Newbie DIY speaker help required

sunilraj

Member
Joined
Dec 7, 2009
Messages
35
Points
8
Location
Bangalore
Hello All,

I am newly joined member and a DIY enthusiast. I am planning to construct a ported floor standing tower using a Full range driver and a woofer. The idea of combining a woofer (not subwoofer) with full range driver is to have enhanced bass. My intension is not use cross-over network for my speaker system. Can I connect the woofer in series with the full range driver? Will hi-freq feed to woofer degrade the sound quality? And is it safe for woofer being driven with hi-freq signals?

Will appreciate your inputs.
 
I somehow missed that caveat completely. Sunil, if the genres you listen to are not that demanding, I think you could get by with just the FR in a specifically designed cabinet. What FR driver do you have in mind?
 
Why?

10 characters

I believe cross-over networks can add some amount of noise that may degrade the audio quality. My idea is to have pure audio signal (without any external influce from filter circuit) from the amp being feed to the sound drivers.
 
I somehow missed that caveat completely. Sunil, if the genres you listen to are not that demanding, I think you could get by with just the FR in a specifically designed cabinet. What FR driver do you have in mind?

You are rite, but the problem is I do not find quality FR's in bangalore:sad:. I have limited budget to import Fostex or Jordan. I got couple of Philips FR's from a friend of mine. The spec shows fedility range from 100hz to 18khz.
So I thought putting a woofer may be good idea.

Your inputs?
 
You are rite, but the problem is I do not find quality FR's in bangalore:sad:. I have limited budget to import Fostex or Jordan. I got couple of Philips FR's from a friend of mine. The spec shows fedility range from 100hz to 18khz.
So I thought putting a woofer may be good idea.

Your inputs?

Ah I see. Does this driver have a model no.? I remember audere from the forum used a Phillips FR to make an OB set and was quite happy with it. Maybe he would chime in. What budget are you looking at?
 
Where did you read this?

Operating without a XO can damage your drivers. Yes, a passive crossover does degrade SQ to an extent depending on its construction and parts as well as topology.

Operating without one at all is akin to throwing out the baby with the bathwater.

Can it be done? Yes. should it? No. Even a fullrange speaker changes response quite dramatically when placed on a normally sized baffle and some amount of frequency correction is required. A woofer operating in the upper midrange will cause smearing of the image if it does not cut off early enough.

There are ways to totally avoid crossovers, but yours is not one of them.


yep I understand now, that's the reason i am seeking inputs from all the members, these are valuable to my project :).

Here is my new plan,
. Philips Full range driver in the sealed upper compartment of the tower.
. Peerless Woofer in the ported compartment at the bottom of the tower.
. CrossOver designed to cut-off at 3Khz.

I need inputs about sealed compartment for Full range should I go for it, or do both the drivers occuping whole space would perform better?

And I need to understand about baffle correction, why it is needed?
 
I'd think it would be more helpful for you to state your requirements first and then let others help you with the design choices that would help you reach there. If it was me, here's what I'd write down:

1. Total cost (driver, cabinet, finishing, etc.)

2. Kind of music that would be played through the speakers

3. Size of the room that the speakers are intended for

4. Complexity of the cabinet. (The more experienced your carpenter is, the more complex build you can try)

I understand that you already have the FR driver for the project. Do you have any idea about the model? If it's a known model, you have a better chance of finding out it's specs and also check for published designs for the same. Building a cabinet for a driver without knowing anything about it could end up in a less-than ideal pair of speakers. Ofcourse, as in most things YMMV...
 
I'd think it would be more helpful for you to state your requirements first and then let others help you with the design choices that would help you reach there. If it was me, here's what I'd write down:

1. Total cost (driver, cabinet, finishing, etc.)

2. Kind of music that would be played through the speakers

3. Size of the room that the speakers are intended for

4. Complexity of the cabinet. (The more experienced your carpenter is, the more complex build you can try)

I understand that you already have the FR driver for the project. Do you have any idea about the model? If it's a known model, you have a better chance of finding out it's specs and also check for published designs for the same. Building a cabinet for a driver without knowing anything about it could end up in a less-than ideal pair of speakers. Ofcourse, as in most things YMMV...


I am not sure about the model of the drivers. These are laying with my friend which he bought 2 year ago but laying without being used. we do not have the data sheet. The only thing he remembers is 8ohms and 100-18Khz.
I need to just experiment with the enclosure.
And My father him self is a carpenter with over 20years working on high-finish furnitures.
 
