Midrange attenuation l-pad or single resistor?

Sape

Member
2006-05-11 11:38 pm
I am making 3-way and was wondering how to attenuate midrange?
Mid driver is Tangband w5-2106. Need is to get about 1.5-2 db lower.

My question is that how to calculate l-pad impedance when driver impedance is 8.6 - 12.5 ohms between xo-points?
Or is single resistor better option located before crossover?
 
http://www.oaudio.de/en/Loudspeaker-drivers/Tang-Band/5-wide-range-drivers/TangBand-W5-2106-Wide-range-speaker.html

You can do it either way really. A resistor in front tends to have more effect in the passband than at the extremes.

An attenuator generally lowers the whole thing and keeps the shape. you might try 2.2R and 22R.

I would generally prefer a resistor in front, because it raises the overall impedance and makes the speaker an easier load. You might be looking at 2.2R or 3.3R. Nothing will break either way.
 
Last edited:
Hi
How did calculate this and do you mean that l-pad keeps the driver impedance curve the same?

A front resistor raises impedance but keeps impedances intact between xo-points?
An adjustable L-pad is a different animal from a precisely calculated 8 ohm attenuator.

I'm a bit rough and ready with attenuators, using whatever is handy in my parts box. Then I listen to it. For an 8 ohm driver, 2.2 and 22R is about right. or 3.3 and 15R for considerable attenuation. You get a feel for it. There are online calculators, but the idea is to let the filter see an unchanged 8 ohm load even with the resistors.

You model a speaker to know exactly what's going on. Just pick any old 5" woofer like a W130S and see how it works with your filter:
Downloads

Electrically, most woofers are much the same. I haven't looked at your tangband. An attenuator tends to flatten driver impedance, so can be more predictable. A series resistor changes the frequency response where impedance goes high. But like I say, nothing will break, so try whatever works.
 
Last edited:

Sape

Member
2006-05-11 11:38 pm
I am little stupid but is it so that adding say 2.2 and 22R before mid driver (and after xo) still keeps the measured mid driver impedance curve the same-that is from 8.6 to 12.5 ohms in its working area (330hz-3200hz)?
It only attenuates driver output about 2 db?
 
I just modelled it for you. 2.2R in front or a 2.2/22R attenuator is about 2dB reduction. Take your pick!

I have no opinion on what is precisely a better filter. Depends on the precise modelling which really includes the resistors if you do it properly.

Don't let the perfect become the enemy of the good. If the resistors cause problems, you must redesign the whole filter! But they won't.
 
Last edited:

Sape

Member
2006-05-11 11:38 pm
I just modelled it for you. 2.2R in front or a 2.2/22R attenuator is about 2dB reduction. Take your pick!

I have no opinion on what is precisely a better filter. Depends on the precise modelling.

when I calculate xo between mid driver and tweeter the mid driver impedance is now 12.5 ohm.
When I put serial resistor before crossover is midrange driver impedance still 12.5 ohm at xo-point 3200hz?
Same question with l-pad.

Ps. I cant write **** in english?
 
Last edited:
You're making it more difficult than it is. Every ohm in front of an 8 ohm (6 ohm DC) driver attenuates it about 1dB. That's how it works.

If it does too much or too little, you change it. I never try to work out the impedance, because it's not a pure resistance, and leads you astray.

Try the 2.2R. It's easiest. See how it goes. What is this? An off-the-shelf prebuilt crossover?
 

Sape

Member
2006-05-11 11:38 pm
You're making it more difficult than it is. Every ohm in front of an 8 ohm (6 ohm DC) driver attenuates it about 1dB. That's how it works.

If it does too much or too little, you change it. I never try to work out the impedance, because it's not a pure resistance, and leads you astray.

Try the 2.2R. It's easiest. See how it goes. What is this? An off-the-shelf prebuilt crossover?

No, I am just wondering when you calculate crossover you have to know driver impedances at xo-points.

Now because I already made crossovers ready I just want to know how this midrange attenuating (serial or l-pad) affects these midrange driver impedance readings at xo-points? They are 320hz and 3200 hz.

OR do I have to build my crossovers again just to attenuate this midrange?

I can make this pretty complicated but "stupid is what stupid does".
 
OMG, I feel I'm drowning here! :D

This W5-2106 driver looks like lowish inductance, but I can't find that Le voicecoil inductance to hand. Maybe 0.4mH just looking at the curve?

No issue at 320Hz. At 3200Hz it's becoming more inductive. But not too bad.

The rolloff filter at 3200Hz might need bit of resistance in the shunt. Hard to say. How's your filter looking so far?

BTW, I don't calculate crossovers, I model them. But any electrically similar driver (similar DCR and Le) should behave much the same if it's flat across the range of interest.
 

Attachments

  • Tangband W5-2106.JPG
    Tangband W5-2106.JPG
    51.1 KB · Views: 349
Last edited:
It's a wrong question.

The modelling looks like below for something I had to hand. The first is the 2.2R in front. The second is 2.2R and 22R. Maybe the attenuator is doing more at the frequency extremes as expected.
 

Attachments

  • 2.2R resistor in front.PNG
    2.2R resistor in front.PNG
    14.1 KB · Views: 503
  • 2.2-22R attenuator.PNG
    2.2-22R attenuator.PNG
    18.9 KB · Views: 213

Sape

Member
2006-05-11 11:38 pm
It's a wrong question.

The modelling looks like below for something I had to hand. The first is the 2.2R in front. The second is 2.2R and 22R. Maybe the attenuator is doing more at the frequency extremes as expected.

So the 2.2R and 22R is l-pad between xo and driver .It does not alter impedances at all.
I have 1.5R and 39R can I use them here?

Btw you really are a kind person to do this teaching.

Greetings from Kuopio Finland.
 
Let's give you an example. I just did a quick hack of this visaton stock HW 3/130 crossover above.
HW 3/130 NG - 8 Ohm

What actually works about right is the below circuit with an 8", 5" and 1", where I added some attenuation. In fact it's excellent with the right drivers. The W130S is like your tangband, but slightly different Le of 0.6mH.

1.5 and 39R might work just fine after the filter. What 1.5 db? Not very much in fact.
 

Attachments

  • Visaton Stock crossover HW 3-130.PNG
    Visaton Stock crossover HW 3-130.PNG
    9.8 KB · Views: 549
2.2R and 15R is about 4dB.

I was quite rough and ready with it.

3.3R and 22R would probably not have been much different, but higher and better overall speaker impedance.

In your mind you put 6 ohm DC speaker in parallel with 15R to get about 4.5R. then add the 2.2R for about 6.7R overall load, and the speaker gets about 2/3 of the voltage which is 4.5/6.7.

I ignore inductance.
 
Last edited:
We all start somewhere. :)

These visaton projects can just be placed in the projekte folder of Boxsim, and you are up and running in minutes!
http://boxsim-db.de/kategorie/systeme/drei-wege/

Visaton Boxsim

Once you get the idea, you can use any similar drivers (similar Le is important) and fiddle around with resistors to get level right.

Boxsim has an optimiser. You mark components in red and the computer works out the flattest result. It's very neat. So just mark your attenuators, and it adjusts them.
 
Last edited:
Hello there. I also have question regarding single resistor or L-pad but at the tweeter. Tweeter is 8 ohm. High-pass is very simple. There is a series resistor before the crossover, value 3,6ohm, after series a 4,7uF cap and shunt a 0,26mH coil. That's all. My question is shall I use an L-pad after the crossover, right next to the tweeter instead of the single series resistor? Is there any frequency response difference/sound difference/efficiency or any kind of difference between them? Or it really doesn't mean anything which version we use. Thanks.