Originally Posted by Mi10

Nitin,

Fully Agreed. However, think you got me wrong.
My intention was never to use a 3 way crossover here.

1. Midbasses will run Active
2. Woofer will run active
3. For the Tweeter + Midrange - I was thinking of using a 4" 2 way Components (4" Midrange + 1' Tweeter + 2 way Passive Xover @ 3200 Hz) - This is a readymade, of shelf item - (Morel Dotec Ovation 4" Compos)

Alternatively, for the 3rd point, I can source independent Drivers for Midrange and Tweeter and Build my own Crossover

Hope i am clear in my query. Now, let me know which is better and why

okay - with all of the above explained i can see no problems with either method - and it should work perfectly well with no issues in terms of the processing and tuning required

doing your own combination will allow you to possibly buy higher end speakers than a dotech ovation set but if you do buy the set that you suggest then it should work as you envisage it - also to me having that 4" component set will reduce the complexity of having to build your own passive crossovers - so it seems to be an excellent alternative to buying the really high end tweeter for a two way front stage instead

the question then is no longer about crossovers and time alignment - it becomes more of can the morel midrange drivers work free air or do they need enclosures and if they do then will that work up in your a-pillars - if those will work in your a-pillars ...... then absolutely not a single problem that i can see at all - if you are able to keep the midrange and tweeter driver close enough to each other to not need separate amounts of T/A

[Mi10]

Just checked the Mounting Depth and the DP mid range / bass driver is a whole 2" more than the mounting depth of my earlier Hertz EM 100!!!

Now i am having serious concerns if it will fit in the A-Pillar Pod.

So i ll be better taking the other Alternative. Which is, Source independent Mid Range and Tweeter Drivers and Design my own Crossover

I was just looking at the Morel CDM 54
Key specs as follows:
CDM 54
Power - 100 RMS (300 peak)
Fs - 300 hz
Response - 600 Hz-7000 hz
Sensitivity - 91 db
This is having a very steep roll off after 6000 hz @ 30 degrees

I ll most probably be going after the CDM 54. I want a methodology for designing crossovers circuits for following option:

Midrange(BP): 1000 Hz to 5000 Hz @ 6 db slope
Tweeter (HPF): 5000 Hz @ 6 db slope

I understand that adding a simple 'X' farad capacitor in series between the Midrange and Tweeter connection can help me HPF at 'X' frequncy @ 6 db. I dont want a complicated Circuit, i am ok with 6 db crossover even for tweeters (my Tweeter Fs is less than 1 kHz)

However my main question is;
How do i cut off the Midrange response at 5000 hz ?? Do i need an Inductor in series between the Amp and the Midrange ?

Looking forward for your inputs, and Thanks in advance Gurus

[Mi10]

In accordance with the below option; will the attached figure work as intended ?

[navin]

Originally Posted by Mi10

In accordance with the below option; will the attached figure work as intended ?

basically what you have is a 1st order at 4kHz. I'd suggest movig the XO point of the tweeter a bit higher (say 5k). Think 6.8uf for the tweeter and 0.2mH for the woofer.

[Mi10]

Originally Posted by navin

basically what you have is a 1st order at 4kHz. I'd suggest movig the XO point of the tweeter a bit higher (say 5k). Think 6.8uf for the tweeter and 0.2mH for the woofer.

Thanks, Navin.
But why would i do a 0.2 mH for a midrange ?. I can go well down till 6 Khz right ?. In that case, wont i need a 0.1mH Inductor ?


What about a Second order cross over for the tweeter,
say a 4.7 micro Farad in series and a 0.15 milli Henry in Parallel ?Can some one let me know whats the X'over frequency i am looking at in this case, please ?

[Mi10]

Originally Posted by navin

Think 6.8uf for the tweeter and 0.2mH for the woofer.

Originally Posted by Mi10

Thanks, Navin.
But why would i do a 0.2 mH for a midrange ?. I can go well down till 6 Khz right ?. In that case, wont i need a 0.1mH Inductor ?

