Introducing Directiva - An ASR open source platform speaker project

[ctrl]

With the crossover developed by @ctrl, am getting a bit over 90 dB at 2.83V/1m. Am expecting it will stay around that number when Amir measures.

If you set a very slightly decreasing frequency response (about 1dB) in the low frequency range, via crossover, then [email protected]@1m sensitivity is realistic. This is then the sensitivity of the speaker at 2.83V, whether passive or active.

The shelfing filter in the crossover simply raises the low frequency range massively by 6-8dB, but the amplifier power for this must also be generated somewhere. An active crossover does not "conjure up" a 7-8dB sensitivity increase

I've been looking for a small Speaker that can do ~115 dBSPL at 1 meter (above 100Hz) with sensible sensitivity to boot.

If the Directiva is used down to 80Hz with a subwoofer, then up to 106dB is realistic for a short time, maybe a bit more. If the signal covers the full frequency range (20Hz-20kHz), then at 106dB the woofer and the PR reach the limit for linear excursion.
But this requires 400W amplifier power (with 600W 108dB would be possible for a short time).

The SeasDXT tweeter would probably handle this as well, as has already been shown here.

But to reach 115dB regularly, with acceptable amplifier power, the Directiva is not suitable.

For this you should use PA drivers and consider a compression driver for the high frequency. A sensitivity of [email protected]@1m is then possible.
But if you want to cross over at 80Hz to the subwoofer, a 6'' driver will or/and could probably not play low enough. You may need to increase the crossover to the subwoofer to 100Hz.

 

[abdo123]

But to reach 115dB regularly, with acceptable amplifier power, the Directiva is not suitable.

I meant 115 dBSPL @ 0dBFS with a sharp (24 or 48 dB/oct) highpass filter at 100-120 Hz. I'm limited by size aesthetically and I can only do bookshelf speakers (or something that can be sensibly placed on a shelf).

 

[ctrl]

I meant 115 dBSPL @ 0dBFS with a sharp (24 or 48 dB/oct) highpass filter at 100-120 Hz. I'm limited by size aesthetically and I can only do bookshelf speakers (or something that can be sensibly placed on a shelf).

The Directiva has a sensitivity of 83dB [email protected]@1m. You want to achieve 115dB SPL@1m.

The amplifier must produce 32dB additional to reach this mark, for this you need 120V@4Ohm (32 = 20*log(120V/2.83V)), so 3600W@4Ohm (=U²/R).
If you cross the woofer at 120Hz with LR4, then theoretically the excursion of the driver would remain almost in the linear range. But whether the voice coil can survive a transient sound with 3600W, I can not say.

UPDATE:
The power required in real terms is likely to be significantly lower, since the sensitivity of the Purifi driver increases with increasing frequency.
Therefore, you should get by with 1000W and only need it in the range around 200Hz.

 

[abdo123]

The Directiva has a sensitivity of 83dB [email protected]@1m. You want to achieve 115dB SPL@1m.

The amplifier must produce 32dB additional to reach this mark, for this you need 120V@4Ohm (32 = 20*log(120V/2.83V)), so 3600W@4Ohm (=U²/R).
If you cross the woofer at 120Hz with LR4, then theoretically the excursion of the driver would remain almost in the linear range. But whether the voice coil can survive a transient sound with 3600W, I can not say.

oof I was under the impression the sensitivity would be 90dB @2.83V.

 

[ctrl]

oof I was under the impression the sensitivity would be 90dB @2.83V.

See my updated post above.

The Purifi driver reaches 90dB SPL@1kHz@1m in the infinite baffle, but at lower frequencies and after installation in an enclosure, the sound pressure drops rapidly.

UPDATE: Screwed up the voltage/power conversion. Therefore changed the text.

In the left picture you can see the simulated frequency response of the Directiva speaker, which slightly decreases towards low frequencies. At 2.83V the sensitivity is about 83dB SPL@1m.

But the same power is not demanded in the whole frequency range - 2.8V (1.2W, 6.4Ohm) at 40Hz and 0.9V (0.25W, 3.6Ohm) at 1kHz (see center picture, red arrows).

If someone now excites the Directiva with a 1kHz sine wave and measures the voltage/power at the amplifier, he just gets 0.9V. Okay, the sensitivity is determined at 2.83V, so this someone increases the voltage at the amplifier until 2.83V@1kHz sine is reached, measures the sound pressure level of the speaker and comes to incredible >90dB SPL sensitivity at 2.83V@1m.
But of course the speaker now needs about 8V at 40Hz - which far exceeds the 2.83V limit.

 

[abdo123]

See my updated post above.

The Purifi driver reaches 90dB SPL@1kHz@1m in the infinite baffle, but at lower frequencies and after installation in an enclosure, the sound pressure drops rapidly.

