Quick observation:
My midrange has Le of 0.54, and Re of 7.8, the calculated Zobel is 9.75Ohm, and 5.7uF.
However, during actual testing of the speaker, a more desirable result came from increasing the resistance, 18 Ohms & 5.6uF. (I know the cap value should have been changed, I'm just exploring options)
Here's the question:
If I only want to 'slightly' correct the impedance rise of the midrange, I obviously can't change Le, but can I use a higher Re in the calculation, thus giving me 'less' of an impedance correction.
Midrange is the Accuton C158-8-085
I am trying to cross to the Tweeter at 2,800 Hz, and with the 'proper' Zobel, the mid drops too soon, even with smaller inductors.
My midrange has Le of 0.54, and Re of 7.8, the calculated Zobel is 9.75Ohm, and 5.7uF.
However, during actual testing of the speaker, a more desirable result came from increasing the resistance, 18 Ohms & 5.6uF. (I know the cap value should have been changed, I'm just exploring options)
Here's the question:
If I only want to 'slightly' correct the impedance rise of the midrange, I obviously can't change Le, but can I use a higher Re in the calculation, thus giving me 'less' of an impedance correction.
Midrange is the Accuton C158-8-085
I am trying to cross to the Tweeter at 2,800 Hz, and with the 'proper' Zobel, the mid drops too soon, even with smaller inductors.
Remember that the voice coil (VC) inductance is not a true inductance, so tweaking of the Zobel is sometimes required. One method is to calculate the effective inductance, L_VC, so that
R_VC + i omega L_VC = Z0
where Z0 is the measured (complex) impedance at omega = 2 pi f = 2 pi (2800 Hz).
Then compute C_Zobel based on this effective L_VC. In other words, adjust the capacitor in the Zobel to achieve an optimal result.
A midrange filter is usually a 4 element network, but a 6" mid with a big bump at 4kHz doesn't leap out as very suitable for a 2.8kHz crossover to the tweeter.
accuton® Carefully selected loudspeaker drivers.
You'd possibly be able to notch it a little at 5kHz or so, but I think it will still sound dreadful. A Zobel is not going to help much either. Impedance correction here is usually done with something like 9R and 1uF, a lot less than the Zobel capacitor 5.6uF value, because you only want it to kick in at crossover.
A well behaved 4" or 5" mid will be a lot easier. Why? Mainly because the cone resonance will be a lot higher, and you don't have much series inductance to get some steep filtering with a midrange filter. It's also worth mentioning that a bass/mid is not the same animal as a specialist midrange. A specialist midrange has flatter response, has a stiffer surround, and has lower voicecoil inductance from a shorter voicecoil. 🙂
accuton® Carefully selected loudspeaker drivers.
You'd possibly be able to notch it a little at 5kHz or so, but I think it will still sound dreadful. A Zobel is not going to help much either. Impedance correction here is usually done with something like 9R and 1uF, a lot less than the Zobel capacitor 5.6uF value, because you only want it to kick in at crossover.
A well behaved 4" or 5" mid will be a lot easier. Why? Mainly because the cone resonance will be a lot higher, and you don't have much series inductance to get some steep filtering with a midrange filter. It's also worth mentioning that a bass/mid is not the same animal as a specialist midrange. A specialist midrange has flatter response, has a stiffer surround, and has lower voicecoil inductance from a shorter voicecoil. 🙂
Attachments
Last edited:
I agree that crossing the C158-8-085 to a tweeter at 2800kHz is unwise. However, to suggest that a typical midrange filter has only 4 elements is somewhat optimistic! I would say 6-8 components is more typical for a 3-way.
Can I ask what tweeter is being used?
I'm never too old to learn.
System7, thank you. I just figured something out. The capacitor value can correspond with the area you want corrected. The 1uF figure you gave is very close to the C of my LCR. I will go over my notes to see my exact LCR values.
I am using an infinite 'T' slope crossover, so the cone breakup isn't noticeable. I have a very steep drop.
jcandy, I am using the Accuton D20 3/4" diamond tweeter. It doesn't like to be crossed as low as 2500 Hz, making my first speaker with the C173 midrange not perfect. This is the reason I changed from the C173 to the C158.
Crossover parts are mostly Duelund CAST PIO's (caps and inductors), Mundorf 20W resistors.
