G'day Mike
If the DEQX is as good as it looks, it could have saved me a lot of time over the years! You are right about a Unity / Synergy type horn being ideal for this type of automated eq. My frequency response is very uniform throughout the room so adjusting for one point wound not be too much of a compromise.
"basic Trigonometry" Mmm... you'll need a little more than that. I think Paul probably understands.
Cheers
William Cowan
Not that I have any time to work on anything in the near future, but watching from the sidelines is getting a bit old. So, I finally sat down with hornresp to do some modeling, and promptly ran into a question or two on OD horns.
When modeling multiple drivers (2S 2P in this case), are ALL parameters interpreted as per driver? I'm specifically wondering about the volumes - Vrc / Vtc and Ap1 / Lpt.
Below is a very quick hack-together model mostly taken from P.B's model with the Goldwood tweeters here, Audio Psychosis • View topic - Under Dash Horn Alternative but with the driver params adjusted for the chamberless RS52 numbers from HTGuide. This actually models OK, but I'm not sure the volumes are right and/or make any sense - this is mostly since he has Ap1 == Sd, which doesn't fit my understanding of what Ap1 was supposed to be.
When modeling multiple drivers (2S 2P in this case), are ALL parameters interpreted as per driver? I'm specifically wondering about the volumes - Vrc / Vtc and Ap1 / Lpt.
Below is a very quick hack-together model mostly taken from P.B's model with the Goldwood tweeters here, Audio Psychosis • View topic - Under Dash Horn Alternative but with the driver params adjusted for the chamberless RS52 numbers from HTGuide. This actually models OK, but I'm not sure the volumes are right and/or make any sense - this is mostly since he has Ap1 == Sd, which doesn't fit my understanding of what Ap1 was supposed to be.
Attachments
William,
Paul definitely understands how basic this "basic trig" actually is! If you want to work out a Synergy horn in terms of the physical form, getting pieces that actually fit together, then you either need to be very good at resolving "basic trig" in 3 dimensions, or be good at CAD. What you end up with is pieces that have funny angles. eg the main parts of a 60 x 60 have angles that certainly aren't 60 degrees. If you can think in three dimensions that will make sense. Also the edges must be cut to precise angles ... you end up with angles like 34.7 degrees. From this point onwards I'm aiming to get all the horns I built done vis CNC. I have a friend who has offered to help, and this is where my prototypes started - I wanted to be sure that firstly, it would work, and secondly, that I understood how the pieces fit together. I had to build two of them manually to understand them well enough to get it done via CNC. Working out how to get this done via CNC could prove interesting. If I can get all the pieces cut then it's a weekend project to assemble and get it working.
It's fun to build a few manually, but after a while the time really starts to add up.
Paul definitely understands how basic this "basic trig" actually is! If you want to work out a Synergy horn in terms of the physical form, getting pieces that actually fit together, then you either need to be very good at resolving "basic trig" in 3 dimensions, or be good at CAD. What you end up with is pieces that have funny angles. eg the main parts of a 60 x 60 have angles that certainly aren't 60 degrees. If you can think in three dimensions that will make sense. Also the edges must be cut to precise angles ... you end up with angles like 34.7 degrees. From this point onwards I'm aiming to get all the horns I built done vis CNC. I have a friend who has offered to help, and this is where my prototypes started - I wanted to be sure that firstly, it would work, and secondly, that I understood how the pieces fit together. I had to build two of them manually to understand them well enough to get it done via CNC. Working out how to get this done via CNC could prove interesting. If I can get all the pieces cut then it's a weekend project to assemble and get it working.
It's fun to build a few manually, but after a while the time really starts to add up.
Care for one of these, I've been keeping them close?
You certainly did well with yours, a very neat job especially with mitreing the corners as well.
I've been considering using butt joints for security during assembly.That still wont absolve it from the maths, but I like your idea of using CAD.
You certainly did well with yours, a very neat job especially with mitreing the corners as well.
I've been considering using butt joints for security during assembly.That still wont absolve it from the maths, but I like your idea of using CAD.
Attachments
Calculating compound intersections.
Don't know if this will help but there is a low cost software package (Excel based) covering a wide range of compoung angle calcs. I'm going the CAD route now but purchased this some time ago. Basic trig is still needed to complete practical designs but I was impressed with the wide range of situations covered. I don't have any links with this company.
Install Crown Molding Lika a Pro! Master Compound Miter Cuts and Angles!
