The data sheet provided by B&C does not list all of the T/S parameters for this driver. Anyone have them?
In order to cross over at the recommended frequency of 1.6kHz, I need a notch filter(s). There are two resonance peaks for this driver, according to their data sheet. So I need at least: Fs, Qes, Qts, and Re. I sent an email to them on Friday but have not heard back yet. Maybe it's holiday in Italy or something. I'm sure they'll get back to me eventually.
Also, does the electrical impedance change when a driver is coupled to a horn? I bought the ME20 horn to go with this driver.
In order to cross over at the recommended frequency of 1.6kHz, I need a notch filter(s). There are two resonance peaks for this driver, according to their data sheet. So I need at least: Fs, Qes, Qts, and Re. I sent an email to them on Friday but have not heard back yet. Maybe it's holiday in Italy or something. I'm sure they'll get back to me eventually.
Also, does the electrical impedance change when a driver is coupled to a horn? I bought the ME20 horn to go with this driver.
Electrical impedance diffinitely changes when you put a compression driver on a horn. This is part and parcel to the high efficiency making the horn loading a significant part, as opposed to the typical woofer where low efficiency means that air load is fairly insignificant.
Thiele Small figures for a compression driver? They aren't typically given since woofer box modeling doesn't apply. (The system is not well represented by 2nd order highpass analogies.) What were you trying to do with them? How do you kow what notch filters you might need without measuring the driver on the particular horn you will use it with?
David S.
Thiele Small figures for a compression driver? They aren't typically given since woofer box modeling doesn't apply. (The system is not well represented by 2nd order highpass analogies.) What were you trying to do with them? How do you kow what notch filters you might need without measuring the driver on the particular horn you will use it with?
David S.
Well, it looks like I have to measure it while mounted on the horn. Assuming there is still a resonance peak in the same region as shown in the datasheet, then I'm going to need a notch filter as I wrote about above. I need the T/S parameters to design the series notch filter. The formulas are:
C= 0.1592/(Re x Fs x Qes), L= (0.1592 x Qes x Re)/Fs, and R = Re + (Qes x Re)/Qms
Huh, just realized I need Qms. Dang.
I was going to describe my DE250 (mounted) impedance but I'll just attach this screenshot. I have alternated between using two and three notch filters, two for the main peaks and one purely for the response at the peak of hearing frequencies.
I can't say it will be ideal to use formulas here. The frequencies are so close that they'll interact and you need to design them as a team, and then you'll probably want to play with them in tailoring your response.
I can't say it will be ideal to use formulas here. The frequencies are so close that they'll interact and you need to design them as a team, and then you'll probably want to play with them in tailoring your response.
Attachments
Thank you! Yeah, this looks difficult, for sure. I have WT2 and TrueRTA so I can do a decent analysis I hope. I don't want to go active on this little project because I have a sweet 50 watt stereo integrated amp that I want to keep. I'll go active on the 3 way horn loaded system I have planned.
Speaking of WT2, I also have a pair of ME10's mounted on DE10's for another project. So, I checked out the impedance plot and T/S parameters for this combination. They are different from the manufacturers spec, but only Fs is very different. I had emailed B&C asking for the Fs of the DE10 driver, and they replied with 1.7kHz, which is about where the peak is in their plot. So, I measured Fs of 2.5kHz with the ME10 horns installed and the driver sitting on my desk aimed at the ceiling. One thing I don't get is the inductance change. I measured 0.4mH, and the data sheet says 0.1mH. I didn't think the horn would effect the inductance of the voice coil. Do compression drivers have to broken in like woofers?
I can't test the DE250 + ME20 yet since they have not arrived.
Speaking of WT2, I also have a pair of ME10's mounted on DE10's for another project. So, I checked out the impedance plot and T/S parameters for this combination. They are different from the manufacturers spec, but only Fs is very different. I had emailed B&C asking for the Fs of the DE10 driver, and they replied with 1.7kHz, which is about where the peak is in their plot. So, I measured Fs of 2.5kHz with the ME10 horns installed and the driver sitting on my desk aimed at the ceiling. One thing I don't get is the inductance change. I measured 0.4mH, and the data sheet says 0.1mH. I didn't think the horn would effect the inductance of the voice coil. Do compression drivers have to broken in like woofers?
I can't test the DE250 + ME20 yet since they have not arrived.
It looks like the peaks shifted also. This is starting to feel like mud wrestling with a greased pig.
It's not so bad when you get the hang of it!
I hope you don't think (as some misinformation would have you...) that an active system sidesteps dealing with driver issues, this would limit your success with active systems.
Compression drivers are a little different to dome tweeters. I cross mine over at 650Hz (first order).
They didn't shift, my filters grazed the sides of them. It was a juggling act. I can't remember whether that's the best I could do with common values or I have fudged them a little to mod the response since the last revision. I highly recommend a crossover simulator. For this part alone you'll want impedance magnitude and impedance phase for each driver.
OK, which crossover simulator do you recommend? I also have WinSpeakerz, which is fine for boxes, but doesn't simulate crossovers as far as I know. I can calculate the values no problem, but simulating them with an actual driver impedance curve is something I don't seem to have.
Yes, I was under the impression that active systems didn't care about driver impedance, resonances, etc. Can you suggest some reading for me on this?
Yes, I was under the impression that active systems didn't care about driver impedance, resonances, etc. Can you suggest some reading for me on this?
Wow, B&C recommend a crossover point of 1.6kHz for the DE250. Which horn are you using with it? There must have been a significant shift in the frequency response due to the horn in order for you to cross over that low. The response in their data sheet shows a steep cut off at about 900 Hz.