Just finished a chapter on filters from my Digital Signal Processing Text book.

I think i figured out the reason why you had suggested 0.2 mH for the Mid bass / Midrange.

is it becasue of this reason (pls correct me, if i am wrong)
Since its a first order Xover (6 Db) it will still send in signals (at a reduced db) till its next octave. Example: If i use my Xover frequncy as 3 Khz, @ 6 db slope/ocatve, it will send frequncies till 6 Khz (however the loudness (db) of the signal reduces constantly till it reaches 6 Khz.

Please tell me if this is a right understanding.

If this is right, then

Can i have a second order cross over for the tweeter,
say a 4.7 micro Farad in series and a 0.15 milli Henry in Parallel

and a 0.22 mH Inductor for the Mid Range ?

this translates into the following crossover points
Mid Range - LPF 2.9 Khz @ 6 db slope (Signal starts going down constantly till 5.8 Khz, after which my Tweetes, pick it up)
Tweeter - HPF 5.8 Khz @ 12 db slope (Signal starts coming in at 3 Khz @ -12 db - which i think is safe enough; As my Fs of the tweeter is well below 1 Khz)

Please let me know Guru's. My head is spinning.

is it becasue of this reason (pls correct me, if i am wrong)
Since its a first order Xover (6 Db) it will still send in signals (at a reduced db) till its next octave. Example: If i use my Xover frequncy as 3 Khz, @ 6 db slope/ocatve, it will send frequncies till 6 Khz (however the loudness (db) of the signal reduces constantly till it reaches 6 Khz.

okay - im not going to get into the coils and caps part if it

but lets only talk about the crossover slope and how it works

say for simplicity's sake you have a midrange speaker which has a frequency response from 200hz to 8000hz this means that it can play from 200hz to 8000hz but at the absolute limits you cant really play it loud - so to reduce that you want to be one octave away from the limits so you decide to bandpass between 400hz and 4000hz - so the speaker will play in a linear manner from 400hz of 4000hz let us say for example again at 100db SPL

now once again lets concentrate the example to the upper low pass of the speaker at 4000hz - if you use a first order filter the frequencies above 4000hz will reduce by 6db when the curve reaches 8000hz ........ so while the speaker plays 400hz to 4000hz at 100db you will still be getting 8000hz playing but 6db lower ie at 94db

then if you used a second order filter the slope then becomes 12db/octave so while 400hz to 4000hz play at 100db you will have 8000hz playing at 88db - and if you used a third order filter this then makes the slope 18db/octave so that 8000hz will still be playing but at 82db and with a fourth order filter the slope is 24db/octave and hence that 8000hz will still be playing at around 76db

when that 4000hz plays at 100db and the 8000hz plays around 20db lower you dont hear it at all - so the higher slopes are generally considered better

but the story doesnt end there because changing the slope also changes the phase as well - and phase is not only just the difference between changing positive and negative - thats polarity - and swopping polarity only changes phase by 180 degrees - phase has more to do with the arrival times of the wave form to its destination ie your ear - its all about whether the wave is in a crest or trough and how high up the crest or how low down the trough and this is measured in degrees - so differing slopes will change the phase as well - so you cant just say im going to have one slope at 6db/octave or another at 12db/octave because you will still have the same identical frequencies coming from both the drivers because the tweeter plays that range at a rate thats reducing according to the slope and the midrange also plays the same frequency at a rate thats reducing according to the slope - have those at locations that have way differing pathlengths then those frequencies could land up at the destination ie your ear totally out of phase and cause cancellations of those particular frequencies that you wont be able to EQ out ie even if you raise that particular EQ band to become louder the null causes it to cancel even more and that frequency basically ceases to exist in your system