In the left picture you can see the simulated frequency response of the Directiva speaker, which slightly decreases towards low frequencies. At 2.83V (roughly 1W) the sensitivity is about 83dB SPL@1m.

But the same power is not demanded in the whole frequency range - 1.2W at 40Hz and 0.25W at 1kHz (see center picture, red arrows).

If someone now excites the Directiva with a 1kHz sine wave and measures the voltage/power at the amplifier, he just gets 0.25W. Okay, the sensitivity is determined at 1W, so this someone increases the voltage at the amplifier until 1W@1kHz sine is reached, measures the sound pressure level of the speaker and comes to an incredible 88 or even 90dB SPL sensitivity at 1W@1m.
But of course the speaker now needs at least 6W at 40Hz.

View attachment 137924

What is the sensitivity like above 100Hz? since it's the topic of conversation.

 

[ctrl]

What is the sensitivity like above 100Hz? since it's the topic of conversation.

Still 83dB.
If you accept some sound pressure level loss around 100Hz (and thus ripple), then you can tune the speaker to a maximum of 84-85dB sensitivity @2.83V@1m.

 

[RickS]

Ok, why are we quoting different sensitivity numbers for Directiva?

It boils down to different conditions for which the number is generated. This presents a challenge to comparing measurements for different speakers. The differing conditions are distance, baffle size, drive level, impedance, and frequency bandwidth.

Let’s start be looking at the Purifi woofer. It is rated at 88.7 dB at 1 meter away mounted on an “infinite” baffle driven at 2.83V from 300-800 Hz. Looking at this plotted, you get the following…

Note, from Purifi posted measure, the impedance varies from about 2 ohms to 4 ohms from 300 to 800 Hz. So, you can see why talking about a sensitivity measure that conditions are important. As @ctrl mentioned, Directiva has a finite baffle and is mounted in a box with a passive radiator. This increases the impedance about 1 more ohm from 300 to 800 Hz. For SPK5 with port tuning and a passive crossover, the impedance ranges from 5-8 ohms from 300 to 800 Hz.

So the question becomes with varying impedances of the different designs and SPK5 with a passive crossover vs a Directiva's active one, can you have a sensitivity rating that compares them meaningfully? I think you can see why it may be difficult. In the next post, will explore this further...

 

[RickS]

A little diversion today as I hijacked my neighbor's concrete contractor to fix some issues around my house! But now back to the Directiva sensitivity rating discussion.

So, I thought it would be good to start with a visual. This one compares the SPK5 with Directiva r1...

The green trace is SPK5 and the sensitivity is around the 83 dB mark just as with Amir's review. For Directiva r1, I used a 4 ohm dummy load and set the amp output to 2.83V and measured at 1 meter using a 1 k Hz sinewave signal. You can clearly see that the SPL exceeds 90 dB from 300 to 3000 Hz. Yes, there is gain added here from the active crossover, but there are also filters lowering the gain comparably in the target frequency range. As stated earlier, I expect that when Amir review Directiva, the Klippel will have comparable SPLs.

So, is it accurate to rate the Directiva sensitivity so much higher? IMO, as long as the conditions are stated, I have measured two different speaker designs under comparable conditions. If I had not been asked to measure Directiva r1 sensitivity, unless the Klippel is doing something much different to set the amplifier voltage, the only difference is that we would be having this discussion after Amir's review. @ctrl and I have discussed this topic at length and, since we have a difference of opinion, we agreed that this would be a good place to start a open dialog.

As many of you already know, comparing sensitivity between speaker brands is fraught with inaccuracies. At this stage, all I can offer is consistency within my test rig. I have reviewed sensitivity testing procedures from multiple source and they are all different or incomplete in some way. Am always open to improving, so let the constructive debate commence!

 

[mdsimon2]

I think I probably side with @ctrl here.

Seems to me the most reasonable way to compare active speaker sensitivity to passive would be to use an input level such that the maximum amplifier output was 2.83V (I would probably use the speaker instead of a dummy load but hopefully the amp is well behaved and it doesn't matter). I would run a frequency sweep in REW and record the amplifier output with an audio interface. At these output levels you wouldn't even need a voltage divider (although probably not a bad idea to make sure a spurious pop doesn't fry the interface).

Using 1 kHz seems arbitrary and does not represent the area of maximum amplifier output. I like the idea of showing the actual acoustic response rather than just reporting one number as well.

Michael

 

[Selah Audio]

90dB sensitivity? You're beginning to sound like Tekton advertising.

 

[Selah Audio]

I meant 115 dBSPL @ 0dBFS with a sharp (24 or 48 dB/oct) highpass filter at 100-120 Hz. I'm limited by size aesthetically and I can only do bookshelf speakers (or something that can be sensibly placed on a shelf).

