Can you design a speaker with no F-R plot?

[niacin]

Paging through the Parts Express website, I notice a lot of drivers with no F-R plots available. Even some pro-PA drivers, like the ones from Pyle. And many of the regular mid-woofers from Dayton and Pyramid have no published F-R plots at all.

How can you design a speaker with no F-R plot? Yet the speakers are still there in the catalog--so somebody is obviously buying them and using them. Just for suckers?

Sometimes a speaker just has a rating like: "frequency response 70-3500 Hz." Is there some sort of standard at work here, like one can assume the response is no more than +/- 3dB over this range or something?

Does anybody here design with speakers with no F-R plot, and if so, how did you do it?


--Niacin

[johninCR]

Most of the speakers I've built were using drivers for which the only specs I have are the nominal ohms and power handling because they printed on the back.

Down here in Costa Rica, there is very limited driver selection and shipping and import duties more than double the cost of mail order drivers. The only significant driver selection is at the electronics supply houses that keep a wide selection of drivers for TV's, radios, etc. I use the most sensitive instrument available before I buy, my ears. Then I buy a few different ones that I like the best and bring them home for further testing which consists primarily of tone sweeps and listening to music. I haven't spent for than $7 on a driver buying like this and have narrowed it to 6 different ones that I use the most. My favorite happens to be the cheapest, a wide range 4" at $1.50 each. They are reasonably flat from 100hz up to 9800hz, don't have a nasty breakup, and have a nice smooth tone.

I use these drivers to build OB line arrays. I add in a titanium bullet tweeters meant for cars which after tweaking sound very nice with great efficiency, wide dispersion, and none of the harshness that they have out of the box. For the bass I use 6" TV drivers. I usually cut the mids below 150hz to protect the drivers from over excursion with a single capacitor. For the woofers I usually use a first order roll-off starting at anywhere between 60hz and 100hz depending upon how low they need to go and how many I use in relation to the number of mids. The super tweeters need only a cap and an Lpad.

The beauty of OB's in addition to the great natural clear sound, is that I can do cardboard or styrofoam test baffles to determine the ratio of the drivers I want to use and get a good of idea of what they will sound like. I wait to finalize the crossovers until they're mounted on the permanent baffle.

The result is speakers for under $100/pr that always impress people with their looks and sound. The units that I build with 12 or more drivers, I would put up against any store bought speaker costing under $1,000 , plus they have well over 100db of sensitivity so any amp will drive them.

[synergy]

you only need the Qts, Fs, Vas and Re to simulate the bass response of speakers

the other T/S parameters are useful for refining the accuracy of the design but not essential as it is generally assumed in simulations that the speaker will play flat above the desired tuning frequency

as to whether or not F-R plots are useful is a highly debateable matter certainly highlighting a peak in response further up they can be
however the flaw in using them to choose speakers is that you have to know the enclosure, room and methods used to test them to be able to make a critical judgement as these will affect the result greatly

for example it's pratically impossible to have a truely anechoic chamber down to 20Hz and often the very simulations we'd achieve from the using T/S parameters ourselves are spliced onto the bottom of the graph

[Hybrid fourdoor]

Originally posted by niacin
Paging through the Parts Express website, I notice a lot of drivers with no F-R plots available. Even some pro-PA drivers, like the ones from Pyle. And many of the regular mid-woofers from Dayton and Pyramid have no published F-R plots at all.

How can you design a speaker with no F-R plot? Yet the speakers are still there in the catalog--so somebody is obviously buying them and using them. Just for suckers?


--Niacin

Parts Express has graphs of the more popular drivers. This may include ones that are good sellers, or buyout ones that are cheap (but have horrible responses).

Many places don't offer graphs, and I'm gratefull that Parts Express takes the time meausre so many drivers. Its certainly not laziness on anyones part.


Also you have the ability to do the same at home. This may be better, as you get a specific plot and not a generalized plot.

[SY]

Can you design a speaker with no F-R plot?

Sure you can. And it will suck.

How can you design a speaker with no F-R plot?

Even with a plot, you can't. The plots shown in data sheets are not terribly representative of actual drives and certainly not representative of your drivers on your baffle. If you want to do a speaker design (as opposed to building someone else's design), at minimum you need to measure frequency response yourself, on your baffle, in your box, at various angles, and measure impedance.

[Timn8ter]

1. Model with T/S parameters
2. Prototype
3. Measure
4. Tweak
5. Finalize

[johninCR]

Originally posted by SY


Sure you can. And it will suck.

Originally posted by Timn8ter
1. Model with T/S parameters
2. Prototype
3. Measure
4. Tweak
5. Finalize

I disagree. Measure, measure, measure. I say listen, listen, listen, both to music and to test tones and sweeps. The parameters don't tell you the bottom line and that is whether or not a speaker sounds good in accordance with your personal taste.

I just don't see the point in going through all the trouble, just so you can figure out the size box you want. Then you jump through a bunch of construction hoops trying to make your drivers sound like they're not in a box, which is next to impossible.

I say forget about the specs and forget about the box. Listen and tune and tweak as you go. You can enjoy listening to your test baffles as you build your permanent ones and come up with some more tweaks as you build. I find it more rewarding and better sounding than crunching numbers and hoping for the best when it's done.

Just 2 cents from a full boxless convert.

[Salas]

The most honest thing from a manufacturer would be to state: 'indicative frequency plot' and give the measurement parameters like IEC baffle or box dimensions if used. Also smoothing used, measurement system used, angles, mic distance etc. And all this will only suggest some stuff higher, above power range.
Most honest guys are PHL. They just dont give any plot! They know they are adressing engineers with gear and experience.
They know such a plot would mean nothing to an engineer.
FR plot is meaningfully taken only when the driver is mounted where it is about to stay. Impedance also.

[tiroth]

Originally posted by salas
They know such a plot would mean nothing to an engineer.
FR plot is meaningfully taken only when the driver is mounted where it is about to stay. Impedance also.

I completely disagree. There is a lot to be said for measuring in situ. But even 'professional' sound engineers do not have infinite time to try one of every driver in existence. Looking at an honest FR, waterfall plot can tell you a lot about how a driver will behave. You can rule out a driver that is "100-4000 Hz" and has a big breakup mode at 2k, even if all of the data is marginal.

The more data available, the easier it is to narrow the choices. Isn't it better to demo the 10 best drivers based on measurements and listening input (from the grapevine) than pick 100 at random? i bet the former achieves better sound quality as well as being cheaper.

[johninCR]

Originally posted by tiroth


I completely disagree. There is a lot to be said for measuring in situ. But even 'professional' sound engineers do not have infinite time to try one of every driver in existence. Looking at an honest FR, waterfall plot can tell you a lot about how a driver will behave. You can rule out a driver that is "100-4000 Hz" and has a big breakup mode at 2k, even if all of the data is marginal.

The more data available, the easier it is to narrow the choices. Isn't it better to demo the 10 best drivers based on measurements and listening input (from the grapevine) than pick 100 at random? i bet the former achieves better sound quality as well as being cheaper.

I'd bet it's more getting info "from the grapevine" and listening to free samples than it is hunting through claimed specs and measurements. For the big manufacturers they probably don't even get to make the choices except maybe from a small group that is selected by an accountant based on frequency range, Znom, power handling, and PRICE.