several times before I caught on. After S1 and S2 you have the option of typing c, e or h for the flare you want. L12 should appear by the third box when you choose c(onical).boydon_lepasci said:
Hi Boydon and aznboi3644,
Just to clarify - L12 will only appear as the label for the "third box" when zero is entered into that box.
In the default record number 1, the L12 box is labelled "Con" and is alongside the F12 box. The L23 box is labelled "Exp" and is alongside the F23 box. The L34 box is just below the L23 box.
Hope this helps.
Kind regards,
David
Hi David,
Over on the Collaborative Tapped horn project thread, we are trying to understand the relationship between the driver throat and the area at S2.
In Danley's Unity Summation Patent (US 6,411,718 B1) he said that the driver throat area and it's cooresponding horn area should match:
"...while the (mid range) drivers have a throat area of 10 sq in and enter the horn where the horn area is 10 sq in."
Does Hornresp expect the hole in the baffle of a tapped horn to be equal to the area at S2?
Thanks in advance,
~Don
Over on the Collaborative Tapped horn project thread, we are trying to understand the relationship between the driver throat and the area at S2.
In Danley's Unity Summation Patent (US 6,411,718 B1) he said that the driver throat area and it's cooresponding horn area should match:
"...while the (mid range) drivers have a throat area of 10 sq in and enter the horn where the horn area is 10 sq in."
Does Hornresp expect the hole in the baffle of a tapped horn to be equal to the area at S2?
Thanks in advance,
~Don
Yesterday I finally got around to do some displacement measurements on my tapped horn prototype. I used an ACH-01 accelerometer tacked to where the dust cap meets the cone, which is about where the voice coil is attached. The accelerometer was hooked up to the PC sound card (Juli@) through an amplifier with a gain of 10, and the acceleration was measured with ARTA using a sine sweep at Fs=32khz to get good low frequency accuracy. The signal was fed to the horn through a power amp, giving 2.83Vrms at the terminals.
The impulse respone from ARTA was then exported into another program I have written for the purpose. It opens the impulse response, FFT-transforms it and multiplies the acceleration by sqrt(2)/(2*Pi*f)^2 to get the one way peak displacement, and corrects it for accelerometer sensitivity.
I had some problems uploading the screenshots of the results, so I put them on my webpage.
The Hornresp simulation
Measured displacement
As you can see, the Hornresp predictions are very close to reality.
Bjørn
The impulse respone from ARTA was then exported into another program I have written for the purpose. It opens the impulse response, FFT-transforms it and multiplies the acceleration by sqrt(2)/(2*Pi*f)^2 to get the one way peak displacement, and corrects it for accelerometer sensitivity.
I had some problems uploading the screenshots of the results, so I put them on my webpage.
The Hornresp simulation
Measured displacement
As you can see, the Hornresp predictions are very close to reality.
Bjørn
V
Vek
AndrewT said:
Vek is right.
This is a trick I learned from a fellow speaker builder. It has to do with the way FFT works. If you have, say, 1024 FFT bins and sample at 192kHz, you will get a resolution of 192000/1024 or 187.5Hz per bin. If you sample at 32kHz, you get a resolution of 31.25Hz per bin. So by using a lower sampling frequency, you get better low frequency resolution for the same FFT size.
For a tapped horn where we are mainly interested in the low frequency behavior, and with as high resolution as possible. Using a low sampling frequency, we don't have to throw away resolution in a frequency range that we are not interested in.
BTW, FFT size for the displacement measurement is about 32k.
I hope this answers your question.
Bjørn
As one of the first people to ever try (unsuccessfully) to imitate the DTS-20, I can definitely appreciate the struggle you went through to understand the true nature of a tapped horn (at least the simple ones that the diyers are making - Tom is on another level with his tap positions). Anyway, I think that there is/was a general assumption that 'tapping' the driver into the both sides of the line increased baseline spl and lowered excursion. The white papers basically explain it all, but I don't think many people really get what is happening - myself included to some point.
This is what I have come to understand - the tapped horn does not gain any baseline spl simply by virtue of being inside the line. It is other virtues of the tapped horn design that allow the way above average spl for a given (appropriate) driver.
