Jim,
Well done - the Saturn speakers are looking good. Despite not being exactly the same color tone, they still complement the room nicely; and if I recall you mentioning a few post back, the color of Cherry wood will deepen and possibly become similar to the Purifi waveguide speaker.
You might have mentioned already, but may I ask, what is the height of the speaker along with the stand?
Well done - the Saturn speakers are looking good. Despite not being exactly the same color tone, they still complement the room nicely; and if I recall you mentioning a few post back, the color of Cherry wood will deepen and possibly become similar to the Purifi waveguide speaker.
You might have mentioned already, but may I ask, what is the height of the speaker along with the stand?
Beautiful build! In your post #560 it looks like you are using LR4 acoustic filter slopes. I have been simulating a kind of similar project with those Bliesma drivers in Vituix recently (although paired with an 8" woofer) and based on that I have drawn conclusion that acoustic LR2 crossover slopes could produce a more even and narrower directivity than acoustic LR4. I think the M74 could handle a 600Hz acoustic LR2 highpass, but I could be wrong on that. Do you have any plans to compare LR2 vs LR4 acoustic crossover slopes soundwise? Would love to hear your thoughts on this.
Here is the latest filter I am listening to. I started with a filter which was nearly flat on-axis, and subjectively adjusted the high frequency balance downward a bit. I smoothed out some of the wiggles and bumps in the on-axis and sound power curves. I now have a -0.3 dB per octave tilt on axis, and I have matched the target curve +/- 1 dB.
The directivity is nearly constant over a very broad range. The Directivity Index (DI) is 5.3 dB +/- 1.1 dB from 430 Hz to 8.7 kHz. Even at 16k the directivity has risen to just 7.4 dB.
The resulting Predicted In Room (PIR) response is flat +/- 1 dB from 500 Hz to 16 kHz. I don’t place as much importance on the PIR curve as some other people do. I believe that the on-axis curve, the PIR curve, and the sound power curve all need to be managed to get speaker response with high potential. If these three curves are all very nice, the subjective voicing process goes quickly, and adjustments are few, and small.
I talked about this in post #267 Link 267. The driver has a natural 2nd order rolloff at 500 Hz… and this is without any additional electrical filtering. I could not figure out a way to achieve a 2nd order acoustic high-pass response which gave me acceptable low frequency protection of the driver. Bliesma advises that the minimal electrical filtering is 400 Hz 2nd order Butterworth (Q=0.707), and I am using a 450 Hz Q=0.707 filter… this results in a 4th order acoustic high pass. So I am not sure how to achieve a 2nd order high pass… ?
j.
The directivity is nearly constant over a very broad range. The Directivity Index (DI) is 5.3 dB +/- 1.1 dB from 430 Hz to 8.7 kHz. Even at 16k the directivity has risen to just 7.4 dB.
The resulting Predicted In Room (PIR) response is flat +/- 1 dB from 500 Hz to 16 kHz. I don’t place as much importance on the PIR curve as some other people do. I believe that the on-axis curve, the PIR curve, and the sound power curve all need to be managed to get speaker response with high potential. If these three curves are all very nice, the subjective voicing process goes quickly, and adjustments are few, and small.
I have been simulating a kind of similar project with those Bliesma drivers in Vituix recently (although paired with an 8" woofer) and based on that I have drawn conclusion that acoustic LR2 crossover slopes could produce a more even and narrower directivity than acoustic LR4. I think the M74 could handle a 600Hz acoustic LR2 highpass, but I could be wrong on that. Do you have any plans to compare LR2 vs LR4 acoustic crossover slopes soundwise? Would love to hear your thoughts on this.
I talked about this in post #267 Link 267. The driver has a natural 2nd order rolloff at 500 Hz… and this is without any additional electrical filtering. I could not figure out a way to achieve a 2nd order acoustic high-pass response which gave me acceptable low frequency protection of the driver. Bliesma advises that the minimal electrical filtering is 400 Hz 2nd order Butterworth (Q=0.707), and I am using a 450 Hz Q=0.707 filter… this results in a 4th order acoustic high pass. So I am not sure how to achieve a 2nd order high pass… ?
j.
Only way I can think of is maybe apply a really steep high filter (LR4 or LR8?) at ~200Hz to protect the Bliesma and allow the natural rolloff to be untouched above that. This might wreak havoc with other aspects your carefully crafted XO/EQ, tho.I am not sure how to achieve a 2nd order high pass… ?
Or if you're only interested in checking out the idea suggested by @SpaceEsel , apply LR4 high pass at 100~150Hz for protection & test/listen at modest levels.