so designing crossovers is not as simple as just placing a set of caps and inductors in series with the speaker - you have to look at the complete picture and how well the speaker plays at its limits and at point distortion will set in and what slope actually works better with a particular driver - for example you could be using that same speaker we used as an example above and with a steeper slope ie at 24db/octave it might even be able handle playing up to 6000hz at that 100db but at a 6db/octave slope you might need to actually crossover at 3000hz

all of this is theory though - yet to be honest i would be far happier if you used second order filters for your crossover ie a 12db/octave crossover - which makes it a bit more complex but it does help a lot IMHO in preserving your speakers

this is why if you want a three way front stage plus sub (effectively a 4 way setup) its far better to be doing it off a proper processor which gives you the 4 way crossover and adjustable slopes as well as the T/A to compensate for the pathlength differences - i reckon a JBL MS8 or the audison bit one is ideal but since you already have the headunit performing most of those functions for you - you will have to compromise somewhere along the line

i understand that the speaker depth was also a problem and that the smaller driver is hence more effective but maybe you could also consider rebuilding that A-pillar and just buying that component set instead - it makes it easier in terms of the crossover because you do get a crossover designed for the particular parameters of those speakers but then again if i was faced with your predicament i would also want the CDM54 or CDM88 and a good tweeter (like the ones you said you were after) instead of the lower end component set

oh and one more small thing - it depends on the tweeter you get but for example the piccolo has an impedance of 6 ohms not the typical 4 ohms so you have to also take that into account for the capacitor you use (caps and inductors if you go to a higher order filter)

in order to know more about passive crossovers - dont relate it too much to digital signal processing - an excellent reference for crossover building is still "the loudspeaker design cookbook" by Vance Dickason and you can also google lots of webpages on speaker crossover design - those most of those will have tables that give you the circuit design and the different values of the caps and inductors required for each slope as well

[Invinsible]

In short to what Nitin has mentioned:
1. Decide on what drivers you want on the basis where you going to place them and what freq you want to run them on.
2. If you plan to do the 3way plus the sub. decide on what flexibility you can have in terms of freq range and ease of installation.
3. For example I would choose to run midbass with 63hz HPF at 18db to 500hz LPF at 12db. With Midrange and tweeter running at Apiller choose for midrange that go down to 300hz and run them at 500hz HP at 12db to 4khz at 12db. Tweeters could cover above 4khz at 12db or 18db HP.
Most male vocal and femal vocals can go upto 6khz so good idea is to keep the tweeter as close as possible with the midrange.
4. Also keep in mind as Nitin mentioned there is phase shift that may occur which 90deg at 6db slope, 180 deg at 12db, 270deg at 18db, 360deg at 24db slope. Whatever slope you decide make sure you keeps this mind as it will play a large role on the final result will sound. The LPF set on midbass should be in phase with HPF on the midrange and likewise midrange with tweeter.
5. To make it simpler and run all active you could invest in a processor like Rf360.2 which could give you the flexibility to run mid and tweeter active on it with midbass and sub running active on the Headunit. You can get a good deal on it at ebay. The options and detail i had discussed on the clarion dxz78g thread.
6. Designing a xover is not as simpler as it looks. If not done right could be harmful for the drivers in long run and could be more complicated in final tuning.

[navin]

Originally Posted by Mi10

But why would i do a 0.2 mH for a midrange ?. I can go well down till 6 Khz right ?.
What about a Second order cross over for the tweeter

Originally Posted by Mi10

Just finished a chapter on filters from my Digital Signal Processing Text book.

Example: If i use my Xover frequncy as 3 Khz, @ 6 db slope/ocatve, it will send frequncies till 6 Khz (however the loudness (db) of the signal reduces constantly till it reaches 6 Khz.

Can i have a second order cross over for the tweeter,
say a 4.7 micro Farad in series and a 0.15 milli Henry in Parallel

1, You have alredy read more than I have. So stop reading and start doing.

I know of David Weems and Vance Dickansons books. A old friend even made a copy for me but I lost them before I got around to reading them. Kya Karun I am like that only.