Tom
Don't know if this will help but there is a low cost software package (Excel based) covering a wide range of compoung angle calcs. I'm going the CAD route now but purchased this some time ago. Basic trig is still needed to complete practical designs but I was impressed with the wide range of situations covered. I don't have any links with this company.
Install Crown Molding Lika a Pro! Master Compound Miter Cuts and Angles!
Tom
Hi guys,
I have been playing with my 'basic trigonometry' in Excel and decided that, as warned by William and Paul, it is a unique kind of mind that can do this sort of design by maths alone.
Fortunately, through the power of modern computing, trigonometry practice can be largely thrown out the window in favour of CAD modeling (this does not mean you don't need to study it kids). It would be great if we could now just 3D Print it but the CNC route seems more appealing at least for the next few years.
I have made a model of a 60x60 horn that sits somewhere between a SH50 and SM60 design at ~62 liters.
Features:
1. 60x60 degree horn opening to somewhere around 65 degrees at the mouth.
2. Approx. 40cm height for the first horn segment (~2200cm2 area at the mouth)
3. Approx. 10cm height for the second horn segment (~4400cm2 area at the mouth).
4. Wall thickness 18mm.
I have attached a couple of images showing the horn.
Hopefully something like this can be used with 4x Celestion 4" closed back drivers and some bass drivers in the 10" range (somewhere between the SH50 and SM60).
Next steps:
1. I am hoping that Paul is able to provide the HornResp driver parameters for the Celestion 4"
2. I need to do some reading on how to model the '6th order bandpass' type box for bass drivers. I believe with HornResp it needs to be modeled as a Tapped Horn or a Compound Horn. Has anyone played with (or worked out) this type of modeling?
Cheers,
Mike
I have been playing with my 'basic trigonometry' in Excel and decided that, as warned by William and Paul, it is a unique kind of mind that can do this sort of design by maths alone.
Fortunately, through the power of modern computing, trigonometry practice can be largely thrown out the window in favour of CAD modeling (this does not mean you don't need to study it kids). It would be great if we could now just 3D Print it but the CNC route seems more appealing at least for the next few years.
I have made a model of a 60x60 horn that sits somewhere between a SH50 and SM60 design at ~62 liters.
Features:
1. 60x60 degree horn opening to somewhere around 65 degrees at the mouth.
2. Approx. 40cm height for the first horn segment (~2200cm2 area at the mouth)
3. Approx. 10cm height for the second horn segment (~4400cm2 area at the mouth).
4. Wall thickness 18mm.
I have attached a couple of images showing the horn.
Hopefully something like this can be used with 4x Celestion 4" closed back drivers and some bass drivers in the 10" range (somewhere between the SH50 and SM60).
Next steps:
1. I am hoping that Paul is able to provide the HornResp driver parameters for the Celestion 4"
2. I need to do some reading on how to model the '6th order bandpass' type box for bass drivers. I believe with HornResp it needs to be modeled as a Tapped Horn or a Compound Horn. Has anyone played with (or worked out) this type of modeling?
Cheers,
Mike
Attachments
I'd need to double check my model, but that *sounds* correct. As I understand it, when you go from one woofer to multiple woofers in hornresp, you need to adjust the volume of the rear chamber (VRC) the volume of the throat chamber (VTC) and the volume of the port that's attached to the throat chamber.
In other words, just because you double the number of woofers, hornresp doesn't assume that you want to double the volume of the chamber in front of and behind the woofer. You have to do that yourself.
Also, on a subjective note, I have the Goldwood paper cone tweeter here, which is very similar to the Misco rdc3a that John H is using, and also the Dayton aluminum dome, which you are using.
IMHO, the Dayton sounds a heck of a lot cleaner. But the Goldwood is still quite promising for a few reasons:
#1 - incredibly affordable. $26 for four.
#2 - it's quite sturdy. I listened to it unbaffled with no xover and it didn't sound strained
#3 - While the Dayton is audibly cleaner, I have to think that four of the Goldwoods, mounted on a horn, would have an excursion that's next to nothing. And that keeps distortion down.
#4 - Probably my favorite thing about the Goldwood is that it's *incredibly* small. The depth on the Goldwood is significantly shallower than even the Tangband 2". I would be surprised if the depth is under 4cm.
#5 - The Dayton is clean, but it's quite a bulky driver, nearly as large as the Pyle 5". Also, the Dayton will require modification of the back chamber.
#6 - As noted in a previous post, the impedance peak of our driver goes DOWN as the ratio of air to MMS goes *down*. In other words, a driver with very low MMS works nice on a horn. The Goldwood driver has practically no surround, and I'll bet that keeps the MMS vanishingly low.
