Jump to content

Tweeter instrument tests


Recommended Posts

Here are the main patet-up plots from the earlier thread withOUT any comments.

I will soon receive a JBL tweeter to add to the instrument tests. I have been told the CP25 is a "clone" of the JBL tweeter and may test the same. I will also be retesting the Beyma and the T-35, but this time at lower levels. The Beyma CP25 tweeter may be exonerated. Tests using a Fluke 87 true RMS meter in the minimum and maximum hold modes connected directly to the tweeter in one of my speaker shows a maximum level of about 0.4 V RMS while my power amp shows a level of 3 Watts maximum. This is really crankin' for me! I will be starting at 0.5V and moving down to see what happens to the distortion at more normal levels. The previous thread was starting to conclude that the test levels where to high but the thread was locked before any further investigation cold be done. It may be that the higher distortion levels into the CP25 may be still low enough to not be a problem and be an intentional trade-off to provide better dispersion as a benefit. All good engineering is a balance of compromises.

Here is the on-asis plots of the 5 tweeters.

AL K.

post-2934-13819583154354_thumb.gif

Link to comment
Share on other sites

  • Replies 43
  • Created
  • Last Reply

Top Posters In This Topic

Edited: There are two incompatible definitions of THD and my background led me to the wrong one for audio. Thank you Al for bringing that to my attention. I have edited the posting to correctthat and I am using Al's estimate of the Beyma throat diameter.

In his book Acoustics, Leo L. Beranek, derives a formula to calculate second harmonic distortion (in percent) for infinitely long exponential horns. He notes, since most of the distortion occurs near the throat, the formula is correct for finite horns. Horns with other shapes would have somewhat higher or lower values depending on whether they expand less or more rapidly than exponential.

Measurements of distortion also include contributions within the driver. Beranek's formula applies to the horn portion and depends on two factors: the ratio of the frequency to the cutoff frequency and the power density in the throat.

According to the Beyma CP25 datasheet, at 5kHz the output is about 104 dB and h2 is about 75 dB so that h2 is about 29 dB below the fundamental tone at 5 kHz. Converting to percent, 10^(-29/20) = 3.5%. The response drops rapidly below 1.8 kHz so that should be near the cutoff frequency. We can use Beranek's formula to estimate the power density required to have h2 = 3.5% with f/fc = 2.8 and it gives about 0.53 mW/cm^2.

A drawing of the tweeter is given although the throat is not dimensioned. Al estimated the throat to be about 1/4" so the area is about 0.32 cm2 and the power in the throat about 0.17 W (acoustical).

The above posted result has the 10 kHz peak about 18 dB below the 5 kHz peak or about 12.6% h2 to fundamental. In Beranek's formula, this requires about 2.2 W acoustical power. Assuming an efficiency of about 10% which would be ballpark for a driver rated 104 dB SPL, the drive power to the tweeter would be about 22 watt, very near its CW power rating of 25 watts. While this does not sound like much power, the tweeter typically receives about 10% of the program, so in use a multi-driver loudspeaker would be delivering about 22 acoustical watts and a pair would deliver 44 acoustical watts. According to Wikipedia, a jackhammer sound power level is about 1 acoustical watt.

Perhaps tweeters with 10% of the program should be tested at a lower power that takes into account typical listening levels and overhead for peaks. This raises the question, what is the preferred volume forum members use for listening? Perhaps I should add...when the wife is out!

Link to comment
Share on other sites

Edited as in the previous post:

Looking at the Selenium datasheet we have a throat diameter of 25 mm, a cutoff near 1 kHz, and h2 about -29 dB at 5 kHz. Doing the same calculations as in my previous post, to deliver the same power as the Beyma (0.17 W acoustical) Beranek's formula estimates the h2 distortion at about 3.5%. This is in contrast to the -53 dB or 0.22% posted above. Thus, based on vendor data and Beranek's theory, both the Beyma and Selenium should have similar second harmonic levels. The Selenium sensitivity is listed as 109 dB so it would be expected to require 5 dB less drive than the Beyma, but that should have no effect on the distortion.

Is anyone aware of calculations for the intermodulation products and the dependence on throat size, cuttoffs, whatever?

Link to comment
Share on other sites

Nat,

Based on the calculations you just did, suggest a realistic level and frequency to do harmonic distortion testing. I was going to use a single 5 KHz tone at 0.5 V RMS. That amounts to 31 mW of power assuming the tweeter is 8 Ohms.

