HornResp, tapped pipes....

[Brian Steele]

Hi Everyone:

Have any of you tried using HornResp to model a "tapped pipe" (imagine a tapped horn with constant Sd), then built one to see how closely it matches the predicted output?

Here's one idea I've come up with:

Tapped pipe parameters:
S1=S2=S3=S4 = 290 cm^2
L12=L34 = 11 cm
L23 = 177 cm

Driver params:
Sd=129.84
Cms=1.02X10-3
Mmd=7m82
Re=3.55
BL=4.02
Rms=0.96
Le=0.19
Size = 6.5"

Build Approach:

1. split the pipe into four equal sections A, B, C, D, each with cross-section 14.5x20.

2. Arrange segments to form a rectangular box with x-section as follows:

AD
BC

3. Vent A into B, B into C, C into D with vent=290cm^2. Leave end of D open.

4. Mount driver on panel between A and D, midpoint of driver located 11 cm from exit.

Looks like it would be simple to do, the only drawback being a rather large (~2.6 cu.ft.) box for one 6.5" driver. Hornresp does suggest though that I'd be able to achieve 95dB/1W/1M with an F3 point around 33 Hz with this alignment into 2PI space...

[diy speaker guy]

Originally Posted by Brian Steele

Hi Everyone:

Have any of you tried using HornResp to model a "tapped pipe" (imagine a tapped horn with constant Sd), then built one to see how closely it matches the predicted output?

Not a tapped pipe, but I have heard several diy tapped horns and measured and compared at least two of them to the hornresp predictions. I don't have a reliable spl meter so can't comment on that, but the frequency response was very very close to predicted. In a couple of them the dip right before the two big peaks was a bit (and in one case a lot) bigger than predicted, other than that wysiwyg from design to measurement.

[Brian Steele]

Originally Posted by diy speaker guy

Not a tapped pipe, but I have heard several diy tapped horns and measured and compared at least two of them to the hornresp predictions. I don't have a reliable spl meter so can't comment on that, but the frequency response was very very close to predicted. In a couple of them the dip right before the two big peaks was a bit (and in one case a lot) bigger than predicted, other than that wysiwyg from design to measurement.

Thanks! My curiousity has gotten the better of me so I'm building this tapped pipe today, just to see if its output is anywhere near what HornResp suggests. HornResp suggests a clean and loud 40 Hz from this 6.5 driver in 2PI space. Should be interesting if I achieve this in reality.

[jonpike]

Don't know if you've seen this site, where he has a few different tapped horns of the Danley Sound Labs style... (follow white paper link) He also uses Hornrsp for calculating, maybe that can give you some guidelines...

You mentioning a 6 1/2" driver made me remember his TB W6-1139 horn projects.

http://www.volvotreter.de/th.htm

Interested to hear what you're using, and how it turns out. I don't know what to expect from a tapped pipe...

[Brian Steele]

One word - damnnnn!!!

More later - I'm still working on all the photographs and measurements...

[Brian Steele]

Just some pics detailing the construction of my "proof of concept" tapped pipe..


All the parts, cut and ready to assemble...



Step 1...



Step 2...



Step 3...



Step 4...



Step 5...



Step 6...



The last piece to attach. I opted to use screws and weather stripping to attach this part, rather than permanently glue it into place. This will allow me to easily access all sections of the pipe if I have to do so during the "tweaking" phase.



This shows the size of the 6.5" tapped pipe in comparison to my current HT subwoofer - an ACI SV10 in a 2.5 cu.ft. cabinet tuned to 21 Hz and driven by an AVA250 subwoofer amplifier. Yes, it's big for a subwoofer that contains only one 6.5" driver, but not that big. To the left is one of my "Mini-Me" sats - one of the few DIY full range projects I've done.

[Brian Steele]

Test Results so far.

First, frequency response of the tapped pipe driven by the subwoofer channel from my HT system (LP filter 100 Hz @12dB/octave). In green is the response predicted by HornResp (without filtering)




Second, frequency response of the tapped pipe driven by the subwoofer channel from my HT system (LP filter 100 Hz @12dB/octave). This time I've included a simple passive 12dB/octave filter (140uF cap, 4.5mH coil). I also restricted the mouth of the pipe a bit (using a baffle and the end of a 3" Madisound flared vent) to show the huge response change when this is done (F3 drops, but SPL also drops significantly). Again in green is the raw response predicted by HornResp (without filtering)




Finally, the impedance of the raw system (measured by WT3). In green is the impedance predicted by HornResp.

[Wolf]

Originally Posted by Brian Steele

Test Results so far.

