Hi all,
this is my first post here.
I would like to build my first 2-way speaker (I built a small fullrange last year), and need some advice.
In order to test some wood working techniques I wanted to build an inexpensive but good enough speaker before spending a lot in parts for an expensive one (good candidate could be the Zaph Audio SR71).
I searched on the web but didn't found anything that I can build; on the excellent Zaph's site I found a couple projects interesting (ZMV5), but in Italy I can't find the drivers needed, and having the drivers shipped from the States would be uneconomical.
Looking in the Zaph's site I found a woofer that could match my idea of a good enough and inexpensive driver - the Peerless 830656 for 21 Euro each, coupled with maybe a Vifa tweeter, in a roughly 10 l vented enclosure.
Anybody has a design with this woofer? I tried to design a crossover, but I'm not sure I did it correct.
Thanks, Ralf
this is my first post here.
I would like to build my first 2-way speaker (I built a small fullrange last year), and need some advice.
In order to test some wood working techniques I wanted to build an inexpensive but good enough speaker before spending a lot in parts for an expensive one (good candidate could be the Zaph Audio SR71).
I searched on the web but didn't found anything that I can build; on the excellent Zaph's site I found a couple projects interesting (ZMV5), but in Italy I can't find the drivers needed, and having the drivers shipped from the States would be uneconomical.
Looking in the Zaph's site I found a woofer that could match my idea of a good enough and inexpensive driver - the Peerless 830656 for 21 Euro each, coupled with maybe a Vifa tweeter, in a roughly 10 l vented enclosure.
Anybody has a design with this woofer? I tried to design a crossover, but I'm not sure I did it correct.
Thanks, Ralf
Member
Joined 2003
My design is with either a 10L 55Hz tuning or a 12.3L 50Hz tuning enclosure. The baffle is the same 195x333 mm, and drivers are located 130 and 260 mm from bottom.
I chose a Vifa BC25TG15-04 tweeter and ended with this crossover:
woofer: 2nd order 1.35 mH (0.2 Ohm) and 4.7 uF
tweeter: 3rd order 6.8 uF, 0.15 mH, 15 uF, for the L-pad 2.2 Ohm in series and 1.5 Ohm in parallel.
Ralf
I chose a Vifa BC25TG15-04 tweeter and ended with this crossover:
woofer: 2nd order 1.35 mH (0.2 Ohm) and 4.7 uF
tweeter: 3rd order 6.8 uF, 0.15 mH, 15 uF, for the L-pad 2.2 Ohm in series and 1.5 Ohm in parallel.
Ralf
Member
Joined 2003
The component values are not of much interest if there is no accompanying data. Unfortunately I don't have the time to run through a bunch of simulations to see what you've done.
Basically what I am asking is the process that you went through to determine the crossover. How were these values chosen?
-Baffle step compensation?
-driver phase taken into account
-acoustic offsets between drivers taken into account (z-axis)
Can you provide any simulated or measured data showing how you expect the speaker to perform?
Basically what I am asking is the process that you went through to determine the crossover. How were these values chosen?
-Baffle step compensation?
-driver phase taken into account
-acoustic offsets between drivers taken into account (z-axis)
Can you provide any simulated or measured data showing how you expect the speaker to perform?
Hi !
Do you have all the speaker ? I will make a similar design with the DX25TG05-04 in a fortnight using a 190x300x(250mm deep) box 8L@55Hz. If you use the same tweeter i can share my design
. It will be crossed 2kHz LR4.
For your design (3kHz LR4) you should use instead
woofer: 2nd order 2.7 mH (0.3 Ohm) and 3.3uF
tweeter: 3rd order 4.7 uF, 0.22 mH, 15 uF, for the L-pad 3.3 Ohm in series and 1.5 Ohm in parallel.
I think these values are better than yours but not optimum.
