What's the specs to the horn you built? Flare rate, mouth size (cm^2), throat size (cm^2), length (cm), is it exponential, tractrix, hyperbolic, etc. This will determine where you might want to tune the back chamber to. Let's start there, and then I can walk you through the tuning process. Lastly, be careful with where and how high an order high pass you use with that horn. You can really kill the sound with a bad filter choice.
Rgs, JLH
Hi
Well it's all in the thread there ( which is rambling , I admit ) but briefly :
122Hz ( don't ask ! ) Tractrix , truncated to 3500cm2 mouth area, 90cm2 throat ( ~ 4" dia effective, but square ) . Mouth is about 75cm x 46 cm so ~1.6:1 ratio.
Flat sides are 18mm birch ply, curved sides 2x 9mm 'bendy' MDF with filler in the grooves on each layer .
From hornresp ( see attached .txt file ) I'm looking at somewhere between 3.5 and 5L back-volume depending on whether I use the B&C or the Fane driver . If I do a proper 'nice' chamber , it will be when I get the Fane drivers - the B&C can have a 'lash-up' for now .
One question I just need the answer to , from someone who's built horns, is "what should it sound like without the back-chamber" , as it currently sounds nothing like the driver did on its own, out in the open - which was rather nice. All the 'snap' and fine upper-bass tone has gone, since I mounted it onto the horn .
MJ
Well it's all in the thread there ( which is rambling , I admit ) but briefly :
122Hz ( don't ask ! ) Tractrix , truncated to 3500cm2 mouth area, 90cm2 throat ( ~ 4" dia effective, but square ) . Mouth is about 75cm x 46 cm so ~1.6:1 ratio.
Flat sides are 18mm birch ply, curved sides 2x 9mm 'bendy' MDF with filler in the grooves on each layer .
From hornresp ( see attached .txt file ) I'm looking at somewhere between 3.5 and 5L back-volume depending on whether I use the B&C or the Fane driver . If I do a proper 'nice' chamber , it will be when I get the Fane drivers - the B&C can have a 'lash-up' for now .
One question I just need the answer to , from someone who's built horns, is "what should it sound like without the back-chamber" , as it currently sounds nothing like the driver did on its own, out in the open - which was rather nice. All the 'snap' and fine upper-bass tone has gone, since I mounted it onto the horn .
MJ
Last edited:
Hey MJ,
What you have built has close to no resemblances at all to a Tractrix. It is actually not far from pure exponential. The more you cut a TX from the mouth the more it equals an exponential flare.
This doesn´t mean your horn can´t be made sound good, though.
You must add a back chamber, it will sound bad whatever flare the horn has....
What you have built has close to no resemblances at all to a Tractrix. It is actually not far from pure exponential. The more you cut a TX from the mouth the more it equals an exponential flare.
This doesn´t mean your horn can´t be made sound good, though.
You must add a back chamber, it will sound bad whatever flare the horn has....
I'm well aware of what it is, Lars, I spend a while deciding what to make. It's a truncated tractrix , and as you say , probably not far from an exponential of the same mouth and throat size. I was trying to get something as good as possible , of a compromise mouth area, and length also a consideration, that was OK for my room. It happened that as well as I could simulate , the Tractrix curve gave significantly lower ripple than the exponential, on the sim at least, so there is a difference . It's something of a test-bed really, as I wanted to first build it as rectangular ( the easiest for me to build ) and do some evaluation .
Your comment about horns sounding bad without a back-chamber is very helpful . That's all I need to know for now - I will be starting work on a Mk.1 back-chamber this afternoon, built from scrap 3/4" birch ply . I'll allow for 5 litres max, and work down from there on tests .
MJ
Your comment about horns sounding bad without a back-chamber is very helpful . That's all I need to know for now - I will be starting work on a Mk.1 back-chamber this afternoon, built from scrap 3/4" birch ply . I'll allow for 5 litres max, and work down from there on tests .
