Many threads and posts have been written about Magnpan active filtering, but I did not know any about active-passive filtering.
I equipped my MGIIIa with a 2 way active passive filter, and the obtained results are the best I had since years, far (far-far...) away from the original filtering, and far better from a basic active filtering based on a modified behringer DCX 24/96.
- Measures shows that the response curve stays flat, phase response is improved in the bass area (no more self), and the impulse response is improved by at least 15%.
+++ Listening is very dynamic and neutral, and the already very good tri-dimensionnal performance and the Maggies is improved !!
- The simulation of the filter was performed with a freeware software. After tests, I choose a 12db/oct in place of the original 12db/oct - 18db/oct. But this is my personnal taste.
- Only good resistors (vishays) and polystyrene caps were used. This is a main issue : with standard components, the result was not so dynamic and neutral.
More generally, I'm surprised that this solution is NOT more often considered, when a high musicality and dynamic are researcherd.
These kind of filter are cheap (about 20Euro), easy to build, performant.
The only obstacle is that you absolutly need a simulation software. Because of the correlation between the differents steps of the filter, the classic RC calculation do not work properly.
Why are they forgotten in the Hifi world... ?
Mathieu
MGIIIa + AR SP16 + Cary valve amp (medium-treeble) + Carver PM 1.5 (bas)
I equipped my MGIIIa with a 2 way active passive filter, and the obtained results are the best I had since years, far (far-far...) away from the original filtering, and far better from a basic active filtering based on a modified behringer DCX 24/96.
- Measures shows that the response curve stays flat, phase response is improved in the bass area (no more self), and the impulse response is improved by at least 15%.
+++ Listening is very dynamic and neutral, and the already very good tri-dimensionnal performance and the Maggies is improved !!
- The simulation of the filter was performed with a freeware software. After tests, I choose a 12db/oct in place of the original 12db/oct - 18db/oct. But this is my personnal taste.
- Only good resistors (vishays) and polystyrene caps were used. This is a main issue : with standard components, the result was not so dynamic and neutral.
More generally, I'm surprised that this solution is NOT more often considered, when a high musicality and dynamic are researcherd.
These kind of filter are cheap (about 20Euro), easy to build, performant.
The only obstacle is that you absolutly need a simulation software. Because of the correlation between the differents steps of the filter, the classic RC calculation do not work properly.
Why are they forgotten in the Hifi world... ?
Mathieu
MGIIIa + AR SP16 + Cary valve amp (medium-treeble) + Carver PM 1.5 (bas)
Attachments
Last edited:
Thanks vary much for your post .Magnepan made a box for what your doing for the MG3a an it gets bad press now thay have moved away from even biwiring...Thay had a pot so you could set the top to the bottem i think,the rebbons are so good that ever a so so active crossover can make the sound good.
Hi Mathieu,
I don't quite understand what you mean by "2-way active passive"?
The picture on the R (of the XO box) suggests that it is a Passive Line Level XO (PLLXO) - is this correct?
If so then, yes, I guess you could call this "active" in the sense that the amplifiers are directly connected to the speaker ... so it is a passive implementation of an active approach!
I have done the same thing with an active XO - except that I went 3-way active. I suggest you would find it difficult to achieve a good result if you tried to install a PLLXO for the (internal) 12dB mid/ribbon XO as well.
However, you could change slopes to all-6dB (given that the 3.7 has this) - this is my next experiment with my Maggie 3-ways (currently, I have stock slopes - 18dB bass LP/12dB mid HP/12dB mid LP/12dB ribbon HP).
Regards,
Andy
I don't quite understand what you mean by "2-way active passive"?
The picture on the R (of the XO box) suggests that it is a Passive Line Level XO (PLLXO) - is this correct?
If so then, yes, I guess you could call this "active" in the sense that the amplifiers are directly connected to the speaker ... so it is a passive implementation of an active approach!
I have done the same thing with an active XO - except that I went 3-way active. I suggest you would find it difficult to achieve a good result if you tried to install a PLLXO for the (internal) 12dB mid/ribbon XO as well.
However, you could change slopes to all-6dB (given that the 3.7 has this) - this is my next experiment with my Maggie 3-ways (currently, I have stock slopes - 18dB bass LP/12dB mid HP/12dB mid LP/12dB ribbon HP).
Regards,
Andy
Last edited:
Dear all,
Indeed the new filter is a PLLXO, but only 2 ways, 'cause I see no advantage in having the ribbon tweeter connected directly to a power amp...
I tried also a 18db/oct PLLXO (as suggested in the original Magnepan filter) but there was not any advantage. This must be connected to the fact that, for the same slopes, Passive XO and PLLXO do not have the same phase shift.
Curves show improvement both ion the dynamic range and pahse shift.
The listening feeling confirm both.
I did also made a first attemp of PLLXO years ago, without any good results.