Do you have all the speaker ? I will make a similar design with the DX25TG05-04 in a fortnight using a 190x300x(250mm deep) box 8L@55Hz. If you use the same tweeter i can share my design
. It will be crossed 2kHz LR4.For your design (3kHz LR4) you should use instead
woofer: 2nd order 2.7 mH (0.3 Ohm) and 3.3uF
tweeter: 3rd order 4.7 uF, 0.22 mH, 15 uF, for the L-pad 3.3 Ohm in series and 1.5 Ohm in parallel.
I think these values are better than yours but not optimum.
This was the method I used.
I played a bit with WinISD in order to have an idea of a good tuning.
I used Zaph's frequency response for the woofer, and Vifa's one for the tweeter, and the manufacture's for the impedance, creating so 2 base frd and 2 base zma files. I extracted the phase data using this function in the Response Modeler (RM) program, and obtained 2 zma files with phase information.
I then used the RM in order to obtain the bass section with my enclosure, and also the baffle step and diffraction. For the tweeter obviously only the diffraction effect. I created so a frd file with bass gain from the enclosure and baffle step effect for the woofer, and a frd file with diffraction effect for the tweeter.
I then used the Passive Crossover Designer program in order to design the crossover using the frd and zma files. As a z offset between woofer and tweeter I used 20 mm.
The frequency response is attached
I played a bit with WinISD in order to have an idea of a good tuning.
I used Zaph's frequency response for the woofer, and Vifa's one for the tweeter, and the manufacture's for the impedance, creating so 2 base frd and 2 base zma files. I extracted the phase data using this function in the Response Modeler (RM) program, and obtained 2 zma files with phase information.
I then used the RM in order to obtain the bass section with my enclosure, and also the baffle step and diffraction. For the tweeter obviously only the diffraction effect. I created so a frd file with bass gain from the enclosure and baffle step effect for the woofer, and a frd file with diffraction effect for the tweeter.
I then used the Passive Crossover Designer program in order to design the crossover using the frd and zma files. As a z offset between woofer and tweeter I used 20 mm.
The frequency response is attached
Attachments
Member
Joined 2003
By Response Modeller (RM), you mean the Minimum Phase Response Modeller? I have no experience using this spreadsheet, but I do think that there is something wrong with your graphs.
Looking at the phase, this looks far too smooth to have been extracted from Frequency response data. Also, the pointy peaks and troughs look quite odd. Baffle Step Compensation could be increased some.
I'm not sure how you would figure out baffle step and diffraction with the response modeller, but then again I haven't played around with it to see what it can do. [Edit: I've looked into this spreadsheet, and it will give some idea of diffraction and baffle step, but it a somewhat simplified view, with no off-axis capabilities. Baffle Diffraction Simulator is a much more powerful tool]
It looks like you've used a simulated impedance for your calculations. There is really no need for this unless the impedance peak caused by the port is in the affected frequency range of the crossover. The original measured impedance will be more accurate for crossover calculations.
Here's the process I go through using the FRD tools:
-trace frequency response and impedance from measured data
-simulate this response on the measurement baffle, save resulting response
-simulate the woofer in-box using Unibox, save resulting response
-simulate baffle diffraction using your intended baffle using Baffle Diffraction Simulator, save resulting response
-combine results and extract phase information using Frequency Response Combiner. Save final response including phase data.
-Input Final frequency response and impedance data into Jeff Bagby's Passive Crossover Designer.
There are some guidelines that you may want to follow at these sites:
http://www.rjbaudio.com/Audiofiles/FRDtools.html
http://www.geocities.com/woove99/Spkrbldg/DesigningXO.htm
Looking at the phase, this looks far too smooth to have been extracted from Frequency response data. Also, the pointy peaks and troughs look quite odd. Baffle Step Compensation could be increased some.
I'm not sure how you would figure out baffle step and diffraction with the response modeller, but then again I haven't played around with it to see what it can do. [Edit: I've looked into this spreadsheet, and it will give some idea of diffraction and baffle step, but it a somewhat simplified view, with no off-axis capabilities. Baffle Diffraction Simulator is a much more powerful tool]
It looks like you've used a simulated impedance for your calculations. There is really no need for this unless the impedance peak caused by the port is in the affected frequency range of the crossover. The original measured impedance will be more accurate for crossover calculations.
