1)Thanks Linn!
2) Hi Rvrazvan, design was already narowed to JBL 2360a and BMS 4590 components.
Still crossover and woofer has to be determined.
Right now I am studing passive 2nd order crossovers. Idea is to run passive with the coaxial and active with the woofer. I guess it will look like a frankenstein speaker for a while...
Best.
2) Hi Rvrazvan, design was already narowed to JBL 2360a and BMS 4590 components.
Still crossover and woofer has to be determined.
Right now I am studing passive 2nd order crossovers. Idea is to run passive with the coaxial and active with the woofer. I guess it will look like a frankenstein speaker for a while...
Best.
The BMS4590 are at home!!
During the week I am bringing the Horns 2360A.
My plan is to start testing the Horn-Driver in MONO with the first initial setup:
Computer -> MiniDSP Ch1 -> DAC Ch1 -> 1.5watt 45 SET Amplifier ->BMS High
-----------> MiniDSP Ch2 -> DAC Ch2 -> 1.5watt 45 SET Amplifier ->BMS Low
Filters for High Frequency BMS will be:
1)High pass 18db/oct at 6.3khz
2)Low pass 18db/oct at 20khz
Filters for Low Frequency BMS will be:
1)Low pass 18db/oct at 6.3khz
2)High pass 18db/oct at 400hz
This is my first experience with horns, so I am a bit concerned about not destroying the BMS diaphragms. Only Thing I am aware of, is to not go too low in frequency.
I am thinking 300hz would be a lower limit for BMS low.
and 6khz would be a lower limit for BMS highs.
I plan to measure sound with a Berhinger calibrated microphone using HornResp.
And of course... listen to music.
Any advice will be much appretiated.
Cheers!
During the week I am bringing the Horns 2360A.
My plan is to start testing the Horn-Driver in MONO with the first initial setup:
Computer -> MiniDSP Ch1 -> DAC Ch1 -> 1.5watt 45 SET Amplifier ->BMS High
-----------> MiniDSP Ch2 -> DAC Ch2 -> 1.5watt 45 SET Amplifier ->BMS Low
Filters for High Frequency BMS will be:
1)High pass 18db/oct at 6.3khz
2)Low pass 18db/oct at 20khz
Filters for Low Frequency BMS will be:
1)Low pass 18db/oct at 6.3khz
2)High pass 18db/oct at 400hz
This is my first experience with horns, so I am a bit concerned about not destroying the BMS diaphragms. Only Thing I am aware of, is to not go too low in frequency.
I am thinking 300hz would be a lower limit for BMS low.
and 6khz would be a lower limit for BMS highs.
I plan to measure sound with a Berhinger calibrated microphone using HornResp.
And of course... listen to music.
Any advice will be much appretiated.
Cheers!
Last edited:
Hello guys,
I am coming back with great news. I have set up and have been listening since last Friday and during the whole weekend. Sound is unbelievable, I am so happy with the results. Horns are a different game (compared with cone drivers). Directivity is great, clarity is great, dynamics is superb, it doesn´t sound harsh and the diffraction horn reaches the highs with no problems up to 20khz without severe beaming.
On the listening side, music sounds great, I just couldn’t stop listening during to music during the whole weekend. Voices are natural and trumpets and bronzes sound so great. There are no signs of compression, actually I discover with this experiment how compressed sound is in more traditional cone speakers. It’s so clear that I can hear how the sound engineer is moving the volume knobs in the different tracks of some recordings.
There are still lots to do, in particular I am having issues with Woofer integration. Even with this, sounds is superb and I could experience the potential and the horn performance. Here there is a measurement on axis done with Holmimpulse:
This is the tri-amplified MONO setup I used as a first test:
Highs: JBL 2360A -> BMS4590 HF -> 45 Globe Cunningham Single Ended No feedback, Tango XE20s transformers (right channel)->MiniDSP 2x8
Mids: JBL 2360A -> BMS4590 MF -> 45 Globe Cunningham Single Ended No feedback, Tango XE20s transformers (left channel)->MiniDSP 2x8
Woofer: Omega Pro 18a in U-shape Open baffle -> 41hz.com 10W Class D amplifier -> MiniDSP 2x8.