I am not sure about the model of the drivers. These are laying with my friend which he bought 2 year ago but laying without being used. we do not have the data sheet. The only thing he remembers is 8ohms and 100-18Khz.
I need to just experiment with the enclosure.
And My father him self is a carpenter with over 20years working on high-finish furnitures.

Ah I see. Good luck with your project. :)
 
Hai sunil,

You say the FR range is 100hz to 18 khz, and are going to cross over at 3 khz. Are you not chopping off the natural response of the FR. I feel you should cross over at 250hz or below, and allow the FR to do its job above that.
Just my 2 cents, and I think people like Cranky will guide you in a better way.

N.Murali
 
At a certain frequency, depending on the dimensions of the baffle and the driver, and the location of the driver on the baffle, the output increases by 6dB. This frequency is known as the baffle step frequency, and is easily calculated using the program called Edge (Google it).

A baffle step filter corrects this anomaly and restores the frequency to anechoic.

One way of correcting baffle step without a crossover is to have *very* wide and large baffle. IIRC for a 4inch driver one would need a baffle ~48 inches x 48inches to achieve this. This is usually not practical for most of us, hence the use of a network.

If you're using the Philips 8" fullrange driver, I have commented extensively on it in jaudere's FR project thread, have a look. Your intention of cutting it off at 3KHz is correct - I would actually cross it over even lower. However doing this with no data and no measurements will be a challenge at best.


May be my Philips FR is same as the one used by jaudere (8ohms 25w rms) and his project may help to device mine.
Mean while I tried to design an RLC lowpass filter for the cross-over, but the effective impedance calculates to around 4.2omhs at 3Khz (and lower for below), but my amp rating is for 8omhs (min) for max output of 100w rms. The lower effective impedance may overheat my amp. Any solution for this?

And the wiring scheme is parallel.
 
Last edited:
Hai sunil,

You say the FR range is 100hz to 18 khz, and are going to cross over at 3 khz. Are you not chopping off the natural response of the FR. I feel you should cross over at 250hz or below, and allow the FR to do its job above that.
Just my 2 cents, and I think people like Cranky will guide you in a better way.

N.Murali

hello murali,
I am open to all possiblities before I finalize on the design. Experts like you and cranky can suggest the best out of all possible design.
 
is your problem.

The only reason to put an R in there would be to control inductiance rise in the woofer/FR at HF, but the Philips has very little of it, so a zobel is not strictly necessary. a textbook RLC will put a resistor in parallel with the VC, causing an impedance dip (I can comment only after the schematic tho). Use a regular CLC||R (C series, L series, C || with R for zobel) and you'll be fine. The parallel LC is not required for a driver like the Philips as it extends well on both sides of the crossover.

Effective impedance of 4.2ohm for a 8 ohm nominal is borderline, but your amp should be able to handle it unless it's a terribly limited design. The driver alone would dip to 5.5-6 ohms in some places, so an 8 ohm nominal amp should be able to work with it.

I just cannot understand, but i know where i live!:)
 
Unfortunately, my friend met with an accident and hospitalized, hence would not able to concentrate on my project. I will have to start again on this and come with initial design to get feed back from cranky and others.

Will comeback soon...
 
is your problem.

The only reason to put an R in there would be to control inductiance rise in the woofer/FR at HF, but the Philips has very little of it, so a zobel is not strictly necessary. a textbook RLC will put a resistor in parallel with the VC, causing an impedance dip (I can comment only after the schematic tho). Use a regular CLC||R (C series, L series, C || with R for zobel) and you'll be fine. The parallel LC is not required for a driver like the Philips as it extends well on both sides of the crossover.

Effective impedance of 4.2ohm for a 8 ohm nominal is borderline, but your amp should be able to handle it unless it's a terribly limited design. The driver alone would dip to 5.5-6 ohms in some places, so an 8 ohm nominal amp should be able to work with it.


Hello Cranky,
I have decided not to use R for the filter circuit. Since I am planning to cut the LPF between 2 - 3Khz and the FR will be using complete signal, I can afford cut-off error for Woofer due to its inductance.

Now, if I want to use two FR's, do series connection is suggested? I have listened with 2 FR's connected in series and then in parallel (same music being played), I did not notice any difference in audible quality and sound level in both cases. However I read that series connection degrades hi-frequency respose. I understand the motor effect of one coil may influence the other, but how worse it could be for a philps driver? Can you comment on this.
 