MJ
Without a properly tuned back chamber it will be weak sounding and have no snap to it. Exactly what you are hearing now. That chamber will have to get down to a liter or less. Remove the woofer from the horn. Line the walls of the chamber with expanding foam. Only leave enough room to fit the woofer back in. If you over do it with the foam, you can just cut some of it out. This will get you close. Fine tuning requires a function generator, a series resistor and an AC volt meter. Use these to find the resonance of the horn. If the center frequency is below 120Hz, then add more foam. If it is above 125Hz, try adding a little bit of poly fill to drop it back down around your flare rate. Keep adjusting until you get between 120Hz and 125Hz. The sound will be a night and day difference.
Rgs, JLH
Hey MJ,
Just out of curiosity I did a flare diagram using radius. Green is 127Hz Tractrix, blue is your truncated Tx and red is 134Hz Hypex T=0,9.
The flares aren´t all true as I only used seven datapoints, but good enough.
I haven´t compared the Hypex against your, but it would be interesting to sim both in HR.
Just out of curiosity I did a flare diagram using radius. Green is 127Hz Tractrix, blue is your truncated Tx and red is 134Hz Hypex T=0,9.
The flares aren´t all true as I only used seven datapoints, but good enough.
I haven´t compared the Hypex against your, but it would be interesting to sim both in HR.
Good stuff !
Thanks for the profile plots Lars - there's actually more difference than I thought , beyond 55cm axial length . My own sims and results for Hyp vs Tra were on this page :
audio-talk :: View topic - Something stirs in the Undergrowth
I have also attached the results below in case you can't see them on our forum .
Bear in mind that the Hyperbolic for this sim was a T=0.6 , and the hyperbolic was about 10cm longer than the Tractrix - they both had the same end mouth size . They also both use JLH's measured Fane data .
John , couple of things -
1 Litre seems amazingly small - Hornresp is indicating somewhere between 3 and 5 litres depending on whether I use the B&C or the Fane - and 1 litre cripples the frequency response - why the big theory/vs/practice discrepancy ?
On the resonance measurement - typical suggested series-R value ? ... and, am I looking for a dip in the AC voltage across the R ?
Cheers
Mark
Thanks for the profile plots Lars - there's actually more difference than I thought , beyond 55cm axial length . My own sims and results for Hyp vs Tra were on this page :
audio-talk :: View topic - Something stirs in the Undergrowth
I have also attached the results below in case you can't see them on our forum .
Bear in mind that the Hyperbolic for this sim was a T=0.6 , and the hyperbolic was about 10cm longer than the Tractrix - they both had the same end mouth size . They also both use JLH's measured Fane data .
John , couple of things -
1 Litre seems amazingly small - Hornresp is indicating somewhere between 3 and 5 litres depending on whether I use the B&C or the Fane - and 1 litre cripples the frequency response - why the big theory/vs/practice discrepancy ?
On the resonance measurement - typical suggested series-R value ? ... and, am I looking for a dip in the AC voltage across the R ?
Cheers
Mark
Attachments
Mid Bass Driver/Horn Notes
Let me know your thoughts and drivers I can simulate.[/QUOTE]
Drivers
Check this one out, it is nearby and of good quality:
P-Audio, SD12-1000EL 12” LF Driver
http://www.paudiothailand.com/pdf/products/SD12-1000EL5.pdf
Australia
Cannon Sound and Light Pty Ltd.
Mr. Daniel Trevor
Email: admin@cannonsound.com
This is what I use, but they are further away
McCauley 6224 10” LF Driver
McCauley.com : Products: Components > 6224: Overview
Horns
If you haven’t already, study the following articles to start.
http://www.volvotreter.de/downloads/Edgar-Midrange-Horn.pdf
Midbass Horns
http://www.jblpro.com/catalog/support/getfile.aspx?doctype=3&docid=201
For an all-horn system consider using the Klipsch Jubilee as a design guide.
Without use of room corners, bass horns are contraindicated.
