PE Dayton Reference Series

[JonMarsh]

Do you think tha'ts because the rear window is reasonably tall, in spite of the large area for the spider plate and magnet? Are you going to try the RS180's "officially", or stick with the CA18 for your design?

I hope to do some measurements of these pretty soon- maybe this weekend, if not, definitely the weekend following.

I've wondered about making a version of the Arvo dipole more like my first proto, with the 11" woofers, possibly Splinter sound drivers, but with the RS180 on top. OTOH, it's not like I need another permutation! Since MFK is doing one, and you're experimenting with this also, I'll just sit back and see what you guys come up with, while I finish tuning the current Arvo's and start the serious construction drawing for the Saint-Saen. I ran the excursion/SPL calcs on what I think will be my first build concept for those, and it's looking good for my target SPL of 107 dB. (dual TC2+ on each side (OK down to ~35 Hz for 4 assuming low bass fed to both channnels), ~70 Hz crossover to 8 Extremis 6 (OK to 50 Hz for 8 @ 107 dB), then the RD50's, then JP2's.)

[cjd]

If you're using the PE published data files, do yourself a favor and run them through the Frequency Response combiner and do an Extract Minimum Phase, export the result and use that. You can also use this tool to combine BDS results with the driver response.

So far I have found this to yield surprisingly good results - using the data straight from PE can introduce trouble (been there too).

Have a trio of in-progress MTMs (RS150 + Seas 27TDFC) that are currently running open baffle with a tentative crossover designed with PE's data files. Soon as the mess of kitchen renovation is cleaned up I'll get back to these - including doing measurements and comparing with the results I got using PE's numbers plus the various FRD Consortium tools.

oh yeah... the phase bit finally prompted me to join up here... after lurking for who knows how long. back to the shadows for me now.

C

[JonMarsh]

Well, welcome to HT Guide! Don't be bashful, and when you have some results you're comfortable sharing, please, by all means do!

I haven't been using the PE data, mostly because I haven't trusted it very much, and it's not relevant to my baffle applications. I've take your route once and had pretty decent results; LspCAD 6 should support some of that, too, but at this point I trust BDS more, having used it on several projects and gotten repeatable results verified by measurements.

My biggest concern is getting good nearfield data to support judging what the true highest crossover frequency I can use may be, though I have my intuitions, anyway. For the center channel project with RS180's, I'm not expecting the crossover to be over 1400 Hz, with SS D2904/6000-01. Those are just the cutest little things, and allow really tight driver spacing, and can go fairly low with the right crossover.

~Jon

[Hank]

(dual TC2+ on each side (OK down to ~35 Hz for 4 assuming low bass fed to both channnels), ~70 Hz crossover to 8 Extremis 6 (OK to 50 Hz for 8 @ 107 dB), then the RD50's, then JP2's.)

:drool:

BUT, $$$ :rant:

[capslock]

Let me guess, which of the many, many manufacturers it could be...?

Actually, I work for the company that makes the litho lenses of which we'd like to sell more to your employer.

Greetings,

Eric

[JonMarsh]

Yeah, you know how it goes, Eric. When the semi guys catch a cold (sniffle), the equipment makers get pnuemonia!

Yeah, there aren't very many large world class German semiconductor companies, are there?

Mit fruendlich grüßen,

Jon

[Davey]

Jon,

The spider area of both drivers is much the same and creates outlets to the side that are nearly identical. The area behind the spider is where the large difference is, and I think that's of minimal concern.

Yes, I'm going to keep using the CA18 drivers in my design since they're similarly priced and work well. I've sent one of the spare CA18 drivers to Mark K. to include in his next go'round of 7" driver tests. That will be informative since we can see how it performs relative to the RS180 in his objective tests.

I don't see any reason why the RS180 drivers couldn't be used, but it would require a more complicated crossover. Also, my innovative mounting scheme would have to be changed considerably.

I think the basic design I came up with is sound and will accomodate a number of different inexpensive drivers. However, I doubt there will be much interest in building this design since folks interested in dipole designs are generally looking to spend bigger dollars and trying to achieve something much closer to state-of-the-art. But, it is a good exercise and shows the relative differences between dipole and regular box speakers.

Cheers,

Davey.

[JonMarsh]

Jon,

The spider area of both drivers is much the same and creates outlets to the side that are nearly identical. The area behind the spider is where the large difference is, and I think that's of minimal concern.

Yes, I'm going to keep using the CA18 drivers in my design since they're similarly priced and work well. I've sent one of the spare CA18 drivers to Mark K. to include in his next go'round of 7" driver tests. That will be informative since we can see how it performs relative to the RS180 in his objective tests.

I don't see any reason why the RS180 drivers couldn't be used, but it would require a more complicated crossover. Also, my innovative mounting scheme would have to be changed considerably.

