Yeah passive xo is very nice as long as it is feasible. Here in your application it saves some money on the amplification and reduces hiss on the tweeter. DSP on the low xo saves money on expensive xo parts for low frequencies and gives flexibility and opportunity to control the whole system even though there was passive xo between mid and tweet. Extra bass SPL capability can be harvested for extended low frequency extension with the DSP and separate amplification for example.
Have you considered cardioid mid for the setup? Or thought about your listening situation how to address possible issues or is there any? Whats behind reasoning on the crossover frequency to bass? These are some stuff I've been experimenting with so kind of interesting to me at least!
My system has xo around 200Hz not sure what is exactly since didn't look
It is 15" to 8" to 1" in a waveguide setup. I thought initially to make passive xo between mid and tweet but unfortunately the drivers I eventually bought have so bad response around crossover region that I don't see how to do the xo like DSP can so just went with DSP all around. Cardioid mid just about makes it to the 200Hz and the bass enclosure is so big the sound changes if xo is taken much over 200Hz so it was a nice fit. More over, the bass enclosure is quite flat so within 1/8wl to the wall, no SBIR problems. I'm placing the thing against front wall for practical reasons. ~200Hz xo also enables the mid to be raised higher to balance between floor and ceiling reflections (in simulator), but his is different for different height rooms. Hence modular design. I'm able to raise or lower the ways to reduce / average out vertical reflections as well as toe in the mid tweet section independently of the big bass that is flat against the wall. I've got a goal and designing against it is not too hard, taking it further into sound quality without additional costs other than time and effort. It looks like going for sound quality is making the system too big so might try another design after this one is finished.
In general, if DSP is allowed in the system it unlocks many compromise situations in the system and more importantly frees designer thinking like anything was possible xo wise without any extra cost. This opens up whole new set of opportunities for a design and driver selection for example. As the system develops it might be possible revert back to passive parts!
---
Thought a minute outside about the topic and thread. I admin I haven't read it through but as it is 12 year old already it suggest people are not interested making a reference speaker in general, like best for some application, I think. Everyone are interested on various things like the concepts or learning or just exploring stuff, and very few are interested to tackle the system design problem in general and happened to land on the 3-way land at least for the moment. Maybe some shot past this phase and some didn't arrive yet Maybe there are many, but I'm lost and just don't see it? Maybe someone started commercial venture and didn't share the secrets on for the public.
Reasoning from the thread longevity without conclusion there has not been attempt to make a reference design other hand perhaps some random goals and everyone are doing it on their own or transition to something else.
Have you considered cardioid mid for the setup? Or thought about your listening situation how to address possible issues or is there any? Whats behind reasoning on the crossover frequency to bass? These are some stuff I've been experimenting with so kind of interesting to me at least!
My system has xo around 200Hz not sure what is exactly since didn't look
It is 15" to 8" to 1" in a waveguide setup. I thought initially to make passive xo between mid and tweet but unfortunately the drivers I eventually bought have so bad response around crossover region that I don't see how to do the xo like DSP can so just went with DSP all around. Cardioid mid just about makes it to the 200Hz and the bass enclosure is so big the sound changes if xo is taken much over 200Hz so it was a nice fit. More over, the bass enclosure is quite flat so within 1/8wl to the wall, no SBIR problems. I'm placing the thing against front wall for practical reasons. ~200Hz xo also enables the mid to be raised higher to balance between floor and ceiling reflections (in simulator), but his is different for different height rooms. Hence modular design. I'm able to raise or lower the ways to reduce / average out vertical reflections as well as toe in the mid tweet section independently of the big bass that is flat against the wall. I've got a goal and designing against it is not too hard, taking it further into sound quality without additional costs other than time and effort. It looks like going for sound quality is making the system too big so might try another design after this one is finished.In general, if DSP is allowed in the system it unlocks many compromise situations in the system and more importantly frees designer thinking like anything was possible xo wise without any extra cost. This opens up whole new set of opportunities for a design and driver selection for example. As the system develops it might be possible revert back to passive parts!
