Hornresp

I week or so ago I posted about the difference in port velocity numbers between HornResp and WinISD. I have since received my hot wire anemometer and was able to do a quick test on an enclosure I had sitting around.

The enclosure I tested is a home theater subwoofer using a JBL GTO1014 driver in 2.0 cu ft net tuned to 23 Hz. The port is 10.5" x 1.875" for a port area of 19.6875 sq in (127.02 sq cm). I measured port velocities from 13 to 35 Hz at 20 volts and 26.2 volts.

The anemometer I have is a Testo 405i. You can see more about it here: testo 405i Hot-Wire Anemometer Smart Probe | Portable devices - Operation via smartphone | Portable devices | Product Type | Testo, Inc

Here is the inputs to HornResp I used:
qrK16yk.png


Here is the peak port outlet velocity:
rcEJfiH.png


Here is the peak port velocity graph out of WinISD for the same enclosure:
BucREgS.png


Here are the results from the measurements:
vWP0aDb.png

The anemometer measure RMS velocity so to get the peak I multiplied the anemometer measurements by 1.41. I measured at both 20 volts and 26.2 volts but I only entered the values from WinISD and HornResp into the spreadsheet for 26.2 volts.

My measurements came in significantly below what both HornResp and WinISD predicted. I suspect my meter measures a little on the low side since its not optimized for bi-directional flow, it would be nice to know just how much lower its measuring, but I have no way of determining that.
 
I week or so ago I posted about the difference in port velocity numbers between HornResp and WinISD. I have since received my hot wire anemometer and was able to do a quick test on an enclosure I had sitting around.

The enclosure I tested is a home theater subwoofer using a JBL GTO1014 driver in 2.0 cu ft net tuned to 23 Hz. The port is 10.5" x 1.875" for a port area of 19.6875 sq in (127.02 sq cm). I measured port velocities from 13 to 35 Hz at 20 volts and 26.2 volts.

The anemometer I have is a Testo 405i. You can see more about it here: testo 405i Hot-Wire Anemometer Smart Probe | Portable devices - Operation via smartphone | Portable devices | Product Type | Testo, Inc

Here is the inputs to HornResp I used:
qrK16yk.png


Here is the peak port outlet velocity:
rcEJfiH.png


Here is the peak port velocity graph out of WinISD for the same enclosure:
BucREgS.png


Here are the results from the measurements:
vWP0aDb.png

The anemometer measure RMS velocity so to get the peak I multiplied the anemometer measurements by 1.41. I measured at both 20 volts and 26.2 volts but I only entered the values from WinISD and HornResp into the spreadsheet for 26.2 volts.

My measurements came in significantly below what both HornResp and WinISD predicted. I suspect my meter measures a little on the low side since its not optimized for bi-directional flow, it would be nice to know just how much lower its measuring, but I have no way of determining that.

Looking at your post I'd be inclined to say that the graphs are in pretty close agreement.

Try changing the size of your horizontal and vertical axis to match that of the Hornresp graph and I think that you will find almost perfect agreement.
 
Looking at your post I'd be inclined to say that the graphs are in pretty close agreement.

Try changing the size of your horizontal and vertical axis to match that of the Hornresp graph and I think that you will find almost perfect agreement.

I totally agree the shape of graphs out of WinISD and HornResp are very similar to each other. They also are very similar in shape to the port velocities I measured. Where I have concerns is the difference in the simulated values between WinISD, HornResp, and what I measured. For example, on 26.2 volts at 21 hz I measured 7.57 meters per second RMS. That works out to a peak velocity of 10.71 m/sec. WinISD predicts 15.17 m/sec and HornResp predicts 22.19. That's a pretty big difference. When designing enclosures I try to keep port velocities under a certain speed and its hard to know which numbers to trust.
 
I totally agree the shape of graphs out of WinISD and HornResp are very similar to each other. They also are very similar in shape to the port velocities I measured. Where I have concerns is the difference in the simulated values between WinISD, HornResp, and what I measured. For example, on 26.2 volts at 21 hz I measured 7.57 meters per second RMS. That works out to a peak velocity of 10.71 m/sec. WinISD predicts 15.17 m/sec and HornResp predicts 22.19. That's a pretty big difference. When designing enclosures I try to keep port velocities under a certain speed and its hard to know which numbers to trust.

