Tweet
Saint-Saëns planar- AKA SMJ-40- time for a project now?
This project concept has an unusual history and evolution. It started out in early mid 2021 based on the idea of a short narrow column bass system, originally targeting the recently introduced Dayton E180HD-44.
This dual magnet long excursion driver seemed to represent some very innovative thinking- and was certainly "interesting" looking, to say the least... and the early graphs showed it should be easy to work with up to 400 or 500Hz for a crossover point. The impedance curve gave some hints that this driver really was designed like a subwoofer, with a significant rise in the impedance between 125Hz and 1kHz. So, distortion might not be what we want in the upper midrange for a two way.
I acquired 4, and tested them, and discovered (as many have) that instead of having monotonically falling HD above 100 Hz, it's rising, due to inductivity modulation. So those parts got set aside, and I began looking at others, including several PuriFi models, and even pre-orders 8 of the PuriFi 8" parts. But while the 6.5" PuriF parts had physical parameters suited to small enclosures, the 8" parts are closer to monkey coffin drivers...
Hmmm, looks like we'll need to do a comprehensive data sheet review...
At the same time, I started looking at AMT style drivers for the tweeters, and was hearing things about B&G planers being taken over by GRS. But also I was impressed by the half dozen PuriFi cone midranges I'd bought as samples, initially just for the Isiris update project.
A lot going on with house buying and moving, and then in May of 2022 Steve sent me a video of the PS Audio FR30, which was quite a new implementation for them compared with speaker systems they'd shown previously with side mounted woofers near the floor, midwoofer/low mid link drivers, etc. Someone new was doing some new thinking there. And the LF section looked just like what I had in mind for the as yet unnamed project.
While PS Audio was cagey about showing the planar drivers in any of their videos, except for one slip up I noted, GRS was introducing their B&G based lineup, in the US through Parts Express, with much better detailed data sheets than B&G every offered. Hmmm... something to check out.
Being somewhat familiar with Magneplanars from the old days, I was wondering just how the diaphragms and overall motor systems were tuned and implemented. And of course, that woofer driver needed to be defined, also. I could see the "resemblance" to a certain family of Peerless woofers, but these days via their volume requirements for purchases Peerless has basically pulled out of the DIY market.
In the end, a Dayton woofer showed a remarkable balance of desirable woofer characteristic bordering on subwoofer turf, as it actually is part of a series of subwoofer drivers.
This one plot was taken near field at a drive level target for one driver equal to the four driver series/parallel array at 91 dB (at 2.83VRMS drive for the net 4 ohm array):
Though this test included a PR, due to QNF microphone placement the PR contribution (on the back of the cabinet) is not a significant factor in the measurement.
A variety of enclosure sizes and PR alignments and drivers were evaluated, and the example chosen at this time uses 4 RSS210HF-4 with 4 Seas SL26R XM003 passive radiators, with mass added to bring the PR mass to 350g. The calculated results look very promising with the target SPL of 110dB being achieved with 40W per woofer...
Note, the maximum excursion is about 60% of the RSS210HF rating. So this is not on the ragged edge...
An alternative PR close in performance is the Dayton RSS265PR; I chose the Seas part because of their reputation for assiduous monitoring of production compliance linearity, which is necessary for low distortion at high output levels.
Now, about those planar drivers...
It should be noted that the PT2522 tweeter is offered in two formats; one open back dipole, and one with a closed back. The differences gives some clues about the challenges working with these planar parts, both the midrange and tweeter. The midrange is only offered in an open back parts. Well after development evaluations started, GRS introduced some mounting panels, which helps significantly with testing and implementation.
Factory DS SPL plots for PT2522 (dipole)
Factory DS SPL plots for PT2522C. ("c" for closed back)
My experience measuring these parts was "interesting". I concluded that neither was an optimal solution- the factory 2522c parts did not measure like the DS graph, though the dipole 2522 more resembled the factory curve, but with a more extreme dip. But what these charts do display with some fidelity is the variation in response level depending on where the diaphragm resonance occurs, which can be measured with some accuracy looking at the impedance curve. My conclusion was that the "optimum" rear volume would be heavily damped and somewhere between open back and the very shallow factory cup, which results in a series of four experiments in search of what might be the best alignment, as conventional modeling tools would not handle this.
The midrange planer in dipole mode as from the factory presents the same fundamental issues- and in fact I had some success in modeling it and determining lumped parameters, but not fully accurately modeling the overall response profile.
This measured impedance curve on one sample was quite similar, and correlates well with some observed behavior in the data sheet SPL plot.
That also resulted in several test configuration, once again taking four test articles to home in on what seems to be the best compromise for SPL, even though I could accurately predict the Fb for a given volume. The "raw" response in dipole mode shows a 10dB range over the desired operational range from 400-500Hz to 2-2.2kHz. How can one get to a more optimum response level in this range?
This version has been selected as the POC target for system integration, and already verified in a POC crossover design.
The impedance curve of the POC target configuration:
Next I'll talk about the planned Proof Of Concept enclosure build.
