If I understand you correctly, first of all you are planning to correct the woofers outside the MFB loop, and then apply MFB to the (mildly?) corrected system? I suppose you then keep the most distorted part of the woofers new passband outside the loop. I.m.o. that defies the purpose of MFB. Excursion is still excursion, whether one applies MFB or not.
Actually I have my doubts whether that is possible at all, but maybe that remark shows my lack of fundamental understanding control theory and stability issues...
Actually I have my doubts whether that is possible at all, but maybe that remark shows my lack of fundamental understanding control theory and stability issues...
Dipole woofers lack efficiency at low frequency. So it will have a roll off.
Musik from my source will then have frequency correction, so i implemented dipole correction for 20 years ago. And it works great! Giving 15dB more headroom for the Quads making them more dynamic. etc, etc,,
Dipole woofer in itself is superior to other low frequency generators since the interact with the room much better acoustically, and has so many advantages sound-vice so i will never go back to anything else.
However the high excursion at low frequency needed to play at decent volume gives distortion, playing at low volume they blend in well with (modified) QUAD ESL-63, but playing Infected Mushrooms at 100dB the coloration becomes apparent from the subs. Current feedback reduces distortion, and in combination with Piratelogic MFB I expect to reach 20-25dB less distortion in total. I have seen other people reaching this.
Musik from my source will then have frequency correction, so i implemented dipole correction for 20 years ago. And it works great! Giving 15dB more headroom for the Quads making them more dynamic. etc, etc,,
Dipole woofer in itself is superior to other low frequency generators since the interact with the room much better acoustically, and has so many advantages sound-vice so i will never go back to anything else.
However the high excursion at low frequency needed to play at decent volume gives distortion, playing at low volume they blend in well with (modified) QUAD ESL-63, but playing Infected Mushrooms at 100dB the coloration becomes apparent from the subs. Current feedback reduces distortion, and in combination with Piratelogic MFB I expect to reach 20-25dB less distortion in total. I have seen other people reaching this.
It is very tempting to start a discussion about monopoles vs dipoles in the 20-80 Hz range, but I will resist that... Correcting dipole roll off outside the MFB loop will not give you the 20-25 dB distortion reduction you aim for. Furthermore, a 20-25 dB distortion reduction will require a somewhat more sophisticated loop circuitry than the basic R17/C10 integrator in your specific case, with the -uncorrected- relatively high dipole roll off @100Hz. You simply need more local loopgain than the basic integrator circuit can realize without running into insufficient phase margin@unity gain problems.
Have a look at the att.s 137 and 147 of this thread, posted by our unsurpassable Bolserst.
Have a look at the att.s 137 and 147 of this thread, posted by our unsurpassable Bolserst.
Attachments
Working on getting you the all new and improved EVE 2020.
- Adjustable accelerometer bias voltage
- Input buffer
- Low noise pxe buffer
- Pxe phase switch
- Through-hole loop shaping parts & rails bleeders
- Max rails +/- 60V
- Open/closed loop mode switch
- led loop limiter
- Symmetric line out
- ne5532 opamps
- 50x50 mm
- JST headers
Last edited:
Footnote about OB subs and MFB. I had a 15-inch woofer on a large panel with an irregular perimeter for a few years. While I like the OB sound for mains a lot, not sure I found anything special about sub's sound quality or performance in a room otherwise. (Much easier to change boxes around to good locations than large Ops.)
1. The output from an OB sub includes a substantial "free" boost at the natural free-air resonance of the driver. Sure can change the non-EQ FR plots. This is not discussed in fevered discussions of Linkwitz Transform but it ought to change the LT, at least in modelling. (I'd like to hear where this shows up in esl63's fabulous system?)
2. The real obstacle is that you need immense cone motion with an OB sub if it is to go low freq. With MFB, you can get error correction, of course. But if the cone starts going seriously out of its operating range, hopeless for MFB to be anything but a disaster.
B.
1.
1. The output from an OB sub includes a substantial "free" boost at the natural free-air resonance of the driver. Sure can change the non-EQ FR plots. This is not discussed in fevered discussions of Linkwitz Transform but it ought to change the LT, at least in modelling. (I'd like to hear where this shows up in esl63's fabulous system?)
2. The real obstacle is that you need immense cone motion with an OB sub if it is to go low freq. With MFB, you can get error correction, of course. But if the cone starts going seriously out of its operating range, hopeless for MFB to be anything but a disaster.
B.
1.
Last edited:
Well, Ben,
The plots of ESL 63 do not show the boost you are referring to. Modern low Qts drivers will have a very controlled roll off. Exactly where there are large excursions, MFB will immensely improve the system if substantial feedback is applied at least.
Please correct me if I am wrong, but I still assume one would need over 30 dB loopgain to have any useful correction of the OB woofers. Maybe one of the experts ( Bolserst?) could chime in here: I am still at the beginning of my MFB learning curve and looking at a steep hill...
The plots of ESL 63 do not show the boost you are referring to. Modern low Qts drivers will have a very controlled roll off. Exactly where there are large excursions, MFB will immensely improve the system if substantial feedback is applied at least.
Please correct me if I am wrong, but I still assume one would need over 30 dB loopgain to have any useful correction of the OB woofers. Maybe one of the experts ( Bolserst?) could chime in here: I am still at the beginning of my MFB learning curve and looking at a steep hill...
Puzzling there is no boost at resonance. Maybe gets lost in the room eigentones and LT boost. The free-air cone motion peak is not gradual when I measure a sub in free-air.
MFB for OB if you are aiming for real low freq might be tricky. It can be done, I suppose, but a stretch unless you have enough cone surface to avoid big cone motions.
B.
Thats why i have the "Infra sub" below 27Hz, so i decouple the dipoles and pressurize the room at frequencies where I got no resonances du to the long wavelengths.
Am working to include comments, tips & tricks from this thread - thank you all! - into the updated EVE2020 manual, while reading I bumped into bolserst post #148 which says:
With regards to C1 and R2, these are for keeping rf artefacts from entering the servo loop which could proof troublesome especially when long leads are used between EVE and the Starbass.
The non inverting setup around IC1a was chosen to allow for a more flexible change of pxe.gain, using an inverting setup would require adjustment of C6 each time R14 is changed. Initially I had IC1a gain fixed and pxe level set by means of a potmeter at the output of IC1a but found the current solution less noisy and easier to work with.
With regards to C1 and R2, these are for keeping rf artefacts from entering the servo loop which could proof troublesome especially when long leads are used between EVE and the Starbass.
Last edited:
Done, click here : https://www.diyaudio.com/forums/subwoofers/355995-piratelogic-eve-servobass-module.html#post6241608
the new eve 2020 manual is also there.