Blog

Blog

Saturday, January 20, 2007

Time flies when your site is getting stale...What can I say, I'm really busy with life-4 kids, full time career as a physician. I know it's a bit of a broken record. Still, rest assured I'm working on projects. Coming shortly is another 7" distortion comparison. It was motivated by Zaph's tests. Generally, our test results more or less mirror each other, with some notable exceptions. Again, this appears to be the case. The Usher 8945P and the Scan revelator are coming out on top in the nonlinear distortion dept. No revolutionary changes. For the most part, I only tested midlevel or higher drivers. No poor performers/budget drivers. I will present the data a bit differently, in more of a narrative format. Sometimes I think that I post way too many graphs and that some folks just get confused. I've always resisted posting too much in the way of summary comments, but I am tempted to come up with a leaderboard. Hey, it's my site and if I can't have an opinion on my site, what's the point.

I've only got a couple of projects this year, not counting the laminate flooring in the family room, Uh, 700 sq feet, and 500 sq feet of tile in the living, dining and bathrooms, more landscaping...but I digress. Oh, speakers. Well, that too. I keep wanting to start some projects, but people keep sending me drivers to test. For 2007 though, I only have 3 projects. I'm working on a dipole 2 way. I'm trying to get close to full range output out of 4 RS225's per side passively crossed to a 27TBFC/G. There is active dipole eq. I've got a pair of NaO mini boards and I'm going to try to modify these for use. John K is receptive and willing to help. I've just got to finish up the outdoor measurements. Unfortunately that laminate floor thing-I've been given a "finish it by the superbowl or else," ultimatum. My wife is actually really sweet. And, really hot. So, what can I say? laminate before dipole panel. Other projects on the books for 2007 include a 3 way with the Peerless HDS for a mid, the new peerless HDS tweeter, and the current woofer of choice, the L26. I also have been toying with a Peerless HDS two way in the PE 3/4 cabinet. Unfortunately, it looks like there have been some production changes with the Peerless HDS exclusive since the early production runs of 2005. It means I've got to order a new pair from MAD and retest/use those. Plus a new pair for the 3 way. Eeek! Anyway, I'll get a pair, and throw them in a PE 3/4 cabinet and do the measurements. Then recycle them for the 3 way...

A couple of other random thoughts.

As usual, there are posters commenting that nonlinear distortion doesn't matter. I admit, it's a bit complex, but many distortions are very audible. I'd recommend that, instead of posting their opinion, they buy themselves Praxis or SE and actually listen to some signals at different levels of distortion and then post their opinion. While it's true that you can take a driver with higher nonlinear distortion and get around this by using multiple drivers, in the average 2 way, in complex passages with a wide frequency spectrum distribution, the resultant distortion is audible. Look again at my nonlinear distortion primer at the differences between Danny's 5" woofer and the SS8554. Not a fair comparison you say! That's exactly my point. Even at moderate listening levels, the 5" has distortion audible enough to change the character of the sound. Sure, you can use the 5" in a three way and get around the distortion. But the point is, that for most two ways, even listening at moderate levels introduces some distortion that has an audible signature. If you're choosing a two way, choosing a woofer with lower nonlinear distortion is clearly the correct choice. You can argue about whether small differences in ranking are significantly different, but there is a big difference between the top third and bottom third of drivers.

I'm shooting for another DIY in the fall. We'll wait till the weather cools in, say, October. I will slowly work on SE and lspCAD presentations, as well as praxis. We can pull out the LCD projector and have at it. I'll also bring praxis to test folks projects. I'll try to enlist Charlie and a couple others to start now on the planning and get commitments early for projects and equipment. Good planning and lots of announcements and prodding seems to get a fair number of folks and quality projects. Look for a posting in the next month or two. We'll need someone to help coordinate who's bringing what, equipment and sountracks, refreshments etc, location. Remember to post your thought's on the NCdiy/charlie's board. Currently there are two hosts for the NCdiyaudio. It's a long story. Kinda like the Morel thing...except, hopefully, no acrimony. I'm not sure where the pointer is currently for the NCdiyaudio main site is. We'll get it all sorted out. If anyone wants to work on the Peerless project and wants some help, let me know. Like I said above, I'm interested in a two way, second order Peerless HDS exclusive/HDS tweeter based on a PE 3/4 box. I think it would have wide DIY appeal and could be a great unit. Personally, though, the 3 way and dipole take precedence. Now if someone wanted to help, build the baffle, buy the drivers, and bring them over, well, I'm willing to help...

