@Kornbread said:
Could delay the further woofer using dsp?
The front woofer and the tweeter would need to be delayed?
Wondering what would happen if you delayed the inside woofer just a smidge and/or use the dsp to smooth out the overall woofer response by adjusting only the inside woofer.
wouldn't that increase the interference wavelength, as it would take longer to interfere with the backward wave from the front woofer. Delaying the front woofer and tweeter would allow both the front woofer and the back woofer wave to reach the listener / mic at the same time and no 180 out of phase interference - i think
I suppose that DSP could solve this ripple problem. But then, due to the DSP signal processing time lag, I would also have to DSP the tweeter signal as well. And I do not want to do that. DSP changes the sound quality of high frequencies in a way that I do not like. I would much rather stick with a passive solution.
@ani_101 said:
Bill, do you have an measurement of the ripple or can you post a zoomed in image or indicate the ripple. I suppose if you disconnect the second (inside) driver the 1600 hz ripple go away?
In theory, it should. I don't have a measurement of this ripple. I have a 4" long test tunnel that I used to measure the TS parameters of the two isobarik woofers. I could mount this tunnel on a small box and then take an isobarik 1 meter FR measurement. Then I could mount a single woofer on the same box and take a single driver 1 meter FR measurement. Then add both curves into OnmiMic and compare. The impedance would be different for the two measurements. The single woofer would be a 4 ohm load. The isobarik woofer would be an 8 ohm load. This would affect the low frequency box alignment somewhat, but the higher frequencies should be comparable. I will give this a go and report back.
@ani_101 said:
Would a 2.5 way work for isibarik? The front woofer playing upto the hf, the back woofer only the lows? Would that be weird?
But then mark (from meniscus) had one woofer ported and one sealed playing the same freq...
I would think that this arrangement could be make to work. Put some additional inductance on the internal woofer to roll off the highs faster. You could also try using two different woofers in isobarik configuration. The external woofer would be a higher quality unit crossed at 3k or so. The internal woofer would be a subwoofer with very limited high frequency performance. You would have to very carefully match output levels of the two drivers, or this could turn into a real mess!
For the test, I used one of my completed Zonker cabinets. I ran two sets of wires out the port hole. One set of wires going to the inner woofer and the other set going to the exterior woofer. These woofers (SD215A-88) have dual 8 ohm voice coils, which I connected in series for 16 ohms. So each woofer was 16 ohms for the individual measurements. For the isobarik woofer tests, I paralleled the two 16 ohm drivers for 8 ohms.
I ran the following OmniMic FR tests at a distance 26 inches on axis to the woofer:
1) Both woofers connected in phase
2) Both woofers connected out of phase
3) Exterior woofer only connected (interior woofer became an internal passive radiator)
4) Internal woofer only connected (exterior woofer became a sound deadening cover)
I also ran DATS V2 impedance curves for all of the above measurements
Here is a comparison of #1 and #3.
Here is an impedance curve for #1 and #3
The SPL differences at high frequencies were much smaller than I had anticipated. And they were not at the frequencies that I had calculated based on the 4" isobarik tunnel (1695Hz). This is probably because the tunnel air gap is very resilient and is lined with 1/2" fiberglass. The tunnel probably acts as a complex resonating chamber on its own. Just speculating. Not what I expected. (If you want me to post the other curves too, let me know).
Thanks Bill. This is interesting. I also noticed the difference isn't much between 1 and 3 to complicate things to try a 2.5 way. I should remember to test out both woofers connected together and then each of them individually.
I am curious to see the out of phase graph, is there a significant amount of cancellation compared to graph 1?
@ani_101 said:
I am curious to see the out of phase graph, is there a significant amount of cancellation compared to graph 1?
Thanks
Ani
Yes, it makes a big difference. As you can see in the graph below, the out-of-phase isobarik woofer FR gets clobbered below 100Hz. And it peaks up a little in the 150Hz to 300Hz area as well. This peaking is probably the comb filtering type response created by changing separation distances. This seems to indicate that isobarik tunnel type woofer coupling is mainly effective at the low frequencies, where the wavelength is huge compared to the dimensions of the connecting tunnel. As you move higher in frequency, the wavelength becomes small compared to the tunnel dimensions and therefore the effect is less predictable and much less in magnitude. As you suggest, this is good news because 2 way or 2.5 way designs will not be impacted very much by the tunnel.