The crossover settings:
HF: High Pass BW 24 db/oct at 6300hz
MF:
Low Pass BW 24 db/oct at 6300hz
High Pass BW 24 db/oct at 500hz
LW:
Low Pass BW 24db/oct at 500hz
This is a photo of the 45 SET amplifier:
I will post a list of the things that need to be resolved, horn photos, etc., in a next post. Now I am happy.
Best,
Carlos
I am coming back with great news. I have set up and have been listening since last Friday and during the whole weekend. Sound is unbelievable, I am so happy with the results. Horns are a different game (compared with cone drivers). Directivity is great, clarity is great, dynamics is superb, it doesn´t sound harsh and the diffraction horn reaches the highs with no problems up to 20khz without severe beaming.
On the listening side, music sounds great, I just couldn’t stop listening during to music during the whole weekend. Voices are natural and trumpets and bronzes sound so great. There are no signs of compression, actually I discover with this experiment how compressed sound is in more traditional cone speakers. It’s so clear that I can hear how the sound engineer is moving the volume knobs in the different tracks of some recordings.
There are still lots to do, in particular I am having issues with Woofer integration. Even with this, sounds is superb and I could experience the potential and the horn performance. Here there is a measurement on axis done with Holmimpulse:
An externally hosted image should be here but it was not working when we last tested it.
This is the tri-amplified MONO setup I used as a first test:
Highs: JBL 2360A -> BMS4590 HF -> 45 Globe Cunningham Single Ended No feedback, Tango XE20s transformers (right channel)->MiniDSP 2x8
Mids: JBL 2360A -> BMS4590 MF -> 45 Globe Cunningham Single Ended No feedback, Tango XE20s transformers (left channel)->MiniDSP 2x8
Woofer: Omega Pro 18a in U-shape Open baffle -> 41hz.com 10W Class D amplifier -> MiniDSP 2x8.
The crossover settings:
HF: High Pass BW 24 db/oct at 6300hz
MF:
Low Pass BW 24 db/oct at 6300hz
High Pass BW 24 db/oct at 500hz
LW:
Low Pass BW 24db/oct at 500hz
This is a photo of the 45 SET amplifier:
An externally hosted image should be here but it was not working when we last tested it.
I will post a list of the things that need to be resolved, horn photos, etc., in a next post. Now I am happy.
Best,
Carlos
Last edited:
Carlos,
A few points:
1. A 10 dB difference in level makes something sound twice (or half) as loud, your mid/high levels are about twice as loud as your lows.
2. The 10 dB dip in the 4-5 hHz range might be indicative of a polarity reversal (though I would expect a deeper notch if reversed), or some delay may need to be added to the mid or HF to time align them.
3.The 35 dB hole at 325 Hz, wow. Open baffle wraparound cancellation?
Again, time alignment or polarity may be part of this problem.
Normally I'm not a fan of crossing over a relatively small diaphragm compression driver as low as 325 Hz, but the woofer's response may require an exception to that rule, and the BMS and the JBL horn are designed to work that low.
As it is, you are running 10 times the power through the mid/highs (or 1/10th the power to the woofer) as you should, once the mid/high level is normalized to the woofer dropping the crossover point should not be a problem.
4. A finer vertical scale resolution would be better, optimally you should be able to get response to fit within a +/- 5 dB window, having an 80 dB window makes the huge variations seen look rather small.
5. Looking at the individual response of each driver would be helpful, as would looking at their phase response.
Art
Last edited:
Art, Carlos...
check the gating on the Holm measurement. The gate is at 1.2 kHz (that small arrow on the x-axis). That means two things:
1. Data below 1.2 kHz is invalid.
2. The resolution of the response above 1.2 kHz is 1.2 kHz. Which means Holm will draw a line for you from 1.2 kHz to 2.4 kHz, but there is no data in that range.
I don't know how much experience you have with Holm or taking measurements. If you don't have much, read the documentation that comes with Holm. It is useful. And ask questions here.
Some things to note when making measurements:
1. The area should be fairly clear of other obstructions. Outdoors is best.
2. When sound bounces off the floor (and other surfaces) and gets to the mic, it produces peaks and dips in the response. You can "gate" the impulse response such that the later arriving reflections don't get into the measurement. There is a slider in Holm that lets you do that on the impulse response.