System7, thank you. I just figured something out. The capacitor value can correspond with the area you want corrected. The 1uF figure you gave is very close to the C of my LCR. I will go over my notes to see my exact LCR values.
I am using an infinite 'T' slope crossover, so the cone breakup isn't noticeable. I have a very steep drop.
jcandy, I am using the Accuton D20 3/4" diamond tweeter. It doesn't like to be crossed as low as 2500 Hz, making my first speaker with the C173 midrange not perfect. This is the reason I changed from the C173 to the C158.
Crossover parts are mostly Duelund CAST PIO's (caps and inductors), Mundorf 20W resistors.
Billy,
designers use Zobels whenever they feel it helps them shape frequency
response curve. If not, they don't. Measure it and see how it looks. Then
you will know what's to be done.
designers use Zobels whenever they feel it helps them shape frequency
response curve. If not, they don't. Measure it and see how it looks. Then
you will know what's to be done.
The problem occurs in the calculation of the impedance and phase correction network in conjunction with how the Le and Re of a speaker is measured.The oft used dubious Zobel formula is the culprit here.The Smith & Larson Woofer Tester is the simplest method of quickly and accurately obtaining the values of the components at a specific frequency to form a Boucherot cell .This network corrects the impedance and phase angle to give the desired resistive load.
Midrange is SEAS MCA15RCY, tweeter is Monacor DT-254. Crossover frequencies are 500Hz, 4Khz pure 2nd order filters in this 3-way speaker.
I used for Zobel at the midrange 1,5ohm+10uF because people and speaker designers mostly used these values. Somebody used 1,8 ohm+9,2uF other used 1,2ohm+8,2uF or 1,5ohm+12uF.
But according to an online calculator need 7,2ohm+14uF.
My question why people not used values according to the suggested calculation? Other question with my values I cannot damage the midrange or the amplifier? Am I safety?
What exactly the R and C influence? Is it really better to keep the suggested values or not?
Thank you.
I used for Zobel at the midrange 1,5ohm+10uF because people and speaker designers mostly used these values. Somebody used 1,8 ohm+9,2uF other used 1,2ohm+8,2uF or 1,5ohm+12uF.
But according to an online calculator need 7,2ohm+14uF.
My question why people not used values according to the suggested calculation? Other question with my values I cannot damage the midrange or the amplifier? Am I safety?
What exactly the R and C influence? Is it really better to keep the suggested values or not?
Thank you.
Last edited:
The quoted formula for a Zobel network to correct a speaker's impedance is nonsense . The various on-line calculators that rely on it are also in default. You need to find the Re and Le at the crossover frequency and calculate the network using vector analysis, or if you can get hold of the S & L Woofer Tester ( post #7 above ) use it to obtain the compensation values.
This is where simulation really helps.
The calculators for Zobel filters are pretty imprecise, and subject to fudging.
It's important to understand you have 2 goals:
1 - Ensure the proper function of the low pass filter
2 - Get correct phase matching.
To do 1, you pretty much just have to make sure the impedance of the driver appears flat-ish. Increase the capacitor until the effect of the filter is low enough so that the Le of the driver is out of the way of the low pass filter. Adjust the resistance to adjust the new shelf at the top.
Ok, so you have the "perfect" Zobel. Still, the uF can be adjusted to improve the phase matching if you are having trouble otherwise. It's a little tweak to save for the end.
The calculators for Zobel filters are pretty imprecise, and subject to fudging.
It's important to understand you have 2 goals:
1 - Ensure the proper function of the low pass filter
2 - Get correct phase matching.
To do 1, you pretty much just have to make sure the impedance of the driver appears flat-ish. Increase the capacitor until the effect of the filter is low enough so that the Le of the driver is out of the way of the low pass filter. Adjust the resistance to adjust the new shelf at the top.
Ok, so you have the "perfect" Zobel. Still, the uF can be adjusted to improve the phase matching if you are having trouble otherwise. It's a little tweak to save for the end.
While I agree those formulas are sloppy as hell, you can get ideal with nothing more than XSim and a little trial and error. 🙂
BTW, I wrote about the Zobel here, with charts showing what happens when you get it right, and wrong and a link to a good simulation file to try things with.
