John,
Those were just approximations based on the 40 Hz. I didn't actually model the horn. It was merely to illustrate the very large mouth one would need for the numerous driver array. Additionally, too, the large flare needed to support 40 Hz for just a single driver. Check your CIR to see if it is 1 or greater.
I have a foreshortened midbass horn for a 15" driver that has a 9" x9" throat, and exponential flare out to an ~17" x 26" mouth at 23" long. Rear chamber is probably 30 liters or so. That is a very short horn for 80 Hz cut-off, but good to perhaps ~95 Hz with one side on the floor. You can do that sort of thing with a 15" driver in an exponential horn with compression.
If you model it, it will show a pronounced peak from 100 to 200 Hz, with not much but roll off either side. It is more extended above 250 Hz than shown, so HornResp won't include this.
Your experimental length of 4.7 ft axial length doesn't seem that far off if the combined Sd is close to a 15". But your Fs is still too high, right?
You probably need something in the ~50 Hz range for a 40 Hz horn.
How you model the throat will have a large bearing too on the other dimensions. That is, are you slot loading with some compression, just using the outside dimensions of the drivers stacked up, or actually using a "throat" into the horn for each device. Check if SD=S1, or is less. It will make a difference.
Tim
Those were just approximations based on the 40 Hz. I didn't actually model the horn. It was merely to illustrate the very large mouth one would need for the numerous driver array. Additionally, too, the large flare needed to support 40 Hz for just a single driver. Check your CIR to see if it is 1 or greater.
I have a foreshortened midbass horn for a 15" driver that has a 9" x9" throat, and exponential flare out to an ~17" x 26" mouth at 23" long. Rear chamber is probably 30 liters or so. That is a very short horn for 80 Hz cut-off, but good to perhaps ~95 Hz with one side on the floor. You can do that sort of thing with a 15" driver in an exponential horn with compression.
If you model it, it will show a pronounced peak from 100 to 200 Hz, with not much but roll off either side. It is more extended above 250 Hz than shown, so HornResp won't include this.
Your experimental length of 4.7 ft axial length doesn't seem that far off if the combined Sd is close to a 15". But your Fs is still too high, right?
You probably need something in the ~50 Hz range for a 40 Hz horn.
How you model the throat will have a large bearing too on the other dimensions. That is, are you slot loading with some compression, just using the outside dimensions of the drivers stacked up, or actually using a "throat" into the horn for each device. Check if SD=S1, or is less. It will make a difference.
Tim
Tim,
My thought process was as follows:
Since it's a rear horn, length is a big issue to get proper phasing where the front radiation takes over. I want to keep it at 1/2 wavelength without going too short. I settled on 120hz, 1/2 wavelength 143cm.
I wanted to keep the throat/sd ratio above .3 , but on the low side to keep the overall size of the horn small and extension is the goal not overall efficiency. Also, I want to keep the narrow width of the throat in combinations of 12mm &/or 18mm plywood thickness to easily make a spacer to ensure it is uniform throughout the length. I settled on a throat of 570 cm2 or 42% of the 1360 sd of 15 5" drivers I plan to use now and a width starting at 30mm.
For the flare rate, I just copied the hypex flare of the Jensen Imperial which also has 4 segments and a large compression chamber.
I have 2 areas of flexibility at the end. The mouth size and the compression chamber volume.
Bracing is needed anyway and I read that very narrow passages is bad. 15 braces through the 1st 2 segments gives me 16 passages that start at 3cmX12cm and end at 10x12 after the 2nd segment, where they will flow seamlessly into 8 passages and 4 for the final flare. That way I can cut 3 pieces of plywood that are dividers throughout the horn, 4 that run through the 3rd segment and 8 that cover only the 1st 2 segments. This won't be too difficult. Just cut the pieces, then cut out 12mm slots to slide the 3 vertical panels into making assembly quite easy. Once I get the first panel in, glued and sealed, the braces will form a skeleton for the rest of the horn.
One thing I don't understand is what to look for in the acoustical impedance graphs. The other is the effects of a large chamber. In playing around with chamber size my best response is with a large chamber, but I end up with an upper frequency cutoff based on throat and chamber that is lower than 40hz. ????
My thought process was as follows:
Since it's a rear horn, length is a big issue to get proper phasing where the front radiation takes over. I want to keep it at 1/2 wavelength without going too short. I settled on 120hz, 1/2 wavelength 143cm.
I wanted to keep the throat/sd ratio above .3 , but on the low side to keep the overall size of the horn small and extension is the goal not overall efficiency. Also, I want to keep the narrow width of the throat in combinations of 12mm &/or 18mm plywood thickness to easily make a spacer to ensure it is uniform throughout the length. I settled on a throat of 570 cm2 or 42% of the 1360 sd of 15 5" drivers I plan to use now and a width starting at 30mm.