---
Thought a minute outside about the topic and thread. I admin I haven't read it through but as it is 12 year old already it suggest people are not interested making a reference speaker in general, like best for some application, I think. Everyone are interested on various things like the concepts or learning or just exploring stuff, and very few are interested to tackle the system design problem in general and happened to land on the 3-way land at least for the moment. Maybe some shot past this phase and some didn't arrive yet
Maybe there are many, but I'm lost and just don't see it? Maybe someone started commercial venture and didn't share the secrets on for the public.Reasoning from the thread longevity without conclusion there has not been attempt to make a reference design other hand perhaps some random goals and everyone are doing it on their own or transition to something else.
Last edited:
I think there has been some confusion in this thread about the differences among project goals, customer needs, requirements, and specs.
It starts with a project goal, which in the case of a loudspeaker is often so obvious that it need not be stated. The goal for most DIY loudspeaker projects is to design/build a speaker which sounds good and meets the end user's (the customer's) needs. It is such an obvious thing that is usually no point in stating it. For a commercial speaker, the goal is typically centered on satisfying some market need. If the marketing people say there is a demand for very tall, thin, line array speakers in the $1500 range, then the goal is to satisfy that demand and make a x.y% profit.
Customer needs: the desires and expectations of the customer, both subjective and objective, both qualitative and quantitative. "I want a car that is fun, and fast"... that is subjective and qualitative. "I want a car that fits in my garage"... that is quantitative. " I want a speaker which sounds good with kick drum and vocals, plays loud, and fits in my room without having to move the other furniture". That is a very valid expression of customer need. Customers often cannot express their needs in terms of technical performance measures. They can say they want a speaker which is distortion free, for example, but they cannot say what level of 3rd HD they will accept at a given frequency. Often times customer needs are implied or unstated.
Requirements: Objective quantitative statements about the product, such as the size, shape, technical performance, cost. It is the designer's job to translate customer needs into requirements. "I want a car that is fast" might translate into "the car shall accelerate from 0 to 100 km/hr in 7 sec or less". For a loudspeaker, the requirements will be qualitative statements about the size, weight, and performance of the speaker. "the speaker shall be at least 35" tall, but no taller than 45"... "The speaker shall have a bass extension of -9 dB at 32 Hz" ... "The speaker shall have no more than 1% THD from 200 Hz to 10 kHz at 95 dB @1 M"... Requirements might also be statements about the aesthetic qualities, such as "this speaker shall be glossy white".
Spec or specifications: The list of features, parameters, and drawings which define the product form fit & function. Specs allow multiple companies to produce products which are interchangeable with each other. Bolts, light bulbs, connectors, tires, etc are built to specs. A DIY kit speaker is essentially a spec which we follow and build to.
So in this thread, we had a list of customer needs, but we failed to translate them into technical requirements. The lack of a goal was not the problem. The problem is a disagreement on the methods needed to meet the customer needs. The customer wants "good bass", and to this designer that means a sealed box, but to that other designer it means a dipole system. Every aspect of performance faced a similar disagreement. So the requirements could never be written.
j.
It starts with a project goal, which in the case of a loudspeaker is often so obvious that it need not be stated. The goal for most DIY loudspeaker projects is to design/build a speaker which sounds good and meets the end user's (the customer's) needs. It is such an obvious thing that is usually no point in stating it. For a commercial speaker, the goal is typically centered on satisfying some market need. If the marketing people say there is a demand for very tall, thin, line array speakers in the $1500 range, then the goal is to satisfy that demand and make a x.y% profit.
Customer needs: the desires and expectations of the customer, both subjective and objective, both qualitative and quantitative. "I want a car that is fun, and fast"... that is subjective and qualitative. "I want a car that fits in my garage"... that is quantitative. " I want a speaker which sounds good with kick drum and vocals, plays loud, and fits in my room without having to move the other furniture". That is a very valid expression of customer need. Customers often cannot express their needs in terms of technical performance measures. They can say they want a speaker which is distortion free, for example, but they cannot say what level of 3rd HD they will accept at a given frequency. Often times customer needs are implied or unstated.
Requirements: Objective quantitative statements about the product, such as the size, shape, technical performance, cost. It is the designer's job to translate customer needs into requirements. "I want a car that is fast" might translate into "the car shall accelerate from 0 to 100 km/hr in 7 sec or less". For a loudspeaker, the requirements will be qualitative statements about the size, weight, and performance of the speaker. "the speaker shall be at least 35" tall, but no taller than 45"... "The speaker shall have a bass extension of -9 dB at 32 Hz" ... "The speaker shall have no more than 1% THD from 200 Hz to 10 kHz at 95 dB @1 M"... Requirements might also be statements about the aesthetic qualities, such as "this speaker shall be glossy white".
