jc's tractrix bass bin

[jwc]

I just noticed something. Again...I haven't looked at this in awhile.

You asked about "filling the gaps". Now that was some wood glue initially....(I put on too much), then wood filler. Sanded several times....lots of hours...days. Then I use this white stuff...a primer. Forgot the name..3-4 coats. Sanded more between. Then painted black...sand in there too.

The PL Premium was for the construction process. Sorry I didn't make that clear.

jc

[Tractrix]

I'm curious about this design. It looks to me that you actually have two horns, connected mouth to throat. The first horn(s) are very short, one for each driver, and they are merged to the throat of the larger tractrix horn. Can I ask where you got this idea? How did you design the very short horns? Are they exponential or some other shape?

Thanks for posting all of the photos for the horn. I've been working with an Excel spreadsheet to design a similar horn, but with only one driver.

I've read that tractrix horns generally do not provide good bass response. You say that you designed these as full size because you wanted to be able to put them anywhere, but what if I was committed to placing them against a wall for 1/4 space and designed a 1/4 size tractrix horn. Would the bass response be better or worse than a 1/2 or full size horn?

I'm planning on using the 12PE32 driver since I need extension up to 500 Hz. How do you change the parameters for the horn calculation when you use 2 drivers like you have done?

[Marvel]

The eminence 3012HO or the 3012LF. If you were to XO the high end 400-500Hz.....you could use either. If you wanted to see if the bass bin would get out to ~800Hz....go with the HO version.

Those would be what I would try next.

jc

addendum. You know....I would have to calculate for sure....but I was thinking the cabinet volume may not be enough for the 3012HO to get down to 70Hz. I would need to calculate that......

I've been reading a bunch, but still can't understand how to calculate the back chamber volume for a given driver. Can you point me in some direction where I might be able to understand it? The Edgar articles haven't been of much help to me. You can PM me if you would like.

TIA

Bruce

[PrestonTom]

Bruce: The back volume can calculated from a simple formula that I can't find right now. It is in the Delgado & Klipsch (JAES, 2000) article and is referenced from some stuff by DB Keele. Both those can be found online.

[Tractrix]

There are couple of formulas on this page:

http://www.rocketsciencecanada.com/rocketsciencecanada/Sound/Horns/Info_Design.asp

[DrWho]

I personally use hornresp and adjust the rear volume until the lowest impedance peak flattens out in the simulation...which works well when you start deviating from standard rule of thumb horn size equations. Although I must confess that I've never tried comparing the equations to the hornresp model - I should probably do that.

On my next horn build, I'm going to use the Dayton WT3 woofer tester to measure the impedance and then adjust the rear volume until the impedance flattens.

[Marvel]

Thanks a bunch, everyone. Plenty of new reading material for me now... [Y]

[jwc]

I build a horn then see what my "constrictions' are for the rear volume. If the woofer chamber is in a folded horn...then there are limitations. This is what happens when you build a horn...then look for drivers to mate to it.

I usually go "backwards"....by calculating a driver when it "annuls" in a particualr design.

Fs x Qts (1 + VASliters / cabinet volumeliters) = annul frequency

jc

[jwc]

I'm curious about this design. It looks to me that you actually have two horns, connected mouth to throat. The first horn(s) are very short, one for each driver, and they are merged to the throat of the larger tractrix horn. Can I ask where you got this idea? How did you design the very short horns? Are they exponential or some other shape? Thanks for posting all of the photos for the horn. I've been working with an Excel spreadsheet to design a similar horn, but with only one driver. I've read that tractrix horns generally do not provide good bass response. You say that you designed these as full size because you wanted to be able to put them anywhere, but what if I was committed to placing them against a wall for 1/4 space and designed a 1/4 size tractrix horn. Would the bass response be better or worse than a 1/2 or full size horn? I'm planning on using the 12PE32 driver since I need extension up to 500 Hz. How do you change the parameters for the horn calculation when you use 2 drivers like you have done?

This would be very hard for me to describe in words....and lengthy. It you were sitting next to me with my drawing it would be easy.