This dual magnet long excursion driver seemed to represent some very innovative thinking- and was certainly "interesting" looking, to say the least... and the early graphs showed it should be easy to work with up to 400 or 500Hz for a crossover point. The impedance curve gave some hints that this driver really was designed like a subwoofer, with a significant rise in the impedance between 125Hz and 1kHz. So, distortion might not be what we want in the upper midrange for a two way.
I acquired 4, and tested them, and discovered (as many have) that instead of having monotonically falling HD above 100 Hz, it's rising, due to inductivity modulation. So those parts got set aside, and I began looking at others, including several PuriFi models, and even pre-orders 8 of the PuriFi 8" parts. But while the 6.5" PuriF parts had physical parameters suited to small enclosures, the 8" parts are closer to monkey coffin drivers...
Hmmm, looks like we'll need to do a comprehensive data sheet review...
At the same time, I started looking at AMT style drivers for the tweeters, and was hearing things about B&G planers being taken over by GRS. But also I was impressed by the half dozen PuriFi cone midranges I'd bought as samples, initially just for the Isiris update project.
A lot going on with house buying and moving, and then in May of 2022 Steve sent me a video of the PS Audio FR30, which was quite a new implementation for them compared with speaker systems they'd shown previously with side mounted woofers near the floor, midwoofer/low mid link drivers, etc. Someone new was doing some new thinking there. And the LF section looked just like what I had in mind for the as yet unnamed project.
While PS Audio was cagey about showing the planar drivers in any of their videos, except for one slip up I noted, GRS was introducing their B&G based lineup, in the US through Parts Express, with much better detailed data sheets than B&G every offered. Hmmm... something to check out.
Being somewhat familiar with Magneplanars from the old days, I was wondering just how the diaphragms and overall motor systems were tuned and implemented. And of course, that woofer driver needed to be defined, also. I could see the "resemblance" to a certain family of Peerless woofers, but these days via their volume requirements for purchases Peerless has basically pulled out of the DIY market.
In the end, a Dayton woofer showed a remarkable balance of desirable woofer characteristic bordering on subwoofer turf, as it actually is part of a series of subwoofer drivers.
- Very linear cone behavior out to 4kHz, with breakup mode at 5kHz; much better behaved than the RS225 woofer, with its large dip before the breakup region.
- Useful Xmax of 9 mm one way.
- Three ring motor with good midrange distortion characteristics also.
This one plot was taken near field at a drive level target for one driver equal to the four driver series/parallel array at 91 dB (at 2.83VRMS drive for the net 4 ohm array):
Though this test included a PR, due to QNF microphone placement the PR contribution (on the back of the cabinet) is not a significant factor in the measurement.
A variety of enclosure sizes and PR alignments and drivers were evaluated, and the example chosen at this time uses 4 RSS210HF-4 with 4 Seas SL26R XM003 passive radiators, with mass added to bring the PR mass to 350g. The calculated results look very promising with the target SPL of 110dB being achieved with 40W per woofer...
Note, the maximum excursion is about 60% of the RSS210HF rating. So this is not on the ragged edge...
An alternative PR close in performance is the Dayton RSS265PR; I chose the Seas part because of their reputation for assiduous monitoring of production compliance linearity, which is necessary for low distortion at high output levels.
Now, about those planar drivers...
It should be noted that the PT2522 tweeter is offered in two formats; one open back dipole, and one with a closed back. The differences gives some clues about the challenges working with these planar parts, both the midrange and tweeter. The midrange is only offered in an open back parts. Well after development evaluations started, GRS introduced some mounting panels, which helps significantly with testing and implementation.
Factory DS SPL plots for PT2522 (dipole)
Factory DS SPL plots for PT2522C. ("c" for closed back)
My experience measuring these parts was "interesting". I concluded that neither was an optimal solution- the factory 2522c parts did not measure like the DS graph, though the dipole 2522 more resembled the factory curve, but with a more extreme dip. But what these charts do display with some fidelity is the variation in response level depending on where the diaphragm resonance occurs, which can be measured with some accuracy looking at the impedance curve. My conclusion was that the "optimum" rear volume would be heavily damped and somewhere between open back and the very shallow factory cup, which results in a series of four experiments in search of what might be the best alignment, as conventional modeling tools would not handle this.
The midrange planer in dipole mode as from the factory presents the same fundamental issues- and in fact I had some success in modeling it and determining lumped parameters, but not fully accurately modeling the overall response profile.
This measured impedance curve on one sample was quite similar, and correlates well with some observed behavior in the data sheet SPL plot.
That also resulted in several test configuration, once again taking four test articles to home in on what seems to be the best compromise for SPL, even though I could accurately predict the Fb for a given volume. The "raw" response in dipole mode shows a 10dB range over the desired operational range from 400-500Hz to 2-2.2kHz. How can one get to a more optimum response level in this range?
This version has been selected as the POC target for system integration, and already verified in a POC crossover design.
The impedance curve of the POC target configuration:
Next I'll talk about the planned Proof Of Concept enclosure build.
Comment