3. The mic should be calibrated for its response.
check the gating on the Holm measurement. The gate is at 1.2 kHz (that small arrow on the x-axis). That means two things:
1. Data below 1.2 kHz is invalid.
2. The resolution of the response above 1.2 kHz is 1.2 kHz. Which means Holm will draw a line for you from 1.2 kHz to 2.4 kHz, but there is no data in that range.
I don't know how much experience you have with Holm or taking measurements. If you don't have much, read the documentation that comes with Holm. It is useful. And ask questions here.
Some things to note when making measurements:
1. The area should be fairly clear of other obstructions. Outdoors is best.
2. When sound bounces off the floor (and other surfaces) and gets to the mic, it produces peaks and dips in the response. You can "gate" the impulse response such that the later arriving reflections don't get into the measurement. There is a slider in Holm that lets you do that on the impulse response.
3. The mic should be calibrated for its response.
Art,
I'm glad you were the one to say all that as I would have felt bad to be the one to say it! I agree that the low frequency output is so depressed that I just couldn't understand his raves about the sound? That hole between the lows and mids is huge and I don't see why he is cutting even an 18" off that low if he has that much of a gap. I would rather have a little cone breakup than that much of a hole. I still don't see much point in using the same driver for both the mids and highs on two different horns, the response curve is terrible between 2K and 7K that is just not an acceptable response curve. I think you are being nice saying plus and minus 5db is okay but I could accept that if he was close to accomplishing that. I doubt highly with the fast slopes that the crossover polarity is inverted or that it wouldn't be a narrow notch, it is just to wide to be that with a 24db slope. I would shift the bass cutoff up 200 cycles and see if the hole fills in and would look to see if that is a problem with the open back cabinet, one of the reasons I don't like that idea in the least, it just doesn't work out often unless he plays with the distance to the front wall and probably moves it very close to the wall. I would be looking real hard at a small cone midrange to clean up the mids and then a single horn on the top. Of course this is all opinion based on his FR curve he posted. Who knows how much the response is affected by his choice of tube amplifier and the impedance curve of the drivers?
I'm glad you were the one to say all that as I would have felt bad to be the one to say it! I agree that the low frequency output is so depressed that I just couldn't understand his raves about the sound? That hole between the lows and mids is huge and I don't see why he is cutting even an 18" off that low if he has that much of a gap. I would rather have a little cone breakup than that much of a hole. I still don't see much point in using the same driver for both the mids and highs on two different horns, the response curve is terrible between 2K and 7K that is just not an acceptable response curve. I think you are being nice saying plus and minus 5db is okay but I could accept that if he was close to accomplishing that. I doubt highly with the fast slopes that the crossover polarity is inverted or that it wouldn't be a narrow notch, it is just to wide to be that with a 24db slope. I would shift the bass cutoff up 200 cycles and see if the hole fills in and would look to see if that is a problem with the open back cabinet, one of the reasons I don't like that idea in the least, it just doesn't work out often unless he plays with the distance to the front wall and probably moves it very close to the wall. I would be looking real hard at a small cone midrange to clean up the mids and then a single horn on the top. Of course this is all opinion based on his FR curve he posted. Who knows how much the response is affected by his choice of tube amplifier and the impedance curve of the drivers?
Carlos is using a dual diaphragm annular co-ax BMS4590 on a single JBL 2360 horn for mid/high.
There is a known timing issue between the output of the two diaphragms in the BMS4590 , I just don't recall off hand which needs to be delayed. The delay needed is only a fraction of a millisecond, but at the mid/hi crossover frequency, wavelengths are only a fraction of a millisecond.
The JBL 2360 is quite a deep horn, unless the BMS4590 is located in line with the 18" woofer voice coil, it will need several ms of delay to "get in the ballpark".
When are you going to start using paragraph breaks
?Art
I have no experience with Holmresponse, ( I can help people with Smaart) but everything you mentioned in post 248 makes sense.
I don't know what a "small gate" compared to a "large gate" looks like in this instance, but "filling in the gaps" in 1.2 kHz increments seems like a lot of information is missed.
Even so, there are definitely some gross issues to be dealt with in shaping up the response of Carlos' system.
