Help please crossover guys.

[martyh45]

I have been playing with the crossover in the A15. Varying the first series resistor helped with the too hot HP but dulled the treble too much. Faced with the prospect of learning complex theory and then applying it I decided to try something easier first :-) I simulated the crossover using an 8 Ohm resistor as a load and varied the value of the various components and looked at what they did to the transfer function. I tweaked away with standard values until I had a curve that looked like the inverse of what I was hearing. It will be at least spring before I can measure what I build but I did find a plot of the raw horn / driver frequency response.

So now for the questions?. Is looking at the voltage transfer function of the crossover a valid way to predict changes to the system FR? I realize that an 8 ohm resistor is not a very good model of the driver/ horn combo but is it close enough for this type of work? Lastly, assuming I am on the right track will the differences in phase angle shown in the attached plot have audible consequences or are they close enough to the stock to be unnoticeable. FWIW specified stock XO frequency is 630 Hz

The attachments?. First is the phase and voltage transfer of the various models. Stock with SW1 open (closed is high frequency emphasis position) Modified XO with switch in both positions. (I am using the switch in the modified version to switch in a bypass capacitor for the first series resistor)
http://i142.photobucket.com/albums/r...ps1fd619cc.jpg

And? the Publised FR and Z of the Horn/ Driver Combo
http://i142.photobucket.com/albums/r...psb42e348e.png

Thanks in advance
M

[cradeldorf]

I don't have a clue what your saying, but as far as the phase is swinging in your first picture, have you tried reversing the HF leads?

[Alien_Shore]

two things:
1) what simulator are you using to predict the transfer function of your mods?
2) the simulation will be greatly improved if you provide it with the actual impedance data and FR data of your drivers

I have been doing much the same kind of work here with various Altec components, but I have the imped and FR data loaded into Jeff Babgy's PCD. Jeff told me that you cannot predict FR from just looking at the impedance of your system...he's a guru in this, so I believe him, but I am also still searching to gain a better understanding of why this is. Specifically, for example there are calculations one can perform to calculate attenuation by knowing the dB cut value and a tweeter's nominal impedance. So in my mind there is a definite mathematical link between imped and FR of a system.

Do you have a schematic of your XO that you're working with?

[martyh45]

Hi Mike,
I am using Circuit lab. I have no Idea how to feed the program FR data from the drivers or if it is even possible. I basically listened to them and adjusted the value of R2. I simulated the voltage transfer function of the HP section and tried to correlate what I was hearing at different values to what I was seeing in the plots. I got to a point where I thought the voice range sounded right but the treble was too depressed. I adjusted the voltage transfer to follow the curve I was getting at the preferred value of R2 but changed values to get less attenuation over 5K. The good news is I was able to scrounge the parts I was missing from a local buddy so I built the junkboxmods version linked below. Just finished about an hour ago and my initial impression is that it did what I wanted it to do. So far so good, not sure what potential impact the change in phase response would be but imaging is fantastic. Feel free to play with these schematics on the site. I’ve saved private versions. One note, the resistors in series with the coils are the copper resistance of the coil not a separate component.

https://www.circuitlab.com/circuit/s...5-junkboxmods/
https://www.circuitlab.com/circuit/2kkt7h/a15hppublic/

[martyh45]

I don't have a clue what your saying, but as far as the phase is swinging in your first picture, have you tried reversing the HF leads?

So far no, they were unopened until I got them so I am reasonably confident that the polarity was correct.

[Alien_Shore]

Hi Mike,
I am using Circuit lab. I have no Idea how to feed the program FR data from the drivers or if it is even possible.

Great site, Marty - thanks for sharing that. I need to carve out some time to play with that. I see PCD being complementary to Circuit Lab. The advantage PCD will give is that it incorporates the actual FR and Impedance measurements from the drivers and allows highly accurate simulation of your system, including phase. You also have other speaker-specific variables you can adjust such as driver/baffle offset, so you can model off-axis responses. I'm still learning to use the tool, but it produces amazingly accurate models. Best of all, it's free if you already own MS-Excel.

[zelgall]

Ditto on PCD. I just downloaded it a couple of days ago and found that my Woofer Tester 2 can export the FRD and ZMA files PCD uses. I ran my 604-8G's and was able to fine tune my crossovers slightly, smoothing out a mid peak I was getting.
Published data is OK for a starting point but they're likely either averaged or optimistic.