A Speaker Maker's Journey: Crossover Basics - The Zobel
A Speaker Maker's Journey: Crossover Basics - The Zobel
Last edited:
Thanks for comments.
Well I really don't have expereince on this. Can somebody help me out with a simulation?
I was thinking maybe a 1,5ohm is too low? Probably better idea to use around 7ohm instead to be safety and follow the online calculator value instead?
Well I really don't have expereince on this. Can somebody help me out with a simulation?
I was thinking maybe a 1,5ohm is too low? Probably better idea to use around 7ohm instead to be safety and follow the online calculator value instead?
karoly 7603, the resistor values that you mention in your first post do not appear to be anywhere near a ball park figure. perhaps if someone can supply a graph of the impedance and phase curves vs frequency for the SEAS MCA15RCY then, the network values can be calculated for you. I would add that the 4 kHz crossover to the Monacor tweeter seems a little high given the off axis frequency response curves for the mid-range driver. Perhaps a 2.5 to 3kHz crossover point would be better.
Note Speaker Dave's words of wisdom in posts 18,22 & 29 of this thread:
Tweeter Zobel, or more accurately, RC network.
Tweeter Zobel, or more accurately, RC network.
The resistor value should be at or above the minimum impedance of the driver.
Why don't you look at the link I sent and play with the speaker example??
Why don't you look at the link I sent and play with the speaker example??
Thanks for the comments. Here somebody used the Seas MCA15RCY
Poor Man'
or here http://gxiberras.free.fr/hifi/Astasia/Filtre/filtre_Astasia_III_89dB.GIF
and here jii-LS03 "Lumiukko" - Blogit - Hifiharrastajat.org
also here http://www.3dar.ru/publication/3 WAY PROJECT.pdf
I downloaded the LM-1 zip and also the Xsim software.
I have never seen this before but I like it. I was playing with the values but this is not relevant for me at this moment but roughly I see what is doing.
Probably was my misstake to see other people's crossover. I think I will use around 7ohm resistor and I let the current 10uF capacitor.
Poor Man'
or here http://gxiberras.free.fr/hifi/Astasia/Filtre/filtre_Astasia_III_89dB.GIF
and here jii-LS03 "Lumiukko" - Blogit - Hifiharrastajat.org
also here http://www.3dar.ru/publication/3 WAY PROJECT.pdf
I downloaded the LM-1 zip and also the Xsim software.
I have never seen this before but I like it. I was playing with the values but this is not relevant for me at this moment but roughly I see what is doing.
Probably was my misstake to see other people's crossover. I think I will use around 7ohm resistor and I let the current 10uF capacitor.
Last edited:
Hi VaNarn, really better a bit lower cross, like you told. A 3Khz is looks optimal.
Otherwise in which software is possible to simulate and create the crossover? I also would like to learn. Is the Xsim suitable?
I looked at your reference material and it appears you are trying to adapt these to suit your xover needs. It is not the way I would go about it. To follow a published plan involves following them accurately, using the same parts ( or a supplied kit of parts ) as the original. However you really need to listen to a finished system first , to know if it suits your tastes in music and other requirements such as looks. I prefer using crossover networks using the minimum amount of parts to avoid interactions between them. This means starting with the low order text book crossover circuits and adding impedance correcting networks ( Zobel ) and L pads as required. Low order xovers do need care in the selection of the speakers used as the operating conditions can be severe.
After a lot listening and simulation test Zobel values at the woofer 8,2ohm+7,5uF and the midrange 8,2ohm+6,8uF.
I hear as I increase the resistance the crossover point is better for the ears at the midrange/tweeter. At 9,1ohm or even 10ohm the simulation shows flat frequency response. Can I go higher than 8,2ohm?
At the woofer bigger value seems a bit flatten/smoothing out the hump. Can I go here also up to 10ohm? Peoples, calculators, forums suggests max. the drivers impedance. All the drivers impedance are 8 ohm. Thanks.
I hear as I increase the resistance the crossover point is better for the ears at the midrange/tweeter. At 9,1ohm or even 10ohm the simulation shows flat frequency response. Can I go higher than 8,2ohm?
At the woofer bigger value seems a bit flatten/smoothing out the hump. Can I go here also up to 10ohm? Peoples, calculators, forums suggests max. the drivers impedance. All the drivers impedance are 8 ohm. Thanks.