Calling Ed Frias (AR.Com ?)

[poneal]

Hi Ed,
I have a measuring setup now and after completing my latest two way I decided to measure the AR.Coms. What I noticed was a 10db dip at around 3000hz. Almost like a reversed polarity on the tweeter. I had these crossovers made by madisound. I will try and reverse the polarity later in the week when I get time but was wondering if you had designed them like this or maybe the crossover got wired wrong. Any help would be appreciated. Regards, Paul.

[kexodusc]

Paul, I know the Madisound board had a few discussions about this in the past. For awhile there were some really iffy batches of tweeters, not sure if that's the deal here.. When I had a pair of mine measured at the university's DIY club mine were -2.5 dB around the same region (a bit lower than 3000 Hz, I think). It didn't really follow the general shape of the SW plot I did when plugging the drivers and xo into Speaker Workshop (actual measurements for the drivers), but they did measure -3/+2 dB so that's not bad at all. I got way way more bass response than expected though.

Did you flush mount the drivers? Wouldn't a small distance (even 1/4") affect phase too?
-10 dB seems way too low though, I think you would notice that right away when listening.

Sounds like the same issue Danny Ritchie was talking about a few years back that led to the new xo design. This seems to be an intermittent problem, everynow and then you get someone who reports this. Wonder if it's the Madisound xo's? They've deviated from Ed's specs before? Might be worthwhile to open them up and verify all the elements are right.
As I recall, they were shipping them with a 7 ohm resistor for awhile. Not sure that'd cause the dip you're experiencing, but it wouldn't hurt to double check Madisounds work.

[poneal]

Yeah, I was expecting a 2-4db dip because I usually end up with a dip like that in this area. The impedance looked fine as I measured it too. I even ran a pulse test to make sure I had everything set up right. Same FR plot though. 10db is just way to much IMO. I will check the other 4 and see if they measure the same. If so then I think Madisound either wired them wrong or I got some bad drivers. After designing my other two way, I was thinking that they sounded a little dull and this seems to point out why. Regards, Paul.

[poneal]

Danny Richie. Here is a link to the FR at GR that looks almost exactly like the one I did this weekend. He has a modified crossover to smooth it out.

http://www.gr-research.com/edspeaker.htm

[kexodusc]

Danny's xo looks promising and might be a fix. I know Ed's latest xo was suppose to fix this too, so I'm tempted to believe there might be something else at play.
Oh well, maybe you'll come up with some mods of your own?

I'm doing my own mod to the ar.com's hopefully this weekend. I just finished redoing the PeeCreek design (same Peerless woofer) in a dual chamber reflex box as a housewarming gift for my sister (I don't have the room and prefer the Ar.com's,lucky her). I increased the volume to about 21 litres. Wow. You almost get a sealed sound with the DCR design. Better power handling, and a more full bass sound. It seems tighter too. Quite the improvement. The DCR box I did for the sub in my cousins new Dodge Charger (nice car btw) benefited even more...I'm becoming a big fan of this.

By adding just 50% volume I'm calculating about 8-10 Hz lower bass response. Power handling won't suffer anything noticeable at all. This is pretty good considering the top chamber retains the original volume, so the woofer only "sees" the 10.58 L box. If these work the same as the PeeCreeks, there won't be any noticeable changes in tone to the speaker, just more bass. I suppose I could make the cabinets a tiny bit bigger but I'd prefer to keep that top chamber at the 10.58L volume.

We'll see. This might be a pretty cheap, easy mod for some folks to consider if they don't have a sub or want to make towers...It's a bit harder to build, but cheap, and the benefits are nice.

[poneal]

you're gaining a lot of knowledge on box building. You will be doing TL soon. Have you scoped out a project using the RS drivers yet? If you want to test them out, use my RS150 + MDT20 design. They are definitely different sounding. Very detailed but not to the point of harsh. The bass is a diffent sounding bass. Very tight and fast. You will not believe the difference in sound. It takes awhile to get used to but it grows on you. Regards, Paul.

[Swerd]

Dennis Murphy also made a corrected crossover for the AR.coms. Someone sent him a pair, which he measured and designed new crossovers which flattened out the midrange dip, included BSC and filtered out the hot tweeter response in the 10-20 kHz range.

I have ordered the parts for the new crossovers and will compare them to the Madisound ones I've had for about 2 years. See the attached files.

[Swerd]

I have a measuring setup now. . .Paul
I've been toying on and off with the idea of getting design software and a speaker measuring set up, but I've been baffled about what direction to take. Now that you've taken that step, would you mind telling us what you have, what it cost, and what your experience has been. Thanks.

