Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
73.12.71.57
In Reply to: RE: Magnepan 2.5R crossover help posted by kappa546 on April 23, 2015 at 06:53:14
Changing the tweeter cap to 25 uf will lower the HP -3db point but when I ran it on my spreadsheet, there was not noticeable change in the output curve. You can see from the Bode diagram that there is a significant drop in the output of the MG2.5 above 200 Hz but then it is somewhat flat. This is the electrical curve only and does not necessarily reflect the speaker output.
As far as polarity, the tweeter is inverted relative to the woofer. If it is wired in the same polarity there is a huge drop at ~1000 hz (blue line).
If one speaker is wired backward compared to the other than as you note you will have a diffuse image but there will not be a dropout.
The 2 uf cap is in parallel with a 2 ohm resistor both of which are in series with the tweeter to raise its impedance which a ridiculously low 1.45 ohms. At the crossover frequency (1000 hz) and lower the added impedance is about 2 ohms (same as the resistor alone) and as the frequency rises it drops reaching 1 ohm at 40khz.
Follow Ups:
I located a small typo in my spreadsheet which caused it to give an erroneous picture. In fact the output curve is shifted upward with the larger cap. Here is the changed Bode:
Thanks for the input guys. I've already looked at the electrical response diagrams for the Magnepans and that only tells you part of the story. The bass panels have a severely rising response (or dipole dropoff, take your pic), hence the chosen crossover point by the factory. In my living room I actually have flat-ish output from ~40-50hz up to about 200 (the peaks and valleys in this range are definitely room and placement related), then there is a lot more energy from there peaking at around 500hz. Then it drops from that peak about 10db at 1.2khz and the ribbons start kicking in strong at 4khz. This is just a quick measurement with Mathaudio, haven't done a more comprehensive measurement with REW yet.
I know the resistors are there to bring the ribbon impedance up from 1.45 ohms, but I still don't get the point of the 2uF cap.
When I talked about speaker polarity i meant switching the whole speaker inputs (- at +, + at -), not individual drivers. I know this model has the mylar back so that's probably why it's flipped. The suggestion I read insinuated Magnepan had many 2.5R that were shipped with something wired wrong or labeled incorrectly which caused severe cancellation at the crossover. I can also confirm this with previous measurements (forgot to screen shot it) because the null centered at 1.2khz was like 20db down! This caused a very hollow, far away from the stage type of presentation. Once i flipped polarity at the speakers plugs the music sounded much less hollow. I don't know how the crossover managed this, it's the kind of thing I'd expect if I was flipping the polarity of one driver on one speaker, but not when flipping the whole speaker. I guess I could switch one drivers polarity to see the effects, but I haven't taken the socks off yet and won't really have the time for that until this summer most likely. I have much more experience in active setups than I do fiddling around with passives (that's probably obvious by now), but I won't be taking that step with the Maggies just yet.
The presence of the cap causes a slight rise (I estimate 0.5db) in the tweeter response over the range 15K-20K (and the rise continues beyond). I am pretty sure you would not notice it's absence but apparently Magnepan felt it was beneficial. It was also used with the 2.6.
It flattens the response for canine listeners. :)
Dave.
The tweeter is more sensitive than the midbass panel - we guestimated a 5-6 db difference - so the Bode diagram understates the tweeter's contribution since despite the 2 ohm resistor its output should be at about the same level as the midbass panel.
So the lower XO from the higher val cap should be more pronounced than the Bode diagram shows.
I mentioned that the Bode diagram was probably not an accurate representation of the actual speaker output, but I fail to see why a change in the cap would be more apparent with a more sensitive speaker. The crossover point would be lower with a more sensitive speaker but the relative shift with a change in caps would be unaffected. Am I missing something?
BTW, I found a small error in my spreadsheet and in fact the output curve is shifted upward in the with the larger the cap. Mea culpa.
But it would make the difference more noticeable than the bode diagram would indicate. Hopefully it is just enough to partially fill the dip in his response so it would be a bit smoother.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: