DIY Tweets that are on par with Electrostats

[annunaki]

Hello All, it's been awhile since I created a topic. Anyway I am getting into real planning stages for a set of DIY full active reference speakers.

Going back to some extended listening sessions I had done a number of years ago with some Martin Logan Prodigy speakers and some B&W 801 & 802 I really enjoyed the open sound of the top end on the MLs but preferred the lower mid/bass dynamics on the B&W.

I want to try and get the best of both if possible in building a 3 way reference pair.

I have been considering the SB Acoustics Sartori line for mid/midbass and am looking into dual CSS SDX 10 for the bass section or dual Ultimax 8" for the bass section.

The tweeter is where I am running into an issue. I am not opposed to standard dome style tweeters but know that they don't always match that same capability as a "stat" for open neutrality.

What are some options you may suggest I look into?

I appreciate the help in advance!

 

[Swerd]

I have been considering the SB Acoustics Sartori line for mid/midbass and am looking into dual CSS SDX 10 for the bass section or dual Ultimax 8" for the bass section.

The tweeter is where I am running into an issue. I am not opposed to standard dome style tweeters but know that they don't always match that same capability as a "stat" for open neutrality.

In general, let your choice of woofer narrow down your choice of mid range driver, and let your mid range driver narrow down your choice of tweeter.

First, choose the two woofers and run all the calculations to choose box volume and dimensions. You should be able to estimate the sensitivity from that. Look for estimated sensitivities that require the least padding down, preferably none, of the mid range and/or tweeter drivers. Aim for similar sensitivities among all drivers in the design.

Both sub woofers you mentioned have sensitivities of about 84 dB for a single driver, probably 87-88 dB if you use two. (As a guess, I think they might be better used as separate sub woofers with a small 2-way speaker, rather than part of a 3-way speaker.)

The frequency response curve of the manufacturers spec sheet should give you a good idea of the high frequency performance of the woofers. Look for a frequency where the on-axis response and the 30-45° off-axis response differ by 3 dB or less. This should be the crossover frequency to the mid range.

Both the CSS SDX10 and the Ultimax 8 are designed as sub woofers and the spec sheets don't show an off-axis frequency response. Here are two examples with 10" non-sub woofers, the aluminum Seas L26RFX/P and the coated paper Seas CA26RE4X. The aluminum L26 has a sensitivity of 88 dB and the coated paper CA26 is 91 dB. Both spec sheets show frequency response curves measured at 0°, 30°, and 60° angles, as well as the frequencies where break up happens. In a 3-way, I'd aim for a crossover frequency at or below 500 Hz.

I don't know which mid range driver you want, so I picked two different 5" Sartori drivers as examples, the Sartori MR13P-4 (91 dB sensitivity) and the Sartori MW13PNW-4 (89 dB sensitivity). The MR13P-4 has a smoother frequency response than the MW13PNM-4, see the dips at 1500 Hz. Both might be crossed over to a tweeter at about 4000 Hz.

After all that, you can choose a tweeter, knowing the sensitivity of the mid range driver and the crossover frequency. Look for a smooth frequency response on- and off-axis above the crossover frequency, and a sensitivity that is close to that of the mid range drivers. If you aim to crossover at about 3-4 kHz, you have a very wide choice of dome or ribbon tweeters. Some may actually be inexpensive. Generally you can guess a tweeter's lowest frequency that avoids distortion by multiplying it's Fs by 3. As the Fs goes lower, the tweeter's price goes up. In a 2-way speaker this becomes important, but in a 3-way you should be crossing over to the tweeter well above any low frequency distortion.

As a guess, I'd consider the Dayton RS28-F dome or one of the Fountek ribbons, such as the NeoCD3.0. But there are many other tweeters that also might do well for you.

 

[Swerd]

Mid range drivers that might be good are the Tang Band W4-1337SD 4" titanium., or a similar Tang Band 4" with a paper/bamboo fiber cone.

Dennis Murphy has used the titanium 4" as a mid driver in the Philharmonic Slims Tower, and in the the Salk Song3, he used the paper/bamboo 4".

Those 4" mid drivers have good off-axis dispersion up to 4 kHz or higher (?), allowing excellent imaging across the entire mid range. Their high frequency performance easily allow using somewhat less expensive ribbon tweeters.

 

[shadyJ]

It could be that 'open neutrality' that you enjoy about stats is simply its dipole dispersion pattern. You don't need a electrostatic panel to have a dipole pattern. Look at Legacy's higher end designs. You can still do dipole with dome tweeters. And the nice thing is, with a design like that you can easily switch to bipole if you wanted, unlike an electrostatic panel. Examine the design for this thing.

If it were me, I would go for a very wide and uniform monopole dispersion pattern. It would not be cheap, but you are talking about making a high-end speaker. One thing I would do is, to avoid floor bounce, is to put mid bass generation near the floor. So have the low frequency driver take up to 300 Hz and mounted very low, near the ground.

 

[Swerd]

If it were me, I would go for a very wide and uniform monopole dispersion pattern. It would not be cheap, but you are talking about making a high-end speaker. One thing I would do is, to avoid floor bounce, is to put mid bass generation near the floor. So have the low frequency driver take up to 300 Hz and mounted very low, near the ground.

I'm with you about a monopole design.

What you're describing is essentially the Salk Song3. It uses a potent Sartori 7.5" woofer (reaches down to 33 Hz), with a 4" Tang Band mid, and a RAAL ribbon tweet. It seems like annunaki wants a bigger woofer. With a bigger cabinet, that has the possibility of going deeper.

 

[j_garcia]

http://forums.audioholics.com/forums/threads/would-wiring-4-super-tweeters-per-speaker-box-increase-life-of-tweeters.106754/page-8#post-1180724

 

[annunaki]

Let me jump back in here and clarify my goals

I am looking to design/build a reference speaker that is capable of >=95% of the N801/N802 ML Prodigy performance at around $1,200-$1,500 all in driver cost. I want full range performance out of them as well which is why I was thinking the dual SDX-10/dual Ultimax 8" on the low end of things and blending it with a Satori 6.5" or dual 6.5" (since they are capable down into the upper ranges of the subs). Depending upon design a 3.5 way design would be an option to as I will need significant mid-bass/mid-range to keep up.

I have heard great things about a RAAL/Satori combo before so not sure if that would get me there or not as I have not heard any to date.

Thanks for the replies so far. Ultimately it is electrostat type openness/transparency with great dynamics like the N802/N801.

 

[highfigh]

It could be that 'open neutrality' that you enjoy about stats is simply its dipole dispersion pattern. You don't need a electrostatic panel to have a dipole pattern. Look at Legacy's higher end designs. You can still do dipole with dome tweeters. And the nice thing is, with a design like that you can easily switch to bipole if you wanted, unlike an electrostatic panel. Examine the design for this thing.

If it were me, I would go for a very wide and uniform monopole dispersion pattern. It would not be cheap, but you are talking about making a high-end speaker. One thing I would do is, to avoid floor bounce, is to put mid bass generation near the floor. So have the low frequency driver take up to 300 Hz and mounted very low, near the ground.

What do you think about down-firing?

 

[shadyJ]

What do you think about down-firing?

Down-firing is fine for omnidirectional bass frequencies, especially subwoofer frequencies. I think above subwoofer frequencies, as sound becomes more directional, it has to be more carefully integrated into the design of the speaker.

 

[lsiberian]

It could be that 'open neutrality' that you enjoy about stats is simply its dipole dispersion pattern. You don't need a electrostatic panel to have a dipole pattern. Look at Legacy's higher end designs. You can still do dipole with dome tweeters. And the nice thing is, with a design like that you can easily switch to bipole if you wanted, unlike an electrostatic panel. Examine the design for this thing.

If it were me, I would go for a very wide and uniform monopole dispersion pattern. It would not be cheap, but you are talking about making a high-end speaker. One thing I would do is, to avoid floor bounce, is to put mid bass generation near the floor. So have the low frequency driver take up to 300 Hz and mounted very low, near the ground.

Some of us like our speakers crazy. That's why all my fronts are bipolar.

 

[TLS Guy]

Mid range drivers that might be good are the Tang Band W4-1337SD 4" titanium., or a similar Tang Band 4" with a paper/bamboo fiber cone.

Dennis Murphy has used the titanium 4" as a mid driver in the Philharmonic Slims Tower, and in the the Salk Song3, he used the paper/bamboo 4".

Those 4" mid drivers have good off-axis dispersion up to 4 kHz or higher (?), allowing excellent imaging across the entire mid range. Their high frequency performance easily allow using somewhat less expensive ribbon tweeters.

A reference design needs power handling in spades. Those Tang Band drivers are nowhere near adequate for anything I would call a reference design.

I would recommend that you look at the SEAS Excel range that I have been very pleased with. Even with those for a reference speaker I would go MTM so you have two voice coils on the mid.

For reference speakers in the midrange, I'm very keen on metal cone drivers. These drivers maintain good performance over time. The cone is also a heat sink and so you you get less thermal dynamic compression. This should always be a major consideration for a reference speaker.