I ordered xover parts and mounted them on the inside of the back panel:
I abandoned my custom port/stand (epic fail) and installed a 10.125" long by 2.4"ID port on the back:
On-Axis frequency response looks really good. Excellent right/left channel match:
Here is the revised final xover. R1 of 8 ohms was too bright sounding. R1 of 10 ohms sounded a little dull. But trimming R1 to 9.4 ohms sounded "just right" to my old ears.
@ugly_woofer said:
I like the steep crossover slopes Bill, nicely done.
Thanks, Nick. I kept the cost down by using Daytons I-core's and NPE's on the woofer branch. It is sounding pretty good so far and the tweeters seems to be able to handle the power when I crank them up into the 90dB+ range. I'll run some 95dB distortion plots and post them later.
@Wolf said:
I wonder if mine will have the same issue with the offset back to back arrangement.
I reviewed your tandem drawings again. You have a 3 x 10" tunnel/chamber with the isobarik woofers positioned side by side. Will you be putting damping material of any kind in the tunnel space between the two driver magnets? If so, this could potentially help to dampen out any issues, I would think.
@Kornbread said:
Once I get close it is soooooo hard to make those smaller decisions.
The differences were only about 0.5 dB between the 9.4 and the 10 ohms resistor. I soldered two 333 ohm 1 watt metal films resistors across the 10 ohm to get the 9.4 ohm value:
Finally finished! It got up to 50 degrees yesterday and 55 today (a real December heat wave for Wisconsin), so I set the finished speakers outside for clear coating and a few pics. Last week, I sealed the exposed particle board edges with several coats of Crystalac, which dried clear. Yesterday, I gave them about 7 to 8 light coats of semi-gloss clear lacquer (two rattle cans). The Crystalac undercoating fogged a little when top coated with lacquer, but cleared again after about 6 hours or so.
I painted the red horseshoe shaped racing stripe around the speaker with a small paint brush. It is centered on the tweeter and covers up the glue line left by making my own handi-panels out of 12" wide shelving stock:
The port/stand system was a failure because I miscalculated the length & tuning, so I pluged the port hole with a 3/4" thick plate covered with foam rubber on both sides:
I made a removable port with a square mounting flange. The port was 10.125" long for a 32Hz tuning. The best NF port measurement came from positioning the port so that 50% of the port was inside the box and 50% was outside the box:
Here I fixed my router boo-boo by mixing up a batch of saw dust and wood glue. I troweled it into postion by hand and used a section of the plastic cone protector, that came with the speaker, as a form to hold the glue in place until it dried. After it set for several days, I routed it again, this time being careful not to let the router bearing fall into the woofer recess.
Did you try putting a 90 on the port inside? You can then change the location inside the cabinet by varying the lengths before and after the 90 or by rotating the 90.
@squamishdroc said:
I like that you left the edges raw - looks good👍👍
Thanks, squamishdroc. I remembered your suggestion a while back about the look of the raw edges, so I decided to give it a try. I think it turned out well with very good contrast against the dark laminate. The secret to getting this look is to first seal just the edges with 7 or 8 coats of Crystalac. Crystalac dries clear and does not yellow the particle board like lacquer, Sealcoat sanding sealer, or wood glue. Once the Crystalac was dry, the top coats of lacquer did not soak in and cause the particle board to turn yellow.
In my pictures, if you look at the particle board on the port frame, it has a deeper yellow cast compared to the edges. This is because I did not use Crystalac on the port frame. I simply sprayed the port frames with several coats of lacquer, which soaked in and gave the frames a yellow cast.
@ugly_woofer said:
Did you try putting a 90 on the port inside? You can then change the location inside the cabinet by varying the lengths before and after the 90 or by rotating the 90.
Yes, I tried 90 degree elbows in an attempt to position the internal port terminus in the same position, but I could not get it to work. I flipped the back 180 degrees also to get the internal terminus in the same position, but no luck. For some odd reason, I could only get it to work with a straight pipe length, 10.125" long, pointed directly toward the internal trap opening. The internal trap is about 5.5" long, stuffed with denim.
In the future, what I need to do is to build an experimental box with a large, overhanging removable back. Then I could easily move the port around and determine the best position before cutting any holes in the real back panel.