I've got distortion tests of the 7100 compared with the 6600. They are, as of yet, unpublished. I'm witholding them until DLR and Zaph take a look at the 7100. DLR's got the 6600 and 7100 now, and supposedly Zaph is getting a pair of 7100's to evaluate, so we should have a bit of data on those as well.

Anyway, look for a 7" comparo in a couple of days. Part one...I'll add a couple of extra drivers if I can get a hold of them. Shortly thereafter, some data on the 7100.

Sunday, July 9, 2006

The midrange comparison is up. Wow, the RS2 is a nice mid. Of course, now I want Darren to tame the resonant peak and make the flange smaller. You know, for $68 a pair, I expect a lot!!! This mid is a bargain, like the rest of the RS series. Overall, all these drivers are very good to excellent and you can't go wrong with any of them.

We're working on a DIY around October of this year. So start your projects now, if you haven't, for the fourth or fifth (I can't remember) annual Northern CA DIY.

Saturday, June 24, 2006

Working on adding the Accuton C44 and SS12m mid to my midrange tests. The most fair comparison is the C44/MDT55/RS52, since they are all 2" domes. The SS is kind of separate, but I'll throw it in and see where it lands. Finally, I'll get that midrange comparo off my chest. I kind of wanted to get out of all these driver tests...but the Tymphany visit and the RS52 have sort of kicked my rear into gear.

Something odd happened today. I was on the Best Buy site this afternoon, about to buy one of those Insginia boxes, and right before I hit the credit card submit, I woke up and thought, "what the #$%^ am I doing?" Really. It's just because it has a coaxial and a curved box. I think we should all badger MAD or PE to come up with a set of curved back boxes, right? Jeez, maybe I should buy the Insignia just for the boxes...no, wait...nip that thought in the bud. No, you can't make something for $35. Unfortunately that doesn't mean they don't stink. Look at Zaph's FR. Not very good. Someone really, really needs to come up with a good coaxial and a cool set of curved boxes...

Wednesday, June 21, 2006

Whoa, just did preliminary tests on the RS52 compared with an MDT55. Turns out it's a very good driver. Flatter in the passband than the morel, with better low end extension. Has that resonant peak you've gotta deal with, but otherwise it appears to be an excellent dome mid. I will try to post, and obtain another pair for testing.

Sunday, June 18, 2006

You'd think I died. The site's becoming stale, as they say. Anyway, I'm still around, but very busy with other aspects of my life. I do have some projects on the plate. I have the sample drivers from Tymphany, including the Aircirc 6600 tweeter, a 7" revelator, and a 5" revelator. Unfortunately, the 5" revelator appears not to be available through MAD or solen, so, uh, whoops. Didn't mean to do that... (also the 5" revelators may only be cosmetically different-still waiting for an update from the folks at Tymphany.)

I also got a the plans to John k's mini and some extra boards. I'm going to try to build a set by October. I'm also going to try a variation after I build John's version -- one with a pair of RS225's and and the reigning king of budget tweeters, the Seas 27TBFC/G

So what happened to the RS225Seas tweet? Well, I just don't need this. What am I going to do with it if I build it. The reality is that the RS22528A is a great pair of speakers. If I had need for another pair of boxes like this, I would redo this. But building another set would cost what? $300-400 bucks. I'm more interested in other projects right now. Sure, I could model one for nothing, since I have the drivers on hand and one test box, but that's not the same as building and listening.

Monday, February 13, 2006

I will post a version of my RS22528A crossover that incorporates 1.5dB less bafflestep but should otherwise be very similar. I'm also thinking of doing a version with a Seas 27TBFC/G, a crossover of 1.4kHz, and a B5 slope using John K's spreadsheets if anyone is interested. If no one's interested, well then, never mind. Frankly, the design is very good, and I'd leave it alone, but it's only 50%-60% that any given RS28A will make it to 1.3kHz. The RS28A is a great tweeter, but it isn't as consistent down lown as the Seas 27 series-or, that's my guess. But I'm not actually sure. See below.

PS Thanks for all your votes for the Typhany meeting. I'm honored. We'll see how this all plays out.

PPS I've been slowly adding to my collection of RS28A drivers. It might seem, from this data, that there is too much variability in the RS28 low end response. That's not a fair conclusion. The vast majority of RS28's can be used at 1.6k and above, at it's rated low end. And, if you find one that's not, PE is very good about returns and exchanges. The reality is, I'm asking a heck of a lot to get to 1.3-1.5k with a 1" dome. Now are we really sure all the Seas 27TDFC, TDC, and TBFC/G can go this low? I haven't had the opportunity to test 20 of them, so I'm not sure either. Still QC of the Seas is very good. I will investigate the Seas units more over time. It may be that, if you need a really low xover, you've got to test ANY unit you use. That may be the price to pay for a 1.4k xover.