This new line from Eton is tempting, especially this 4" unit. Power handling is on the low side at 50 watts, so 2 would give you 100 watts which is still a bit lean.

However the SEAS Excel woofers have a very potent motor system which is now tried and true.

I used the 7", which gives you 200 watts where the rel power is. That is a huge issue with a lot of speakers, especially three ways, that not nearly enough power handling is allowed for in the midrange.

You have lots of choices for the woofer. Make sure you select a low Q woofer (not above 0.35) and select one with good power handling that meets the size of cabinet you want at the F3 you desire.

As far as tweeters are concerned, again power becomes an issue. In tweeters I have a bias against metal tweeters. The top end RAAL ribbon would have the required power handling. In general planar tweeters and ribbons do not.

A really good soft dome tweeter is hard to beat. There are good choices from SEAS and Scanspeak in particular. I have been very pleased with my SEAS Excel Crescendos. I have heard high end speakers using the Scanspeak Beryllium tweeters, but they seem to spit at me a little compared to my soft domes.

All I can tell is that these tweeters never seem to put a foot wrong.

Since you are going active, you don't have to worry about driver sensitivities.

Try and keep the crossover points out of the speech discrimination band as far as you can. The 7" SEAS needs to cross at 2.5 KHz. However you would be able to cross these drivers at 200 or 250 Hz and that would give you a mid pass band of just over three octaves which is fine.

If you are designing a reference speaker then it not only has to have an exemplary frequency response, dispersion and transient response, but enough power to give that real you are there power performance.

I don't think we live that far apart and you are welcome to come over to hear the SEAS drivers in my TL design.

 

[TLS Guy]

Let me jump back in here and clarify my goals

I am looking to design/build a reference speaker that is capable of >=95% of the N801/N802 ML Prodigy performance at around $1,200-$1,500 all in driver cost. I want full range performance out of them as well which is why I was thinking the dual SDX-10/dual Ultimax 8" on the low end of things and blending it with a Satori 6.5" or dual 6.5" (since they are capable down into the upper ranges of the subs). Depending upon design a 3.5 way design would be an option to as I will need significant mid-bass/mid-range to keep up.

I have heard great things about a RAAL/Satori combo before so not sure if that would get me there or not as I have not heard any to date.

Thanks for the replies so far. Ultimately it is electrostat type openness/transparency with great dynamics like the N802/N801.

Be careful of too many crossovers, especially close together.

I would use two 10" drivers with an Fs of around 20 Hz. They should be able to meet the mids. and scrap the 6.5 Sartori.

Or else use a sub driver, and then the 7" SEAS Excels will meet them.

Since you are going active a nice thing you can do if you choose woofers with a more extended range is to give the BSC of the mids to one of the woofers. This really off loads the mids and cleans things up a lot.

 

[Swerd]

For reference speakers in the midrange, I'm very keen on metal cone drivers. These drivers maintain good performance over time. The cone is also a heat sink and so you you get less thermal dynamic compression. This should always be a major consideration for a reference speaker.

This new line from Eton is tempting, especially this 4" unit. Power handling is on the low side at 50 watts, so 2 would give you 100 watts which is still a bit lean.

Funny that you mentioned the Eton magnesium mid range driver. Jim Salk and Dennis Murphy have just shown (at the Axpona show) a new 3-way, the Song3 Encore. It uses the Eton mid range, a beryllium tweeter, and a 9½" Satori woofer that achieves 25 Hz (Fs = 24.5 Hz) in a ported cabinet.

 

[TLS Guy]

Funny that you mentioned the Eton magnesium mid range driver. Jim Salk and Dennis Murphy have just shown (at the Axpona show) a new 3-way, the Song3 Encore. It uses the Eton mid range, a beryllium tweeter, and a 9½" Satori woofer that achieves 25 Hz (Fs = 24.5 Hz) in a ported cabinet.

Yes, that Eton driver looks ideal for a domestic speaker, although I think I would use two of them. I suspect the tweeter is Scanspeak. That is one of the best tweeters around, although I'm not convinced it is better than a really good soft dome.

That Satori driver looks like another winner. That is the first decent replacement for the KEF B 139. Its acoustic response and T/S specs are remarkable similar, except that unit does not resonate violently at 900 Hz. The T/S specs are right in the sweet spot for a traditional TL. Those drivers should work well with very low order crossovers. I see Jim claims phase coherence at crossover.
I bet that is a nice sounding speaker.

Do you know if that is a true TL, or another mass loaded effort?

 

[Swerd]

Yes, that Eton driver looks ideal for a domestic speaker, although I think I would use two of them. I suspect the tweeter is Scanspeak. That is one of the best tweeters around, although I'm not convinced it is better than a really good soft dome.