My mistake was to misconsider the importance of the quality of the components.
=> PLLXO must be 'better' than the amplifier and preamplifier which are connected before... Otherwise the passive Magnepan filter gives more transparent results, especialy in the medium range.
I alspo tried active XO -a modified Behringer 3400 and later a DCX2496-, but the sonic result was'nt at my taste.
Now what next ?
- I changed the medium stock selfinductance against a 'flat wire' one, and gained even more dynamic.
- A sub is under construction, using a DCX 2496 + Carver power amp and a 18' JBL louspeaker in 120L B.Reflex. Then I will add a high pass 12db/oct cut in the PLLXO at about 45Hz. I already experienced this on my Acoustat with good results.
Have a nice day everybody !
Indeed the new filter is a PLLXO, but only 2 ways, 'cause I see no advantage in having the ribbon tweeter connected directly to a power amp...
I tried also a 18db/oct PLLXO (as suggested in the original Magnepan filter) but there was not any advantage. This must be connected to the fact that, for the same slopes, Passive XO and PLLXO do not have the same phase shift.
Curves show improvement both ion the dynamic range and pahse shift.
The listening feeling confirm both.
I did also made a first attemp of PLLXO years ago, without any good results.
My mistake was to misconsider the importance of the quality of the components.
=> PLLXO must be 'better' than the amplifier and preamplifier which are connected before... Otherwise the passive Magnepan filter gives more transparent results, especialy in the medium range.
I alspo tried active XO -a modified Behringer 3400 and later a DCX2496-, but the sonic result was'nt at my taste.
Now what next ?
- I changed the medium stock selfinductance against a 'flat wire' one, and gained even more dynamic.
- A sub is under construction, using a DCX 2496 + Carver power amp and a 18' JBL louspeaker in 120L B.Reflex. Then I will add a high pass 12db/oct cut in the PLLXO at about 45Hz. I already experienced this on my Acoustat with good results.
Have a nice day everybody !
Last edited:
You need to do the experiment (ie. connect the ribbon directly to an amp) before you will understand the advantage.
However, I suspect you would lose too much signal on the mid panel by having a HP and LP 12dB passive line-level filter. IE. active is the only way to go, for 3-way.
However, if you switched to 6dB slopes (like the 3.7 has) you might get a successful result.
I'm not surprised - these are pro XOs. You can build a very good-sounding analogue active XO using Rod Elliott's P09/P81 PCBs.
Regards,
Andy
Just found this thread. Very interesting!
I own a pair of MGIIIa's and have heard how great they sound with a really good active XO (I have heard them driven through a vintage Levinson LNC-2 with MGIIIa specific frequencies).
I have tried a cheap (Behringer) active XO but the magic that is my Aleph 3 is lost as soon as the Behringer is put between the pre and power amps.
Have heard from the owner of the LNC-2 that the Bryston and other units he tried do not come near the old Levinson.
What are your suggestions for a top-quality active XO? Pref. DIY!
I own a pair of MGIIIa's and have heard how great they sound with a really good active XO (I have heard them driven through a vintage Levinson LNC-2 with MGIIIa specific frequencies).
I have tried a cheap (Behringer) active XO but the magic that is my Aleph 3 is lost as soon as the Behringer is put between the pre and power amps.
Have heard from the owner of the LNC-2 that the Bryston and other units he tried do not come near the old Levinson.
What are your suggestions for a top-quality active XO? Pref. DIY!
Use Rod Elliott's P09/P81 PCB, with P05 PS - see here:
ESP Projects Pages - DIY Audio and Electronics
Nominally, these are either 12dB or 24dB filters but you can change the bass LP filter to 18dB if you want, quite simply.
Regards,
Andy
I had a look, but these are opamp based. Generally, I am not such an opamp fan and would prefer discrete components. Any non-opamp suggestions?
I agree with you, Albert, in the sense that discrete "sounds like" it would be better sounding ... but I have been looking for a discrete DIY XO for over 10 years and have not found one!
Marchand makes analogue XOs - see here:
electronic crossover, active crossover, discrete opamp, custom amplifier, passive preamp
... but I believe these are also opamp-based - except for their XM126 which is tube-based (but quite pricey!
).Regards,
Andy
PLLXO ?
Dear all,
You are not convinced of the advantages of the PLLXO...
On my side, I would not have written the post if the transformation of the MGIII has not been so dramatic.
For about 40$, you can get the resistors (takman or similar), caps (polystyrene) and the casing and RCA connectors.
- Soft Micro cap 10 is free, the biggest work was to optimize the simulation. it took me some weeks but i do not regret it.
In the simulaiton I considered the amps internal resistance to be 30 to 50kohm. This is why I added the 30kohm before the amp, in order to smooth the amp resistance.
- After testing the 0.25 uF cap was replaced with a 0.15uF. Medium is then low-cut at about 450Hz.