Spec or specifications: The list of features, parameters, and drawings which define the product form fit & function. Specs allow multiple companies to produce products which are interchangeable with each other. Bolts, light bulbs, connectors, tires, etc are built to specs. A DIY kit speaker is essentially a spec which we follow and build to.
So in this thread, we had a list of customer needs, but we failed to translate them into technical requirements. The lack of a goal was not the problem. The problem is a disagreement on the methods needed to meet the customer needs. The customer wants "good bass", and to this designer that means a sealed box, but to that other designer it means a dipole system. Every aspect of performance faced a similar disagreement. So the requirements could never be written.
j.
hifijim, Nice, proper terminology!
So, we would need a specification but for that we would need the requirements and to make those we would need a curator or the leader who would make the requirement list out of the needs? Is this what you have been doing, collecting the needs?
But for now the customer and the curator have been the same thing (all of us)? I mean number of ways and list of drivers would be in the specification, but cost limit would be in the needs section like "i need cheap and good loudspeaker"? Yeah thats what ive been trying to say, we have started wrong end of the process. Luckily 3 ways is fine in specification for the goal of good sound so the topic title should be fine for reaching goal of good sound.
But both moondog and MrHifiTunes have expressed while good sound is fine they don't specifically go for after quality if I understood correctly. I understand moondog has goal of fine enough sound and fun with the hobby, and MrHifiTunes goal is to learn while designing without goal (a goal having open requirements) to see how things work because it is interesting and fun I presume. Both fine goals/motives but not in line with each other I think, other than the fun part.
Should we try and find common goal, instead of squeezing performance try and have a fun project? Good sounding speaker that is fun to build and listen to?
I want fun speakers that are loud, proper dance bass and not shrill highs like in the small town club which is also the pizzeria and buss station! I also have them in the livingroom all the time so can't be too big or black or smelly because the wife is allergic I listen mostly Jazz and other acoustic music, but just now I'm blasting Cult! And its not a joke but the real needs!
So, we would need a specification but for that we would need the requirements and to make those we would need a curator or the leader who would make the requirement list out of the needs? Is this what you have been doing, collecting the needs?
But for now the customer and the curator have been the same thing (all of us)? I mean number of ways and list of drivers would be in the specification, but cost limit would be in the needs section like "i need cheap and good loudspeaker"? Yeah thats what ive been trying to say, we have started wrong end of the process. Luckily 3 ways is fine in specification for the goal of good sound so the topic title should be fine for reaching goal of good sound.
But both moondog and MrHifiTunes have expressed while good sound is fine they don't specifically go for after quality if I understood correctly. I understand moondog has goal of fine enough sound and fun with the hobby, and MrHifiTunes goal is to learn while designing without goal (a goal having open requirements) to see how things work because it is interesting and fun I presume. Both fine goals/motives but not in line with each other I think, other than the fun part.
Should we try and find common goal, instead of squeezing performance try and have a fun project? Good sounding speaker that is fun to build and listen to?
I want fun speakers that are loud, proper dance bass and not shrill highs like in the small town club which is also the pizzeria and buss station! I also have them in the livingroom all the time so can't be too big or black or smelly because the wife is allergic
I listen mostly Jazz and other acoustic music, but just now I'm blasting Cult! And its not a joke but the real needs!
Last edited:
Here's my take on what I thought to be the best compromise I could achieve with a 3-way given my admittedly limited knowledge (first DIY speaker after 1 year of research).
My design goals were as below:
My design goals were as below:
- Linear frequency response ; natural tonal balance
- LF extension
- Low horizontal directivity
- Low THD/IMD
- Low stored energy (fast CSD incl. in LF)
- High sensitivity
- Proper phase integration across drivers +/-1oct. and 0-45degrees off-axis
- High electrical impedance / low phase
- Low vertical directivity
Attachments
Last edited:
^nice, you have put some thought and effort on it! If you have fixed all things feel like fixing and the diffraction still bothers you then experiment with some addon roundovers perhaps?