I believe the tweeter is made by an American company, but I cannot remember the name. Dennis Murphy had told me, some time in the past year, that he had tested some new beryllium dome tweeters that he liked. It was not made by Scanspeak, Focal, or any of the usual recognizable names.

The sensitivity of the Eton mid is similar to that of the woofer and tweeter. The Eton is expensive, and I think Jim was trying to keep the price of the Song3 Encore from going higher than it is, $5900 plus FedEx shipping for a pair of 97 lb. towers.

That Satori driver looks like another winner. That is the first decent replacement for the KEF B 139. Its acoustic response and T/S specs are remarkable similar, except that unit does not resonate violently at 900 Hz. The T/S specs are right in the sweet spot for a traditional TL. Those drivers should work well with very low order crossovers. I see Jim claims phase coherence at crossover.
I bet that is a nice sounding speaker.

Do you know if that is a true TL, or another mass loaded effort?

Salk makes other similar 3-ways with the 7½" Satori woofer and either a Tang Band (Song3) or Accuton 4" (Song3-A) mid driver. Both use a RAAL ribbon tweeter, and both use a standard bass reflex cabinet, not a MLTL. I'll guess the 9½" woofer also uses a similar alignment. In the photo (above) Dennis sent me, you can see the rectangular port at the lower edge of the cabinet of the speaker with one woofer. It is advertised as available with either a front or rear port.

If Jim claims "phase coherence at the crossover frequency", it's likely a 4th order L/R crossover that Dennis uses so often. That puts drivers 360° out of phase with regard to time, but one full cycle out of phase equals in phase if you consider phase only.

 

[Swerd]

The beryllium dome tweeter in the Song3 Encore could be made by Transducer Lab. I'm not sure which model, either the N28BER-A or -G.
http://www.transducerlab.com/index.php?pag=pro

Here is the N28BER-G

And here is a close up of the Salk Song3 Encore mid range and tweeter

 

[TLS Guy]

The beryllium dome tweeter in the Song3 Encore could be made by Transducer Lab. I'm not sure which model, either the N28BER-A or -G.
http://www.transducerlab.com/index.php?pag=pro

Here is the N28BER-G

And here is a close up of the Salk Song3 Encore mid range and tweeter

This is the Scanspeak, so it probably is the Transducer Lab with a mesh grill which does not show on their site.

The FR is better then the Scanspeak. It comes with or without ferrofluid. It is a very low Q tweeter so could be driven down below Fs.

I think you could use very low order crossover with the drivers in the Song 3.

I suppose technically a fourth order LR filter is phase coherent in a way, except the coherence is a whole cycle apart, and so is certainly NOT time coherent. That is why in audio and in crossovers we need to use the term time and not phase. It actually expresses the same thing. However if you express it as time it forces you to realize that unless there is no time delay or advance then it is actually not phase coherent. This may sound like semantics, but I think it is an important point and makes for far greater clarity of thought.

I have a design on the books for a two way, that only has three components in the crossover and one of them is a resistor. That is almost time perfect. Trouble is I have no need of the speaker and no takers. You could turn it into a three way with that Satori driver and add probably only one or two components more.

Drivers are appearing with much wider bandwidth. And so we should be able to design and build better speakers with far less time aberration.

 

[Swerd]

… so it probably is the Transducer Lab with a mesh grill which does not show on their site.

Perhaps the mesh grill is how Transducer Lab provides a custom version of their driver for Salk. If it isn't in the catalog, it must be "custom made" .

I suppose technically a fourth order LR filter is phase coherent in a way, except the coherence is a whole cycle apart, and so is certainly NOT time coherent. That is why in audio and in crossovers we need to use the term time and not phase. It actually expresses the same thing. However if you express it as time it forces you to realize that unless there is no time delay or advance then it is actually not phase coherent. This may sound like semantics, but I think it is an important point and makes for far greater clarity of thought.

I've had similar discussions before with Dennis.

He believes that efforts at making a crossover "phase coherent" within one octave of the crossover frequency are worthwhile, where there is no difference between 0° out-of-phase and 360° out-of-phase. He says it can be demonstrated that listeners can audibly detect differences in sound when phase correct speakers are compared to other similar speakers with non-phase coherent crossovers.

Efforts to preserve true time coherence are more difficult (expensive). He also says when comparing a time coherent design with off-set drivers and 1st order (acoustic) crossovers to a 4th order phase coherent crossover but no off-set drivers, listeners do not seem to hear a difference.

