Hi All!
I am building my first 3.5Way floorstanding speakers. I selected the drivers, designed the enclosure and built a test version, using scrap MDF that I had laying around my shop. The idea was to validate that it is a feasible project. Drivers are:
- Tweeter: Peerless XT25SC90-04
- Mid: HiVi Research DMB-A soft dome
- Mid Bass: Vifa P17WJ-00-08
- Woofers: Generic Aiyima chinese 8" woofers. I will be in the US by the end of the year and will get Dayton Audio DC200-8 8" Classic woofers. Box volume was already defined using Dayton's drivers.
The test box is built and I've just took a few measurements of each driver. I used REW in my macbook connected to a cheap car amplifier and a dayton audio umm-6 calibrated USB microphone, at 1m and aligned with the Mid dome (which is the height my ears will be when listening).
I also took measurements in my dedicated listening room, where I will put the speaker. In this case, the microphone was about 2,7 meters away and the speaker was toed out about 15 degrees.
Now comes the question. For each measurement, I generated a set of frd files, imported on VirtuixCad and designed a crossover. I expected the measurements to be different, but the differences that I found seem to be a little too much (do not mind the average SPL - i've used different volumes in each measurement).
This is the XO based (alpha draft version 😂) on the in-room measurement (do not mind the fact that it has only one woofer - I measured both woofers in parallel in a single sweep):
This is the XO (also alpha draft version 😂) based on the outdoor measurement:
I understand that reflections, room modes, etc would change the frequency response (that is noticeable specially below 200hz), but the huge phase difference that I am seeing - which is changing drastically crossover points and driver polarity - is bothering me a lot.
What could be causing this? Maybe the distance between the speaker and the microphone, which changes a lot the time arrival of each driver?
One last info: I am using in-room measurements because the speaker will be there for a long time. I will be using it in this room to listen to music only (for movies, the current setup that you see in the pictures will be used). So i am aiming for a passive XO that is tailor made for my room.
I would love to hear your insights.
Thanks!
I am building my first 3.5Way floorstanding speakers. I selected the drivers, designed the enclosure and built a test version, using scrap MDF that I had laying around my shop. The idea was to validate that it is a feasible project. Drivers are:
- Tweeter: Peerless XT25SC90-04
- Mid: HiVi Research DMB-A soft dome
- Mid Bass: Vifa P17WJ-00-08
- Woofers: Generic Aiyima chinese 8" woofers. I will be in the US by the end of the year and will get Dayton Audio DC200-8 8" Classic woofers. Box volume was already defined using Dayton's drivers.
The test box is built and I've just took a few measurements of each driver. I used REW in my macbook connected to a cheap car amplifier and a dayton audio umm-6 calibrated USB microphone, at 1m and aligned with the Mid dome (which is the height my ears will be when listening).
I also took measurements in my dedicated listening room, where I will put the speaker. In this case, the microphone was about 2,7 meters away and the speaker was toed out about 15 degrees.
Now comes the question. For each measurement, I generated a set of frd files, imported on VirtuixCad and designed a crossover. I expected the measurements to be different, but the differences that I found seem to be a little too much (do not mind the average SPL - i've used different volumes in each measurement).
This is the XO based (alpha draft version 😂) on the in-room measurement (do not mind the fact that it has only one woofer - I measured both woofers in parallel in a single sweep):
This is the XO (also alpha draft version 😂) based on the outdoor measurement:
I understand that reflections, room modes, etc would change the frequency response (that is noticeable specially below 200hz), but the huge phase difference that I am seeing - which is changing drastically crossover points and driver polarity - is bothering me a lot.
What could be causing this? Maybe the distance between the speaker and the microphone, which changes a lot the time arrival of each driver?
One last info: I am using in-room measurements because the speaker will be there for a long time. I will be using it in this room to listen to music only (for movies, the current setup that you see in the pictures will be used). So i am aiming for a passive XO that is tailor made for my room.
I would love to hear your insights.
Thanks!
Last edited:
The phase difference you talk about doesn't seem to pertain to the crossover. If it did I presume you'd have shown the phase panel. In that case the issue isn't phase but group delay. I'd advise you to keep working on the in room bass response (not in a crossover simulator). Consider a multi-sub setup.
Not sure what kind of help you're asking for here, so generally speaking if your room or speaker has acoustic anomalies then the fix generally has to be an acoustic one.. a redesign of your speaker to fit the room. You can try to compensate electronically if you wish but you'll have to work it out as you go. Perhaps work with an equaliser.
Hi Allen!
Maybe I jumped into conclusions too early. I've been looking to both the measurements and the phase doesn't seem to be that off between measurements. I also checked group delay information and it seems to be within sane values for both designs (< 2ms from 200hz up). I am not really concerned with low frequency differences, as LF behavior changes a lot depending on where the speaker is placed.