According to diffraction simulator the wide baffle with smaller drivers on it, without roundovers, will have diffraction interference on axis all the way from ~baffle width to tweeter beaming, or where the roundovers start to be effective. Maybe use some big plumbing pipe or cardboard tube as slap on fix to first try if there is any audible difference or is it just wiggle on the graphs. roundover radius close to lenght from driver center to the current edge would pretty much kill the diffraction interference and one centimeter radius wouldn't probably make much difference. Use something big enough easily available stuff.
Another option is to make quick minimal box/baffle for the mid and put that on top of the speaker, the tweeter can be freestanding on top of that or left where it is currently. Flip the bass box upside down if it keeps the woofer mid distance a bit closer to current setting so no much change other than the baffle size for mid and tweet. Any difference?
Have fun!
According to diffraction simulator the wide baffle with smaller drivers on it, without roundovers, will have diffraction interference on axis all the way from ~baffle width to tweeter beaming, or where the roundovers start to be effective. Maybe use some big plumbing pipe or cardboard tube as slap on fix to first try if there is any audible difference or is it just wiggle on the graphs. roundover radius close to lenght from driver center to the current edge would pretty much kill the diffraction interference and one centimeter radius wouldn't probably make much difference. Use something big enough easily available stuff.
Another option is to make quick minimal box/baffle for the mid and put that on top of the speaker, the tweeter can be freestanding on top of that or left where it is currently. Flip the bass box upside down if it keeps the woofer mid distance a bit closer to current setting so no much change other than the baffle size for mid and tweet. Any difference?
Have fun!
Last edited:
Im really confused here.
I thought is was clear that the GOAL isn't a reference speaker in the sence of best possible to this day.
THE GOAL is to make a set of 3-way speakers with the same drivers.
This for educational purpose. To allow newbies to explore the differences. But In the same time a well thought out design.
Maybe I say something stupid here, but it would be good if one explain why he make certain design decisions. That's the educational part of the project.
Of course not every design is possible, but this can be pointed out why it is not possible.
all other conditions are posted before.
Like I said, if you get 3 drivers. What is the best you can build with it?
Then would be nice if specialist in different discipline propose something.
(narrow/wide tower ; TL,BR, closed) (serial/parallel XO) Time-aligment, etc etc...
This is the open part but i hope those with expertise could fill the blanks.
--- sidenote---
Like I said my interest goes for a design with wide baffle, TL and serial XO. If someone can point me to that kind of project, please do.
I thought is was clear that the GOAL isn't a reference speaker in the sence of best possible to this day.
THE GOAL is to make a set of 3-way speakers with the same drivers.
This for educational purpose. To allow newbies to explore the differences. But In the same time a well thought out design.
Maybe I say something stupid here, but it would be good if one explain why he make certain design decisions. That's the educational part of the project.
Of course not every design is possible, but this can be pointed out why it is not possible.
all other conditions are posted before.
Like I said, if you get 3 drivers. What is the best you can build with it?
Then would be nice if specialist in different discipline propose something.
(narrow/wide tower ; TL,BR, closed) (serial/parallel XO) Time-aligment, etc etc...
This is the open part but i hope those with expertise could fill the blanks.
--- sidenote---
Like I said my interest goes for a design with wide baffle, TL and serial XO. If someone can point me to that kind of project, please do.
Best for what purpose? or which since you call for multiple designs? you are not seeing there needs to be a goal to design for? I mean the needs, goal is good sound.
Ok we could design loudspeaker that is best for:
-smallest possible form factor
-cheapest possible
-under the television
-into a bookshelf
-floorstanding against the wall
-big dedicated listening space but soft volume
-gift for hard rocking uncle
...
you are not seeing there needs to be a goal to design for? I mean the needs, goal is good sound. Ok we could design loudspeaker that is best for:
-smallest possible form factor
-cheapest possible
-under the television
-into a bookshelf
-floorstanding against the wall
-big dedicated listening space but soft volume
-gift for hard rocking uncle
...
Last edited:
These look great and seem pretty sensible as far as cost vs performance. Do you care to share more details on the crossover? Also at what frequency do you see the diffraction issues? I thought the wave guide was supposed to mitigate much of the diffraction issues?
thanks, just trying to learn.
Speaker design is all about compromises. I think we need to know which part can be compromised to build a "reference speaker".
@tmuikku
This was mentioned before wasn't ?
medium size room, listening distance 2,5-3m was posted long time ago.
