Dayton Measured in Stereophile. What, What

[Dennis Murphy]

The measuring distance, the position of the mic relative to the woofer and tweeter, and input level all can affect resolution, the depth and position of any cancellation null between drivers, and overall smoothness. But it can't explain differences in tweeter level like we're encountering here. I did manage to get my PE Omnimic system going, and I measured the NHT and the unmodified Dayton, and then I used Praxis to measure the unmodified Dayton. The mic position and distance was the same. Although my two mics don't agree perfectly above about 12 kHz, they do agree on evrything else. Obviously these little tweeters aren't tweeting enough. I can only assume PE sourced them from a different factory way over there, and they just aren't as sensitive. My mod would probably work on new purchases, but not old.

 

[Chu Gai]

Well it would be interesting to see what PE says in response to your inquiry. As far as being new, who knows if they made an intentional change or just ran into a bad batch of tweeters?

 

[Dennis Murphy]

Well, I wrote a long annotated e-mal to the PE tech department. We'll see if they have anything to say. It's a shame this discrepancy exists, because the modified speaker is really quite nice.

 

[Chu Gai]

You know, Zaph also listed a modification.

 

[Dennis Murphy]

You know, Zaph also listed a modification.

Hmmmm Wish I had noticed that sooner. I modeled his crossover with my measurements, and it worked pretty well despite the difference in tweeter output. The reason is that Zaph is running his tweeter hot at the top, and it's just a tad cool with my tweeter. The main difference in our approach is that I use shaping resistors in the parallel circuits to optimize phase relationships and get the woofer break down a little more. I don't know how audible this would be.

 

[Chu Gai]

Any chance you can further explain that thing with the resistors? Also, what did you think of him suggesting a variovent or a DIY version.

 

[Dennis Murphy]

Any chance you can further explain that thing with the resistors? Also, what did you think of him suggesting a variovent or a DIY version.

There isn't much to explain. Placing a resistor in line with, say, a parallel cap in the woofer circuit gives you much more flexibility in shaping the knee of the crossover slope. If you're trying to target a 4th order Linkwitz-Riley low pass slope, the resistor (in conjunction with the cap) will control the intial roll-off so that you don't overshoot and end up with a acoustic roll-off that resembles a right triangle legs more than a smooth continuous curve. That keeps the woofer in phase with the tweeter (which would also use a resistor in line with its parallel inductor) in the crossover region, and the two drivers will sum flat in phase. You can still get a flat response when the drivers are not in phase--it depends on the individual shapes of the crossover slopes--but in my experience the result will not be as transparent. But that's all very controversial. I just know that I've always ended up with a neutral and transparent sound when I've followed the rules to the letter.

 

[gman76]

Obviously these little tweeters aren't tweeting enough. I can only assume PE sourced them from a different factory way over there, and they just aren't as sensitive. My mod would probably work on new purchases, but not old.

View attachment 11141View attachment 11142View attachment 11143

Dennis, I recently bought the B652's and my ears agree with your first plot. The tweeters almost seem as if they're not hooked up. Your plot shows that the output drops like a rock above 5kHz. Either PE substituted the old tweeter for a newer cheaper (arguably non-working) tweeter or it is hooked up out-of-phase or the 3.3uF cap in series the driver is somehow not properly passing the highs. I'm very tempted to return the B652s. Zaph and others provided freq response plots that seemed reasonable for a cheap speaker, but with a tweeeter installed, the plot that exhibits highs that are completely gone is just wrong. Zaph took measurements many yrs ago, so it is quite possible PE subbed the tweet out. Granted I paid only $30 for the pair, but I bought them based on Zaph's results and simple mod that he said made it a passable sounding speaker. While I havent added a xo yet, I'm finding it hard to believe these speakers can sound halfway decent.

I'm wondering if you or Chu can take measurements with (1) the tweeter phase reversed and (2) the tweeter disconnected. I'm having a hard time believing that the tweet is wired wrong but it is weird that you show such a dramatic null at 9kHz. I'm wondering if anyone out there has both the "old" (2009) and "new" (2012) model, and would care to comment.

 

[gman76]

B652 impedance vs freq

If anyone cares ... I took a number of impedance measurements of the B652 without modding. Sorry PE, I wouldn't call this an 8 ohm speaker, it's more like 5 ohms. Sorry can't load the excel graph.

freq Impedance
15 3.92
25 4.00
35 4.13
50 4.49
70 5.11
92 5.35
150 4.64
200 4.36
300 4.23
400 4.26
500 4.31
650 4.46
820 4.71
1000 5.03
1500 5.85
2000 6.64
3000 8.23
3500 8.70
5000 9.17
10000 7.35

 

[maximoiglesias]

A good low cost option is Sony SS-B1000, as surrounds of course..................

 

[Dennis Murphy]

Dennis, I recently bought the B652's and my ears agree with your first plot. The tweeters almost seem as if they're not hooked up. Your plot shows that the output drops like a rock above 5kHz. Either PE substituted the old tweeter for a newer cheaper (arguably non-working) tweeter or it is hooked up out-of-phase or the 3.3uF cap in series the driver is somehow not properly passing the highs. I'm very tempted to return the B652s. Zaph and others provided freq response plots that seemed reasonable for a cheap speaker, but with a tweeeter installed, the plot that exhibits highs that are completely gone is just wrong. Zaph took measurements many yrs ago, so it is quite possible PE subbed the tweet out. Granted I paid only $30 for the pair, but I bought them based on Zaph's results and simple mod that he said made it a passable sounding speaker. While I havent added a xo yet, I'm finding it hard to believe these speakers can sound halfway decent.

I'm wondering if you or Chu can take measurements with (1) the tweeter phase reversed and (2) the tweeter disconnected. I'm having a hard time believing that the tweet is wired wrong but it is weird that you show such a dramatic null at 9kHz. I'm wondering if anyone out there has both the "old" (2009) and "new" (2012) model, and would care to comment.

Hi. I'm really rushed right now, but I can tell you that I reversed the tweeter polarity and measured without the tweeter. That's not it--it measures best the way it is. I wrote to Parts Express, but they never replied. I don't think they want to spend a lot of time on this puppy. They probably switched suppliers for the tweeter, and are stuck with the inventory.

 

[gman76]

I wrote to Parts Express, but they never replied. They probably switched suppliers for the tweeter, and are stuck with the inventory.

Ya, one that looks the part but does essentially nothing. Sorta fraudulent on PE's part, if you ask me.