Site Links
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Categories
In this Discussion
HF Peaks in Mids
Something I run across rather frequently is peaking at the very top end of a midrange (12khz+). Usually it's down low enough I don't concern myself with it too much. But this time it was something I was borderline concerned about. This is using a M4N as the mid but this problem isn't exclusive to this design/driver.
For reference the tweeter level is about 88db in this frequency range for this measurement, so this high frequency peaking is around -30db down. However I can knock it down to about -40db with a LCR across the mid. I guess my questions are two fold:
1) Would you bother treating it when it's -30db?
2) Is there a better way to address this than with the LCR? I haven't been able to defeat it with otherwise with just component value adjustments so far.
The LCR is only like $5 in parts that I have on hand, but I was trying to keep parts count to a minimum as this is just a little two way I'm doing for my wife. But my questions are both in relation to this design and more generally as well.
Comments
Unless your wife can hear it, there isn't much need to mess with it, IMO Mr. Impious.
My suggestion is to listen to just the mid with and without the LCR. If you can hear the ringing without the LCR, then use it, if not eliminate it.
30db is pretty far down.. Not usually down far enough to be considered inaudible, but it can depend how high the accompanying distortion is. If the distortion is low enough it may be a non-issue. Though some folks can be more sensitive to certain frequency ranges. That could be one that your wife can pick out, but you or I can't.
Thanks for the feedback!
Listen is always good advice. If you wanted to know more about that breakup mode and how much attention it needs by measurements alone you should measure distortion and waterfall. If that is ringing like a bell or is a huge distortion spike the notch might be a really good investment.
It is tough when dealing with untrained/inexperienced ears. Most folks will just accept the current status quo because they haven't experienced better to know better exists. Even if they are hearing things you can't, there is no way for you to know unless you change it and they are able to notice the difference. Though if you make it to that point, that is when the fun realy starts.
The primary breakup of the M4N is really at 6kHz, I'd address that above anything else. Use a parallel notch rather than series, as the last components before the driver, it will perform and sound better, trust me. You filter response as-is is -60dB at 10kHz, there's not much point in doing much more there. The LCR addition in your circuit isn't doing a whole lot, the upper frequencies are already notched by the two LC combos you have there.
Filter response with and without the LCR:
I'd go with something like this:
Comparison of filter response, yellow dashed is your filter above.