If the driver is cleverly placed in a tapped horn, you can eliminate the 3rd harmonic (IIRC that's the one) - other than the cosmetic differences, this is the only thing that separates a tapped horn from a BIB style pipe horn - so that as long as you use only limited bandwidth, no stuffing is needed for a relatively smooth response over that limited bandwidth. (But not many diyer's are even using the 3rd harmonic trick to gain the last bit of bandwidth.)
This was a huge revelation to me, even though GM mentioned something very similar over a year before and I glossed over it. But I still could not come to terms with the published specs of the DTS-20 until...
A couple of months ago, TD explained the numbers. IIRC he said something to the effect of the numbers being "conservatively exaggerated" compared to the marketing numbers employed by other similar companies in the market. Then everything made sense.
The tapped horn is better (ymmv) than a BIB (in some aspects) because it does not need stuffing.
The tapped horn is better (ymmv) than a 6th order bandpass (in some aspects) because it completely avoids port compression.
A tapped horn is not louder or have less excursion simply because the driver is inside the line, it is louder and has less excursion because it is an intricately designed machine. It's a BIB style pipe horn without the rippled response (within it's passband), so no need for stuffing or corner loading, and it has no port compression. Nothing mystical going on here at all. I wish it didn't take me 3 years to grasp that, especially since the white paper says it all, and it was the first thing I read on the subject.
But I'm sure you already knew that, assuming it's even correct, which is never a guarantee coming from me.
The plot thins
Now we have two quite different THs that come very close to matching the Hornresp/ AkAbak predictions, I don't see much room for there being a generic problem with the simulations. On its own my measurement could have been a fluke, but that is not so likely with two independent verifications of the model.
Therefore if people do "step up to the podium" demonstrating a measured discrepancy, it is more probable that it is something quite specific in the TH design that causes the problem, such as wrong driver or horn parameters, or perhaps just an error in the measurement.
(For others apart from Ian: The other measurement was my post under the title "displacement measurement" in the collaborative tapped horn thread.)
Ken
iand said:
Now we have two quite different THs that come very close to matching the Hornresp/ AkAbak predictions, I don't see much room for there being a generic problem with the simulations. On its own my measurement could have been a fluke, but that is not so likely with two independent verifications of the model.
Therefore if people do "step up to the podium" demonstrating a measured discrepancy, it is more probable that it is something quite specific in the TH design that causes the problem, such as wrong driver or horn parameters, or perhaps just an error in the measurement.
(For others apart from Ian: The other measurement was my post under the title "displacement measurement" in the collaborative tapped horn thread.)
Ken
Kolbrek said:
Hi Bjørn,
This is fantastic news indeed! Thank you so much for taking the trouble to investigate something that has puzzled Ian ('iand') and myself for several months now. It is really great to finally learn that the Hornresp displacement predictions for a tapped horn are not so bad after all
.Feedback such as yours is absolutely invaluable to me - thanks again for your excellent work!
Kind regards,
David
David McBean said:
Hi Bjørn,
This is fantastic news indeed! Thank you so much for taking the trouble to investigate something that has puzzled Ian ('iand') and myself for several months now. It is really great to finally learn that the Hornresp displacement predictions for a tapped horn are not so bad after all
Feedback such as yours is absolutely invaluable to me - thanks again for your excellent work!
Kind regards,
David
What is interesting about Bjørn's measured results is that the peaks are very close to predictions (allowing for some small errors in tuning frequencies) but the minimum-displacement dips are not.
I can think of two possible reasons for this:
1. Losses in the tapped horn acoustic system lowering the Q at resonance, so cone travel is not reduced to almost zero.
-- David, can you comment on this?
2. Cone flexure so that even if the average travel over the cone is close to zero the driven centre of the cone (where the accelerometer is mounted) moves one way while the outer section flexes the other way under the opposing tapped horn air pressure.
-- Bjorn, can you make any measurements with different accelerometer positions to check if this is happening?
Cheers
Ian
P.S. Of course it's the maximum displacement that really matters, but it would still be useful to understand the reason for any differences between simulation and real life