I clock my internal port tubes from vertical to (subjectively) reduce the port resonances. I ended up pointing the port at the top corner of the box on several projects. I don't have data to share but think this would be a good investigation.
I have wondered if having a port begin near a corner would change the tuning somehow. My guy tells me the pressure inside the box should be fairly constant, but I'm not certain of that.
@Tom_S said:
I have wondered if having a port begin near a corner would change the tuning somehow. My guy tells me the pressure inside the box should be fairly constant, but I'm not certain of that.
@Tom_S said:
I have wondered if having a port begin near a corner would change the tuning somehow. My guy tells me the pressure inside the box should be fairly constant, but I'm not certain of that.
Placement of the port at the wall like a slot port or at the corner will change the length compared to center of the wall for the same box tuning.
Thanks for the reference, Ani. I wish I had read this thread before I constructed my "port/stand". I made my port/stand about 3 to 4 inches too long because I did not account for the interior walls, as suggested by Jeff B.
@Tom_S said:
I have wondered if having a port begin near a corner would change the tuning somehow. My guy tells me the pressure inside the box should be fairly constant, but I'm not certain of that.
Regarding constant pressure, my current understanding is that the pressure inside the box is constant, but only at frequencies were the wavelength is long compared to the longest internal dimension on the box (i.e., in and around the tuning frequency). But when the wavelengths are shorter, then all kinds of strange peaks and dips start to show up (i.e., 100Hz to 1kHz).
@Tom_S said:
I have wondered if having a port begin near a corner would change the tuning somehow. My guy tells me the pressure inside the box should be fairly constant, but I'm not certain of that.
Regarding constant pressure, my current understanding is that the pressure inside the box is constant, but only at frequencies were the wavelength is long compared to the longest internal dimension on the box (i.e., in and around the tuning frequency). But when the wavelengths are shorter, then all kinds of strange peaks and dips start to show up (i.e., 100Hz to 1kHz).
After thinking about this a little more, I need to correct my previous post. The pressure inside a bass reflex speaker enclosure is not constant, but varies substantially above and below the tuning frequency. At the tuning frequency (Fb) the pressure inside the box (Helmholtz resonance) is so high that it actually stops the woofer cone. Above and below the tuning frequency, the pressure inside the box is very low. This is why the voice coil vent holes in the first production runs of the CX150-8 coaxial speakers caused a huge buzzing noise at resonance when mounted in a bass reflex enclosure, but did not buzz at all when mounted in a small sealed enclosure and crossed at 400Hz. Here is the PETT thread: http://techtalk.parts-express.com/forum/tech-talk-forum/1340640-dayton-coaxial-w-silk-dome
It's not obvious from this image, but cabinet pressure is in kPa at 0.5kPa per division, so about 1.8kPa at the peak. This is a random 12" driver in 60L tuned to 30Hz and provided 100W input power.
Comments
wouldn't that increase the interference wavelength, as it would take longer to interfere with the backward wave from the front woofer. Delaying the front woofer and tweeter would allow both the front woofer and the back woofer wave to reach the listener / mic at the same time and no 180 out of phase interference - i think
I suppose that DSP could solve this ripple problem. But then, due to the DSP signal processing time lag, I would also have to DSP the tweeter signal as well. And I do not want to do that. DSP changes the sound quality of high frequencies in a way that I do not like. I would much rather stick with a passive solution.
In theory, it should. I don't have a measurement of this ripple. I have a 4" long test tunnel that I used to measure the TS parameters of the two isobarik woofers. I could mount this tunnel on a small box and then take an isobarik 1 meter FR measurement. Then I could mount a single woofer on the same box and take a single driver 1 meter FR measurement. Then add both curves into OnmiMic and compare. The impedance would be different for the two measurements. The single woofer would be a 4 ohm load. The isobarik woofer would be an 8 ohm load. This would affect the low frequency box alignment somewhat, but the higher frequencies should be comparable. I will give this a go and report back.
I would think that this arrangement could be make to work. Put some additional inductance on the internal woofer to roll off the highs faster. You could also try using two different woofers in isobarik configuration. The external woofer would be a higher quality unit crossed at 3k or so. The internal woofer would be a subwoofer with very limited high frequency performance. You would have to very carefully match output levels of the two drivers, or this could turn into a real mess!