I just finished reading your thread on the AR.coms at Madisound. Dennis and Danny pretty well covered it. The crossovers I posted are the same ones Dennis mentioned that he did for the Peerless tweeter. I had asked his opinion on replacing that tweet with others such as the Morel MDT-20 or Danny Richie's T2. Dennis said the high freqency shrillness in the Peerless was easily corrected, although he slghtly favors the Morel in that price range over the other choices, it wasn't worth replacing them. For what it's worth, Dennis said he likes the Peerless CSX series woofers in general, and wishes that Danish Sound Technolgy's new owners would continue their production.

Richard

[poneal]

As far as setup goes. I use SpeakerWorkshop for the software. It was free and I was using it before I even got my mic and mixer. I just had to use the .frd tools to get BSC, MinPhase, etc., before importing them into SW. I then ordered the ECM8000 mic ($50) and the UB802 Berringer Mixer ($50). RomanB then sent me his calibration files for the ECM8000+UB802. He had his sent of and actually measured. I then bought a mic boom from RS for $15. It was a floor model and had some scratches but I didn't care--I was looking for cheap. 25' mic cable XLR ends ($20) and a SoundBlaster Live 24 ($35) soundcard. My on board sc didnt work with SW or at least I couldn't get it work. I made a cable impedance jig using a RS 1/8" to 1/8" phone plug cut in half and reference resistor soldered in place. Various cables from RS. 1/4" to 1/8". RCA to 1/8". and 1/4" to RCA mono for left out to amp. Had an existing receiver I use for amplifying the signal. So it's not really a lot of cash outlay to get the equipment, but what is hard is getting the cables, soundcard, and SW setup properly. This took me about 2 weeks with lots of help and emails back and forth. Once that was done, the measurement part was easy. If you go this route, let me know and I'll be happy to help you out. Regards, Paul.

[Swerd]

Thanks a lot. When I decide if I want to go that route, I'll take you up on that.

Right now, I'm still working on better wood box construction and veneering techniques, and relying on DIY speaker designs published by others. As you can guess, I've learned that Dennis Murphy is a great resource. He lives nearby me, and is very willing to teach. Unfortunately, he is going to retire soon and move to Arizona. There is always the Madisound board.

[poneal]

Blasphemous. I really hope he doesn't do that. That would be a great loss to all of us.

[Swerd]

No, he is going to retire from his day job here in Washington, DC and move to Arizona. :D

I think he'll work even more on DIY speakers.

[EFE Speakers]

Paul,
Something is wrong if you see a 10db drop at 3k! I don't know what your program recommends for the mic set up, but I have very different measuring methods than most use, but that still wouldn't account for a 10db drop at any freq!

First, unhook tweeter and measure the speaker with only the woofer operating, see (with the same measuring method) where the woofer starts to rolls off at. Then unhook the woofer and try just the tweeter and see where it rolls off at, this should give you an idea of which driver might be the problem. Then you'll have to figure out if it is the driver or the crossover rolling off to soon.

If you recall a few months ago, one poster posted his frequency chart for his DIY's and they measured very good, as my latest crossover was designed to. The problem may be the newer Peerless drivers I've been hearing about, but that can easily be determined by measuring them before they go into the speakers. If both of your speakers measure the same, it may well be the new drivers. Measuring the driver's natural roll off and then comparing it after it goes through the crossover will give you a good idea where the problem is.

I measure each driver that comes out of the box before I install it in the speaker, making sure the pair I use match perfectly. If I find anomalies, I match two drivers with the same anomalies and then adjust the crossover (if necessary) to compensate for those anomalies! Sometime the anomalies disappear when the drivers connect to the crossover. I know most DIY builders simply assemble their speakers and measure the final results, but I have found that too many components (especially drivers) often vary and need some fine tuning (in the crossover) to compensate.

Each step of building for me is measuring ahead of time to assure the final product will not be off because of a combination of small component anomalies. Each driver, each cap, each resistor and each inductor is measure before I assemble the speakers, then I know the two will at least match when I make the final measurement. If their is an anomaly in the final measurement, I usually will need to adjust the crossovers exactly the same on both speakers because both match and reveal that same anomaly. This matching process can only be done if you have the equipment to measure impedance, capacitance, resistance and inductance. For you guys that do a lot of DIYing, it may pay to buy a good multimeter for this purpose.

Let us know what you find the problem to be, thanks!

Ed Frias
EFE TECHNOLOGY Speakers

[poneal]

Here is a link to the on-axis individual response and over total response.

http://home.stx.rr.com/poneal/ar.jpg