Friday, December 2, 2005

Well, so Zaph posted a ribbon review over at www.zaphaudio.com. As usual, it ruffled some feathers on the MAD board, and I'm sure elsewhere. I've been a bit hesitant to jump in, but I'll throw in a couple of comments. In general, I'm disappointed by the pro ribbon crowd. I don't object to someone saying that subjectively it sounds better, but they offer no concrete reason why a ribbon might be superior. Now, my ribbon tests have been somewhat limited. I've got and played with the G2si, phillips RT?8 (boy, it's been so long I can't remember it's designation), a neo 3, an older Raven R-1 (with the outboard transformer), and an as of yet untested RAW acoustics short ribbon (sorry Al-I'll get to it. I promise...).

What bugs me is that the pro ribbon crowd insists that ribbons sound better because they are more transparent, detailed, blah, blah and get offended if you tell them that it might be just distortion.

What can I say confidently?

Ribbons cost more than domes, in general.

Ribbons are more fragile than domes.

Ribbons distort more than domes.

Ribbons seem to have an extended on axis frequency response.

Ribbons do not have markedly smoother FR linearity/linear distortion than a similar high quality dome

Ribbons have a wider horizontal dispersion

Ribbons have a narrower vertical dispersion

What are the fundamental properties of a driver?

1. Frequency response. Linear distortion and CSD are the same side of this coin and DO NOT offer any independent information from a high quality FR curve. They may present the data differently and highlight different features, but this is all the same fundamental information

2. Nonlinear distortion

3. Power response

As far as I'm concerned, everything else is either secondary, or, just a reflection of the above. All that stuff about impulse and step response, rise time, low moving mass, well, all of that is reflected in the above.

I would be careful of what you read. Many posts on the boards are just flat out wrong. I'm really not interested in getting into a shouting match with anyone, so I've not responded. One of these days we'll all have to get together and develop a "ten most common misconceptions repeated over and over on the boards" list.

Now, I'm not saying I know everything. There are a lot of topics I'm silent on because I don't know them that well. End of story. Unlike some posters that seem to have an opinion on every subject under the sun.

But what about this idea that some distortion adds detail? This seems to be a hotly contested point. As usual, the audiophiles just can't seem to fathom the idea that their perceived detail is due to distortion, since they have better ears than everyone else. Well, it is in fact a very plausible idea that has parallels in our visual processing. Image processing software sharpens images and add apparent detail by introducing a very specific type of distortion. I do this all the time in my image editing programs and it's very effective. No one looks at the photos and says, "gee wiz, look at all that distortion." Unless you use too much of it, then it's quite apparent. So what exactly do these image editing programs do? Well, they basically find borders between areas of light and dark, and they exaggerate or distort them. The whites get whiter, the blacks get blacker. It's a bit of an oversimplification, but that's what's happening more or less. And, it's very effective.

It's not that much of a stretch to think that our ears could respond in a similar way. Think about it for a moment. At low and moderate levels, the driver reproduces relatively cleanly. But transient peaks are "enhanced" with extra nonlinear distortion and they are perceived more sharply and with more detail.

Actually, it could be a good idea, if it were adjustable. This is the key point. In a good image editing program, you can adjust this with a high degree of flexibility. If we now go back to audio, the main problem is that this is an effect fixed on SPL. Listen too low, and not enough "sharpness." Now turn the volume up. It will start to sound strained. "Oh, it's the recording." Really. Try this out. Take out your favorite "detailed" recording and listen to it at your typical, but relatively loud listening level. "Oh yea, that's good." Now, increase the volume by 3-6 dB. Does it sound more realistic and concert like, or does it sound strained? In fact, does it just sound loud? Perceived loudness is sometimes due to distortion. I mean, if you can talk over your system, it isn't that loud.

Distortion is important. Don't be fooled by the posters that say otherwise. Look at SL's site. "But what about Dr. Geddes, he said below a certain level distortion isn't that important." Absolutely. This is not a reason to build a short ribbon crossed second order to a 5" driver at 2k. Duh, look at the kind of systems Dr. Geddes designs--very high output ones. I suspect most of the ribbon diy loudspeakers wouldn't last a millesecond in Dr. Geddes' house before the ribbon deformed.