Here Jubilee build. Detailed construction drawings are available on the web as well
Hornloudspeaker Magazine: Comprehensive list of Klipsch plans, vintage/heritage line
DIY Klipsch Jubilee
Regards,
WHG
Let me know your thoughts and drivers I can simulate.[/QUOTE]
Drivers
Check this one out, it is nearby and of good quality:
P-Audio, SD12-1000EL 12” LF Driver
http://www.paudiothailand.com/pdf/products/SD12-1000EL5.pdf
Australia
Cannon Sound and Light Pty Ltd.
Mr. Daniel Trevor
Email: admin@cannonsound.com
This is what I use, but they are further away
McCauley 6224 10” LF Driver
McCauley.com : Products: Components > 6224: Overview
Horns
If you haven’t already, study the following articles to start.
http://www.volvotreter.de/downloads/Edgar-Midrange-Horn.pdf
Midbass Horns
http://www.jblpro.com/catalog/support/getfile.aspx?doctype=3&docid=201
For an all-horn system consider using the Klipsch Jubilee as a design guide.
Without use of room corners, bass horns are contraindicated.
Here Jubilee build. Detailed construction drawings are available on the web as well
Hornloudspeaker Magazine: Comprehensive list of Klipsch plans, vintage/heritage line
DIY Klipsch Jubilee
Regards,
WHG
Theory and real life don't always line up. That's why you have to build it to find out. The typical series resistance would be 1K. You want to raise the output impedance of your function generator so you'll know when the the resoance peaks. You measure across the resistor and adjust the frequency to the lowest voltage reading. You want this peak to occur around 120Hz to 125Hz for your horn. Keep adjusting the rear chamber volume until you get there.
Rgs, JLH
Thanks Allen and JLH
I've got the tools to get this optimised now , thanks for the advice .
I made a sturdy wooden box yesterday, about 5 litres internal, without the driver - it's just going to be clamped/taped into place over the driver with a gasket of some sort . Just need to do a bit of skimming of edges & cosmetic rounding-off . Should be possible to get some initial tests later this afternoon .
I can screw/glue various wood blocks and/or polystyrene in there to reduce the volume & see what happens .
I must say the horn sounded better last night for some reason . I got a 100Hz high-pass and a 800Hz low-pass in the amp , and put a brace across the mouth ( some thin brass tube ) . Got a bit more tone and slightly more dynamics. It does integrate well with the 288-H and Azura 425 above - that's one issue ticked-off the list !
Starting to feel this will work out now .
MJ
I've got the tools to get this optimised now , thanks for the advice .
I made a sturdy wooden box yesterday, about 5 litres internal, without the driver - it's just going to be clamped/taped into place over the driver with a gasket of some sort . Just need to do a bit of skimming of edges & cosmetic rounding-off . Should be possible to get some initial tests later this afternoon .
I can screw/glue various wood blocks and/or polystyrene in there to reduce the volume & see what happens .
I must say the horn sounded better last night for some reason . I got a 100Hz high-pass and a 800Hz low-pass in the amp , and put a brace across the mouth ( some thin brass tube ) . Got a bit more tone and slightly more dynamics. It does integrate well with the 288-H and Azura 425 above - that's one issue ticked-off the list !
Starting to feel this will work out now .
MJ
Hmmm....
Well, some good news and some bad news.
I got my back-box finished and strapped on to the speaker flange, seal not quite perfect, I'll work on that in the next 2-3 days . Playing the speaker, the definition and dynamics seem to have improved , tone also a little . There's still something 'nasty' in there though, putting my ears on edge from time to time. Piano sounds a bit 'cracked' . That was there before .
So, I currently have about 5.3 litres minus the driver volume , so maybe 4.5 litres ?
I tried the electrical resonance test as JLH described . Signal generator into the amp ; 1k resistor in series in the 'red' speaker wire . Voltmeter set on AC volts connected across the 1k resistor .