It is one horn really. 90 sq in throat. You could think of two 45 sq in throats that fire into a intial short horn about 14cm...that have a tractrix flare expansion based on a 90 sq in throat, Fc 100 full space.

The explanation for the full space......that would be lengthy and a bunch of jibberish if I were to explain all that. I could have done 1/2 space and use a smaller Fc.....but can't remember why I talked myself out of that when I drew this up long time ago. One thing though........I like the idea of pulling the speakers away from the wall/corner.

jc

[Tractrix]

This would be very hard for me to describe in words....and lengthy. It you were sitting next to me with my drawing it would be easy.

It is one horn really. 90 sq in throat. You could think of two 45 sq in throats that fire into a intial short horn about 14cm...that have a tractrix flare expansion based on a 90 sq in throat, Fc 100 full space.

The explanation for the full space......that would be lengthy and a bunch of jibberish if I were to explain all that. I could have done 1/2 space and use a smaller Fc.....but can't remember why I talked myself out of that when I drew this up long time ago. One thing though........I like the idea of pulling the speakers away from the wall/corner.

jc

OK, thanks. What made you decide that the initial short horn should have an Fc of 100Hz? Is that a full size expansion? I thought this was an 80 Hz horn over all? I'm confused... sorry.

[Tractrix]

I personally use hornresp and adjust the rear volume until the lowest impedance peak flattens out in the simulation...which works well when you start deviating from standard rule of thumb horn size equations. Although I must confess that I've never tried comparing the equations to the hornresp model - I should probably do that.

On my next horn build, I'm going to use the Dayton WT3 woofer tester to measure the impedance and then adjust the rear volume until the impedance flattens.

I see from trying this out in Hornresp that if you adjust the rear volume to make the electrical impedance resonance point become flat, then the SPL response in the lower frequencies goes down. If I increase the chamber volume just a bit above that point, the bass response improves. I suppose there is a compromise in there? What if we used a passive RLC notch filter to cancel the impedance peak instead?

[chrapladm]

I am still wondering what the other horn designs you had mentioned earlier that you said where easy to build but maybe nutty or something.

I would love to still try and build a horn like this but honestly dont know if I have the skills for the build. But I do like to learn so who knows.

Did you ever try the 3012HO's?

[DrWho]

I personally use hornresp and adjust the rear volume until the lowest impedance peak flattens out in the simulation...which works well when you start deviating from standard rule of thumb horn size equations. Although I must confess that I've never tried comparing the equations to the hornresp model - I should probably do that.

On my next horn build, I'm going to use the Dayton WT3 woofer tester to measure the impedance and then adjust the rear volume until the impedance flattens.

I see from trying this out in Hornresp that if you adjust the rear volume to make the electrical impedance resonance point become flat, then the SPL response in the lower frequencies goes down. If I increase the chamber volume just a bit above that point, the bass response improves. I suppose there is a compromise in there? What if we used a passive RLC notch filter to cancel the impedance peak instead?

Well the impedance peak actually makes it easier for SS amps to drive the speaker (less current needed)....a zobel network would just dissipate extra power. I believe the goal of reactance annulling is to make sure the pressure in front of and behind the driver is the same - which in turn reduces distortions related to the flexing of the diaphragm.

Btw, we ended up leaving the rear volume on our horn a little big since the LF extension wasn't quite as good as the model predicted - I think because the wavefront coming out of the horn wasn't coupling well to the 1/2 space environment. In other words, it matched more closely the full space response, but that's also how it ended up being used in the application we put it into...

[Tractrix]

I have been told that the rear chamber calculation in Hornresp follows the Leach (or is it Keele?) formula, which is only appropriate for Hyp-Ex horns. It's not clear if it applies to a tractrix horn. Also, it seems that anything that affects the acoustical impedance of the horn also affects the ideal rear chamber size. Tractrix bass horns are known for not extending down to their cutoff frequency in real life. That's why they are mostly used in midrange and tweeter horns.

I wasn't referring to a Zobel network. I was mistaken about the notch filter.

