A good tweeter to match 5" SB Acoustics Magnesium Woofers

[xmax]

As you can see the third harmonic is higher at 3K (the most sensitive area of our hearing)
with the resistor only the B&K mic has low enough distortion to see that. If you look
at the Via wave you can see the how low I like the 3rd harmonic. The test conditions should not have
anyone concerned about reflections. The other 2 mics are the UMM-6 and the
USB OMNI.



Maybe you guys measure speakers for different reasons than I do.
Although looking at it again the third harmonic is so low I don't think it's an issue.
I will return to this test at another date in the mean time think for yourselves and please don't drink the kool-aid.

[xmax]

Again I understand the importance of the gate setting but I have not had to change it. I will redo the
test to minimize the chance of reflections skewing the test I will also use REW if you prefer and set the
gate and mic distance how ever you like. (I imagine the mic fairly close and all other surfaces as far away as
possible)

Although since any reflections skewing the first test would be the same with and without the resistor,
I believe the 3rd harmonic distortion will be highish regardless of whoever does the test with any reasonable
gate/mic ratios.

[ryanbouma]

Xmax, it is great that you admit you dont know how to tweak the gate time. Shows maybe you're not as arrogant as you have appeared to be. But to be perfectly fair and honest, if you don't know how to do that, you don't know how to measure a speaker. The gate/window time is pretty much the single most important setup issue with any speaker measurement. IMO its more important that drive voltage, microphone type, software type, amplifier, etc. The only thing more important would be getting the distance from reflections as far as possible and mic distance. Which those things actually dictate the gate time, which is why they are more important. I am with the others that disagree with you about how a resistor makes distortion rise and such. At this point it is clear you have not proven anything despite the comparison graph you posted, which was not very helpful even before we knew it was possibly not reflection free. Sent from my SGH-I337M using Tapatalk

[xmax]

Still no answer on the window/gate.

In your example, the filter Fc increases (can't say precisely what point you considered to be Fc) as I see it, the level at 3K is practically unchanged, the response below it is down relative to 3K, almost -6db all the way to noise, so how is that lowering the tuning? The level above 3K is commendably flat with the resistor.

dlr

To be perfectly honest I'm not sure how to tweak the gate settings on the software, I remember when I first got it, I talked to the designer ,
he went over a few guidelines with me about keeping reflections out of the equation and I have been using it ever since with great success.

I will be happy to retest it with whatever conditions you like. I also encourage others to do the exact same test to see what they get. If nothing
else, all this bickering has got people thinking about the concept and when it can be used successfully.

Last but not least I encourage people to run a test like this and just listen to the tweeter with and without a post filter resistor of 6.2 ohms
with the tweeter volume corrected if needed. I'm pretty sure most anybody will hear what I am saying.

[dlr]

"Why does 3rd harmonic in the response with the resistor increase below 3K when the drive signal level is down, yet decrease above it?"

Because of the transfer function allows more low frequency information through because the 6.2 ohm resistor essentially changes the
the tuning frequency of the filter lower, a simple analogy would be using a filter for a 8 ohm tweeter designed for a 4 ohm tweeter.
I have came to this conclusion from experience..

Still no answer on the window/gate.

In your example, the filter Fc increases (can't say precisely what point you considered to be Fc) as I see it, the level at 3K is practically unchanged, the response below it is down relative to 3K, almost -6db all the way to noise, so how is that lowering the tuning? The level above 3K is commendably flat with the resistor.

dlr

[xmax]

I'm not trying to change your mind. I'm interested in the measurements and interpretation. The dip is 5K, not 4K, my mistake on that, but I don't see how a 2nd order distortion dip such as that can be diffraction related. It's in the 2nd order harmonics, there's a dip at 2K. 5K is not a 2nd order harmonic. It's not 3rd, either. In fact, not a harmonic of any kind to the dip centered at 2K.

You've also not answered the question of the time window used. The graph has no indication of any sort of time limitation, the response shown goes down to 10Hz which wouldn't even be possible in essentially any anechoic chamber extant that I've read about given even their limitations. Everything below about 400Hz looks to be in noise, not useful, the two arrows at 150 and 300 provide focus on anomalies that look also to be of no use, the 5K dip in 2nd harmonic is odd as I would expect it to be at 4K (again, maybe I'm interpreting it wrong), all of which makes me question the distortion results overall.

How can the 5K be a 2nd harmonic of the diffraction at 2K? Why does 3rd harmonic in the response with the resistor increase below 3K when the drive signal level is down, yet decrease above it? Provide more detail, it could help answer questions. I suspect I'm not the only one with those questions.

And again, what was the time window used? What kind of time window was used? Certainly you know those. That will have a significant impact on any measurements.

dlr

"Why does 3rd harmonic in the response with the resistor increase below 3K when the drive signal level is down, yet decrease above it?"

Because of the transfer function allows more low frequency information through because the 6.2 ohm resistor essentially changes the
the tuning frequency of the filter lower, a simple analogy would be using a filter for a 8 ohm tweeter designed for a 4 ohm tweeter.
I have came to this conclusion from experience.