There is no difference from an electrical point of view, except that a 16 ohm impedance from two drivers in series may not allow the amplifier to operate at optimum conditions.

Two sound sources with identical output cannot be placed close to each other due to lobing. The highest frequency the two can reproduce is limited by their distance from each other and the distance from the listener. It is possible to reduce this effect by using a tall vertical line, but even then this is a fundamental limitation.

You can simply connect the two and series and play back some music. As you move your head along the axes, the sound will change considerably. This results in a narrow sweet spot and a defective (non-linear) frequency response.

If it is not audible to you, it probably doesn't matter anyway :)

When multiple woofers are used, it is common practice to cut one off at a low enough frequency so as to prevent lobes. Most of the time, this is called a x.5 way configuration (for a full range with a second one being cut off, it would be a 1.5 way, but the most common is 2.5 way). For very large woofers being cut off very low, it is not a problem to wire them any way, as the frequency can be low enough to prevent lobing - under 200Hz it is a non-issue.



Hi Cranky,
Actually the drivers I have is not 8ohms 25w rms, but 4ohms 25w peak ( will be around 14w rms). I did the measurements using the DMM (arrived yesterday) and found this data. The cone started to have excessive movement with constant sine wave applied (volume twisted gradually) between 2-3 volts range.
With 100w/ch the drivers will be meager to handle it, so had to combine multiple drivers to distribute the same. Also today I enquired about philips drivers in SP road here in bangalore, the only available stuff is 4ohms 15w rms.

How about connecting two FR in parallel and woofer in series with them, ie the parallel FR combo is connected to LP crossover network in series.
 
I'm not talking about electricity, I'm talking about radiated sound field.

Have a look at this: Lobbing What is it in layman's terms? - diyAudio

The simplest analogy is to think of a pond or pool when you throw in a stone. When one stone is thrown, the ripples are smooth and uniform. When you throw two, the peaks and troughs cancel each other at various points in their path. Now think of the two stones as your two drivers, and the pond as the air in your room. I don't know if there is a simpler way to explain it, and even this is gross oversimplification, that is it.

If the drivers you are using are not sufficient to achieve the required listening level, you can:

1. Use a different speaker, or

2. Cross the speaker over higher and add a subwoofer or bass driver. Cross it as high as the large driver can go without beaming (beaming is also explained in the same post - and is somewhat related).

Adding voltage using a sinewave generator is misleading. Music is not a sine wave, and generally the onset of distortion is audible.

In my experience, I have used the Philips 8" driver and it is sufficient to generate concert levels in mid-size rooms with around 10 watts of clean power.

The sound quality is trash, but it is loud enough for most people. you will never be pushing 100 watts into those speakers - you'll go deaf long before you do.

Cranky,
I thought of below solutions,
1. Place two drivers very close to each other and have cross-over at 3Khz.
Now question is which driver will receive the low end the bottom or the top.

2. Place the second driver at top firing position.

Your thoughts on this?
 
Generally, the 'tweeter' will be at ear height, so that will depend on which driver is closer to your ear when the speaker is placed.

3KHz is too high, I would try moving it to 600-800Hz. Ideally, the intelligent solution would be to calculate the baffle step frequency using Edge or similar tools, and cross over the lower speaker just at that point, so you get full baffle step compensation for good power handling, and the other driver carries on till the top end of the range.

The other solution is to have a constant directivity design with the second driver mounted to the back of the baffle, but this would require the speaker to be placed in the middle of the room.

Top mounted would work for a direct-reflecting sort of setup, but there is a possibility of the reflections overwhelming the clarity.

I guess it will come down to your personal choices.



The edge shows two frequencies, f1 --> 118hz and f2 --> 240hz. I am using the philips 9710 T/S parameters to calculate the tower dimentions of 1300mmx280mm ( I am planning to measure the T/S parameters though).
The edge graph show the peak of 2.5db spanning 500hz to 1000hz after which it rolls down to 0db for higher frequencies.
should I Cut-off at 500hz or 1000hz?
 
For your F1 and F2 frequencies you need 6dB/oct (single inductor) at 240Hz on the lower woofer. This is very very large inductor, so it's better you do it at amplifier level.


Cranky,
Any specific reason to choose 1st order filter, why not 3rd order?
 
A beautiful, well-constructed speaker with class-leading soundstage, imaging and bass that is fast, deep, and precise.
Back
Top