2. Yes a 3kHz 6db/oct Xo will be 6db down at 6kHz and 12db down at 12kHz.....

3. You can mix slopes. Remember I have not see the driver curves. The Final slope will depend on NOT only the electrical slope of the XO but also the acoustic slope of the driver so if the driver is rolling off 6db/oct from 3kHz then the total slope for the midbass (along with the 6db/oct filter) will be 12db/oct. Phase is dependant on this combined slope (provided the drivers are on the same plane). Remember that in a car the woofer and tweeter are not always on the same plane.

Sadly all drivers do not such linear slopes. Thankfully our ears are not that sensitive outside 200-4000Hz (or 300-3000Hz for old foggies like me).

Originally Posted by naughty001

- you have to look at the complete picture
- yet to be honest i would be far happier if you used second order filters for your crossover ie a 12db/octave crossover

"the loudspeaker design cookbook" by Vance Dickason

a. In theory Naughty is right.

b. 2nd order XOs offer a lot more protection than 1st order XOs but at the cost of that phase shift thingy Naughty referred to. Now the twist phase the percieved audible effect shift is not only dependant on the driver's phase response and the XO's phase response but also the relative distance of the tweeter and woofer from the ear. Phase and time are related na? They are not the same but they translate into each other.

c. Confused? dont be. Try the 1st order XO. Increase volume slowly, very slowly. At the first hint of any strain turn it off. Tweeters are apt to blow without any audible warning. If you want ot be safe try a 4.7uf cap instead of a 6.8uf. There might be a slight (1.5db) hole between 3k and 6K if you do but trust me in a car with those ricshaws honking you wont notice.

Originally Posted by Invinsible

6. Designing a xover is not as simpler as it looks. If not done right could be harmful for the drivers in long run and could be more complicated in final tuning.

Fair enough guys but lets not scare people. If I knew all of this theory when I first got hold of a soldering iron (I was 7 or something) I might never have started making all those speakers. Thank God I was ignorant. In a way it helped. I blew a few drivers ...ok I blew up a lot of drivers but every blow up taught me something.

[Mi10]

Many Thanks, Naughty, Invinsible and Naveen for your inputs. I am going to read more on "concept of Phases in multilevel signals" and then get back with a Xover design.


However, as Naveen pointed out, the response characteristics of the driver involved too is very important. The link below gives you the frequncy response of the Driver
(mid range) i am considering.

http://www.morelhifi.com/support/pdf/mobile/CDM54.pdf

Driver the Response seems to have a steep roll off after 6K.
Can you please let me know, if i will ever need a Low pass cut-off (Inductor Circuit), if i am to HPF my tweeters at 6 Khz and above.

Note that both my Midrange and Tweeters will be mounted on the A-pillar with approx 40-50 mm distance between them. and hence can i presume the concept of Phase / Time Alignment (between the midrange and tweeter) might actually be negligible if not null

Thanks again for all your support and guidance

[Invinsible]

Originally Posted by Mi10

Many Thanks, Naughty, Invinsible and Naveen for your inputs. I am going to read more on "concept of Phases in multilevel signals" and then get back with a Xover design.


However, as Naveen pointed out, the response characteristics of the driver involved too is very important. The link below gives you the frequncy response of the Driver
(mid range) i am considering.

http://www.morelhifi.com/support/pdf/mobile/CDM54.pdf

Driver the Response seems to have a steep roll off after 6K.
Can you please let me know, if i will ever need a Low pass cut-off (Inductor Circuit), if i am to HPF my tweeters at 6 Khz and above.

