Expanding a little on the above, with the MTM configuration, the design axis of the loudspeaker is usually taken to be the tweeter axis. If you take your response measurements at 1 metre, then you will be quite a bit off-axis for the two woofers, but exactly on-axis for the tweeter. For a more typical listening distance on the tweeter axis, say 3 metres, the woofer measurement will be less off-axis than when the measurement is done at a 1-metre distance. It is still necessary to include the effects of the woofers being off-axis in the crossover design in order to get the most accurate summation from the simulation.
For the layout of the drivers on the enclosure baffle, the off-axis nature of the woofer radiation is complicated further by the fact that the tweeter is offset to one side relative to the woofers. Hence, the woofers are off-axis both horizontally and vertically relative to the tweeter axis. In this instance, it may be worthwhile choosing the design axis to be centrally located between the two woofers. The reason for that is that the tweeter radiation pattern through the crossover region is still very broad, while that of the woofers is much less so.
Also, it is probably worth keeping in mind that the measured response of the upper woofer alone is likely to be slightly different from that of the lower woofer alone. That's due to slightly different baffle diffraction effects, as the top woofer is close to three edges, while the bottom woofer is close to only two edges. However, these types of effects might be quite small, which is something that the measurements should show.
I'm not sure if there is much benefit to measuring the individual responses of the two woofers. In the system, they are being summed together at the listening point, so a single measurement would do. If more detailed measurements are proposed, then maybe some vertical up/down and horizontal left/right off-axis measurements would be useful, using 5-degree or 10-degree intervals. However, that's a lot of extra data to have to deal with, both in the measuring and the subsequent simulations. It's probably best to get the design as good as possible for the chosen design axis and listening distance.
That's quite correct. However, the change in angles needs to be included in the measured data somehow. If you measure at a more representative listening distance than a nominal 1 metre, then the off-axis radiation patterns of the drivers are already included in the measurements. That will ensure that the results of the simulations computed at the desired listening point will be as accurate as possible.
I have another Newbie question.
Is there a step by step procedure to follow using Xsim (or any other Xover design software) that prevents/reduces having to go back and re-adjust settings that were working well?
e.g.
1. Levels of each driver
2. Select crossover point.
3. Slopes
4. Phase alignment
5. Z offset
6. Notch filters
Is there a step by step procedure to follow using Xsim (or any other Xover design software) that prevents/reduces having to go back and re-adjust settings that were working well?
e.g.
1. Levels of each driver
2. Select crossover point.
3. Slopes
4. Phase alignment
5. Z offset
6. Notch filters
Selecting your crossover point is a part of the acoustic design which you should know before you come to create an electrical network. It's unlikely you will make a significant change to that based on what you do in Xsim.
It can help to have a target slope to follow, this helps things to fit.
It can help to have a target slope to follow, this helps things to fit.
There will be a lot of this, and there are things you can do but can you give examples because there are so many things this could mean...
Thanks Allen.
I'm in the middle of a 2 way, TL cabinet construction (first build in 20 years), so at the moment I'm just playing with Xsim, using the factory frd and zma of my two drivers trying to familiarize myself with how to use it.
I've also watched quite a few you tube videos, but they seem to skim over important bits, or only cover a fraction of what's required in a properly designed crossover.
I guess what I'm really after is a full step by step procedure which so far, I haven't been able to find.
As well as struggling with a new program, I'm struggling to remember the basics of good crossover design.
I've only been playing with the factory frd and zma files.
I intend using frd and zma files that I'll measured once mounted in the enclosure.
It's a FAST system (similar to xrk971's design)
The driver in the TL is a Dayton RS225-8, which appears to start breaking up around 1700Hz.
The upper freq driver is a Vifa TC9FD18-08 who's factory frd is very good, right down to around 150Hz (although not as sensitive as the RS225). The Vifa will be in a sealed box.
Only based on the factory files, I was playing with a crossover around 1000Hz
Thanks for posting the *.dxo file
OK, the thing about a FAST is that it should be relatively easy to cross. You have a wide range of frequencies free of issues and you can use shallow slopes. Acoustically it's easier if you try for a lower cross. You may find a smaller driver can only handle so much bass, limiting how low you go. Also as you go lower you get closer to working with the smaller driver's impedance peak. This doesn't have to be a problem at all, but it can be a complication.
The next thing to do is consider the baffle. The easiest way is to put in a 6dB step by sight. You can use a baffle simulator to show you what it looks like, and you can even combine that curve with the factory curves.
OK, the thing about a FAST is that it should be relatively easy to cross. You have a wide range of frequencies free of issues and you can use shallow slopes. Acoustically it's easier if you try for a lower cross. You may find a smaller driver can only handle so much bass, limiting how low you go. Also as you go lower you get closer to working with the smaller driver's impedance peak. This doesn't have to be a problem at all, but it can be a complication.
The next thing to do is consider the baffle. The easiest way is to put in a 6dB step by sight. You can use a baffle simulator to show you what it looks like, and you can even combine that curve with the factory curves.
Last edited:
Thanks again Allen.
I was using FPtracer to get the Vifa files, and messed up the zma trace (fixed it now). That's why it wasn't in the *.dxo.
The Vifa's imp. peak is around 120Hz.
According to Hornresp, the twin peaks of the RS225 in the TL enclosure are around 16Hz (25ohms) and 44Hz (25ohms).
I'll have a play around with Tolvan's Edge in the next couple of days (another bl**dy program I have to learn how to use
).
If you want to go a little further you can simulate the baffle. (I don't think I managed to express in my earlier post that it is up to you whether you do this.)
Alternatively you can take measurements.
Member
Joined 2003
All filter design tools will utilize the information you provide. In the case of speakers, it's impedance response over frequency that is important, so provide the software with measured impedance of the piezo and measured frequency response and have at 'er.
Crossover design should be approached in the same manner as any speaker, however peizo driver is capacitive loads compared to most other speakers which are inductive loads.
Crossover design should be approached in the same manner as any speaker, however peizo driver is capacitive loads compared to most other speakers which are inductive loads.