I think the basic design I came up with is sound and will accomodate a number of different inexpensive drivers. However, I doubt there will be much interest in building this design since folks interested in dipole designs are generally looking to spend bigger dollars and trying to achieve something much closer to state-of-the-art. But, it is a good exercise and shows the relative differences between dipole and regular box speakers.

Cheers,

Davey.

Well, I don't know that I'd seel the budget dipole contingent short yet- there's a lot of folks that I think would go for that if they ever heard one- and I know many DIY constructors need to save money on drivers- that's the reason I've done variations on some of my designs, and picked tweeters at time that, while not exactly interchangeable, were darn close- like the Vifa XT25 for those short on funds, versus the SS 98000 for the same design with only minor zobel tweaks.

Look at the interest in JohnK's NaO.

So, I suspect there's lots of folks who would be interested in a good reasonably priced dipole system. Especially if they heard it.

A number of guys who have heard the Arvo Part's at ThomasW have been interested in building them- they get hooked on the dipole clarity pretty quickly.

Anyway, I'll be curious to see how your current project turns out, which ever driver's you go with. It's the mark of really good engineering to get good to very good performance overall out of relatively modestly priced components...

~Jon

[RonS]

Davey,

I'd definitely encourage you to continue with the budget design. While the Orion is an excellent speaker, it's cost and complexity are a major hinderance. I think there's quite a few people out there who would love to try a dipole, but just can't, or don't want to, spend the money on an Orion or similar. Plus, a less expensive and simpler design allows people to experience the dipole sound, and then maybe upgrade in the future.

As you know, I'm using fairly inexpensive drivers (XT25, M8A and 1252DVC), but I am going the digital xover route, so that I can upgrade and play around in the future somewhat easily, without loosing my shirt on crossover parts.

Cheers,
Ron

[cjd]

However, I doubt there will be much interest in building this design since folks interested in dipole designs are generally looking to spend bigger dollars and trying to achieve something much closer to state-of-the-art.

Nuh uh. People that know dipole, yes. I brought a set of $9 dipoles to the DIY Chicago event (which used raw PE data and had a nifty dip before XO frequency that, when I re-normalized phase data and adjusted the XO, cleaned up nicely...) The comments? The open, spacious sound got people - really surprised a number of folks, to the extent that it took a number of speakers for the feeling of closed in sound to go away (probably, also the point where the parts used started to really become dynamic. The fact that they had no bass, no mid-bass, not even really any lower midrange, kind-of hurt their reception otherwise, but that worries me so terribly. . . I mean, I burned $9 of parts I had lying around! Then again, the lack of bass also helped un-complicate things such that detail could be rendered cleanly where people are most sensitive.

So... keep at 'em!

C

[Mark K]

Davey,

I too, think there's a role for budget dipoles. You've just got to get a flashy website now...

[aaaa]

RS225 Measurements

Well, after having to mess around with some cable problems, and fighting the %&@%@# binding posts on the new Palladiums (that CE stuff, designed to prevent idiots from shorting things out, makes it impossible to use many spade connectors, such as WBT), I got some basic meausrements done on my new RS225s, inlcuding in the M8ta enclosures.


First, let's look at the free air impedance.



This plots a little bigger than I would normally do, just so you can see the glitch at ~1.6 kHz fairly clearly.





Here, in the enclosure, is the impedance curve; note the difference in scale, so the 1.6 kHz glitch amplitude, though reduced, is still evident.

The LF tuning for the port enclosure combo is quite as expected, with an impedance null at 22-23 Hz, the center of the port tuning.

Now, next step is to look at the VERY near field response, hunting for snarks as it is, in the cone behavior. Mark K and I have both seeen irregularities in very nearfield measurements which are somewhat smoothed out (masked) in one meter measurements, but which in our opinion show problematic behavior. Don't ignore the clues, I say...

This is a near field (1") MLS measurement...



See that sharp dip at 1.5 kHz? That coincides pretty well with the impedance glitch. It also is very, very similar to Mark K's measured results. We'd hoped that the published PE curves were the results of an improved cone, but that doesn't seem to be the case.

Here's a more conventional 1 meter measurement; since it's not on an IEC baffle, it shows the characteristic drop between 800 Hz and 200 Hz. This is why you need baffle step compensation.



Note how the 1 meter resonse is smoothed by contributions from the total cone output- but this doesn't tell the story, with regards to that first cone mode and dip at 1.5 kHz. The dip at 200 Hz is due to adjacent room boundary for where the speaker is located, and using a relatively lont time gate.


Based on these results, I thought some detailed swept sine measurements nearfield would be a good idea, as that will have better resolutionn than MLS.