[Tractrix]

Also, did you design your tractix from the Hornresp program or from a spreadsheet? I noticed a big difference between this spreadsheet:

http://www.volvotreter.de/dl-section.htm

He didn't respond to my request for clarification on the differences. It seems that the spreadsheet preserves the flare of the tractrix at the mouth when downsizing the horn, but Hornresp does not. It merely chops off the tractrix curve to make it smaller. I'm not sure which way is the proper one. I was very excited when I initially designed a 60 Hz tractrix using the spreadsheet because it wasn't very big. But, when I used Hornresp it was much larger, using the same driver. So, I switched to a 100 Hz Hyp-Ex horn, where both the spreadsheet and Hornresp agree. I just have to accept that I need subwoofers...

[seti]

I've heard several of the horns based on jc's formula and his horns. That is what I would use. They all sound great.

[jwc]

At some point....I'd like to call you. A lot of this stuff is heard to explain in a few paragraphs.

I use online calculators....purely for "calculators". Volvoteer's has a lot of other stuff built into it that I ignore. Another I've used is...

http://fullrangedriver.com/singledriver/tractrixcalc.html

Now.....I've compared these two to each other and they are close if not the same.

Hornresp.....I piddled with it for sometime and currently don't use it. I still draw out everything on paper. I used to do the exp or tractrix calculations by hand but came out very similar to the calculators.

I thought Hornresp was cool when I wanted to take a design/idea and figure out what kinda response change I would get for small changes in throat size or whether I'm gonna stick the design in a corner or not. The last three bass horns I designed and put together (not posted), I never touched the Hornresp program. This isn't a recommendation to you....just the way I do it.

jc

[Invidiosulus]

Having the mouth cut off is less correct, so in that instance Erik's spreadsheet is better as you will get less edge diffraction.

Or something like that...

[DrWho]

I wrote my own tractrix calculator in SciLab (free version of Matlab)....hornresp and volvotreer use different assumptions about the speed of sound, but once compensated for, all three simulations completely agree. They all use different input parameters, but once you know how to make sure you're trying to design the same thing, then they will all match. I stick to my SciLab code because I can directly import the final design into SketchUp, which I use for printing 1:1 scale stencils. This is vital for trying to get two curved surfaces to line up nicely at the corners.

For what it's worth, any of these tools are really only good for identifying the acoustic impedance and its expected impact on the performance. However, there is so much more going on in horns that you almost have to ignore the acoustic impedance - it's just something you go back to to make sure your final solution isn't screwing you royal style. Is the difference between 8dB and 10dB of gain from a horn really a big deal?

[Tractrix]

At some point....I'd like to call you. A lot of this stuff is heard to explain in a few paragraphs.

I use online calculators....purely for "calculators". Volvoteer's has a lot of other stuff built into it that I ignore. Another I've used is...

http://fullrangedriver.com/singledriver/tractrixcalc.html

Now.....I've compared these two to each other and they are close if not the same.

Hornresp.....I piddled with it for sometime and currently don't use it. I still draw out everything on paper. I used to do the exp or tractrix calculations by hand but came out very similar to the calculators.

I thought Hornresp was cool when I wanted to take a design/idea and figure out what kinda response change I would get for small changes in throat size or whether I'm gonna stick the design in a corner or not. The last three bass horns I designed and put together (not posted), I never touched the Hornresp program. This isn't a recommendation to you....just the way I do it.

jc

That would be great, after we deal with the hurricane coming through! I appreciate your assistance.

With Volvoteer's spreadsheet, for a 1/4 size 60 Hz tractrix horn using the B&C 12PE32 driver, I get:

throat area: 178.8cm2

mouth area: 6388 cm2

length: 67.1cm

For the single driver website you listed, I get:

(I put in the same throat area)

mouth area: 6145cm2

length: 158cm

for Hornresp, using their system designer, I get:

throat area: 168cm2

mouth area: 6539cm2

length: 161cm

So, I don't know what the heck is going on, something has to be wrong here.