To apply a 12dB HP and yield a 12dB HP initially requires a much higher xover point applied before the driver's inherent rolloff starts. You would lose a lot of bandwidth and effectively convert the midrange to a filler driver approach in this case.
I do not feel an acoustic LR2 Mid HP is warranted nor is it to be a perceived improvement here.
I do not feel an acoustic LR2 Mid HP is warranted nor is it to be a perceived improvement here.
Well done! How do you like the sound?
The dip in the power response at 2-3kHz might give a voicing in the pleasant direction of the reverberant field in the room. The dispersion it's very wide that results to my experience in a more bright/light but very large and natural appearing soundstage if there is some space around the speakers that early reflections are minimized and most part of reflections arrive at the ear >> 10ms after direct sound ("Haas-Effect"). If there is tighter distance to the walls, and so early reflections might play more role for sound perception, I would try to EQ the FR between 4-7kHz -1dB or so if it sound a bit too bright. If not, direct sound and listening window might dominate sound perception and this is very well linearized with your speaker.
I see some potential in phase matching in the transistion range of the drivers. It's hard to get with a 4-way and linear filters as the bandwith of each passband is tight and the low- and highpass behaviour will influence each other. I always tune and optimize for proper phase transition in the drivers transisiton range - not because i believe in some "time-perception", but a proper phase transition brings a stable and symmetric vertical dispersion characteristics, so lobes will not point to different directions dependent on frequency and so the reflections will come from stable directions.
Two examples with 3-ways and Hypex modules:
(DXT/15NBAC/26W Disco/FA123; sorry no relevant vertical spin)
(T25A/Mid120_vHE/WO24P/FA503)
But the most important is still: How do you like the sound till now? 😉
The dip in the power response at 2-3kHz might give a voicing in the pleasant direction of the reverberant field in the room. The dispersion it's very wide that results to my experience in a more bright/light but very large and natural appearing soundstage if there is some space around the speakers that early reflections are minimized and most part of reflections arrive at the ear >> 10ms after direct sound ("Haas-Effect"). If there is tighter distance to the walls, and so early reflections might play more role for sound perception, I would try to EQ the FR between 4-7kHz -1dB or so if it sound a bit too bright. If not, direct sound and listening window might dominate sound perception and this is very well linearized with your speaker.
I see some potential in phase matching in the transistion range of the drivers. It's hard to get with a 4-way and linear filters as the bandwith of each passband is tight and the low- and highpass behaviour will influence each other. I always tune and optimize for proper phase transition in the drivers transisiton range - not because i believe in some "time-perception", but a proper phase transition brings a stable and symmetric vertical dispersion characteristics, so lobes will not point to different directions dependent on frequency and so the reflections will come from stable directions.
Two examples with 3-ways and Hypex modules:
(DXT/15NBAC/26W Disco/FA123; sorry no relevant vertical spin)
(T25A/Mid120_vHE/WO24P/FA503)
But the most important is still: How do you like the sound till now? 😉
I agree. I went back and reviewed my work in post 267. I don't think it is possible to achieve a 2nd order acoustic high pass with this driver. It is simple logic to understand why, but I will demonstrate it anyway.To apply a 12dB HP and yield a 12dB HP initially requires a much higher xover point applied before the driver's inherent rolloff starts. You would lose a lot of bandwidth and effectively convert the midrange to a filler driver approach in this case.
I do not feel an acoustic LR2 Mid HP is warranted nor is it to be a perceived improvement here.
Here is my M74A driver with all of its DSP filtering, overlayed with an idealized bandpass response: 464 Hz LR4 high pass and 3000 Hz 2nd order Q=0.68 low pass
Below is the electrical filter gain that achieves the above acoustical response
Now if I want a 2nd order low pass at 600 Hz, I can achieve it in theory. This is the M74A response with a new filter, overlaid with an idealized 2nd order 600 Hz Q=0.5 high pass and 2nd order 3000 Hz Q=0.68 low pass
However, if we look at the electrical filter gain, we are getting no real protection under 400 Hz (which is approximate Fs). The electrical gain at 100 Hz is only 2 dB lower than the gain at 2000 Hz.
I tried adding a steep filter at 100 Hz, but this affects the phase of the driver and complicates the woofer to mid crossover.
So I conclude that a 2nd order acoustic high pass is not feasible.
j.