Saturday, November 12, 2005

Wow, I don't update my blog much. I've been remarkably busy at work and home, and that really put a crimp in my speaker projects. Thankfully, one of my work deadlines has passed, so it should slow a it and I've got a number of things in the pipeline. The Morel MDT33/Supreme 110 preview just came out. A comparison between the RS125 and RS150 is coming very shortly as well. The testing is done, but I have to generate a web page for it. I sent DLR an AC g2si pair for him to test, and maybe try his lambswool mod etc. So that should be interesting.

I also should have a sub comparison between the RS10 and 12 HF subs, along with a peerless xls 12 and a TC from O-audio. I have been having some testbed limitations for testing subs, but I think I have that worked out. Will I test any others? Not unless you send them to me now...

As far as projects, well, jeeez, if I could ever get the RS225/RS28A finished. How hard is it to wire up two pairs of xovers...And yet, my mind wanders on to other projects.

One project that I've been thinking of doing is a driver frequency response and impedance database on specific baffles. What I want to do is get a small collection of ready made boxes and make a library of driver response curves on those boxes. Obviously, the PE boxes come to mind, since they seem to have the biggest selection of reasonably priced ready made boxes.

How the heck would this work? Well, say I have 4 different PE test boxes. All you have to do is send me a prerouted baffle with your favorite drivers, using one of the 4 boxes, and I would measure the FR curves for you, as well as impedance. I mail your drivers and baffle back. In addition to giving you the data, I make it available here. After a while, we'd have a pretty large database of common drivers on specific PE cabinets for the entire diy community to use.

Specific questions.

1. I'd measure both at a single point, and a separate set of measurements 1m on driver axis. Use whatever measurements suite you.

2. Remember, you don't have to send a cabinet, just the drivers and a prerouted front baffle.

3. If I have the drivers, I can use mine to save on shipping.

4. Will I route the baffle? No. It's the least you can do.

5. Are you in PE's pocket? No. If you would like measurements on a different cabinet and are willing to pay to have it shipped both ways and don't mind some dings, well, send away and I'll measure it. The PE cabs are the defacto premade cabinets, like it or not.

6. I will pay for return shipping for drivers and baffles. I will not pay return shipping on cabinets. Be real!

7. I have a fair number of drivers lying around, as you can guess. I could very easily generate data on a large number of drivers quickly. All you'd have to do is sent a prerouted baffle.

Anyway, let me know what you think of this.

Friday, July 1, 2005

I hope you enjoyed the site redesign. If you don't find it as clear as you'd like, just email me with constructive suggestions and I'll see if I can improve the site.

As you can see, this site so far has leaned towards empirical test data on drivers. I'm at the point where I am going to back away a bit from testing so many drivers and move on towards posting more designs. I am currently emulating multiple xover iterations with lspCAD 6.10 pro for a simple sealed two way I'm working on. The emulation feature is a godsend. I don't know why it took so long. I'm going to see if this Athlon 64 bit PC is powerful enough to run the emulation off the Delta 410 card and run praxis to see a real time FR curve. I'll let you know.

The two way I'm working on is a Dayton RS225/RS28A combo in a PE sealed box of 0.75 cu ft. It has an F3 in the 50 Hz range, give or take. While not earth shattering, it does have impressive bass performance in small to medium rooms. It is designed as a small to medium room monitor. I have an SS8545/SS9500 sealed two way in my guest bedroom. It's F3 is around 60-65 Hz and in this small, square guest room, there is audible bass into the 30 to 40 Hz range. In fact, my ported MLTL which has an SS8555 and is tuned flat to 30 Hz, sounded terrible in the guest room. Too much bass. There is much to be said for tailoring the low frequency curve, at least modestly, to the expected room placement and size.

Why don't I have the SS8554 and SS8545 based loudspeakers posted? Well, I never really ironed out the ML TL 3 way to my satisfaction. The SS8545 minimonitor is not posted for a different reason. It's clearly ready for prime time, so to speak, but the baffle construction is so unique that I'm not sure I can measure it. The front baffle is 2.25 inches at the woofer and 1.5 inches at the tweeter with all sorts of beveled cuts which have been radiused. It would be very hard to post the measurements at this point. I've got a couple others, but for one reason or another they're just not, well, they're not that good!

I do have a project in mind though, a straightforward 2 way that I think might have appeal for the DIY speaker builder. I know what you're thinking, another 2 way, sheesh! Well, this project will be a bit different since it will be based on the drivers that objectively offer the best performance in their price class and can use a ready made cabinet or your own built cabinet. I will also make extensive use of the emulator in lspCAD and, unlike most of the projects on the web, I will have extensive measurements, including distortion and off axis performance numbers. It will be a very well engineered and documented budget two way. Hopefully it will be ready for play at the Northern California DIY in October or November of 2005.