Mathieu
Dear all,
You are not convinced of the advantages of the PLLXO...
On my side, I would not have written the post if the transformation of the MGIII has not been so dramatic.
For about 40$, you can get the resistors (takman or similar), caps (polystyrene) and the casing and RCA connectors.
- Soft Micro cap 10 is free, the biggest work was to optimize the simulation. it took me some weeks but i do not regret it.
In the simulaiton I considered the amps internal resistance to be 30 to 50kohm. This is why I added the 30kohm before the amp, in order to smooth the amp resistance.
- After testing the 0.25 uF cap was replaced with a 0.15uF. Medium is then low-cut at about 450Hz.
Mathieu
Magnepans MOX-1
All you can do is put things in front of the sound,Less is More.
In the 90s i had a pr of the MG3a like you an Windell at Magnepan told me there was no better way to get good sound,To day thay have found the crossover that thay use is the only way to go ... Ok for there new MG1.7s an other .7 Maybe...Your setup work vary good with the MG3a,s...Your on the right path i think,just get caps you like an your there...
All you can do is put things in front of the sound,Less is More.
In the 90s i had a pr of the MG3a like you an Windell at Magnepan told me there was no better way to get good sound,To day thay have found the crossover that thay use is the only way to go ... Ok for there new MG1.7s an other .7 Maybe...Your setup work vary good with the MG3a,s...Your on the right path i think,just get caps you like an your there...
Attachments
![1304102105[1].jpg](/community/data/attachments/231/231826-f7a605544a0628524b057474a841cf21.jpg)
![1304102106[1].jpg](/community/data/attachments/231/231827-b7e4f320545a2276555aa83d753d76f0.jpg)
Ulaahh...
Ulaaahh...
The photo show a nice 6db/oct filter, with an 'nice' resistor coming from nowere...
6db/oct is the most artistical soluiton, on the paper it gives the perfect results...
Reality is completly different : the two speakers must match in many areas, having the same 'tone', the same phase shift, the same impulse response, the same power handling...
... on three octaves at least.
If no, then ... I experienced so many failures in the last 30years that I will not review them.
Basically, there is no way to obtain high level of sound with a 6dbct filter, in respect with a good impulse response and stable stereo image. Few succeded, I admire them !
About MGIII : only having a look to the impulse response of the bass and mid panels of the MGIII show that ther is a total incompatibility. Sorry for the Magnepan staff for whom i've got the highest respect.
The medium pannel is far away quicker than the low panel.
12db gives acceptable results, 18db/oct is 'better in impulse response, but i prefered the less precise but more rich sound given by the 12db/oct.
good to discuss about this ,
Mathieu
Ulaaahh...
The photo show a nice 6db/oct filter, with an 'nice' resistor coming from nowere...
6db/oct is the most artistical soluiton, on the paper it gives the perfect results...
Reality is completly different : the two speakers must match in many areas, having the same 'tone', the same phase shift, the same impulse response, the same power handling...
... on three octaves at least.
If no, then ... I experienced so many failures in the last 30years that I will not review them.
Basically, there is no way to obtain high level of sound with a 6db
ct filter, in respect with a good impulse response and stable stereo image. Few succeded, I admire them !About MGIII : only having a look to the impulse response of the bass and mid panels of the MGIII show that ther is a total incompatibility. Sorry for the Magnepan staff for whom i've got the highest respect.
The medium pannel is far away quicker than the low panel.
12db gives acceptable results, 18db/oct is 'better in impulse response, but i prefered the less precise but more rich sound given by the 12db/oct.
good to discuss about this ,
Mathieu
Basically, there is no way to obtain high level of sound with a 6db
Mathieu
Then how do you explain the fact that for the current models 1.7 & 3.7, Magnepan have used all-6dB filters?
Regards,
Andy
Aarghh...
I m trapped.
i don t know the latest mgs, i can only assume that the impulse response of the low frequency pannel is closer to the mid pannel. Mg people know these pannels like their own children and they found a way to quicken the low pannel...
Basically the big advantage of the mg s is that low and mid pannel use the same mylar film. So at least the 'tone' is the same.
When writing a out my attemts about tuning 6db/oct filters, i was maily thinking about traditional speakers. 20years ago i sweared only by 6db, listening tests brang me bagk o reality...
cheers
mathieu
I m trapped.
i don t know the latest mgs, i can only assume that the impulse response of the low frequency pannel is closer to the mid pannel. Mg people know these pannels like their own children and they found a way to quicken the low pannel...
Basically the big advantage of the mg s is that low and mid pannel use the same mylar film. So at least the 'tone' is the same.
When writing a out my attemts about tuning 6db/oct filters, i was maily thinking about traditional speakers. 20years ago i sweared only by 6db, listening tests brang me bagk o reality...
cheers
mathieu
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.