That's why he passes on trying achieve true time coherence, but does make efforts in his 4th order LR crossovers to achieve symmetrical roll-off curves of both drivers within one octave of the crossover frequency. Those unusual LCR filters he puts in his crossovers are usually aimed to fine tune that.

I have heard Vandersteen 3a speakers which claim time & phase coherence, and uses 1st order acoustic crossovers (very complex filters with many parts) as well as arrangement of the different drivers to achieve similar physical acoustic origins. They are nice sounding, but the imaging sweet spot in the range between the mid range and tweeter is very narrow. If you move your head inches to the left or right, their wonderful imaging vanishes.

In contrast, the imaging that Dennis's designs achieve with 4th order LR crossovers and no time aligning off-sets of driver mounts sound equally wonderful, and remains so for a much wider distance. In my listening room, with two speakers about 10' apart, the sweet spot is nearly one sofa wide, at a listening distance of about 9'. These are 2-way MTM speakers with SEAS Excel W16 6" woofers and a "custom" RAAL ribbon tweeter, crossed at about 2.5 kHz.

 

[TLS Guy]

I think those tweeters would have to come from the factory with a grill. EPA laws here and certainly Europe forbid the possibility of touching Beryllium. That was why Shure had to stop production of their V15 series cartridges. As I'm sure you know, Beryllium is a highly toxic metal and carcinogenic as well.

There certainly is a pecking order with a crossover with smooth FR and good dispersion having pride of place.

Low order crossovers are hard to implement I can tell you that. My rears have first order crossovers at 900 Hz, and 5 KHz the 180 Hz crossover being third order. However it does not really have a sweet spot and sounds good all the way round the room when I play them solo. They have a very good sound stage and high "immediacy" factor about them. I think when it comes to time shifts, people who listen to speakers a lot have tuned out the problem. However I maintain that you should shift time the minimum required to make a decent speaker.

I think I'm more sensitive as I listened so long to full range speakers and still do.

The speakers I'm listening to here at Eagan right now are first order transition to second high pass and first order low pass, making a composite fourth order, or pretty close. The LF crossover is almost entirely acoustic with the cross between the band pass and bass being second order acoustic plus second order electric in the top end and the low pass entirely acoustic fourth order, with the electric component well above the pass band. Again absolutely no sweet spot issues and excellent room coverage. So I'm not convinced that the order per se creates sweet spot issues, but how it is implemented.

I would say that I think I do like full cycle time shifts the least. I absolutely do believe that all things being equal you should aim for designs with minimal time shifts. Of course in my back ground I was highly influenced by Ted Jordan from and early age. But I do think that his researches and teachings have enormous validity. I do think his star is starting to rise again.

The sad fact remains that it is a real challenge to have a nice smooth frequency response and dispersion without creating time aberrations. Achieving the first while minimizing the second is real work.

 

[Swerd]

An interesting conversation topic on this very rainy day on the East Coast. Thanks.

I don't have much else to offer on the subject. It's quite possible that the different opinions and solutions you & Dennis Murphy have arrived at may be due to the different types of drivers you've used.

Perhaps Dennis can chime in. He was in Chicago at Axpona over the weekend, and just flew back the other day.

I think those tweeters would have to come from the factory with a grill. EPA laws here and certainly Europe forbid the possibility of touching Beryllium. That was why Shure had to stop production of their V15 series cartridges. As I'm sure you know, Beryllium is a highly toxic metal and carcinogenic as well.

I actually knew nothing about beryllium toxicity or carcinogenicity, despite what I did for a living. I just read about it on Wikipedia.

The cancer treatment world focuses heavily on coming up with new cancer treatment drugs. This requires clinical trial testing, which requires access to patients willing to join clinical trials. If a disease is rare, such as those caused by beryllium exposure, a clinical trial focusing on that would take too long to complete. As a result, most trials focus on cancer types with plenty of patients. Mesothelioma is another example – fewer and fewer patients are available, and all of them have been heavily pre-treated. That's also why fewer pediatric clinical trials have been done.

In general, I found the US oncology community remarkably uninterested in addressing cancer prevention. How do you design a clinical trial for the prevention of a disease? We of course know all about how smokers who stop smoking are much less likely to develop lung cancer, but what about the other types of cancer that may have genetic, viral, or other environmental causes?

In a sense, we are performing a large scale preventative trial by immunizing a young generation to human papillomavirus (HPV). No one now argues that it won't prevent cervical cancer in women (that evidence is already accumulating), but what other diseases will vaccination against HPV prevent?

I'm getting well off topic and will stop. Nothing causes peoples' eyes to glaze over more than talking about cancer.