First, frequency response of the tapped pipe driven by the subwoofer channel from my HT system (LP filter 100 Hz @12dB/octave). In green is the response predicted by HornResp (without filtering)




Second, frequency response of the tapped pipe driven by the subwoofer channel from my HT system (LP filter 100 Hz @12dB/octave). This time I've included a simple passive 12dB/octave filter (140uF cap, 4.5mH coil). I also restricted the mouth of the pipe a bit (using a baffle and the end of a 3" Madisound flared vent) to show the huge response change when this is done (F3 drops, but SPL also drops significantly). Again in green is the raw response predicted by HornResp (without filtering)




Finally, the impedance of the raw system (measured by WT3). In green is the impedance predicted by HornResp.

What 6.5" driver, and what input-voltage to the driver?
Looks interesting!
Wolf

[Brian Steele]

Originally Posted by jonpike

Don't know if you've seen this site

Yep, seen it! That's the site that increased my interest in tapped horns beyound the curiousity stage. It's while playing around with the horn parameters that I realised that tapped-pipes could be modelled with HornResp as well, and then I REALLY got interested.

Tapped pipes do have some advantages - including (1) they are much easier to construct (no strange angles to worry about) and (2) it's easier to fold them in two dimensions, as per my "proof of concept" example. I suspect that you probably give up a bit of efficiency by opting for a tapped pipe instead of a tapped horn, but I haven't done enough modelling to confirm that.

[Brian Steele]

Originally Posted by Wolf

What 6.5" driver, and what input-voltage to the driver?
Looks interesting!
Wolf

Would you believe that it's a cheap Pyramid 6.5" driver? It was the cheapest driver in their line when I purchased it several years ago. It used WT3-measured specs rather than the published ones in HornResp. Re=3.55 Ohms, Le=0.19mH, Qms=3.03, Qes=0.64, Qts=0.51, Vas=24.6 l, Fs=53.5 Hz. Xmax=??. I did this exercise just to find out if response predictions by HornResp for tapped pipes can be trusted.

Now, imagine what could be done with a decent modern 6.5" driver or even a good 8" one in a tapped pipe . I'm going to try modelling one of the Exodus Audio ones to see what the result might be like.

I didn't measure the test voltage when doing the FR checks, but HornResp suggests efficiency just above 95dB/1W/1M for the alignment.

[jonpike]

Very interesting... Got any design info for tapped pipes? Guess it's basically very similar to the tapped horn concept, but with a straighter line?

Yep, rather attention getting to see 112-115dB or more, out of a TB 6.5"!!
Nothing like a little acoustic impedance transformation to up efficiency.

I keep thinking what a TB W8 neo sub would do in one of these...

[Brian Steele]

Originally Posted by jonpike

Very interesting... Got any design info for tapped pipes?

Not much yet . One of the reasons for this exercise is to get some actual measurements and build experience which can then be used to develop and fine-tune a practical process for designing and constructing tapped pipes.

Originally Posted by jonpike

Guess it's basically very similar to the tapped horn concept, but with a straighter line?

Yup. You can use HornResp to model them - simply keep Sd constant from the start to the end of the "horn" when using the Tapped Horn Modelling Wizard, and play with the other parameters until you get FR and excursion response curves that you think are usable. When using the wizard, make sure to set the "Sd=variable" option, and check the schematic to ensure that the schematic looks like a large and long tube.

Originally Posted by jonpike

Yep, rather attention getting to see 112-115dB or more, out of a TB 6.5"!!

HornResp defaults to 0.5*PI modelling, which may inflate the numbers a bit. Setting it to 2*PI modelling should bring its predicted results more in line with the predicted results for direct-radiator systems from most box-modelling programs.

I'll post up a bit more of my observations later - it's time to get ready for work .

[diy speaker guy]

Nice experiment. How are you overlaying the hornresp predictions and the measurements on the same graph?

Just FYI in case you haven't noticed - when the driver taps at the very beginning and the very end of the pipe (or horn) there is no difference between a tapped pipe and a simple end loaded tl. Your design offsets the driver a few cm from either end but still the differences between having the driver inside (tapped pipe) and outside (close to end loaded transmission line) is going to be fairly negligible.

I wish I had left a removable panel for experimenting with stuffing but I didn't. Either way, here is my first tapped horn using a single tang band w6-1139si, well stuffed and measured outside (the mic is not calibrated). You might have to look close to see the driver tap location. I learned a lot from this and I'll be able to do much better next time.

[Brian Steele]

Originally Posted by diy speaker guy

Nice experiment. How are you overlaying the hornresp predictions and the measurements on the same graph?

I use an Excel workbook I put together especially for that purpose. I export the data from HornResp, WT3 and TrueRTA in txt format, and then cut-n-paste it into the relevant sheets in the workbook.

