Hello everyone, I finally got my hands on a DATS V3 and measured some drivers that I had from an old all in 1 boxed home theater. I did the calibration on the DATS first. After all was said and done I end up with a ported volume of 6.69L for a speaker that is 53.5mm in diameter. Is it just me or is that a large cabinet for such a small driver? The enclosure they came out of was not this large in volume.
Here are the measurements from DATS using ADDED Mass Method:
Fs=216.2
Qts=1.217
Qes=1.573
Qms=5.37
Vas=0.175 litres
Do these numbers seem correct and does an enclosure of 6.69 litres for 2" driver make sense?
Here are the measurements from DATS using ADDED Mass Method:
Fs=216.2
Qts=1.217
Qes=1.573
Qms=5.37
Vas=0.175 litres
Do these numbers seem correct and does an enclosure of 6.69 litres for 2" driver make sense?
I would personally not use DATS data for doing an enclosure. This tool guesses the T/S (according to the designer) as many do.
Outside of the tyoical factory tools — not really available any more and causing some manufactures grief as they try to find other tools that give them simialr consistent results.
But given that, with the Q of the driver as it comes out, the appropriate size box is as large as you can live with. An infinite baffle is appropriate. Given the Fs this is possibly suitable for midTweeter use on an OB above ~400Hz.
dave
Outside of the tyoical factory tools — not really available any more and causing some manufactures grief as they try to find other tools that give them simialr consistent results.
But given that, with the Q of the driver as it comes out, the appropriate size box is as large as you can live with. An infinite baffle is appropriate. Given the Fs this is possibly suitable for midTweeter use on an OB above ~400Hz.
dave
Hi Dave, thanks for the reply though I am not sure I got it. Are you saying this is best suited for an application going down no further than 400Hz. If so, is this what the port tuning should be. This is my first go at it. I have always wanted to make a pair of speakers, and I'm finally getting the ball rolling.
The problem is that there is no closed or bass reflex enclosure that will deliver a clean low-end rolloff with these drivers. Assuming the estimation is anywhere near correct, Qts = 1.2 is simply much too high for that. In a closed box, Qtc is always higher than Qts, and you are generally aiming for a Qtc of 0.707, maybe a bit higher. Bass reflex designs have even stricter requirements, and good BR drivers generally have Qts below 0.5.
Qts that high is not untypical for cheap and/or wideband drivers.
Your drivers may still be useful as mid/high frequency units later, but if you are aiming to make a simple fullrange speaker as your first DIY project, honestly look for something bigger - with fs that high, these are not likely to be of much use below about 250 Hz or so. There are several 3" fullranges that shouldn't break the bank (including the Tymphany TC9FD18-08), you can also go up to 4" or 5" from Tangband.
Qts that high is not untypical for cheap and/or wideband drivers.
Your drivers may still be useful as mid/high frequency units later, but if you are aiming to make a simple fullrange speaker as your first DIY project, honestly look for something bigger - with fs that high, these are not likely to be of much use below about 250 Hz or so. There are several 3" fullranges that shouldn't break the bank (including the Tymphany TC9FD18-08), you can also go up to 4" or 5" from Tangband.
At a half-liter and up, that driver will be around a +3dB bump at 260Hz. And you probably should not ask for lower than 260Hz from a 2-inch more than a foot from your ear. The 6 liter "solution" is a hair smoother and FAR larger. Not likely a good solution unless a gallon jug is handy.
Cave-man approach. Many speakers will be "happy" (as happy as they get) in a box of (D*1.1)^3. Here that is 11ci or 0.19L. Which is essentially the Vas. So Q and Fs shift-up to 1.7 (4.6dB bump) at 300Hz. Which is not -that- much different from the 6.6L "solution" but 1/30th the box volume.
With Q that high, a vent-box resonance won't do a whole lot, the speaker is in charge. You could add a whole lot of volume for very little bass extension or reduction of excursion.
> The enclosure they came out of was not this large
It's no longer a Black Art. Factory speakers are usually "nearly right", or not THAT far away from an optimum.
This DATS must be like the bird-dog who walks past the bird, completely loses the trail, and finally gets tired and stops.
Cave-man approach. Many speakers will be "happy" (as happy as they get) in a box of (D*1.1)^3. Here that is 11ci or 0.19L. Which is essentially the Vas. So Q and Fs shift-up to 1.7 (4.6dB bump) at 300Hz. Which is not -that- much different from the 6.6L "solution" but 1/30th the box volume.
