usually are not designed to be applied to real drivers.
From what I can understand the Braun speaker is 8 ohm nominal and the crossover points are 700 and 5000Hz. It should be rather simple to reverse engineer based on that to confirm the nominal impedance of the mid and high (the low is obviously 8) as there seems to be no lpad just some small serial resistors. Nevertheless I don’t think the crossover will fit well to the other speaker drivers you mentioned even though they are probably 8 ohm as well. Sensitivity and breakup resonances will differ. Phase possibly. If your drivers truly is 4 ohm then it just won’t fit.
Yes but it's not all that matters. In fact, if I wanted 8 ohms but all I could find was 4 ohms then I could still make it work and be happy.
Even this isn't enough. I would not just double my capacitance and halve my resistance etc for 4 ohms.. In fact I would also measure the unique features such as the curvature of the impedance of the new tweeter, it's sensitivity, it's response shape, it's dispersion etc.
Goodbye
And please don’t forget to bring your impedance plots to the shop when you go for speaker’s drivers shopping or, at least, mail them to the seller if you’d want to buy online.
You have measured the DC resistance of each of the voice coils. Keep in mind that the woofer, midrange, and tweeter voice coils will also have, in simple terms, various values of inductance. This will lead to a rising impedance at high frequencies. Hence, the measured 4-ohm resistance is only applicable at 0Hz, and becomes very inaccurate at higher frequencies and around each driver's resonance frequency.
An example driver impedance curve is shown below. Notice that there is a phase component of the impedance, and this must be taken into account for any crossover design. The magnitude and phase of the impedance curve will interact with the components in the crossover network, so textbook values based on a constant resistance will definitely not apply.
For your project to have some chance of success, it is necessary to measure each of the drivers that you wish to use. It will be required to determine the magnitude and phase response of each driver's impedance as a function of frequency. One available option is to use a computer sound-card-based approach using REW: Impedance Measurement. Those measurements will be an intrinsic part of any crossover design/simulation.
Last edited:
You should be using Dayton DATS or a McGyvered up impedance measurement from Room EQ Wizard to measure the driver impedances.
LOL Very cool history, we have the same guy in Lyon, perhaps he went retirement but he has a padawan. I like Artist like that. I started to build my first speaker with them. It is speaker craftsmanship in an other period . Note a brand like SEAS works with him building some custom speaker for him like a 8" Coaxial or a 7" Excel Coaxial. He sells a lot of SEAS drivers and kits. He has the visit of SEAS people, and the shop looks old, small. It is an other network of sellers for the manufactories, others partners 😉
Sorry @presscot return to the subject. It is a reverse engineering problem. In this modern world we have tools to do thing well and easy.
1. A classical design process is : measure -> design -> build according to goal, i use this strategy and it is the best.
2. But I have an other strategy because I don't want to buy the drivers and I want to know how it could be used.
The process is :
Model -> design -> emulate -> build -> measure
Here you should draw the crossover in a tool, model your drivers and see the overall response 🙂
You can use the tools you want. I recommend VituixCad2 if you don't have any tools, I am sure you have a computer 🙂
I builded some speakers like that, and it works pretty well, if the modelisation is precise with all good parameters.
You never know how it sounds till the building and listening.
Enjoy !