I failed to mention >
A very simple, cheap & easy way to add 'top-end' to a rear FR speaker is to simply add a 'classic 1005 style' piezo horn BUT
using it reverse facing WITH its rear cover removed.
This way you get some direct radiation from the 'rear of piezo', but mostly reflected top-end. (experimentation is required)
If it is too bright, you simply use a cheap series capacitor, below the value of 1uF. (like .47 or .22uF)
You might be surprised with the effectiveness of this approach 🙂
Nice idea!
Piezo is best filtered with 50 to 100 ohms parallel to the driver and 0.1 to 0.5 microfarad in series.
It's better filtered, then.
Piezo is best filtered with 50 to 100 ohms parallel to the driver and 0.1 to 0.5 microfarad in series.
It's better filtered, then.
Thanks again.
When you say biflex driver do you mean a very rigid diaphragm (such as tne RS100's aluminium cone) that can only "flex" at the weak spot which is the joint between the cone and the voice coil? Triflex should be something like a lightweight paper (or poly) cone? I also can't open the link you sent but I'd love to read more about this if you have any recommendations.
On a side note, why won't comb filtering from the woofer+tweeter playing in the same range be an issue?
Why does this happen? I struggle to understand how dimensionally Fs(Hz)/dVC(m) can be a bandwidth (Hz).
When you say biflex driver do you mean a very rigid diaphragm (such as tne RS100's aluminium cone) that can only "flex" at the weak spot which is the joint between the cone and the voice coil? Triflex should be something like a lightweight paper (or poly) cone? I also can't open the link you sent but I'd love to read more about this if you have any recommendations.
On a side note, why won't comb filtering from the woofer+tweeter playing in the same range be an issue?
Last edited:
I doubt I will use a piezo tweeter, good ones are hard to come by and often as expensive as 3/4" dome tweeters
Interesting! I always check old links before posting and it was fine, but no good now, nor another one I sometimes used!
Regardless, here it's archived presumably for as long as 'we' send $$$ to fund it: W.E. 755 diaphragm
Obviously the VC joint doesn't resonate at any frequency we can hear because it's so rigid, but the diaphragm attached to it does and while the simple math frequency isn't 100% accurate it's plenty good enough for calculating XO points as humans can't tell the difference between 1400 & 1410/whatever nearby Hz it actually is.
No a biflex has a mechanical XO built into the diaphragm such as the Altec 419, 420 and triflex has three such as the W.E. 755. Note too that drivers are often 'laid up' with layers of different materials, coatings to get the desired flex ('tone') such as the curvilinear profile (pg. 36); a good book to read for the basics of speaker, driver design among others written in more recent times, such as John Eargle's update.
Comb filtering in relation to what? I mean if two sources are within a 1/4 WL of the 'combing' frequencies they will sum good enough for us humans to perceive them as a single source thanks to our very powerful internal processor (brain).
Yeah, early W.E. CDs were back filtered with cotton, so some of the 'old school' 'Hornie's' either remove the rear cover or if needing it for power handling, add a second rear facing horn damped to 'taste'.
It makes more sense now!
-I did some maths and apparently due to the distance I'll be listening from the speaker the extent of this will be negligible (the paths of the woofer and tweeter will be almost equal)
If I get a small tweeter such as the ND20FB-4, and place it as close as possible the CTC distance of the RS100-ND20FB will still be of at least 7cm, thus resulting in comb filtering of about everything over 1300Hz. The tweeter won't be crossed that low, of course, but there will basically be combing across the whole range the tweeter is used in.
-I did some maths and apparently due to the distance I'll be listening from the speaker the extent of this will be negligible (the paths of the woofer and tweeter will be almost equal)
Equal paths (delay) means there should be no comb filtering of the direct sound, (if you keep them close enough in phase).
However, I believe the concern was about lobing, not comb filtering. Overlapping multiple drivers causes different summation in different directions.
Some find that what's important with the upper highs is the amount of sound into the room, and they sometimes point the supertweeter away from the listening position in an attempt to take the focus off the direct sound and possible lobing inconsistencies.
However if you plan to cross in the usual way, then you'll take care of the overlap and you will limit it's band and choose a good frequency to do it.