BTW: Unlike IM distortion, my HP 3563 analyzer will actually calculate distortion in percent for total harmonic distortion.

AL K.

Link to comment
Share on other sites

Nat,

I think something is wrong.

Look at this definition of harmonic distortion:

http://en.wikipedia.org/wiki/Total_harmonic_distortion

It shows the calculation using voltage ratios. The 29 dB down 2nd harmonic distortion would be 10^(-29 / 20 ) * 100 or 3.55% distortion. The scales on my HP 334A distortion analyzer are -10dB/30, -20db/10%, -30dB,3% which confirms this. I think that upsets some of you conclusions. Do you agree?

AL K.

Link to comment
Share on other sites

dkalsu,

The JBL tweeters are here. I haven't opened them yet but the boxes look undamaged.

Marion,

The plots are "dumps" directly from the screen of the HP 3563A analyzer through an HPIB cable to a software plotter emulator. They are on a linear scale because a log scale magnifies the low frequency. Look at the high frequency area is more accurate on a linear scale.

Al K.

Link to comment
Share on other sites

Maron,

Al's use of linear and not logarithmic scales to display frequency and distortion is neither simplistic nor misleading as you imply.

His background is RF engineering and filter design and the instrumentation best suited for that discipline is a spectrum analyzer which usually offers only linear scaled displays for frequency and distortion when observing individual distortion products.

While his HP3563A can measure spectra in the "log" mode, the results if displayed on a log scale, would be misleading. It's far easier to read IMD and spectra on a linear scale than a log scale. The same holds true for distortion products.

As a broadcast engineer cum audiophile, I am used to reading distortion products as so many dB below the fundamental. So for me, Al's plots are easy to read. With a little practice, anyone can look at his plots and quickly determine which tweeter is "best".

Now, if you are plotting swept sine measurements (distortion or amplitude on the Y-axis and frequency response on the X-axis), a log display is easier to resolve than a linear scale because of the log display's higher resolution at low frequencies.

I have three Beyma CP25 tweeters and will be testing them next week with an Audio Precision SYS-2722 coupled to my 1/2" & 1/4" Bruel & Kjaer mics to see if Al has a rogue CP25 with out-of-spec distortion. Maybe I'll post the data in both log and linear formats and let you decide which is simplistic or misleading :>).

Lee

Link to comment
Share on other sites

Here is the first tests on the JBL 2404 tweeter. This is the on-axis plots (black) and 45 deg off-axis vertical and horizontal. The two are virtually identical. This is the obvious dispersion champ!

The front trace is anechoic using a short spike, then ignoring everything after the mike hears it to compute the response.

Al K.

post-2934-13819583194356_thumb.gif

Link to comment
Share on other sites

Nat,

Based on the calculations you just did, suggest a realistic level and frequency to do harmonic distortion testing. I was going to use a single 5 KHz tone at 0.5 V RMS. That amounts to 31 mW of power assuming the tweeter is 8 Ohms.

BTW: Unlike IM distortion, my HP 3563 analyzer will actually calculate distortion in percent for total harmonic distortion.

AL K.

Al,

With efficiencies about 10% that would give about 3.1 mW acoustical power in the throat. For the Beyma with a 1/4" throat, the 2nd harmonic distortion would be about 0.5% or -46 dB. With the larger throat, the Selenium numbers calculated to 0.2% or -54 dB.

Thanks for making the measurements.

nat

Link to comment
Share on other sites

Humans perceive tones logarithmically, so the visual depiction of a log frequency plot will correlate better to the audible perception. It just means you have to interpret the linear scales differently. I have to say it was incredibly misleading at first since the rest of the industry and audio literature almost always uses log plots. And the few occasions they don't, they call it out very clearly so as to not lead to erroneous conclusions from those not looking closely at the graph labels.

Another way to compare the linearity of different drive units is to increase the level until the distortion exceeds some threshold and then record that SPL. It might make sense to record the SPL at multiple different distortion levels (say 0.1%, 1% and 10%) since the distortion mechanisms in speakers tend to be exponential and can have a very different shape to the distortion knee depending on what the source of the distortion is.

Link to comment
Share on other sites

Al,

I have a pair of 2405 that I will be using to build a JBL 4345 clone. I'm really busy at work at the moment - but if I get a chance, I can certainly try to mail you the pair I have. They are the alnico versions. I'm not exactly sure how they will measure - I simply picked them up as they are a favorite among many audio enthusiasts. I'm sure many others would be curious to know how they perform.

Cheers,

D

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...