Anyways, jury is still out on these two. But the Goldwood is a heck of a lot better than I expected. Nice find by JLH, in regards to using cone tweeters. They have a lot of things going for them.
On a side note, I wish I could figure out an easy way to guesstimate the Thiele Small on the Goldwood. My Dayton WT3 won't do it because the impedance peak of the Goldwood is 'mixed up' with the rise caused by inductance.
I took a stab at calculating the T/S by hand using the formulas, but gave up because all of them require a sine wave generator and a multimeter.
I can post the curve later, the Goldwood's graph is very similar to the Misco RDC3A
Thanks Patrick,
I have found three threads with Unity/Synergy type Akabak scripts:
1. This AVS Forum thread shows a 3-way Synergy without crossover.
2. This diyAudio thread shows a 2-way Synergy with a crossover.
3. Your excellent diagram posted on Audio Psychosis a few weeks ago showing how to use the script in link 1.
Are these considered best practice or is there more I should be reading?
Cheers,
Mike
I have found three threads with Unity/Synergy type Akabak scripts:
1. This AVS Forum thread shows a 3-way Synergy without crossover.
2. This diyAudio thread shows a 2-way Synergy with a crossover.
3. Your excellent diagram posted on Audio Psychosis a few weeks ago showing how to use the script in link 1.
Are these considered best practice or is there more I should be reading?
Cheers,
Mike
Last edited:
There are a few Akabak scripts out there, each with their own advantages.
#1 - John H's script is the only one I'm aware of that has a proper crossover
#2 - The one from AVS forum includes the effect of the woofer ports. I can't recall if John's does.
#3 - My model is based on #2, but includes what I believe is a more accurate model of the ports. Basically, the model from #2 simulates the Synergy enclosure as if it's a simple vented box, and it's not. It's a tapered transmission line with an offset vent. The difference is fairly subtle, but it's there. My model is at forum.audiopsychosis.com
Keep in mind these sims are going to get time consuming - I'd guess it would take about 80 hours to develop a proper model, build a horn, and measure it.
#1 - John H's script is the only one I'm aware of that has a proper crossover
#2 - The one from AVS forum includes the effect of the woofer ports. I can't recall if John's does.
#3 - My model is based on #2, but includes what I believe is a more accurate model of the ports. Basically, the model from #2 simulates the Synergy enclosure as if it's a simple vented box, and it's not. It's a tapered transmission line with an offset vent. The difference is fairly subtle, but it's there. My model is at forum.audiopsychosis.com
Keep in mind these sims are going to get time consuming - I'd guess it would take about 80 hours to develop a proper model, build a horn, and measure it.
If you are modelling the woofer on its own, then it can be modelled in hornresp the same way as the mid. You simply make an allowance for the fact that it won't fully model the bottom end if you also intend to use a rear bandpass chamber as well. In that case you would probably have to model it in something else as a 6th order bandpass with some allowance for a transition between horn loading to then at the bottom end running it as a direct radiator. With a practical horn, it seems you get that transition somewhere around 200 Hz, give or take. In some of the Synergy horns you can see hints of that transition. In one of them I mentioned earlier, it seems that you see a lower bandpass peak, then a dip in between the horn loaded range and the tuning of the rear bandpass chamber.
So I guess it depends on whether you want to model everything fully before building or not.
Hi Patrick,
Your understanding is correct.
The Vrc and Vtc values entered on the input parameters screen are the values used by Hornresp.
Kind regards,
David
Final details for the Celestion 4" group buy should be out soon, I'm hoping next week. This won't be an open-ended group buy but we will have a fixed order to fill. So a case of first in, first served. We have enough interest to fill the order. Of course, if anyone pulls out, others may be able to get in as well. If you haven't expressed interest yet, send me a PM, there may still be a chance to get on board.
Regarding the driver parameters, there is a little guesswork involved because I can't measure VAS with sealed back drivers. However, when I get a chance I will supply a measurement of the driver on S1 along with the parameters I have been able to work out. You will then be able to overlay sims with measurements in REW. I have found that I can get a very close match at the top end, which is perhaps the more critical region.
Regarding the driver parameters, there is a little guesswork involved because I can't measure VAS with sealed back drivers. However, when I get a chance I will supply a measurement of the driver on S1 along with the parameters I have been able to work out. You will then be able to overlay sims with measurements in REW. I have found that I can get a very close match at the top end, which is perhaps the more critical region.