What is bothering me is what I had to do to get a reasonable resulting FR. Using the in-room measurement, the polarity of the mid high driver has to be inverted. However, when using the outdoors measurement, the polarity of the tweeter must be inverted instead of the mid high. It seems to be a difference too big for two measurements of the same box with the same drivers. Or is it?
Well, communication is a b*tch! 🤣 I love culture differences - Here in Brazil we are not used to be so straightforward in our questions, we do like to put a lot of info and go through all "ramifications" of the topic. Let me be as clear as I can be: Why am I getting these differences between designs (assuming my measurements are correct)?
It is my first "complex" design and also the first one that I am going trough all these measurements and design steps. As so, and that is why I built a test box, I want to know if it is a feasible project. If the drivers can be used together, if the driver positioning in the enclosure is adequate (aesthetics, a.k.a WAF, has a huge role here), etc....
I am adding REW measurements and VirtuixCAD files here if they are any help, but please, do not feel obligated to go through them. A small correction: The in-doors measurements I've used were taken on-axis, not off-axis as described in my post.
Hummmm, is that so? The way I see it, IMHO as well, it is a 3.5-way, as there is no high pass filter in the mid-bass circuit, so both will play as low as they go. At least that was what ChatGPT told me when I asked if my design should be considered a 3.5-way or 4-way system. 😬
I agree with @presscot this looks like a 4 way. Think would about using a band pass filter on the Vifa woofer. I would also consider leaving out the HiVi dome mid, the Vifa looks like it can cross to the tweeter around 3kHz. I think a passive 4 way will be quite difficult to execute, the problem is the low pass and high pass filters that make up a band pass filter need about 8 to 10 times frequency spread to prevent them from interacting to much. If you go active this would not be a problem. Lastly going from 3 way to 3.5 way would just require adding an inductor in series with the upper Dayton woofer to compensate for baffle step. I believe @AllenB is trying to direct you towards measuring as close to anechoic conditions as possible, thus outdoors and not in the listening room. Then use an equalizer or room treatments to deal with in room response. You should look into getting a microphone stand, the ottoman and the ladder you are using will definitely alter your measurement.
Hi @mtidge ! I can imagine that a passive 4-way is not that easy to design. But one of the reasons that I am going with a 4-way (or 3.5-way, not really relevant here) is the challenge itself. I am almost retired and want something to make me scratch my head a little 🙃
I heard this dome mid at a friend's house and really liked it, the sound was so airy and alive and the imaging was amazing. So I am building this speaker around it. I could drop the Vifa mid bass and let the dayton woofers cross with the mid dome at around 900hz, but then it would be a standard 3-way. Not what I am going for here.
I will be redoing all my measurements, both indoors and outdoors, with extra attention to details. Will also get a near field measurement to get a correct bass response, as described here: Quasi-Anechoic measurements. The idea is to understant what is happening, learn a lot (this is my first measurement based design, all other speakers where based on driver specs only) and have fun.
BTW, I will get a mic stand, thanks!
I heard this dome mid at a friend's house and really liked it, the sound was so airy and alive and the imaging was amazing. So I am building this speaker around it. I could drop the Vifa mid bass and let the dayton woofers cross with the mid dome at around 900hz, but then it would be a standard 3-way. Not what I am going for here.
I will be redoing all my measurements, both indoors and outdoors, with extra attention to details. Will also get a near field measurement to get a correct bass response, as described here: Quasi-Anechoic measurements. The idea is to understant what is happening, learn a lot (this is my first measurement based design, all other speakers where based on driver specs only) and have fun.
BTW, I will get a mic stand, thanks!
You colud also do polar measurements (off axis in 10° steps or even 5°) once you're at it!
Have you tried adding an all-pass to the midbass with the same frequency as your lowpass for the woofers?
Hi All,
Just to close this subject, I took all measurements again and now that difference between in-room and outdoors is not present anymore. Of course the measurements are not the same, but both crossover designs look a lot more alike now.
I guess that I made a mistake when measuring the first time after all...
BTW, I went back to the original design that I had before creating this thread. Instead of front firing dual 8"woofers, I'll go with side-firing woofers, mounted opposed to each other. I'll create a build thread soon. Here is a cr*ppy picture of the test box:
Thanks a lot!
Just to close this subject, I took all measurements again and now that difference between in-room and outdoors is not present anymore. Of course the measurements are not the same, but both crossover designs look a lot more alike now.
I guess that I made a mistake when measuring the first time after all...
BTW, I went back to the original design that I had before creating this thread. Instead of front firing dual 8"woofers, I'll go with side-firing woofers, mounted opposed to each other. I'll create a build thread soon. Here is a cr*ppy picture of the test box:
Thanks a lot!