Here's the process I go through using the FRD tools:
-trace frequency response and impedance from measured data
-simulate this response on the measurement baffle, save resulting response
-simulate the woofer in-box using Unibox, save resulting response
-simulate baffle diffraction using your intended baffle using Baffle Diffraction Simulator, save resulting response
-combine results and extract phase information using Frequency Response Combiner. Save final response including phase data.
-Input Final frequency response and impedance data into Jeff Bagby's Passive Crossover Designer.
There are some guidelines that you may want to follow at these sites:
http://www.rjbaudio.com/Audiofiles/FRDtools.html
http://www.geocities.com/woove99/Spkrbldg/DesigningXO.htm
DcibeL,
I found the Response Modeler here
It seems that the Minimum Phase Response Modeler is a subset of what I used (same author).
The phase has been extracted from the impedance, and if you use a smooth impedance plot also the phase will be smooth. You are right about the woofer impedance, as I modeled it with the basta! program. However using the impedance from the peerless site I found a similar result for the frequency response in the PCD program (even slightly better).
I chose a non complete baffle step compensation expecting some room gain.
I will have a look at the Baffle Diffraction Simulator, but I have a question: what do you mean with "simulate this response on the measurement baffle"? I expect that the measurements are made on an infinite baffle (or something similar) - OK, I read better, I have to subtract the diffraction induced by the baffle used for the measurement.
And a final question: do you think that 20 mm for the acoustic offset between drivers on the z-axis is a good value?
For jerome69: share your design, maybe I will use your because I don't have yet the drivers, and if your is proven, it will be better than mine!
Thanks, Ralf
I found the Response Modeler here
It seems that the Minimum Phase Response Modeler is a subset of what I used (same author).
The phase has been extracted from the impedance, and if you use a smooth impedance plot also the phase will be smooth. You are right about the woofer impedance, as I modeled it with the basta! program. However using the impedance from the peerless site I found a similar result for the frequency response in the PCD program (even slightly better).
I chose a non complete baffle step compensation expecting some room gain.
I will have a look at the Baffle Diffraction Simulator, but I have a question: what do you mean with "simulate this response on the measurement baffle"? I expect that the measurements are made on an infinite baffle (or something similar) - OK, I read better, I have to subtract the diffraction induced by the baffle used for the measurement.
And a final question: do you think that 20 mm for the acoustic offset between drivers on the z-axis is a good value?
For jerome69: share your design, maybe I will use your because I don't have yet the drivers, and if your is proven, it will be better than mine!
Thanks, Ralf
Member
Joined 2003
Sorry I hadn't looked as closely at your design as maybe I should have. As long as you've compared the simulated impedance to the measured and found it to be very close, you will be okay.
Less than full baffle step is usually a good idea, especially if you cannot move the speakers very far from the wall for whatever reason.
For "simulate on measurement baffle" basically what I am trying to get at is create all the simulations that the Frequency Response Combiner is looking for:
-original measured data
-box sim of measurement conditions (if infinite baffle, simulate closed box of 1000L in Unibox)
-box sim of your cabinet
-baffle diffraction sim of measurement box (if infinite baffle just select infinite baffle or 2pi I think it might say)
-baffle diffraction sim of your baffle
Once you combine all the frequency response data, you'll want to use the 'extract minimum phase" feature of FRC. You may find this to differ from the phase you extracted from the impedance quite a bit. For crossover design you are after acoustic phase, not electrical phase.
The good thing about Baffle Diffraction Simulator is that you can simulate off-axis conditions as well, which is important. What might look like good diffraction on-axis, may be poor off-axis and vice versa.
To find the difference between acoustic centers (z-axis), measure from where the voice coil meets the diaphragm. 20mm is probably close. One thing that I will mention is that when using PCD, the driver offsets are in meters, and you'll want to use the tweeter as the on-axis location. So the woofer will have a negative Z-axis value and negative Y-axis value.