The Directiva borrows from my own design with the bass tuning. Here's a review Erin Hardison did with my Purezza - it may help you decide if the woofer works for your system. https://www.erinsaudiocorner.com/loudspeakers/selah_audio_purezza/

 

[tktran303]

Sensitivity, as referenced to 2.83V input = x dB output @1m for a specific frequency(ies) means little in an active speaker.

The active equaliser/DSP and it’s gain structure can take any input signal and boost to any abitrary value.

In an active speaker what makes sense is the maximum SPL; which is ultimately displacement limited. And the input sensitivity, which is related to the voltage gain.

@ctrl is correct.

 

[RickS]

For the record, I happen to see more than one perspective on the sensitivity question. I have no disagreement that @ctrl is technically correct. I am trying to drive out a meaningful discussion in an attempt to answer @abdo123’s original question meaningfully.

For most active speakers, one could only measure sensitivity starting at the input. Since Directiva has external electronics, what I did was to measure it the same way as if it were a passive speaker. Yes, calling it “sensitivity” is inaccurate, but it is not max SPL either. Lacking anything else, I stated my conditions and made the posted measurements.

The sensitivity of the Directiva tweeter is much higher than the woofer. In a passive design, it would be padded down and that higher sensitivity is wasted in the form of heat. One of the advantages of an active crossover is you can use gain to drive the woofer more and waste less of the higher sensitivity of the tweeter. So, the net “sensitivity“ of the overall speaker is higher than the passive case.

 

[abdo123]

The Directiva borrows from my own design with the bass tuning. Here's a review Erin Hardison did with my Purezza - it may help you decide if the woofer works for your system. https://www.erinsaudiocorner.com/loudspeakers/selah_audio_purezza/

below 1% distortion @ 100 dB above 100Hz is a huge achievement for one driver, how large should the baffle be if you want bass to remain flat to 100Hz? instead of starting to roll-off around 300Hz?

 

[abdo123]

The sensitivity of the Directiva tweeter is much higher than the woofer. In a passive design, it would be padded down and that higher sensitivity is wasted in the form of heat. One of the advantages of an active crossover is you can use gain to drive the woofer more and waste less of the higher sensitivity of the tweeter. So, the net “sensitivity“ of the overall speaker is higher than the passive case.

Now that's the thing, you make it sound as if the extra woofer output is coming out of nowhere, but in reality, you're just doing the same as padding the tweeter but with extra steps .

That's why me & @ctrl got confused.

 

[RickS]

Now that's the thing, you make it sound as if the extra woofer output is coming out of nowhere, but in reality, you're just doing the same as padding the tweeter but with extra steps .

That's why me & @ctrl got confused.

I regret the confusion and is why I wanted to state the conditions and clarify. No, there is no defying of physics!

But not sure what you mean by extra steps?

The active crossover could lower the output of the tweeter just as the padding does in a passive crossover. If we wanted to use the active crossover to design the passive one, this could be done. Otoh, as I stated, why not take advantage of the higher tweeter sensitivity by adding some gain to the woofer (if it can handle)?

 

[abdo123]

I regret the confusion and is why I wanted to state the conditions and clarify. No, there is no defying of physics!

But not sure what you mean by extra steps?

The active crossover could lower the output of the tweeter just as the padding does in a passive crossover. If we wanted to use the active crossover to design the passive one, this could be done. Otoh, as I stated, why not take advantage of the higher tweeter sensitivity by adding some gain to the woofer (if it can handle)?

It's a reference from Rick and Morty, you know slavery with extra steps?

 

[D!sco]

I had a thought about theater applications. it may be a bit early for victory laps, but a mini sub to fill in the subsonic frequencies around the R1’s F3 would be a nice touch. It may be expensive, but an MTM model of the R1 with higher sensitivity would be more appealing for larger living room and surround/center channels.

Speaking of range, what is the goal for bass response/port tunings? I read a white paper on EQSS, or equalized quasi sealed systems. This relies on a much lower tuned vent or port to resonate below the intended driver F3, using EQ to boost the slump which occurs between the driver resonant frequencies and the port’s. Say a woofer with an F3 of 60hz is put in a box with a slot tuned to 20hz, you could (attempt to) use the equalizer to flatten out that 60-20hz slump. I haven’t had the chance to try it yet, but as this is a minidsp project it seems like a good place to try.

Love the project so far.

 

[D!sco]

I'm a DIY Audio newbie who's been toying with VituixCAD to slam drivers together like a child would attempt to mate transformers action figures. Do we have enough data on these drivers to provide usable FRD/ZMA files? I'd like to play with passive crossover values and see what's possible.