Note that both my Midrange and Tweeters will be mounted on the A-pillar with approx 40-50 mm distance between them. and hence can i presume the concept of Phase / Time Alignment (between the midrange and tweeter) might actually be negligible if not null

Thanks again for all your support and guidance

CDM54 is one of the finest Midrange driver available. These have slight better off axis response compared to the CDM88. However if you look at the freq response on both, the CDM88 can be played lowered and can be crossed higher compared to the CDM54. For example I would prefer to cross the CDM54 at 6/12db @ 600hz LP and HP at 4khz @ 12db. Incase of the CDM88 I could go down to 400hz @ 6db LP or 500hz 12db LP TO 5000khz @12db or 4000khz@6db.
Freq. response will differ on how you place. Best compromise is to have them at 30 or 45deg off-axis, rather than having them on axis since you placing them on Apiller.

Set the tweeter at 4khz or 5khz at 12db HP. With CDM54 ideally 4khz and with CDM88 at 5khz.
Since the Tweeter is placed next to the mid I would rather choose to cross them at 4khz at steeper slope @18db.

[navin]

Originally Posted by Mi10

http://www.morelhifi.com/support/pdf/mobile/CDM54.pdf
Driver the Response seems to have a steep roll off after 6K.

Look at the 30deg off axis response. It falls about 10db/oct from 4k to 8k. that is what I would consdier it's HF roll off.

Originally Posted by Invinsible

I would prefer to cross the CDM54 at 6/12db @ 600hz LP and HP at 4khz @ 12db.

I agree. I would XO the CDM54 at 700/4000Hz @ 12db/oct. The Fs is 400Hz and I would like to have the LP XO at about 2x Fs to be safe esp since I do not know the capabilites of the driver's suspension (Xmax and all that).

The data sheet does not have phase response so I can tell but Morel drivers tend to have pretty decent phase characteristics in the pass band (in this case 700-4kHz).

[Mi10]

Originally Posted by navin

Look at the 30deg off axis response. It falls about 10db/oct from 4k to 8k. that is what I would consdier it's HF roll off.



I agree. I would XO the CDM54 at 700/4000Hz @ 12db/oct. The Fs is 400Hz and I would like to have the LP XO at about 2x Fs to be safe esp since I do not know the capabilites of the driver's suspension (Xmax and all that).

The data sheet does not have phase response so I can tell but Morel drivers tend to have pretty decent phase characteristics in the pass band (in this case 700-4kHz).

Thanks, Navin and Invinsible

Which means i ll have to design a second order LPF with following values;
L - 0.22 mH
C - 7 µF

Now regarding the Tweeter;

Please find the link below;
http://www.morelhifi.com/support/pdf..._MT22_MT23.pdf

I was having the MT-22 Tweeter in mind.
Can i have a second order HPF circuit for this too ? Which means its a same values

L - 0.22 mH
C - 7 µF

Originally Posted by Mi10

Thanks, Navin and Invinsible

Which means i ll have to design a second order LPF with following values;
L - 0.22 mH
C - 7 µF

Now regarding the Tweeter;

Please find the link below;
http://www.morelhifi.com/support/pdf..._MT22_MT23.pdf

I was having the MT-22 Tweeter in mind.
Can i have a second order HPF circuit for this too ? Which means its a same values

L - 0.22 mH
C - 7 µF

a high pass and a low pass wont be the identical circuit as you outline - it will definitely be different so something is wrong there

also if you really want to learn about phase the best explanation i have come across of it is in a book by F.Alston Everest which is entitled "Master Handbook Of Acoustics" - im sure you will be able to find an e-book version easily

I blew a few drivers ...ok I blew up a lot of drivers but every blow up taught me something.

im not 100% sure that Mi10 would want to blow up the drivers he is thinking of using here LOL

[Mi10]

Originally Posted by naughty001

a high pass and a low pass wont be the identical circuit as you outline - it will definitely be different so something is wrong there

Nitin, i didnt mean the same circuit - I was just refering to the Component values for C and L.

Circuit is obviously different and HPF and LPF (Infact. it ll be exactly opposite with same Components)

LPF - L is series and C in parallel
HPF - C in series and L is parallel

Originally Posted by naughty001

im not 100% sure that Mi10 would want to blow up the drivers he is thinking of using here LOL

I ll be a dead duck, if something happens like that