It seems as though you've done a fair bit of research, and are putting a lot more effort into this that I ever did on my first build. I am confident that it will turn out well.
Less than full baffle step is usually a good idea, especially if you cannot move the speakers very far from the wall for whatever reason.
For "simulate on measurement baffle" basically what I am trying to get at is create all the simulations that the Frequency Response Combiner is looking for:
-original measured data
-box sim of measurement conditions (if infinite baffle, simulate closed box of 1000L in Unibox)
-box sim of your cabinet
-baffle diffraction sim of measurement box (if infinite baffle just select infinite baffle or 2pi I think it might say)
-baffle diffraction sim of your baffle
Once you combine all the frequency response data, you'll want to use the 'extract minimum phase" feature of FRC. You may find this to differ from the phase you extracted from the impedance quite a bit. For crossover design you are after acoustic phase, not electrical phase.
The good thing about Baffle Diffraction Simulator is that you can simulate off-axis conditions as well, which is important. What might look like good diffraction on-axis, may be poor off-axis and vice versa.
To find the difference between acoustic centers (z-axis), measure from where the voice coil meets the diaphragm. 20mm is probably close. One thing that I will mention is that when using PCD, the driver offsets are in meters, and you'll want to use the tweeter as the on-axis location. So the woofer will have a negative Z-axis value and negative Y-axis value.
It seems as though you've done a fair bit of research, and are putting a lot more effort into this that I ever did on my first build. I am confident that it will turn out well.
I had some time to play with Baffle Diffraction Simulator and Frequency Response Combiner, and I recalculated my design. The resulting response is slightly different than before, but it is good enough. However, off axis is worse, and I found that offsetting the tweeter and tilting the speaker turn it again in a good one. So far so good.
My concern is that in any case the result is strongly dependent from the z-axis offset, with the best value now at 15 mm. Probably this is a too low value, and not having in hand the driver I cannot check.
So, is there someone that can measure or know this value? (AFAIK from where the voice coil meets the diaphragm and the woofer's face).
Thanks, Ralf
My concern is that in any case the result is strongly dependent from the z-axis offset, with the best value now at 15 mm. Probably this is a too low value, and not having in hand the driver I cannot check.
So, is there someone that can measure or know this value? (AFAIK from where the voice coil meets the diaphragm and the woofer's face).
Thanks, Ralf
Member
Joined 2003
The datasheet for the Peerless driver shows the following:
I'd guesstimate that from the the voicecoil to the front of the baffle is about half of the 58.2mm measurement, so probably about 27-28mm. Subtract the recess for the tweeter if you have one, and you're probably somewhere between 20-25mm.
An externally hosted image should be here but it was not working when we last tested it.
I'd guesstimate that from the the voicecoil to the front of the baffle is about half of the 58.2mm measurement, so probably about 27-28mm. Subtract the recess for the tweeter if you have one, and you're probably somewhere between 20-25mm.
For the drivers' acoustic center positions, if you surface-mount the woofer, use a 13-14 mm z offset on the midwoofer and 0 mm on the tweeter (that is, the woofer's AC is 13-14 mm behind the tweeter's with respect to the baffle plane). Upon looking at the drivers' cone/dome structure, I think this should be close to the real difference of their minimum phase points. About this issue. take a look at the following thread:
http://techtalk.parts-express.com/showthread.php?t=208562
You don't have to read all the posts there. Reading posts number 6, 7, 12, 15, 17, 19, 26, and 29 should be sufficient to get the right information on this particular subject.
http://techtalk.parts-express.com/showthread.php?t=208562
You don't have to read all the posts there
. Reading posts number 6, 7, 12, 15, 17, 19, 26, and 29 should be sufficient to get the right information on this particular subject.
Member
Joined 2003
Member
Joined 2003
For curiosity sake, could you post a screenshot of your final crossover response? Something like this would be great!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.