For the flare rate, I just copied the hypex flare of the Jensen Imperial which also has 4 segments and a large compression chamber.
I have 2 areas of flexibility at the end. The mouth size and the compression chamber volume.
Bracing is needed anyway and I read that very narrow passages is bad. 15 braces through the 1st 2 segments gives me 16 passages that start at 3cmX12cm and end at 10x12 after the 2nd segment, where they will flow seamlessly into 8 passages and 4 for the final flare. That way I can cut 3 pieces of plywood that are dividers throughout the horn, 4 that run through the 3rd segment and 8 that cover only the 1st 2 segments. This won't be too difficult. Just cut the pieces, then cut out 12mm slots to slide the 3 vertical panels into making assembly quite easy. Once I get the first panel in, glued and sealed, the braces will form a skeleton for the rest of the horn.
One thing I don't understand is what to look for in the acoustical impedance graphs. The other is the effects of a large chamber. In playing around with chamber size my best response is with a large chamber, but I end up with an upper frequency cutoff based on throat and chamber that is lower than 40hz. ????
John,
If I follow you, the upper cut-off was determined by the low end of the direct radiator output. Yes, a good plan, but as I described on the midbass horn, the program will not show the upper range very accurately and you can expect some output above cut-off. My guess is you will have some irregularities if you measure in-room response where they overlap but I don't think it will be that significant or annoying. You may have to experiment some with damping, for example, to absorb the HF. Look at the Jensen Imperial, as you mentioned, or the Altec A7 where both use a rear wave, one in a direct radiator and horn scoop, and one in a front horn with reflex enclosure, but both with bass output tuned well below the direct radiation. It can work with the right drivers and enclosure.
As far as compression and the throat chamber, I would keep the load on the drivers as low as possible to still yield the efficiency I was looking for. Generally, the higher the compression, the higher the driver load, and I'm not sure of the durability of your clearance units. Throat can also effect high frequency roll-off, which is not specifically the issue, but it is in an overall scope.
VRC should be set to zero in a rear horn, right, while the front compression chamber can be determined by the hypex calculator as a double check. It sounds as though some of the parameters have been mixed up. You may have to deviate some from the Imperial. Do it your way, then use the calculators, then use compare.
Acoustical impedance , near as I can tell in this display, is reactance/resistance in a mirror image and can overlap or diverge. There is a scaling factor to be used based on the design, but I don't know much about it. I look at the electrical impedance more closely. As far as horn design, the acoustic impedance is a reflection of the horn load. As it is a resonant device I expect that the irregular tracking of the graph is part and parcel of the resonant nature. I notice that the two should closely overlap and the peak amplitudes(?) should be kept as low as possible. Things really start to get erratic when you intentionally load bad numbers into the data base, showing that the driver is not being properly loaded by the horn.
Tim
If I follow you, the upper cut-off was determined by the low end of the direct radiator output. Yes, a good plan, but as I described on the midbass horn, the program will not show the upper range very accurately and you can expect some output above cut-off. My guess is you will have some irregularities if you measure in-room response where they overlap but I don't think it will be that significant or annoying. You may have to experiment some with damping, for example, to absorb the HF. Look at the Jensen Imperial, as you mentioned, or the Altec A7 where both use a rear wave, one in a direct radiator and horn scoop, and one in a front horn with reflex enclosure, but both with bass output tuned well below the direct radiation. It can work with the right drivers and enclosure.
As far as compression and the throat chamber, I would keep the load on the drivers as low as possible to still yield the efficiency I was looking for. Generally, the higher the compression, the higher the driver load, and I'm not sure of the durability of your clearance units. Throat can also effect high frequency roll-off, which is not specifically the issue, but it is in an overall scope.
VRC should be set to zero in a rear horn, right, while the front compression chamber can be determined by the hypex calculator as a double check. It sounds as though some of the parameters have been mixed up. You may have to deviate some from the Imperial. Do it your way, then use the calculators, then use compare.
Acoustical impedance , near as I can tell in this display, is reactance/resistance in a mirror image and can overlap or diverge. There is a scaling factor to be used based on the design, but I don't know much about it. I look at the electrical impedance more closely. As far as horn design, the acoustic impedance is a reflection of the horn load. As it is a resonant device I expect that the irregular tracking of the graph is part and parcel of the resonant nature. I notice that the two should closely overlap and the peak amplitudes(?) should be kept as low as possible. Things really start to get erratic when you intentionally load bad numbers into the data base, showing that the driver is not being properly loaded by the horn.
Tim
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.