Thanks for your explanation, this seems very plausible to me. I guess acoustic LR2 was a little too optimistic from my side...I talked about this in post #267 Link 267. The driver has a natural 2nd order rolloff at 500 Hz… and this is without any additional electrical filtering. I could not figure out a way to achieve a 2nd order acoustic high-pass response which gave me acceptable low frequency protection of the driver. Bliesma advises that the minimal electrical filtering is 400 Hz 2nd order Butterworth (Q=0.707), and I am using a 450 Hz Q=0.707 filter… this results in a 4th order acoustic high pass. So I am not sure how to achieve a 2nd order high pass… ?
Something else I was thinking about is the use/need of a protection cap for this expensive midrange. I do have slight trust issues with DSP's, as they can fail or I can simply f*** up when playing around with crossover settings. Also there is the unlikely chance of the amp misbehaving and sending out some high voltage signal or loud pop for whatever reason. I'm aware a protection cap would probably also require impedance flattenig at fs, which would complicate things quite a bit. Not sure if I am too paranoid in this regard. What made you decide for or against a protection cap?
Last edited:
It sounds really nice. Very dynamic, very detailed, and clarity is crystal clear. The tonal balance is just about where I like it. I don't want to say too much about it until I have lived with it a few more weeks and made a few more voicing changes.But the most important is still: How do you like the sound till now? 😉
Part of the voicing process for me is making small changes in the delay between drivers. Once the EQ is dialed in and the tonal balance is right, I have found that small changes in delay can bring about noticeable, even significant changes in the perceived level of detail. For instance, increasing the delay on a tweeter can make it sound as if the tweeter was reduced in level by 1 or 2 dB, even though the measured response barely changes. The very high frequencies can sound muffled. I am sure there is some psychoacoustic explanation for this.
With some of my past projects, I have found an exact phase match between drivers sounded the best. But other projects sounded best with a slight mismatch. Usually this involved the tweeter being slightly ahead of the midrange by 20 - 40 microseconds, which is only a few millimeters. But this small difference increased the perception of detail, and enhanced the perception of 3D space.
I have not gotten to this stage yet... My delays were set analytically based on VituixCad simulation, with the goal of optimizing both the on-axis, soundpower, DI, and PIR curves. When I get to that stage of subjective voicing, I will certainly try a perfect phase alignment.
j.
I think the midrange driver is pretty robust. A 74 mm voice coil can soak up a pretty big current spike. The tweeter is actually the delicate one, so any issue of the sort you are talking about would toast the tweeter first.What made you decide for or against a protection cap?
I have never used an inline capacitor to protect a driver, and I have never had an incident. I try to be careful when testing, and perhaps I have just been lucky.
It is probably a wise thing to do on a tweeter. A large cap (22 to 33 uf) in series with the tweeter would offer good protection from low frequency burps. It is easy to incorporate the cap into the simulation. Someday after I burn up a $300 tweeter, I will start doing it myself... 😉
Same here.I have never used an inline capacitor to protect a driver, and I have never had an incident. I try to be careful when testing, and perhaps I have just been lucky.
😆 Perhaps too obvious to mention, but relevant only with active speakers. Another option for any mid or HF drivers that end up needing to be reduced in level to match the woofer is to add an L-pad with 5W or 10W resistors and bring the level back up in the active XO. The resistors will fare fine with music but likely break before the drivers in the case of a serious spike. Maybe? There's probably some other electronic part that can act like a fuse without the signal degradation that those can sometime cause.It is probably a wise thing to do on a tweeter. A large cap (22 to 33 uf) in series with the tweeter would offer good protection from low frequency burps. It is easy to incorporate the cap into the simulation. Someday after I burn up a $300 tweeter, I will start doing it myself... 😉
hifijim wrote: I tried adding a steep filter at 100 Hz, but this affects the phase of the driver and complicates the woofer to mid crossover.
Are you sure that it affects phase in the system? That could be corrected with delay then.
LR2'ish elliptic acoustic slopes retain phase even better than LR, beyond xo. They come quite easily with natural roll-off. Then mid(s) play as fillers because of wide overlapping responses with good phase match. The filler has lower spl than neighbours. https://musicanddesign.speakerdesign.net/Duelund_and_Beyond.html
"Rather, the transfer functions should be used as the acoustic targets for LP, BP and HP drivers and the filters developed to achieve these results. In one considers the midrange driver we see further that given ω2, the midrange to woofer crossover frequency, then what is really needed is a sealed midrange enclosure which yields a Qtc =0.5 at ω2. In this regard, the T/S parameters of the midrange driver will dictate the mid/woofer crossover point. Additionally, the acoustic output of a tweeter is basically that of a 2nd order HP response."
Since M74 is sealed, it should take rather high power, like ATC mid. Manufacturer doesn't give power rating
Are you sure that it affects phase in the system? That could be corrected with delay then.