However, I believe the concern was about lobing, not comb filtering. Overlapping multiple drivers causes different summation in different directions.
Some find that what's important with the upper highs is the amount of sound into the room, and they sometimes point the supertweeter away from the listening position in an attempt to take the focus off the direct sound and possible lobing inconsistencies.
However if you plan to cross in the usual way, then you'll take care of the overlap and you will limit it's band and choose a good frequency to do it.
Thank you. How can I choose a good crossover frequency and slope to avoid lobing off-axis?
Usually I place the tweeter as close as possible to the woofer and then I calculate the distance of one wavelenght between the woofer and tweeter (for example if the woofer and tweeter are 10cm away, 345Hz/0,1m=3450Hz).
I then try to cross lower that that with the shallowest slope that is effective at attenuating the woofer's breakup modes, and that won't run the tweeter too hard. I then check the response and phase on XSim (by inverting the phase of one of the speakers).
In this case I'd start with a simple inductor on the woofer and a second or third order on the tweeter, aiming for a crossover point of around 3kHz.
Is this approach correct or there's more I need to account for?
Usually I place the tweeter as close as possible to the woofer and then I calculate the distance of one wavelenght between the woofer and tweeter (for example if the woofer and tweeter are 10cm away, 345Hz/0,1m=3450Hz).
I then try to cross lower that that with the shallowest slope that is effective at attenuating the woofer's breakup modes, and that won't run the tweeter too hard. I then check the response and phase on XSim (by inverting the phase of one of the speakers).
In this case I'd start with a simple inductor on the woofer and a second or third order on the tweeter, aiming for a crossover point of around 3kHz.
Is this approach correct or there's more I need to account for?
Your approach is good.
It is not always possible to minimise the distance between sources, but that's OK. As GM says, 1/4 wavelength is good, but you can also have success with wider spacing. It is partly a matter of keeping the ratio of on/off axis sound even through and beyond the crossover frequencies. You can do that using polar measurements all the way.. or you can adjust by ear.
It is not always possible to minimise the distance between sources, but that's OK. As GM says, 1/4 wavelength is good, but you can also have success with wider spacing. It is partly a matter of keeping the ratio of on/off axis sound even through and beyond the crossover frequencies. You can do that using polar measurements all the way.. or you can adjust by ear.
Thanks. 1/4 wavelenght is very hard to achieve unless it's on FAST speaker I guess.
Yes I try to be very conscious about the off-axis dispersion of the woofer, basically I try to avoid using it in its beaming range (where on-axis and off-axis start to differ by more than 3dB. Usually when I look for the "off-axis" response I refer to the 30deg one, but if the tweeter allows I can cross even lower so even at 45 degrees the woofer still has good dispersion.
I always did this by heart so I don't know if it's really needed, but it usually gave me results that were good enough for me. I measured the off axis response on the horizontal axis but never in the vertical one.
For example using the graph I linked earlier (courtesy of Dibirama) would it be right to cross at 3k if I wanted it to be perfect? Or 4k if I wanted it to be "good enough" at 30degree? Or should I cross even lower (I'll be using your average run-of-the-mill dome tweeter, preferably a 3/4" unit)
Yes I try to be very conscious about the off-axis dispersion of the woofer, basically I try to avoid using it in its beaming range (where on-axis and off-axis start to differ by more than 3dB. Usually when I look for the "off-axis" response I refer to the 30deg one, but if the tweeter allows I can cross even lower so even at 45 degrees the woofer still has good dispersion.
I always did this by heart so I don't know if it's really needed, but it usually gave me results that were good enough for me. I measured the off axis response on the horizontal axis but never in the vertical one.
For example using the graph I linked earlier (courtesy of Dibirama) would it be right to cross at 3k if I wanted it to be perfect? Or 4k if I wanted it to be "good enough" at 30degree? Or should I cross even lower (I'll be using your average run-of-the-mill dome tweeter, preferably a 3/4" unit)
The vertical is the same until you include the crossover. It is the cross that brings in the lobing, and it is how you do the cross which gives you the control to balance it.
