LCR conjugate filter to flatten the impedance then try a LR2 high pass around 900Hz and see what you get. Hopefully the peaking of the driver and the roll off of the 900Hz filter cancel each other out until the driver starts rolling off around 600Hz giving you an effective acoustic roll off about there.
It's quite possible you could fix it all up with a 2nd order high pass filter alone if the Q was low enough and crossover point were right. This is why I like XSim and the "CircuitBlocks" [sic] menu. It's easy to create a filter and dial it in by directly adjusting the f and Q of the circuit.
I'm sure other tools have similar features but I'm nothing but a humble hacker who likes pushing parts around on a schematic. Hardest part is just importing the data.
Best,
Erik
I'm sure other tools have similar features but I'm nothing but a humble hacker who likes pushing parts around on a schematic. Hardest part is just importing the data.
Best,
Erik
Any chance you could draw a schematic there, Lojzek? Not familiar with the software. But looks good. 
I was playing with AllenB's idea of a 1.2kHz XO for a Fs 550Hz le 0.2mH 90dB Monacor SPP-125 driver.
Monacor SPP-125 5.25 Inch midrange hifi driver from Monacor 10.81
I'm really winging it here, but how does this one model with the FR curve you've imported?
I've seen this sort of response before, and it surely can't be too hard:
M 10 - 8 Ohm
See Visaton say 1-5kHz band. But the M10 files are broken in Boxsim, so I've never got far with it.
I was playing with AllenB's idea of a 1.2kHz XO for a Fs 550Hz le 0.2mH 90dB Monacor SPP-125 driver.
Monacor SPP-125 5.25 Inch midrange hifi driver from Monacor 10.81
I'm really winging it here, but how does this one model with the FR curve you've imported?
I've seen this sort of response before, and it surely can't be too hard:
M 10 - 8 Ohm
See Visaton say 1-5kHz band. But the M10 files are broken in Boxsim, so I've never got far with it.
Attachments
I won't be drawing a XO filter because I have already provided
all the information one needs to know about the circuit from
the PCD snapshot. Everything is clearly designated. You know what
a series RLC in parallel with the driver is. You also know what a 2nd,
and 3rd order electrical is, and finally you know what a parallel RLC
in series with the driver before the HP and LP is.
Since I have discovered FPGraphTracer, thanks to jReave and a fellow
from PETT forum, I'm glad not having to deal with SPLTools.You should
try it. So, are you asking me to trace the M10-8 Ohm for you?
all the information one needs to know about the circuit from
the PCD snapshot. Everything is clearly designated. You know what
a series RLC in parallel with the driver is. You also know what a 2nd,
and 3rd order electrical is, and finally you know what a parallel RLC
in series with the driver before the HP and LP is.
Since I have discovered FPGraphTracer, thanks to jReave and a fellow
from PETT forum, I'm glad not having to deal with SPLTools.You should
try it. So, are you asking me to trace the M10-8 Ohm for you?
^ that looks like a good start for 600Hz up, but you'll still want an LCR to kill the impedance spike at Fs otherwise you'll never be able to shape a woofer into a matching slope.
I'm also uneasy about using filters with Q>1, though to be honest i've never subjectively tested the audibility of them. I'd prefer to use a low Q lowpass to bring the top end down to match the amplitude at 1kHz, rather than using a high Q highpass to bring the 1kHz region up.
I'm also uneasy about using filters with Q>1, though to be honest i've never subjectively tested the audibility of them. I'd prefer to use a low Q lowpass to bring the top end down to match the amplitude at 1kHz, rather than using a high Q highpass to bring the 1kHz region up.
2.5dB improvement isn't really negligible in the scheme of things - especially at 550Hz, it's an annoying frequency.
If you try to do without it either you get a ~5dB peak in the system response at 550Hz, or you cross your woofer lower and cause a hole around 350-500Hz.
I would even go as far as reducing the resistance of the RLC so the impedance of the mid dips below Revc, which has the effect of shaping the Fs peak down even further and give half a chance at hitting a proper rolloff. Even if the RLC dips to 1ohm the amplifier won't see it because of the highpass filter.
If you try to do without it either you get a ~5dB peak in the system response at 550Hz, or you cross your woofer lower and cause a hole around 350-500Hz.
I would even go as far as reducing the resistance of the RLC so the impedance of the mid dips below Revc, which has the effect of shaping the Fs peak down even further and give half a chance at hitting a proper rolloff. Even if the RLC dips to 1ohm the amplifier won't see it because of the highpass filter.
Thanks for the responses & discussion guys. A little more background - this came up in discussion with another member using the above driver in a refurbishment. FWIW the existing driver has 16uF & 2mH, and is Zobelled. I'm probably going to recommend a notch because the woofer this is paired to, for which incomplete info is available, has a hump above 500Hz as well - kill two birds with 3 components...
I'm going to try out FPGraphTracer as well, thanks for the mention;
Kudos to you all
I'm going to try out FPGraphTracer as well, thanks for the mention;
Kudos to you all
I wouldn't use that driver at 650Hz at all because despite of having the response under control you still have an extremely long decay time around the fs and also big phase problems since it turns 90° at the fs. So yes, you might get the response flat with enough effort and the speaker alone but no, it's not worth it because it won't sound good and you'll most likely have problems with the driver below - unless you just want to clobber something together to 'just use it'.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.