Thanks Bill. The measurements are to test out the HF ripple. The box design would need to rely on the Isobarik box model.
Ani,
For the test, I used one of my completed Zonker cabinets. I ran two sets of wires out the port hole. One set of wires going to the inner woofer and the other set going to the exterior woofer. These woofers (SD215A-88) have dual 8 ohm voice coils, which I connected in series for 16 ohms. So each woofer was 16 ohms for the individual measurements. For the isobarik woofer tests, I paralleled the two 16 ohm drivers for 8 ohms.
I ran the following OmniMic FR tests at a distance 26 inches on axis to the woofer:
1) Both woofers connected in phase
2) Both woofers connected out of phase
3) Exterior woofer only connected (interior woofer became an internal passive radiator)
4) Internal woofer only connected (exterior woofer became a sound deadening cover)
I also ran DATS V2 impedance curves for all of the above measurements
Here is a comparison of #1 and #3.
Here is an impedance curve for #1 and #3
The SPL differences at high frequencies were much smaller than I had anticipated. And they were not at the frequencies that I had calculated based on the 4" isobarik tunnel (1695Hz). This is probably because the tunnel air gap is very resilient and is lined with 1/2" fiberglass. The tunnel probably acts as a complex resonating chamber on its own. Just speculating. Not what I expected. (If you want me to post the other curves too, let me know).
Bill
Thanks Bill. This is interesting. I also noticed the difference isn't much between 1 and 3 to complicate things to try a 2.5 way. I should remember to test out both woofers connected together and then each of them individually.
I am curious to see the out of phase graph, is there a significant amount of cancellation compared to graph 1?
Thanks
Ani
Yes, it makes a big difference. As you can see in the graph below, the out-of-phase isobarik woofer FR gets clobbered below 100Hz. And it peaks up a little in the 150Hz to 300Hz area as well. This peaking is probably the comb filtering type response created by changing separation distances. This seems to indicate that isobarik tunnel type woofer coupling is mainly effective at the low frequencies, where the wavelength is huge compared to the dimensions of the connecting tunnel. As you move higher in frequency, the wavelength becomes small compared to the tunnel dimensions and therefore the effect is less predictable and much less in magnitude. As you suggest, this is good news because 2 way or 2.5 way designs will not be impacted very much by the tunnel.
Bill
Thanks Bill!
A little more progress:
I ordered xover parts and mounted them on the inside of the back panel:
I abandoned my custom port/stand (epic fail) and installed a 10.125" long by 2.4"ID port on the back:
On-Axis frequency response looks really good. Excellent right/left channel match:
Here is the revised final xover. R1 of 8 ohms was too bright sounding. R1 of 10 ohms sounded a little dull. But trimming R1 to 9.4 ohms sounded "just right" to my old ears.
I may have to do some experimental measurements when I get the Tandems that far.
InDIYana Event Website
Once I get close it is soooooo hard to make those smaller decisions.
I wonder if mine will have the same issue with the offset back to back arrangement.
InDIYana Event Website
I like the steep crossover slopes Bill, nicely done.
Thanks, Nick. I kept the cost down by using Daytons I-core's and NPE's on the woofer branch. It is sounding pretty good so far and the tweeters seems to be able to handle the power when I crank them up into the 90dB+ range. I'll run some 95dB distortion plots and post them later.
I reviewed your tandem drawings again. You have a 3 x 10" tunnel/chamber with the isobarik woofers positioned side by side. Will you be putting damping material of any kind in the tunnel space between the two driver magnets? If so, this could potentially help to dampen out any issues, I would think.
The differences were only about 0.5 dB between the 9.4 and the 10 ohms resistor. I soldered two 333 ohm 1 watt metal films resistors across the 10 ohm to get the 9.4 ohm value:
Finally finished! It got up to 50 degrees yesterday and 55 today (a real December heat wave for Wisconsin), so I set the finished speakers outside for clear coating and a few pics. Last week, I sealed the exposed particle board edges with several coats of Crystalac, which dried clear. Yesterday, I gave them about 7 to 8 light coats of semi-gloss clear lacquer (two rattle cans). The Crystalac undercoating fogged a little when top coated with lacquer, but cleared again after about 6 hours or so.