If you use a simulator that can do multiple axes then it will simulate the crossover vertical and overall variations. You can do it other ways.. you can measure vertical after you cross, or you can do your own calculations.
we are in same situation, i also would like to add a bit of top end to my RS-100-8 HT speaker, ported 4 ltrs.
I am going to do it wiith a DAYTON ND25FA.
You see a large dip in FR in the background, because the polarity of twtr is reversed.
when normal, i see a +/-2 dB response.
disclaimer: all of it simulated in Visaton's Boxsim software, driver FR from datasheets.
I am going to do it wiith a DAYTON ND25FA.
You see a large dip in FR in the background, because the polarity of twtr is reversed.
when normal, i see a +/-2 dB response.
disclaimer: all of it simulated in Visaton's Boxsim software, driver FR from datasheets.
Attachments
Thank you. Do you have any tips about crossover design on how to avoid lobing in the first place?
I'll try to move to VituixCAD since XSim can't work on multiple axes and that's what I've what always used.
The ND25FA is another tweeter model I've considered but I'll try to fit a smaller one to bring the acoustic centers as close as possible, because high max SPL isn't required for this speaker since I live in an apartment.
I'll try to move to VituixCAD since XSim can't work on multiple axes and that's what I've what always used.
The ND25FA is another tweeter model I've considered but I'll try to fit a smaller one to bring the acoustic centers as close as possible, because high max SPL isn't required for this speaker since I live in an apartment.
do you mean these plots?
I used these simply becaue they were available locally at a hi-fi store.
you can look at ND25FN-4 if you want smaller one and mount them close together.
boxsim is a difficult software to learn, but once you get the hang of it, you can try different things like moving around the drivers , changing box size, etc.
center uses PC83-8 with ND25FA, in MTM, its about 5dB more sensitive. The store ran out of RS100-8, just had 5 of them
Last edited:
Not so sure about that compared with some packages mention in these posts. I switched from WinISD as I knew that there was no way the frequency response it showed for a woofer could be correct. In Boxsim the bass shelf wasn't so clear and the higher frequency response more like what would be expected from the driver in free air. It was worth entering WinISD tuning parameter into Boxsim.
The advice I received was download Boxsim and load some projects from this site and have a poke about in the various tabs to see what is done where.
https://boxsim-db.de/kategorie/systeme/zwei-wege/
The confusing aspect is 2 ways of entering the cabinet details. One common box and another via the facia. Both include diffraction effects. Each driver in the common box entries has it's own volume as well.
The max spl tab comes up with numbers down to max come excursion. Visitron drivers that are loaded usually state a lower figure than the spec sheets give. Probably down to life and distortion.
I'm using version 2. Crossovers. There is a calculator that needs the impedance of the drivers at the frequency it's going to be used at, Before that the spl of the 2 drivers may need correcting. There is a calculator to remove xdB. The resistance of this has to be added to the driver it's applied to if this method is used. These are under extras, The speaker impedance can be read of the plot.
LOL Where things go wrong. This is ambitious. Whole speaker response.
What happened to the hf cut off - slope of the driver impedance at cross over -IMHO anyway.
The crossover
The RC across the woofer - some impedance correction. It has a calculator for that, The R network on the tweeter - putting them there tends to mask driver impedance variation. Probably pointless on a tweeter but also allows more LC to be added across the R's for more correction. There is a calculator to set the dB reduction.
Phase - I feel it needs to be better
The optimiser can only get things to some point. The above is a mix of selective use of it and when cap values are standardised all of it except caps.
😉 The woofer is a woofer to trying to get it to work to 2kHz is bound to be tricky.
That > .82mH / 18uF circuit above is a very reactive load 😳
There are many amplifiers that wouldn't like it at all.
There are many amplifiers that wouldn't like it at all.
Re. comb filtering >
In the realm of high frequencies, (not bass) comb filtering can actually sound quite good,
as can be seen in TOTALLY UNSMOOTHED frequency responses of many popular tweeters 🙂
In the realm of high frequencies, (not bass) comb filtering can actually sound quite good,
as can be seen in TOTALLY UNSMOOTHED frequency responses of many popular tweeters 🙂