I'd stop short of recommending, most are somewhat unique and take some getting used to. I used EXO to answer a question below, I like it because it does good polar plots. SpeakerWorkshop was used above with those impedance plots and I use it for groundwork, and I used Crossover Simulator (xoversim) from the FRDConsortium in the no measurement tutorial on this forum. I used to use it more often and it isn't too hard to get up and running.
That part is not the issue. Speakers don't have flat responses, their acoustic phase varies, there are regions of a drivers response that shouldn't be used, and it all should go together from angles other than just front on....
So if you don't know how to design a good passive crossover then you can't design a good active one. The only significant difference from our point of view is the interaction with the impedance.
One thing I like about the passive crossover is the simplicity with which I can shape the response in almost any way. Digital processing is interesting, but again, if you know what you're doing.
A waveguide (largely conical), and 18" at the mouth.
This is my data without a crossover.
Attachments
Well, with the tools I have I can do a lot already. I read somewhere that just a Zobel can effect the resonant peak in some drivers, so it seems to me that the first step is just to work with that and see what the response looks like.
That's a big waveguide! I can see that the low end of the response is extended a little by your guide, but I don't think this driver can handle much power down there. The ME20 horn seems to have pretty good directivity, if I'm reading the charts correctly.
Interesting. I've heard that LEAP is very competent but I assume it would be a handful to learn on. I'd have assumed that you'd have your hands full creating data files to work with (frd / zma files). On the other hand you could take advantage of a sim that processes measurements for you. I value the ability to create the data files as it allows me to manipulate them, this is obviously just a preference of mine.
It's enough to hear it clearly across the street 😉
A compression driver is a two degree of freedom system, just like a ported enclosure and that is why you get two peaks even when there is no horn. If you took off the phase plug you would get one. The other is the resonance of the radiation mass with the compliance of the air gap in front of the diaphragm. B&C is not going to give you the data that you need, they never do, no one does. But there is a way, except that it gets complicated.
There are three easy conditions that you can do and modeling the unit under these three conditions will give you the values that you need. The conditions are freely radiating (you can ignoire the phase plug flare as its not significant in this short device, its just a port), block the output (now the phase pluig is just a compliance, remove the back (with either or both of the previous condictions). When you get impedance curves in all three (or four) of these conditions as close as possible to a simulation then you have the correct parameters. Of course you use those parameters that B&C provides, but these aren't always the same as what I get.
When the horn adds more peaks (than two) then there is a significant mouth reflection, A "good" horn will still only have two peaks, although the trough between them will move depending on the horn (just like tuning a ported enclosure - different horns add different mass). I have the DE250 parameters somewhere and I will post them if I find them. But my parameters are not TS poarameters, they don't mean anything in this context, they are just straightforward electro-mechanical; masss, volumes, etc.
SPEAK can do a competent job of modeling a compression driver on a horn. Not many other programs can do that.
There are three easy conditions that you can do and modeling the unit under these three conditions will give you the values that you need. The conditions are freely radiating (you can ignoire the phase plug flare as its not significant in this short device, its just a port), block the output (now the phase pluig is just a compliance, remove the back (with either or both of the previous condictions). When you get impedance curves in all three (or four) of these conditions as close as possible to a simulation then you have the correct parameters. Of course you use those parameters that B&C provides, but these aren't always the same as what I get.
When the horn adds more peaks (than two) then there is a significant mouth reflection, A "good" horn will still only have two peaks, although the trough between them will move depending on the horn (just like tuning a ported enclosure - different horns add different mass). I have the DE250 parameters somewhere and I will post them if I find them. But my parameters are not TS poarameters, they don't mean anything in this context, they are just straightforward electro-mechanical; masss, volumes, etc.
SPEAK can do a competent job of modeling a compression driver on a horn. Not many other programs can do that.
Last edited:
Thank you very much for that informative reply!
I guess if you have a way to measure electrical impedance, Le and Re, you can pretty much design a zobel and the crossover, but how do you design a notch filter @ resonance without T/S parameters?
For a high pass crossover less than one octave from the highest resonance of the compression driver and horn, are you supposed to have 2 notch filters, one for each resonance?
This crossover could get complicated really fast.
I guess if you have a way to measure electrical impedance, Le and Re, you can pretty much design a zobel and the crossover, but how do you design a notch filter @ resonance without T/S parameters?
For a high pass crossover less than one octave from the highest resonance of the compression driver and horn, are you supposed to have 2 notch filters, one for each resonance?
This crossover could get complicated really fast.
A zobel is often used where tools/data are lacking, as it can be done from specs. This is where you may need to deviate from that way of thinking. As it turns out, the zobel isn't likely to be necessary regardless. If you can measure your impedance with the driver mounted, you can fix the peaks either by experimentation or with the help of a sim.
Thanks.
I'm having trouble designing a zobel for this midrange horn I have. It is an Edgar 500Hz tractrix with Dynaudio D54 driver. I used the Re of 4.6 ohms, and Le of 0.07 mH, and got Rz = 5.75 ohms and Cz = 2.12 uF. I used 5.7 ohms and 2.1 uF and hooked it up to LT2. As you can see the impedance in the upper frequency response is not flat. How flat is a zobel network supposed to make it?
Also, I have found at least 3 different zobel calculation methods, 2 of which I have tried. The one I tried and used is:
Cz = Le/Rz*2
Rz = 1.25 x Re
Another one is found here:
Passive Crossover Network Design
where
F = Re/(2 x Pi x Le)
Cz = 1/(2 x Pi x F x Re)
Rz = Re
Which produces a different result than the previous one.
Then there is the one described in the Leach paper:
http://users.ece.gatech.edu/mleach/papers/zobel/zobel.pdf
Which appears to be totally different and I actually don't understand how to use what he shows.
So, I'm at a loss here...
Attachments
- Status
- Not open for further replies.