Originally Posted by diy speaker guy

Just FYI in case you haven't noticed - when the driver taps at the very beginning and the very end of the pipe (or horn) there is no difference between a tapped pipe and a simple end loaded tl. Your design offsets the driver a few cm from either end but still the differences between having the driver inside (tapped pipe) and outside (close to end loaded transmission line) is going to be fairly negligible.

I had thought about that. HornResp did indicate however that there would be several dBs of change at the upper end of the passband with only a few cm of movement of the driver from the mouth. Perhaps there's a relation between the amount of change and the length of the pipe?

Also, I don't think that there's any TL alignment that would give an F3 below 50 Hz for this driver, but I really haven't investigated that possibility. The impedance response also doesn't suggest TL behaviour.

You've got a point about the topology though - it does closely resemble a TL.

FWIW, I used close-miking for my FR tests.

[diy speaker guy]

I had thought about that. HornResp did indicate however that there would be several dBs of change at the upper end of the passband with only a few cm of movement of the driver from the mouth. Perhaps there's a relation between the amount of change and the length of the pipe?

It's pretty well documented that a tapped pipe (or horn) with both taps at the extreme ends is the same thing as an end loaded pipe (or horn), that's a fact and you can use hornresp to verify, frequency response, impedance, everything will be identical. The further the taps get from the ends the more difference there will be between the tapped pipe and a tl with driver offset at the same position and the more the impedance will differ.

You've got a point about the topology though - it does closely resemble a TL.

So does the one I posted - a conical horn is just a tl with a positive taper.

[Brian Steele]

One thing that sticks out from the measurements is that I may have gotten the pipe length wrong. HornResp predicts an impedance minimum around 41.5 Hz for a 199 cm pipe, but my tapped-pipe sub seems to have this minimum at 48 Hz, suggesting that the actual pipe length is actually around 179 cm - 20 cm too short. That means that each "section" of the pipe in the enclosure may be short by 5 cm.

Can someone a bit more familiar with horn/TL folding check my assumptions please? I basically took the 199cm pipe, divided it by four, laid them side by side and linked them as shown in the images above. Box size worked out to approx 51 cm x 32.75 cm x 41 cm. The cross-section of the pipe is approx 15.3 cm x 19.7 cm, and it remains constant through the length of the pipe.

[Brian Steele]

FWIW, the parameters of the Dayton Series II 8" driver suggest that it could be a drop-in replacement for my cheap Pyramid (or it would have been, if I'd chosen to keep one dimension of the pipe's cross section wide enough to support its use!). HornResp suggests 108~109dB @ 40W into half-space all the way down to 40 Hz without exceeding the driver's rated Xmax - significantly greater output than my poor cheap Pyramid driver is capable of in the same tapped pipe.

[Brian Steele]

...and yes, the pipe is too short. Looks like the chosen cross-section affects the effective length when you start to fold it. The box I built is 3" too shallow. Now I'm wondering whether to (1) knock the rear off and add 3" to the depth, (2) add a pipe to the front to make up the difference, (3) scrap the box entirely or (4) leave it as it is, as it's just a "proof of concept" design.

[fwater]

Originally Posted by Brian Steele

...and yes, the pipe is too short. Looks like the chosen cross-section affects the effective length when you start to fold it. The box I built is 3" too shallow. Now I'm wondering whether to (1) knock the rear off and add 3" to the depth, (2) add a pipe to the front to make up the difference, (3) scrap the box entirely or (4) leave it as it is, as it's just a "proof of concept" design.

I say option 4. Employ it somehow and build the next one better if you need a sub in another location. Of course, adding length to the end of it will put the driver farther up the line as someone suggseted earlier. So option 2 might be a little more fun...

I sent you a PM not long ago inquiring if you might help crunch some numbers for an enclosure similar to yours but employing some cheap 8 inchers I have laying around. Do you have the time or desire to help?

[Brian Steele]

Originally Posted by fwater

I say option 4. Employ it somehow and build the next one better if you need a sub in another location. Of course, adding length to the end of it will put the driver farther up the line as someone suggseted earlier. So option 2 might be a little more fun...

According to HornResp, moving the driver further away from the throat is going to adversely affect the response in the middle of the passband. I forgot one option - adjust the mouth dimensions. I played around a bit the last time and it affected the response significantly. I might be able to flatten the response (at the expense of some efficiency) that way.

Originally Posted by fwater

I sent you a PM not long ago inquiring if you might help crunch some numbers for an enclosure similar to yours but employing some cheap 8 inchers I have laying around. Do you have the time or desire to help?

Saw it and responded . You'll need to measure the parameters of those drivers first before we can proceed.