Hi Art, Ra7
Thanks for the feedback. I agree mostly, there was good advice there. I didn't think about Dipole cancelation, that is also an option.
The idea of this first setup was to listen to the 2360A - BMS 4590 couple. I intentionally left the 18in woofer with less SPL. I believe final design should have a 15in woofer (not 18) so I was trying to evaluate the horn mostly (first horn ever) to see what we are talking about and what is the potential.
I am not sure yet if the 350hz dip is caused by Open Baffle effect, Time alignment of sources (acoustical centers are at a distance of 130cm at the moment), or it's just an issue with the measurement (gating), so that investigation will be next priority. I also plan to get a few 15in woofers to compare. After correct woofer integration I will start exploring equalization. Well there is a lot to do, first time getting my hands with practical work with horns so everything is tested (experimented) slowly, to confirm mental ideas with practice.
Kindhorman,
Think you are missing the point. This is just a step by step of a guy working with horns for the first time. Someone learning and corroborating theory with experinces. So I am not presenting a final solution but the "process" and its findings.
Horn cualities like Directivity, Clarity and Dynamics can be apretiated before tunning the XO, and yes, understanding the XO is far from finished, I am loving how it sounds.
Note the compresion driver is coaxial, so only one horn is being used at the time.
Next step is investigation of the Low-Mid integration.
Jack,
Not EQ yet, just an initial run with and educated guess of XO points.
Thanks, I am very impressed with the performance of Horns. I will report more later!
Best
Thanks for the feedback. I agree mostly, there was good advice there. I didn't think about Dipole cancelation, that is also an option.
The idea of this first setup was to listen to the 2360A - BMS 4590 couple. I intentionally left the 18in woofer with less SPL. I believe final design should have a 15in woofer (not 18) so I was trying to evaluate the horn mostly (first horn ever) to see what we are talking about and what is the potential.
I am not sure yet if the 350hz dip is caused by Open Baffle effect, Time alignment of sources (acoustical centers are at a distance of 130cm at the moment), or it's just an issue with the measurement (gating), so that investigation will be next priority. I also plan to get a few 15in woofers to compare. After correct woofer integration I will start exploring equalization. Well there is a lot to do, first time getting my hands with practical work with horns so everything is tested (experimented) slowly, to confirm mental ideas with practice.
Kindhorman,
Think you are missing the point. This is just a step by step of a guy working with horns for the first time. Someone learning and corroborating theory with experinces. So I am not presenting a final solution but the "process" and its findings.
Horn cualities like Directivity, Clarity and Dynamics can be apretiated before tunning the XO, and yes, understanding the XO is far from finished, I am loving how it sounds.
Note the compresion driver is coaxial, so only one horn is being used at the time.
Next step is investigation of the Low-Mid integration.
Jack,
Not EQ yet, just an initial run with and educated guess of XO points.
Thanks, I am very impressed with the performance of Horns. I will report more later!
Best
Thanks man, that is very good information, I tended to think that centers would be align by design, that explains a lot about that dip.
I will remeasure again., and will publish the response included the room.
To be honest, I am not sure about how much better gating you can get in the bass region with a 18 inch woofer that is at 30cm from the floor.
This requires further investigation for sure since I was playing arround sliding the gating and that dip moved between the 300-600hz region and then desapear, so I dont trust the automatic gating detection at the moment.
I guess I will have to elevate the woofer and measure components by separate...
Best,
Carlos.
Charly,
Sorry I misunderstood what you were trying to do and that this was your first experience with a horn loaded system. I was wondering why you decided to go with an open backed bass enclosure and a horn loaded section above that? I am having trouble with the two opposed polar patterns.
I would go with a 15" speaker or even smaller if need be to cross over into the horn mid section better. The truth is that you can always add a larger driver for the bottom octave if you want that bass impact that you get so easy from an 18" speaker.
I guess it was the statement of the mids and highs right under the chart that I looked at to fast and thought was two separate horns.......
Have fun learning, you will learn much if you keep at it. I am sure this will only be the beginning of your horn experience.
Sorry I misunderstood what you were trying to do and that this was your first experience with a horn loaded system. I was wondering why you decided to go with an open backed bass enclosure and a horn loaded section above that? I am having trouble with the two opposed polar patterns.
