With this strange impedance curve
What will be the ideal cut-off frequency for this tweeter?
http://www.transducerlab.com/images/pics/6adc8ed0918ebf2e6abdd1343a43ef72.pdf
What will be the ideal cut-off frequency for this tweeter?
http://www.transducerlab.com/images/pics/6adc8ed0918ebf2e6abdd1343a43ef72.pdf
. Seriously, anywhere above 2kHz will be OK.
I think the rule of thumb is 4x Fr (resonant frequency). So:
560 Hz x 4 = 2,240 Hz.
Looking at the curve, it may be easiest (but not required) between 3 kHz and 4 kHz.
The impedance may need to be dealt with separately, I can't read the curve very well, but if it is varying a great deal through the crossover region it may complicate things, especially if you attempt to pad it. Nice catch!
Best,
-E
560 Hz x 4 = 2,240 Hz.
Looking at the curve, it may be easiest (but not required) between 3 kHz and 4 kHz.
The impedance may need to be dealt with separately, I can't read the curve very well, but if it is varying a great deal through the crossover region it may complicate things, especially if you attempt to pad it. Nice catch!
Best,
-E
Last edited:
I don't think that rule of thumb is particularly useful or even practicable. With a tweeter with such a wide flange, you need to get your xover frequency down to minimize lobing. The only way to be sure is to measure distortion vs frequency. In case an analog xover is used, impedance correction is likely required.
Seriously, though. I have the carbon fiber version of your Tlab here and i see no reason to use it below 2khz unless you're forced to (2-way with a 10'' midbass by exemple) and also not over 3khz because the tweeter will (should) make a better result than the mid driver in those frequencies.
But then again, there is no absolute rules. The best is really to give a try. It depends on your midrange, it depends on the xover slope, it depends on your tastes, etc...
But then again, there is no absolute rules. The best is really to give a try. It depends on your midrange, it depends on the xover slope, it depends on your tastes, etc...
In practice there is no 'ideal cut-off frequency', let alone one based solely upon either the impedance curve or driver Fs, just a series of compromises to select from based on what it is you're doing. The impedance curve is not strange as far as I can see -for e.g., many tweeters that use ferrofluid have similar traits. In the case of the TLab units, they don't use ferrofluid but they do have large decompression chamber, which is presumably what cases the visually similar impedance curve.
I'm not familiar with TLab beryllium dome units, but I have briefly fooled with their ceramic dome drivers, which based on their exceptional distortion performance are quite capable in some circumstances of being crossed at 1.4KHz or lower, and I see no reason based on the published spec. to assume this unit is any different. As a loose guideline, it's preferable for the [mid]bass drivers to give up, either through excursion or distortion performance, before the tweeter. That reduces the likelihood of you frying the latter's VC or reaching its excursion limits. Used well, most TLab units lean toward being excursion rather than distortion limited as far as output is concerned. Assuming a low crossover frequency (relative to a given electrical & acoustical order), you'll likely need at least a shunt resistive element to damp the impedance somewhat. If you're needing most of their sensitivity that may not be possible without the OA impedance load becoming unacceptably low sans any series R; likewise if you are pitching for a very low XO point with a higher order filter. In those cases you'll need an LCR shunt to flatten things more effectively.
I'm not familiar with TLab beryllium dome units, but I have briefly fooled with their ceramic dome drivers, which based on their exceptional distortion performance are quite capable in some circumstances of being crossed at 1.4KHz or lower, and I see no reason based on the published spec. to assume this unit is any different. As a loose guideline, it's preferable for the [mid]bass drivers to give up, either through excursion or distortion performance, before the tweeter. That reduces the likelihood of you frying the latter's VC or reaching its excursion limits. Used well, most TLab units lean toward being excursion rather than distortion limited as far as output is concerned. Assuming a low crossover frequency (relative to a given electrical & acoustical order), you'll likely need at least a shunt resistive element to damp the impedance somewhat. If you're needing most of their sensitivity that may not be possible without the OA impedance load becoming unacceptably low sans any series R; likewise if you are pitching for a very low XO point with a higher order filter. In those cases you'll need an LCR shunt to flatten things more effectively.
Last edited:
I support that statement, THD and also separate K2, K3, ... curves at appropriate levels are very helpful to find "lowest useful" crossover frequencies.
Nevertheless even the "first estimations" in this thread seem "OK".
But as also stated here there is a difference between
a) lowest possible crossover frequency due do electromotorical/mechanical limits of a particular driver (tweeter)
b) an appropriate crossover frequency due to the dispersion pattern of a planned multiway speaker and also the "abilities" of the particular woofer or midrange unit, possibly influenced by breakup issues and the like.
While a) is something like a "technical limitation" b) is a "design decision".
To make good speakers one has to make (good as possible ...) design decisions first and then choose the components accordingly IMO.
Some design decisons may then - of cause - be modified a bit due to reality of components available or "affordable" in a certain project.
In case of trying a passive crossover, flattening a tweeter's impedance curve ist helpful IMO, to meet the desired slope.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.