Hi,
Thanks Patrick for the information about Akabak - there is a program that is definitely in need of an interface redesign. I am glad you have named the variables sensibly as it makes it quite easy to read and modify.
The problem with the Akabak script and aiming for a 'flat acoustic power' response is that assumes you are going to use and build a passive crossover inside Akabak. I guess I will need to modify the script to disable all drivers except the bass drivers. My aim is to model the tapped horn/6th order bandpass behaviour of the bass drivers then the DEQX active crossover can manage the rest. HornResp is certainly easier.
I have been playing with my CAD model further with the aim of finding where these drivers will port into the horn. The model helps a lot as it is easier to visualise where all the problems you didn't think about will raise their head.
For example, in the images below I have added virtual drivers (ignore the driver model but the sizes are correct) to my model with the design for:
1x BMS 4555 1.5" Compression driver (38mm diameter)
4x Celestion TF0410MR 4" (120mm diameter including frame)
2x B&C 8PS21 8" (225mm diameter including frame)
I have found that I cannot fit a 10" driver as hoped and an 8" only just fits (as the frame is actually 8.8" according to their specifications). The problem is that as you can see in the pictures because a woofer is round and I am using a 60 degree horn the driver ports will not be close enough to the corners and may interfere with the HF response. It may not be an issue based on some things that were said earlier in this thread - but I would really like to have a play with a prototype as Mr Danley suggested. It almost appears that one of those horrible car audio type 6x9 inch drivers would be ideal as it would be wide enough to reach the edge of the horn while not taking too much vertical space.
Unfortunately all I can do is modelling at the moment as I am overseas in a country where I don't think CNC will arrive for at least a few decades.
Cheers,
Mike
Thanks Patrick for the information about Akabak - there is a program that is definitely in need of an interface redesign. I am glad you have named the variables sensibly as it makes it quite easy to read and modify.
The problem with the Akabak script and aiming for a 'flat acoustic power' response is that assumes you are going to use and build a passive crossover inside Akabak. I guess I will need to modify the script to disable all drivers except the bass drivers. My aim is to model the tapped horn/6th order bandpass behaviour of the bass drivers then the DEQX active crossover can manage the rest. HornResp is certainly easier.
I have been playing with my CAD model further with the aim of finding where these drivers will port into the horn. The model helps a lot as it is easier to visualise where all the problems you didn't think about will raise their head.
For example, in the images below I have added virtual drivers (ignore the driver model but the sizes are correct) to my model with the design for:
1x BMS 4555 1.5" Compression driver (38mm diameter)
4x Celestion TF0410MR 4" (120mm diameter including frame)
2x B&C 8PS21 8" (225mm diameter including frame)
I have found that I cannot fit a 10" driver as hoped and an 8" only just fits (as the frame is actually 8.8" according to their specifications). The problem is that as you can see in the pictures because a woofer is round and I am using a 60 degree horn the driver ports will not be close enough to the corners and may interfere with the HF response. It may not be an issue based on some things that were said earlier in this thread - but I would really like to have a play with a prototype as Mr Danley suggested. It almost appears that one of those horrible car audio type 6x9 inch drivers would be ideal as it would be wide enough to reach the edge of the horn while not taking too much vertical space.
Unfortunately all I can do is modelling at the moment as I am overseas in a country where I don't think CNC will arrive for at least a few decades.
Cheers,
Mike
Attachments
For my build, I used nothing but Google Sketchup and a band saw.
Once I had the horn dimmensioned accurately in Sketchup, I "pulled it apart" and used the programs measurement tools to create cut sheets with the dimensions and miter angles. I drew each piece out on MDF with a ruler and square and then free-handed them on the band saw. It wasn't too tricky. Here's some pics of construction and my Sketchup file.
Expect a major update on my build this weekend.
Once I had the horn dimmensioned accurately in Sketchup, I "pulled it apart" and used the programs measurement tools to create cut sheets with the dimensions and miter angles. I drew each piece out on MDF with a ruler and square and then free-handed them on the band saw. It wasn't too tricky. Here's some pics of construction and my Sketchup file.
Expect a major update on my build this weekend.
Attachments
Mike,
You can fit a bigger driver if you mount in on the corner, instead of the flat of the horn. And your entry port will be in the corner.
http://www.diyaudio.com/forums/mult...e-bandpass-mid-unity-horn-26.html#post2659780
You can also cut into the flared portion of the horn to create a little more room.
Sheldon
Last edited:
Amen.
Seems people have forgot that far more complicated horns were built with simple hand tools decades before computers were ever invented.
Do you have any measurement charts of your unity horns?