Do you expect a Zobel network to improve the bass extension? https://diyaudioprojects.com/Technical/Speaker-Zobel/
Attenuating the Midax and Trebax further may well be beneficial to the overall balance.
Attenuating the Midax and Trebax further may well be beneficial to the overall balance.
The zobel will counter the rising impedance of the woofers as the frequency increases which makes the -3dB point of the low pass be closer to spec therefore making it so the woofers and midrange overlap less in the frequencies they reproduce.
In my experience when I've added a zobel to an existing speaker I built I've usually had to attenuate the midrange some which usually requires attenuating the tweeter as well.
In my experience when I've added a zobel to an existing speaker I built I've usually had to attenuate the midrange some which usually requires attenuating the tweeter as well.
Of course if you achieve the wanted response without the impedance compensation it's just the same.
You asked about using separate components per woofer, it shouldn't make a difference.
You asked about using separate components per woofer, it shouldn't make a difference.
With one woofer the zobel components are 2.2uF 10 ohms.
With the two woofers in series the components would be 1.1uF and 20 ohms.
I don't have a 1.1uF cap, but I do have two .56uF orange drop caps which makes 1.12uF. The measured value is relatively close.
The 20 ohm resistor I can put two 10 ohm resistors in series.
So I'll try that tomorrow and see if I need to drop the level of the midax and trebax.
What I need to do is measure the resistance of the drivers and draw a schematic with the added l-pad components for the midax and tweeter and the zobel for the woofer. That way I can calculate what the attenuation is and calculate what resistors I need. I'll start with 1db lower on the midax and tweeter.
Only issue with measuring the driver impedance is the other KG-4 tweeter I have one of the terminals broke off and trying to solder to the wire didn't work so I cannot measure the impedance of that driver. For the l-pad on that driver I used a combination of a typical l-pad with a series and parallel resistor then a series resistor to bring the impedance up to 15 ohms. The KG-4 tweeter is sensitive enough to where even after the series resistor an l-pad was still needed to bring its level down.
Of course the RCA woofers likely aren't all that efficient given they likely came from a console stereo. Now had they come from a tube console they likely would have been a bit more efficient.
With the two woofers in series the components would be 1.1uF and 20 ohms.
I don't have a 1.1uF cap, but I do have two .56uF orange drop caps which makes 1.12uF. The measured value is relatively close.
The 20 ohm resistor I can put two 10 ohm resistors in series.
So I'll try that tomorrow and see if I need to drop the level of the midax and trebax.
What I need to do is measure the resistance of the drivers and draw a schematic with the added l-pad components for the midax and tweeter and the zobel for the woofer. That way I can calculate what the attenuation is and calculate what resistors I need. I'll start with 1db lower on the midax and tweeter.
Only issue with measuring the driver impedance is the other KG-4 tweeter I have one of the terminals broke off and trying to solder to the wire didn't work so I cannot measure the impedance of that driver. For the l-pad on that driver I used a combination of a typical l-pad with a series and parallel resistor then a series resistor to bring the impedance up to 15 ohms. The KG-4 tweeter is sensitive enough to where even after the series resistor an l-pad was still needed to bring its level down.
Of course the RCA woofers likely aren't all that efficient given they likely came from a console stereo. Now had they come from a tube console they likely would have been a bit more efficient.
Last edited:
It's not a critical circuit and it will work as expected with minor variations. It's another thing to discuss subtle tonal issues or left/right balance.
Agreed. For one speaker close enough will work.
Now for a stereo speaker it's likely close enough would work there as well given the tolerances of the drivers and crossover parts, however for stereo speakers I do tend to pay a bit more attention to part tolerances usually choosing tighter tolerance parts within reason and where appropriate.
I suspect the effect the cardboard box had on the sound of the woofers and lack of a zobel on the woofers likely caused me to have the midax and tweeter louder for it to sound balanced.
The wood baffle and adding the zobel will likely make me lower their levels which will make the bass sound better because the drivers are better matched to each other.
Now for a stereo speaker it's likely close enough would work there as well given the tolerances of the drivers and crossover parts, however for stereo speakers I do tend to pay a bit more attention to part tolerances usually choosing tighter tolerance parts within reason and where appropriate.
I suspect the effect the cardboard box had on the sound of the woofers and lack of a zobel on the woofers likely caused me to have the midax and tweeter louder for it to sound balanced.
The wood baffle and adding the zobel will likely make me lower their levels which will make the bass sound better because the drivers are better matched to each other.
Last edited:
I shouldn't think an additional attenuation of 1 dB would be audible. I suggest an additional 3 dB for starters.
I may hear much less than 1dB if the speaker and room balance is optimal, although there can be a number of factors involved.
It also depends on how much error there is in the level to begin with as to how much change can be easily heard.
It also depends on how much error there is in the level to begin with as to how much change can be easily heard.
Last edited:
Sound does pass through the screen, but I did learn initially with this speaker that the copper screen is not very acoustically transparent as I originally used a piece as a speaker grill and the woofers sounded lower in volume.
Not sure how the screen material will affect the sound in the screen room either as some will pass through and some will not.
Not sure how the screen material will affect the sound in the screen room either as some will pass through and some will not.
I didn't mean to open a can of worms, it may not be a helpful tangent.. but ordinarily the speaker itself is the first place to look for secondary sources of sound, followed by the design of the speaker to be a good fit for the room.
Here's the l-pad. Gonna try the new values and see if it works. The zobel did make me have to reduce the level of the midax and because it was well balanced with the tweeter I reduced both by close to the same amount.
Decided to test a midax with my DATS V3.
Based on the graph, do I even need a zobel for the midax?
If I do the values are 1.1uF and 11.8 ohms.
Based on the graph, do I even need a zobel for the midax?
If I do the values are 1.1uF and 11.8 ohms.
I don't know whether you'd want RLC Zobel compensation for the peak at 1k5, it's worthy of further measurement or simulation.
The impedance peak is relatively small so I'm not worried about RLC compensation.
Would I need a zobel for the rising impedance as the frequency increases or would the impedance stay low enough to where there's no audible effect of adding one?
Would I need a zobel for the rising impedance as the frequency increases or would the impedance stay low enough to where there's no audible effect of adding one?
For that I'd answer the same as my previous post. It is likely to be easier to find a way to do without it than for the peak, should it be an issue.
It's easy enough to test it and see if there's any audible difference. Just need two .56uF caps in parallel and a 10 ohm resistor and two 2 ohm resistors in series.
That won't do. If there's a difference in response then of course that could be audible, but what's the point of settling for an arbitrary response? (even if it's not yet known what it should be).
To make a fair comparison you need to find a way to cross with and without the RC which gives you the same response, and when you do you'll find they sound the same.
I hope my point finds it's way through 😉
To make a fair comparison you need to find a way to cross with and without the RC which gives you the same response, and when you do you'll find they sound the same.
I hope my point finds it's way through 😉