Cheapest budget? Budget is $600 for the drivers.
3 way tower narrow/wide baffle
difference xo's
with without time alignment. (electric and slanted baffle)
flat /sloped FR's
etc ...
Im not the specialist here to determin what is possible with good chose drivers.
it is for sure not unlimited, it is a limit set. Some techical issues will even narrow it down more.
And like posted before we can start off with just a few designs which can be used as base.
This was mentioned before wasn't ?
medium size room, listening distance 2,5-3m was posted long time ago.
Cheapest budget? Budget is $600 for the drivers.
3 way tower narrow/wide baffle
difference xo's
with without time alignment. (electric and slanted baffle)
flat /sloped FR's
etc ...
Im not the specialist here to determin what is possible with good chose drivers.
it is for sure not unlimited, it is a limit set. Some techical issues will even narrow it down more.
And like posted before we can start off with just a few designs which can be used as base.
Last edited:
Exactly why I gave my criteria of performance ranked my importance.
This ranking might very well differ across person.
And indeed it's all a matter of compromise. Just one example:
- Increasing Mid/Tweeter cross over frequency would:
- Decrease tweeter distortion below 2kHz
- Decrease tweeter's excursion below 1kHz which helps reduce IMD
- Degrade horizontal directivity (mid beams too "hard" past 2.5kHz)
- Degrade vertical directivity if Mid Vs Tweeter's center-to-center distance remains the same
- Compromise linearity as mid is not as linear as tweeter above 2.5kHz
- Provide less attenuation on the mid breakup modes / stored energy (see waterfall)
- Potentially increase impedance (mid is 8Ohms ; Tweeter is 4Ohms)
I can see a dip around 1.5kHz on axis due to diffraction: here presented without any round over:
diffraction_Saisho_NoRoundOver.png
Diffraction largely changes directivity too, so it's complex.
Now, because it occurs below 2kHz, it impacts mostly the mid ; and not the tweeter/waveguide.
See my mid response alone; and how response actually goes up around 1,5kHz off axis:
M2_25cm_0_10_15_20_30_45C_20201107.png
If I had a much more narrow baffle, the diffraction would have been less of an issue given that its range of effect would be ~3kHz, ie where waveguide helps by increasing directivity, ie prevents radiating towards the side of the baffle.
Attachments
Sorry being such a jack *** with all the goal stuff but as you see from Martiganes post above there is gonna be roughly one best crossover for a system and the system design depends on what compromises you would like to prioritize.
Here are comments for the variations:
Would you like the wide baffle of narrow baffle better? The wide baffle will require the roundovers, or slants, or the diffraction will be worse than with narrower baffle. Narrow baffle would benefit them too if you can do it. Although, you would have to build it and listen if it the slants mattered to you. Baffle width changes the xo to woofer a bit and optimal driver selection might change with the baffle width as the convenient xo point changes. If you just prefer the looks of either then it is gonna be that. Some one could design both but they would work pretty similar. Baffle step would be at slightly different frequency ( where the system response widens to omni ) and thus the diffraction would change a bit, unless there was roundovers. I don't see much difference between the two, unless 1 metre wide baffle? How wide you'd like it?
You can simulate any crossover topology for tower to see the response, and with a DSP one could hear them all without too much cost. With passive parts maybe few different crossovers can be tested before the price comes painful unless prototyping parts bin is already bought and passive parts for that matter free to try out (see XRK971 threads, he's photos on nice passive xo prototyping setups). Sometimes there is no option but stick to some particular slope or slopes if the drivers have issues. And, at the end you'll endup with one crossover that works the best for the given system anyway. It is not a let down though, it is victory! You want best sound, not particular xo topology just because.
Speakers with direct radiating drivers will all have pretty similar features on the slope / flatness on the main axis vs. power response (Directivity) and the shape changes a bit with varying the driver size and another bit if two cones were different shaped. Usually the listening axis frequency response is targeted flat unless some particular "house curve" is prefered. The power response will reflect the system acoustic direcitivity behavior (the constrcut!). Usually we'd want power response to have some boost on the bas and then smooth slope or straight flat line depending what is the acoustic performance of the construct. All tweeters are roughly 1" so the difference is not much in the response (directivity). You can play with the directivity to one direction around xo region by changing the distance between the drivers. The overall response has some wiggle depending on what drivers you bought and that might require extra filters in xo. I think the advanced passive xo builders can do filters that are far from text book but as with combination of the particular driver they make a particular acoustic slope and possibly fix some issues like make there response slope some or flat or what ever. In this respect you are right there might be some clever crossovers and some brute force ones. But you would have to present measurements of a system for somebody to do a clever crossover for it. Trying to say there is not much to see until measurements, and then there is merely a routine to come up with best xo, hook it up, measure and confirm by listening, revisit as many times as needed.