I would go with a 15" speaker or even smaller if need be to cross over into the horn mid section better. The truth is that you can always add a larger driver for the bottom octave if you want that bass impact that you get so easy from an 18" speaker.
I guess it was the statement of the mids and highs right under the chart that I looked at to fast and thought was two separate horns.......
Have fun learning, you will learn much if you keep at it. I am sure this will only be the beginning of your horn experience.
You are all way off base on what is going on here, in regards to the HF/MF area. Charly has simply adjusted the volume of the highs to sound right in regards to the mids. And has innocently come up with an interesting solution/situation. The real clue is the fact that 20khz is louder than 5khz. Anyone with real world experience with compression drivers would be ready to add a CD boost to the HF. (In my world that CD stands for constant directivity, not compression driver.) As the high frequency of a compression driver tends to roll of in a CD horn. Also, as Tom said a bit back, he should be prepared to do some (lots?) of EQing to make the driver/horn combination flat in response. (Also of note, as Tom has noted here and elsewhere many times, exponential horns (which this is not) have flatter frequency response, because every single frequency has a different dispersion, with higher frequencies being narrower, and getting wider as they go down, and this compensates for the fact that a driver naturally rolls off as it goes higher in frequency.) So, several things.
1)The bump from 1.5K to 3K is the first thing that needs to be addressed. With a parametric EQ in the controller.
2) Normally, with the (BMS) passive crossover, there is no noticeable distractions in the frequency response curve, at the crossover point. The time differences between the mids and highs will probably be more noticeable on a phase curve, but this is not what is causing the discrepancy in the crossover area.
3) Normally, the process for setting the relative volumes on the amps/processor would be to match the volume at the crossover point. This seems to be off by about 8 to 10 db. But, since he has started out this way, it now leaves an interesting possibility. Instead of adding a CD curve to the top end to get the very highs loud enough, Charly has the option of a PEQ cut in the 6K to 10K range to tame this from the other end (of the HF). Meaning, instead of adding a CD boost, he can have a cut to the lower end of the HF to get the same results.
4) Of most concern to me is the dip at 750hz. I say this because while not as bad as the lower dip, it comes in the middle of a band, and neither the horn or driver should have this much variation in it. And it would not seem to have as simple of an explanation as the lower dip. Of interest to me is the fact that it is roughly two times the frequency. So I wonder if there is some kind of cancellation in the mic/room/speaker set up that is going on. (Factors can be important clues in dips or more infrequently, peaks.)
1)The bump from 1.5K to 3K is the first thing that needs to be addressed. With a parametric EQ in the controller.
2) Normally, with the (BMS) passive crossover, there is no noticeable distractions in the frequency response curve, at the crossover point. The time differences between the mids and highs will probably be more noticeable on a phase curve, but this is not what is causing the discrepancy in the crossover area.
3) Normally, the process for setting the relative volumes on the amps/processor would be to match the volume at the crossover point. This seems to be off by about 8 to 10 db. But, since he has started out this way, it now leaves an interesting possibility. Instead of adding a CD curve to the top end to get the very highs loud enough, Charly has the option of a PEQ cut in the 6K to 10K range to tame this from the other end (of the HF). Meaning, instead of adding a CD boost, he can have a cut to the lower end of the HF to get the same results.
4) Of most concern to me is the dip at 750hz. I say this because while not as bad as the lower dip, it comes in the middle of a band, and neither the horn or driver should have this much variation in it. And it would not seem to have as simple of an explanation as the lower dip. Of interest to me is the fact that it is roughly two times the frequency. So I wonder if there is some kind of cancellation in the mic/room/speaker set up that is going on. (Factors can be important clues in dips or more infrequently, peaks.)
Hi Art, I am a fan of crossing this driver over this low. You are right that this combination should work. I do think your suggestion to get around the 325hz dip is the right one. And a good troubleshooting route. I don't think that the caution is needed. I run the neo version of this all the time at 300hz in outdoor, live music situations, with one per side, so don't think there will be any problems in a playback situation inside, at what I would think would be excruciating levels. I think the caution would be with a sudden burst that would probably hurt his ears first, driver second. But this would be more of a consideration with live music/open mics than with playback. Maybe with the tone for the frequency response capture.
So, the right suggestion, but I don't think he needs as much caution as you do.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.