Here's the swept gated sine nearfield to the driver, from 10 Hz to 2 kHz:




This shows the dip quite clearly at 1.5 kHz, as well as the relative cone motion null at ~22 Hz, from the port tuning.



Just for fun, I also measured on the floor near the port exit from the cabinet.




Various bits and pieces in my living room were shaking and rattling during these tests from the gated sine measurement, between about 18 Hz and 35 Hz; for a single small LF driver, there was a lot of clean low frequency content- certainly the deepest I've heard from an 8" woofer! The response above 100 Hz is a combination of port output and reflected sound in the room from the driver; it's hard to do this test and get a clean output of just the port, and the HF port and room respose does a lot of comb filtering.


Conclusions

LF box tuning is right on the money.

Mark K's original meaurements were correct and still valid for production RS225.

Highest crossover frequency which I think I would use with the driver is 1 kHz, with my 8th order equivalent Cauer-elliptical. In my mind, that leaves the Hales Transcendence tweeter (pretty flat down to 550 Hz) or the Millenium Excel (which most, but not all, published plots show roughly flat to 1 kHz and a very slow roll off below that until 500 Hz) as the possible candidates to mate with it.

Since the current cabinets are cut out for the larger tweeters, and either woofer fits, and I'm setup for external crossovers, may as well try making this work, and see what happens, eh?


~Jon

A very intresting thread. Thanks for the data, Jon M & Mark K.

I'm trying to understand what's going on here, and have some questions for Jon M and Mark K: did you do any close up, say 1", measurements off center? And then take data at gradually increasing increasing distances to see how the 1.6KHz glitch fills out? A lot of data I know, but just thought I'd ask in case either one of you have done this.

Monte Kay has also measured the 8" RS on his site. The stored energy data at 1600 HZ is at:

http://www.mfk-projects.com/images/Dayton%20RS/8in/8@1600.gif

There seems to be no evidence of any ill effects from the 1600 Hz close up notch. Indeed, it measures somewhat better than the SEAS Excel at this frequency. As far as I can tell the notch does does not adversely affect the RS' distortion either. Under the circumstances, it seems to me that this glitch is benign in the RS, certainly when compared to the very pronounced high Q peaks measured in certain other metal cone drivers, and ought not preclude the use of this driver out to higher frequencies .

Comments anyone?

[JonMarsh]

I'm trying to understand what's going on here, and have some questions for Jon M and Mark K: did you do any close up, say 1", measurements off center? And then take data at gradually increasing increasing distances to see how the 1.6KHz glitch fills out? A lot of data I know, but just thought I'd ask in case either one of you have done this.

I've done that type of measurement with several other midwoofers (HiVi M8a, SS models), but not yet with the RS225, though it's on the list for the Xmax break. The thing is, for 8" driver cones which ARE pistonic up to 2 kHz, or so, there's no such "glitch". I've seen simlar problems in other drivers, such as a 1 kHz dip of the same magnitude in an MCM 6-1/2" carbon fiber driver I was investigating a year or so ago. It's a very nice part, about 4.5 mm Xmax, and would work well paired up in a line array with something like the RD50, for example, where I could cross it over at about 600 Hz. OTOH, an Extremis 6 is better...

I'm no particular authority, you know, and if you would like to try a system using the R225 higher, it's certainly anyone's right to do so, regardless of what I think. However, the ETC curves on the Dayton RS225 at 1200, 1400, and 1600 don't look that great, even 800 is a bit marginal. So, my own judgement for a system I want to build, is stay away from going that high. Just like I think the 10" RS270 is quite good up to 600 Hz, but falls apart above that.

YMMV.

Everyone has their own goals and criteria- we just try to share our thoughts and reasoning so that it will give folks something to think about in making their own selections and choices for their designs.

~Jon

[Mark K]

Monte Kay has also measured the 8" RS on his site. The stored energy data at 1600 HZ is at:

http://www.mfk-projects.com/images/Dayton%20RS/8in/8@1600.gif

There seems to be no evidence of any ill effects from the 1600 Hz close up notch. Indeed, it measures somewhat better than the SEAS Excel at this frequency. As far as I can tell the notch does does not adversely affect the RS' distortion either. Under the circumstances, it seems to me that this glitch is benign in the RS, certainly when compared to the very pronounced high Q peaks measured in certain other metal cone drivers, and ought not preclude the use of this driver out to higher frequencies .

Comments anyone?

I guess it depends on how you interpret this. I think that ideally, a driver should track the theoretical curve to -30 at least. Monte's graph shows the RS225 doesn't do this. In fact, it's not a very good ETC curve. After -15-20dB, the driver starts releasing energy/ringing. I don't think that this is that bad relatively. Even the excel has trouble and stores some energy. But, ideally, an 8" driver should be kept at or below 1.5k, for a number of reasons.

And we clearly have tweeters that can do this.