I wound the frequency up and down from 40Hz to 110Hz ( one range ) and 90Hz to whatever (eg. 500Hz ) on the other range .
Nothing, zilch, nada .. no sign of any dip in the voltage, even a small one . I went really slowly around the 80-150Hz area too .
All I could see was a very slight gradual rise in voltage as the frequency is increased .
I also tried 270R resistor, to see if it made any difference, but it didn't .
So, I'm a bit confused. Did I make a mistake ?
MJ
Well, some good news and some bad news.
I got my back-box finished and strapped on to the speaker flange, seal not quite perfect, I'll work on that in the next 2-3 days . Playing the speaker, the definition and dynamics seem to have improved , tone also a little . There's still something 'nasty' in there though, putting my ears on edge from time to time. Piano sounds a bit 'cracked' . That was there before .
So, I currently have about 5.3 litres minus the driver volume , so maybe 4.5 litres ?
I tried the electrical resonance test as JLH described . Signal generator into the amp ; 1k resistor in series in the 'red' speaker wire . Voltmeter set on AC volts connected across the 1k resistor .
I wound the frequency up and down from 40Hz to 110Hz ( one range ) and 90Hz to whatever (eg. 500Hz ) on the other range .
Nothing, zilch, nada .. no sign of any dip in the voltage, even a small one . I went really slowly around the 80-150Hz area too .
All I could see was a very slight gradual rise in voltage as the frequency is increased .
I also tried 270R resistor, to see if it made any difference, but it didn't .
So, I'm a bit confused. Did I make a mistake ?
MJ
OK, I decided to try measuring across the speaker instead of the resistor. This should , to my mind, yield a much more pronounced peak instead of a small dip , in the voltage around resonance.
I checked with the back-volume in place - still nothing much , although some small lumps & bumps from 150Hz to 250Hz .
I then took the back-volume off, and got a nice clear peak around 80Hz . Good ! - progress .
So, I reckon with the B&C driver, I may have gone past the ideal volume, too small, and have lost the main resonance entirely . Maybe the driver is taking up more of the volume than I anticipated . Hornresp certainly indicates if I get to 2 litres with the B&C driver on this horn, the resonance peak disappears up its own backside somewhere around 160-170Hz . ;o)
More soon ....
I checked with the back-volume in place - still nothing much , although some small lumps & bumps from 150Hz to 250Hz .
I then took the back-volume off, and got a nice clear peak around 80Hz . Good ! - progress .
So, I reckon with the B&C driver, I may have gone past the ideal volume, too small, and have lost the main resonance entirely . Maybe the driver is taking up more of the volume than I anticipated . Hornresp certainly indicates if I get to 2 litres with the B&C driver on this horn, the resonance peak disappears up its own backside somewhere around 160-170Hz . ;o)
More soon ....
It makes sense to measure voltage across the driver rather than the resistor. Have you been changing the rear volume in Hornresp and looking at the electrical impedance sim to gauge what to expect?
Out of curiosity, where does hornresp predict a rear volume? Are you using the volume that shows after using the wizard?
Out of curiosity, where does hornresp predict a rear volume? Are you using the volume that shows after using the wizard?
Hmm, based on the HR specs listed in another thread, the rear chamber should 'ring' around 200 Hz to make the horn 'ring' at 120 Hz, so simple sealed cab formulas and Prof. Leach's math calculates ~4 L net undamped rear chamber.
Using Margolis-Small's fiberglass insulation stuffing values calculates a ~0.8 L stuffed chamber, so falls in line with JLH's ~1 L foam experience since foam isn't as efficient, hence the seeming discrepancy between theory and reality that in this case is just an 'apples n' oranges' comparison.
Note that a 'not quite perfect' seal makes it ~aperiodically loaded, so will tend to measure like a too small net Vb, ergo until it's properly sealed the measurements aren't accurate enough for tuning adjustments.