LR2'ish elliptic acoustic slopes retain phase even better than LR, beyond xo. They come quite easily with natural roll-off. Then mid(s) play as fillers because of wide overlapping responses with good phase match. The filler has lower spl than neighbours. https://musicanddesign.speakerdesign.net/Duelund_and_Beyond.html
"Rather, the transfer functions should be used as the acoustic targets for LP, BP and HP drivers and the filters developed to achieve these results. In one considers the midrange driver we see further that given ω2, the midrange to woofer crossover frequency, then what is really needed is a sealed midrange enclosure which yields a Qtc =0.5 at ω2. In this regard, the T/S parameters of the midrange driver will dictate the mid/woofer crossover point. Additionally, the acoustic output of a tweeter is basically that of a 2nd order HP response."
Since M74 is sealed, it should take rather high power, like ATC mid. Manufacturer doesn't give power rating
The question being: why does it need LF protection, apart from the extra power it would have to dissipate.I could not figure out a way to achieve a 2nd order acoustic high-pass response which gave me acceptable low frequency protection of the driver.
I am not so much worried about the voice coil taking damage. However, the high dome profile of the M74 A and T models is what worries me. Looking at pictures of this driver, it seems to me like the tip of the dome is very close to the protective grid and could get damaged in case of overexcursion. I am just not sure how easy it is to get this driver into such overexcursion through accidents. In a practical example: would a few seconds of medium to highish levels of bass heavy electronic music kill the M74 without crossover protection?I think the midrange driver is pretty robust. A 74 mm voice coil can soak up a pretty big current spike. The tweeter is actually the delicate one, so any issue of the sort you are talking about would toast the tweeter first.
I have never used an inline capacitor to protect a driver, and I have never had an incident. I try to be careful when testing, and perhaps I have just been lucky.
It is probably a wise thing to do on a tweeter. A large cap (22 to 33 uf) in series with the tweeter would offer good protection from low frequency burps. It is easy to incorporate the cap into the simulation. Someday after I burn up a $300 tweeter, I will start doing it myself... 😉
The tweeter will get a protective cap for sure in my case. The protection cap on the tweeter I am using right now has saved its life probably more than once 😉 But my setup is especially prone to user errors, as I am using Equalizer APO which is bypassed in programs like REW or my DAW when selecting ASIO instead of the Java driver by mistake. If I was using a less agile system like the Hypex plate amps I would be way more relaxed in this regard 😉
Another reason for LF protection that comes to my mind is the close proximity of the dome of the M74 A and T models to the protective grid. If the tip of the dome would be driven into the protection grid by overexcursion it would certainly break. I just don't know enough about this driver to know if this is a realistic scenario or not.The question being: why does it need LF protection, apart from the extra power it would have to dissipate.
Yes that is a factor. I use the Hypex amps with the analog RCA inputs, and this makes it hard to accidently hit a driver with full 0 dB power. I know some people have damaged drivers while using the digital inputs on the Hypex because it is easier to make a mistake and go full scale.The protection cap on the tweeter I am using right now has saved its life probably more than once 😉 But my setup is especially prone to user errors, as I am using Equalizer APO which is bypassed in programs like REW or my DAW when selecting ASIO instead of the Java driver by mistake. If I was using a less agile system like the Hypex plate amps I would be way more relaxed in this regard 😉
Just to be clear, we are talking about the M74A dome midrange... yes? It has an Sd of 50 cm^2 and an Xmax of 1.2 mm 0-pk. The spec sheet does not say what the mechanical excursion limit is, but I am guessing it is in the range of 3 mm. In my opinion, this driver needs an effective high pass filter. Bliesma recommends at least 400 Hz 2nd order Butterworth.The question being: why does it need LF protection, apart from the extra power it would have to dissipate.
j.
Dome displacement doesn’t change anymore below Fs, but you have to take the spectral energy distribution into account. All the same, for use in a standard living room, you probably could use the unit without extra highpass.
At first I thought "He is nuts, that can't be right"... but after running some simulations, a small 3" driver with an Fs = 400, Qes=0.75, Qts = 0.67 in small sealed enclosure... yeah, you are mostly right. Going a step further, with a high pass filter set to a very low value of 120 Hz, excursion is greatly limited, and I don't hit Xmax until 108 dB SPL at 200 Hz. This sort of filter would allow an elliptical crossover such as @Juhazi was talking about...Dome displacement doesn’t change anymore below Fs,
Now I understand your point...Since M74 is sealed, it should take rather high power, like ATC mid.
Lots to think about...
j.
- Home
- Loudspeakers
- Multi-Way
- High Performance 3-way based on Bliesma M74A