---
I want you to try VituixCAD, popup the diffraction tool and export measurements off the default ideal driver on a bafflesize you like. Export them to the drivers on the main window and add all drivers in the crossover window and hook them up to the signal source. Change the polar map to Normalized to see the system dispersion and where the various sized drivers narrow their response. Here is where you want to crossover at, where dispersion is similar for both drivers and doesn't cause too much of a step in the DI or power. In reality there might be other concerns like the driver breakups or distortion with SPL or box issues like resonances, that would make you want to tweak the crossover points a bit.
Nevertheless you can do what ever xo topology you want, export the graphs (right click on any and "export six-pack"). Then, when you have all exported compare the graphs and try to figure out what works and what not?
I bet you can interpret some of the graphs and then question what all the other graphs mean and gather information you seek. Play with delay to see how it affects phase and polar response, change driver polarities and Y X and Z offsets of the drivers, what happens? It takes some time to get started but you'll learn a lot jus by questioning what you see and what happens if you change the baffle width for example!---
Another tip: have you checked out these for example? DIY-Loudspeakers Troels has posted many many systems with rather elaborate descriptions at least on some, you should check them out. There seems to be some with ~8"-~4"-~1" with drivers under 600€, at least the SEAS 3-Way Classic, mkII and the Discovery Bookshelf I checked. There are some floorstanders as well, didn't check if they fit the bill. There must be many other projects with descriptions already done and laid out for you to ask questions from your self and perhaps from the designer. I'm not sure why would we want to do same design again and again? Discussion is fine, but you have to ask questions from what you have observed.
After you have checked them all on that commercial site wonder why there are so many if there was single best in the world and why we would need to make any more on top of it? What is the best speaker he has designed and why? Why there is three versions of The Loudspeaker?
The answer is they are for different applications and for living for the designer and the company selling the parts. Then think which model would you like to have and why? You could build almost any if not all that feel interesting! There are some TL projects as well.
Last edited:
DIYers are not so interested in fashion, they use what works. Baffling should be sized to give results.
Dipole is a radiation pattern of front and back, but not sides. It uses a sophisticated speaker system.
Open baffle might be better described as an arbitrary baffle with no back, and has through history been a fairly common speaker arrangement. It has some of the same qualities as a dipole because the wide radiation into the room creates spaciousness. On the other hand open baffle would probably not be the direction for a reference project.
thank you for the explanation.
How were you able to get away with so little baffle step on the woofer and end up with 90.5db sensitivity?
If we people who build this project and everybody is using the same drivers then why not have OB midrange as one of the build options, with a second order XO [ electrical not acoustic] most drivers can be used quite low, certainly the cheap Peerless can be used OB at 300Hz with a relatively simple XO at ~300
I already proposed for you to build on xrk's 8+4 as a proven performer with a budget midbass; some there already made an MTM (8+4+8) version; also an 8+4 in a TL. You can add a sub under or a tweeter above. In my experience with SS 10F performing very well up to 15kHz there is no real need for a tweeter. Having said that, I myself am experimenting with adding a tweeter to enable using 2 or 3 of the 10F drivers together for higher sensitivity (see my thread). 8+4+1 is also a famous config for proven OB speakers with controlled off-axis dispersion: NaoNote and LX521 with both closed box and dipole bass.
10F/8424 & RS225-8 FAST / WAW Ref Monitor
If the goal for some is to learn about speaker design process they should be leaning on a platform from others with more experience (else there is no value in reinventing the wheel and risking a waste of money).
Last edited:
While no doubt there are many good and excellent 2-way speakers out there a 3-Way speaker of equal quality will always have a clearer midrange. I might even argue that even if the total cost of the different speakers is the same, which necessitates lower quality drivers and XO parts the 3-Way will most often have a better midrange IF the XO is reasonable designed and the midrange covers a wide enough band.