Of course if you didn't build it per the HR specs, then these values are no good. I just noticed you mentioning a larger throat though....... This will increase the rear chamber net Vb, so maybe you did make it too small.............
GM
Using Margolis-Small's fiberglass insulation stuffing values calculates a ~0.8 L stuffed chamber, so falls in line with JLH's ~1 L foam experience since foam isn't as efficient, hence the seeming discrepancy between theory and reality that in this case is just an 'apples n' oranges' comparison.
Note that a 'not quite perfect' seal makes it ~aperiodically loaded, so will tend to measure like a too small net Vb, ergo until it's properly sealed the measurements aren't accurate enough for tuning adjustments.
Of course if you didn't build it per the HR specs, then these values are no good. I just noticed you mentioning a larger throat though....... This will increase the rear chamber net Vb, so maybe you did make it too small.............
GM
Last edited:
Hi GM, that's a considerable difference for a bit of fibreglass. I'm going to google Margolis-Small for my own purposes. I can appreciate a rear chamber being more functional without damping but I'd much rather put something in there.
I think JLH's foam is the hard setting expanding foam.
Just throwing out some ideas here but I'd also be worried about rectification with a just slightly leaky cabinet.
I think JLH's foam is the hard setting expanding foam.
Just throwing out some ideas here but I'd also be worried about rectification with a just slightly leaky cabinet.
1.5 lb/ft^3 of acoustic fiberglass insulation is pretty dense. Regardless, while all my rear chambers were either just lightly lined or ~aperiodically loaded using fine steel wool packing after Bell Labs/W.E., I've built sealed alignments based on it and it was accurate enough for me.
Hard foam? Hmm, dunno why there would be such a difference in net Vb then, especially since M/S's sealed, vented cab HP program/Prof. Leach's design routine is well proven.
Not sure what you mean by 'worried about rectification with a just slightly leaky cabinet.' I mean there are horn alignments where a well stuffed TL as rear chamber works best overall. Indeed, I much prefer it over reactance annulled since it flattens the impedance, making for a simple XO, but then I used low/no compression to get a ~flat on axis response and toe'd them in as required.
GM
Hard foam? Hmm, dunno why there would be such a difference in net Vb then, especially since M/S's sealed, vented cab HP program/Prof. Leach's design routine is well proven.
Not sure what you mean by 'worried about rectification with a just slightly leaky cabinet.' I mean there are horn alignments where a well stuffed TL as rear chamber works best overall. Indeed, I much prefer it over reactance annulled since it flattens the impedance, making for a simple XO, but then I used low/no compression to get a ~flat on axis response and toe'd them in as required.
GM
Keele´s equation also predictes a Vb in the ballpark of 4dm3.
Using Leach´s formulas for reactance annulling(Fo=Fl) and Hypex for B&C in a 134Hz, T=0,9 Hypex horn we unfortunatelly get unrealistic figures for Vb. It also give you an extreme compression ratio. When faking it a little using trial&error, Fo=Fl 350=134!!, all other figures like L, compression, At, Am match up with a Vb of just below 2dm3.
Sounds weird but I use a spreadsheet I made strictly from the AES-papers. David McBean also doublechecked it against HR and they correlate.
Using Leach´s formulas for reactance annulling(Fo=Fl) and Hypex for B&C in a 134Hz, T=0,9 Hypex horn we unfortunatelly get unrealistic figures for Vb. It also give you an extreme compression ratio. When faking it a little using trial&error, Fo=Fl 350=134!!, all other figures like L, compression, At, Am match up with a Vb of just below 2dm3.
Sounds weird but I use a spreadsheet I made strictly from the AES-papers. David McBean also doublechecked it against HR and they correlate.
OK, agreed, unless the driver's parameters dictates such loading, it won't 'be all it can be' without reactance annulling. If you're worried about Xmax though, then tune the horn to 0.707x the flare frequency using ~aperiodic loading with a higher F3 as the trade-off.
GM
GM
- Status
- Not open for further replies.