The info below 500hz is interesting at best, at this point I will test everything again with the exact baffle and non-102 degree fever.
I also invite people to do some testing on their own, frankly I was taken by they amount of people not seeing what I am talking about.

[dcibel]

Yes I do know a thing or 2 about noise, let's say there is noise in the system why is it louder with the resistor?

You're asking what a source of noise could be? The list is pretty long, furnace, ac, fridge, amp hum, or even a recording engineer that can't keep his yap shut can all be sources of noise during a measurement. It is important to keep everything very quiet through the duration of the measurement for distortion, for good reason as you can see the harmonic information is often very close to the noise floor of the equipment.

[dlr]

I imagine the 4K dip is related to edge diffraction the baffle differs from the design a little I have not got around
to cncing a exact baffle. I'm extra busy and just getting over the damn flu. But again I don't think it will make enough difference
to change my mind.

Correct Solid is the 2nd harmonic and dotted is 3rd.

I'm not trying to change your mind. I'm interested in the measurements and interpretation. The dip is 5K, not 4K, my mistake on that, but I don't see how a 2nd order distortion dip such as that can be diffraction related. It's in the 2nd order harmonics, there's a dip at 2K. 5K is not a 2nd order harmonic. It's not 3rd, either. In fact, not a harmonic of any kind to the dip centered at 2K.

You've also not answered the question of the time window used. The graph has no indication of any sort of time limitation, the response shown goes down to 10Hz which wouldn't even be possible in essentially any anechoic chamber extant that I've read about given even their limitations. Everything below about 400Hz looks to be in noise, not useful, the two arrows at 150 and 300 provide focus on anomalies that look also to be of no use, the 5K dip in 2nd harmonic is odd as I would expect it to be at 4K (again, maybe I'm interpreting it wrong), all of which makes me question the distortion results overall.

How can the 5K be a 2nd harmonic of the diffraction at 2K? Why does 3rd harmonic in the response with the resistor increase below 3K when the drive signal level is down, yet decrease above it? Provide more detail, it could help answer questions. I suspect I'm not the only one with those questions.

And again, what was the time window used? What kind of time window was used? Certainly you know those. That will have a significant impact on any measurements.

dlr

[xmax]

I imagine the 4K dip is related to edge diffraction the baffle differs from the design a little I have not got around
to cncing a exact baffle. I'm extra busy and just getting over the damn flu. But again I don't think it will make enough difference
to change my mind.

Correct Solid is the 2nd harmonic and dotted is 3rd.

[dlr]

To be fair, I ran it with the resistor after as you stated. In series with the driver, not parallel across it. The change is much closer to the change in yours with almost no change at 3K with about 5db drop and somewhat more drop as frequency drops. So on the part there is some agreement in results. For SPL that is.




But that provides significant reduction in output below Fs and should not create any sort of peaks as shown, so those must be due to some sort of measurement conditions.

You're pointing out two peaks at the low end in what appears to be a a level below them in a region that is rising with decreasing frequency when the output is decreasing rapidly (ref. the crossover transfer function), not increasing as your measurement seems to indicate, so the noise floor is almost certainly an issue. Those two points, from what I see, have no relevance. They're are likely measurement condition artifact, all of which highlight possible issues with the measurement conditions. It seems to me that everything below 400Hz should be ignored.

One thing I may have missed is what are the two distortion curves? Solid 2nd harmonic and dotted 3rd? What I see isn't consistent with what I would expect given the change in SPL between the two curves. If I'm reading it wrong, do explain. And I'm puzzled with that massive 4K drop in the curve without the resistor in the blue set.

dlr

[xmax]

I can measure it to -120 dBu (I have 2 audio precision analyzers)

http://techtalk.parts-express.com/member/23426-xmax/media

[xmax]

Yes I do know a thing or 2 about noise, let's say there is noise in the system why is it louder with the resistor?

[dcibel]

You'd think a recording engineer would be well versed on the topic of noise floor, and also what sort of output to expect from a tweeter at 150Hz...Also, who gates a distortion measurement? Also, a gated measurement wouldn't have any resolution at the lower frequencies, but that is not shown in the measurement, unless the gate is "Wide open" at something like 200ms+. All that looks like is greater background noise in one of the measurements, maybe some hum coming through haha better check those grounds!

On thing is true, per my first post in this thread, it has been entertaining as I predicted. ;)

[xmax]

Like I said the response in my graph below 500hz are not directly audible just an interesting result.
I think the change in transfer function shown in my test and the original design are doing what
the designer set out to do. Also the resistor is after the filter although I'm not sure it will it will
make much difference. I am most concerned about the distortion being 10dB higher at 1K.
and high in general. Perhaps you should measure it yourself before saying my results are
flawed.

[dlr]

For those who may wonder why there's so little change below 500Hz, here's what the impedance magnitude is for the 2-element 2nd order crossover without the 6.2 ohm resistor.

dlr