With Q that high, a vent-box resonance won't do a whole lot, the speaker is in charge. You could add a whole lot of volume for very little bass extension or reduction of excursion.
> The enclosure they came out of was not this large
It's no longer a Black Art. Factory speakers are usually "nearly right", or not THAT far away from an optimum.
This DATS must be like the bird-dog who walks past the bird, completely loses the trail, and finally gets tired and stops.
Last edited:
The onus is on me here. I should have been more clear. I'm looking to use the speaker mentioned above with the sub-woofer that came in with the box HT system. And if I need to get an inexpensive tweeter for the top end then I will. I'm not looking at doing a full range build.
I'm following 2 books right now, Speaker Building 201 and Loud Speaker Design Cookbook.
Should my approach be backwards? Should I look for a port tuning tho get the size of my enclosure? From what has been suggested, aim for something between 250-400Hz?
I'm following 2 books right now, Speaker Building 201 and Loud Speaker Design Cookbook.
Should my approach be backwards? Should I look for a port tuning tho get the size of my enclosure? From what has been suggested, aim for something between 250-400Hz?
I assume you are not trying to push the state of the art, just trying and learning and enjoying.
You put this speaker in 10-20 cubic inches, say 1.5"x2.5"x3.25", whatever is handy. It will have a slight bump at 300Hz and probably go to the top of the audio band.
Put 47uFd, 68uFd, or 100uFd in series to shave the bump and reject everything below 270Hz. Turn your woofer into a carpeted corner to absorb highs. It may play great.
You put this speaker in 10-20 cubic inches, say 1.5"x2.5"x3.25", whatever is handy. It will have a slight bump at 300Hz and probably go to the top of the audio band.
Put 47uFd, 68uFd, or 100uFd in series to shave the bump and reject everything below 270Hz. Turn your woofer into a carpeted corner to absorb highs. It may play great.
One can measure the impedance and guess/extend that to the T/S or like manually, S+L WT2, or more expensive testers, the ap actaully hunts down the Fs and from there the Fl/Fh and directly calulate the T/S from the real numbers. My previous tester used the guess/extend method and was not really great, but given my use for driver matching (mostly) it was fine.
dave
dave
Hi planet10,
That’s not correct information. I just started using my DATS3 - it uses a continuously varying logarithmic swept sine stimulus (you can hear it). I’m pretty sure it uses the method Farina (proposed in a 2000 AES paper) to perform a cross-correlation of the stimulus and the measured waveform. The cross-correlation produces an impulse response, then a time-windowed FFT of the impulse response produces the fundamental response. The frequency resolution of the response depends on the FFT size, you can get sub-Hz resolution with a 64k sampled FFT. To calculate the impedance the tester probably uses a current sense resistor and divides the voltage by the current. Anyway that’s the way I’d do it. The only thing I’d do different is to run multiple sweeps to reduce the noise using curve averaging, but that could lead to heating effects on small drivers.
Olen
That’s not correct information. I just started using my DATS3 - it uses a continuously varying logarithmic swept sine stimulus (you can hear it). I’m pretty sure it uses the method Farina (proposed in a 2000 AES paper) to perform a cross-correlation of the stimulus and the measured waveform. The cross-correlation produces an impulse response, then a time-windowed FFT of the impulse response produces the fundamental response. The frequency resolution of the response depends on the FFT size, you can get sub-Hz resolution with a 64k sampled FFT. To calculate the impedance the tester probably uses a current sense resistor and divides the voltage by the current. Anyway that’s the way I’d do it. The only thing I’d do different is to run multiple sweeps to reduce the noise using curve averaging, but that could lead to heating effects on small drivers.
Olen
And then it “guesses” the T/S from the the impedance curve. When i used Fuzzmeasure for T/S it uses the same method.
The common manual method, and S+L WT2 actually go looking for the specific FS and F1 & F2. It literally hunts for them. One can watch it do so as it goes back & forth narrowing in on where those are. A much more accurate & useful method.
I still start any design with the factory numbers, the much pricier (and some NLA) kit used produces more useful numbers.
dave
Neither is accurate, however, since they are based on an overly simplified model of the transducer which neglects viscoelastic effects. The WTPro can produce good impedance measurement data for parameter extraction using more accurate models, though.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.