I painted the red horseshoe shaped racing stripe around the speaker with a small paint brush. It is centered on the tweeter and covers up the glue line left by making my own handi-panels out of 12" wide shelving stock:
The port/stand system was a failure because I miscalculated the length & tuning, so I pluged the port hole with a 3/4" thick plate covered with foam rubber on both sides:
I made a removable port with a square mounting flange. The port was 10.125" long for a 32Hz tuning. The best NF port measurement came from positioning the port so that 50% of the port was inside the box and 50% was outside the box:
Here I fixed my router boo-boo by mixing up a batch of saw dust and wood glue. I troweled it into postion by hand and used a section of the plastic cone protector, that came with the speaker, as a form to hold the glue in place until it dried. After it set for several days, I routed it again, this time being careful not to let the router bearing fall into the woofer recess.
Did you try putting a 90 on the port inside? You can then change the location inside the cabinet by varying the lengths before and after the 90 or by rotating the 90.
I like that you left the edges raw - looks good👍👍
Thanks, squamishdroc.
I remembered your suggestion a while back about the look of the raw edges, so I decided to give it a try. I think it turned out well with very good contrast against the dark laminate. The secret to getting this look is to first seal just the edges with 7 or 8 coats of Crystalac. Crystalac dries clear and does not yellow the particle board like lacquer, Sealcoat sanding sealer, or wood glue. Once the Crystalac was dry, the top coats of lacquer did not soak in and cause the particle board to turn yellow.
In my pictures, if you look at the particle board on the port frame, it has a deeper yellow cast compared to the edges. This is because I did not use Crystalac on the port frame. I simply sprayed the port frames with several coats of lacquer, which soaked in and gave the frames a yellow cast.
Yes, I tried 90 degree elbows in an attempt to position the internal port terminus in the same position, but I could not get it to work. I flipped the back 180 degrees also to get the internal terminus in the same position, but no luck. For some odd reason, I could only get it to work with a straight pipe length, 10.125" long, pointed directly toward the internal trap opening. The internal trap is about 5.5" long, stuffed with denim.
In the future, what I need to do is to build an experimental box with a large, overhanging removable back. Then I could easily move the port around and determine the best position before cutting any holes in the real back panel.
I clock my internal port tubes from vertical to (subjectively) reduce the port resonances. I ended up pointing the port at the top corner of the box on several projects. I don't have data to share but think this would be a good investigation.
I have wondered if having a port begin near a corner would change the tuning somehow. My guy tells me the pressure inside the box should be fairly constant, but I'm not certain of that.
Bill - I like the red stripe!
Tom, please see this thread: http://techtalk.parts-express.com/forum/tech-talk-forum/66368-q-slot-port-length
Placement of the port at the wall like a slot port or at the corner will change the length compared to center of the wall for the same box tuning.
Thanks for the reference, Ani. I wish I had read this thread before I constructed my "port/stand". I made my port/stand about 3 to 4 inches too long because I did not account for the interior walls, as suggested by Jeff B.
Regarding constant pressure, my current understanding is that the pressure inside the box is constant, but only at frequencies were the wavelength is long compared to the longest internal dimension on the box (i.e., in and around the tuning frequency). But when the wavelengths are shorter, then all kinds of strange peaks and dips start to show up (i.e., 100Hz to 1kHz).
After thinking about this a little more, I need to correct my previous post. The pressure inside a bass reflex speaker enclosure is not constant, but varies substantially above and below the tuning frequency. At the tuning frequency (Fb) the pressure inside the box (Helmholtz resonance) is so high that it actually stops the woofer cone. Above and below the tuning frequency, the pressure inside the box is very low. This is why the voice coil vent holes in the first production runs of the CX150-8 coaxial speakers caused a huge buzzing noise at resonance when mounted in a bass reflex enclosure, but did not buzz at all when mounted in a small sealed enclosure and crossed at 400Hz. Here is the PETT thread: http://techtalk.parts-express.com/forum/tech-talk-forum/1340640-dayton-coaxial-w-silk-dome
It's not obvious from this image, but cabinet pressure is in kPa at 0.5kPa per division, so about 1.8kPa at the peak. This is a random 12" driver in 60L tuned to 30Hz and provided 100W input power.