I get stung by my phone again. It's a Note 8 and you think with a screen that large I would be able to follow what was what! Sorry. On the idiot box I get to see your point.

Perhaps there is an explanation. I'm pretty sure Bjorn did some pretty stringent comparisons when he was working on his Phd. And from what I can remember Hornresp was pretty darn close to what was measured.
 
I want to know where is my mistake in thew input ?
The rear chamber is not seen in the schematics.

Hi palesha,

You haven't made a mistake, the "rear chamber" (for a tapped horn, specified using the throat chamber parameters, as you have done) is indicated in the schematic diagram by a black dashed-line circle. Atc =1700 cm^2 so the chamber radius is 23.26 cm, which is 3.76 cm larger than the diaphragm radius of 19.50 cm, as shown in the attachment.

The only time that the chamber line will not be seen is when Atc = Sd, but moving the mousepointer over the red diaphragm circle will confirm (by way of the associated message) that the specified chamber is still actually there.

Kind regards,

David
 

Attachments

  • Attach_1.png
    Attach_1.png
    53.9 KB · Views: 58
Hi palesha,

You haven't made a mistake, the "rear chamber" (for a tapped horn, specified using the throat chamber parameters, as you have done) is indicated in the schematic diagram by a black dashed-line circle. Atc =1700 cm^2 so the chamber radius is 23.26 cm, which is 3.76 cm larger than the diaphragm radius of 19.50 cm, as shown in the attachment.

The only time that the chamber line will not be seen is when Atc = Sd, but moving the mousepointer over the red diaphragm circle will confirm (by way of the associated message) that the specified chamber is still actually there.

Kind regards,

David[/QUOTE

Thanks David Sir!
 
Hi thejessman,

The Hornresp particle velocity results are the same as those predicted by AkAbak, and by simulation software written by Dr Bjørn Kolbrek for his personal use.

Kind regards,

David

Thanks for the response. I'm not trying to rock the boat and please know I've very grateful for your contributions to the audio community. However just because Hornresp give the same numbers as AkAbak doesn't really provide proof that they are accurate. I guess what I'm looking for is if anyone else has attempted to measure actual port velocities and if so what their results are.

So far what I have to go on is my measurements have come in significantly lower than what Hornresp predicts. If the actual port velocity was as high as Hornresp predicted there should be some significant port chuffing, but there wasn't. What Hornresp predicts is also significantly higher than what BassBoxPro, as well as WinISD, have predicted. I'm sure you can understand where my skepticism comes from.
 
Thanks for the response. I'm not trying to rock the boat and please know I've very grateful for your contributions to the audio community. However just because Hornresp give the same numbers as AkAbak doesn't really provide proof that they are accurate. I guess what I'm looking for is if anyone else has attempted to measure actual port velocities and if so what their results are.

So far what I have to go on is my measurements have come in significantly lower than what Hornresp predicts. If the actual port velocity was as high as Hornresp predicted there should be some significant port chuffing, but there wasn't. What Hornresp predicts is also significantly higher than what BassBoxPro, as well as WinISD, have predicted. I'm sure you can understand where my skepticism comes from.

You might want to inquire about your test instrument. Is it calibrated? Is there a margin of possible error?

Have you tried different placements for measuring the air velocity from the port?

There are always different possibilities.

If you search within this thread I believe that you will find Bjorn's observations as to if the air velocity numbers are indeed verified.
 
You might want to inquire about your test instrument. Is it calibrated? Is there a margin of possible error?

Have you tried different placements for measuring the air velocity from the port?

There are always different possibilities.

If you search within this thread I believe that you will find Bjorn's observations as to if the air velocity numbers are indeed verified.

Its certainly possible my test instrument isn't giving me entirely accurate measurements. The manufacturer specs say it should be + or - 5%. Also I'm not using it exactly as intended since its in a rapidly changing airstream so there could be some error there as well.

I have tried different placements within the port and I do get about a 10% change in measurement depending on where the probe is placed.

I took your suggestion and searched this thread. I searched for the term "velocity" and got over 200 hits, but only one of those was by Bjorn and it didn't pertain to what I looking for. I would really like to find any info Bjorn has on matching up simulated particle velocity with real world measurements.
 
Thanks David Sir!

Hi palesha,

Not a problem :).

In my reply yesterday I inadvertently specified Ap as the input parameter, rather than Ap1 as you had done. The schematic diagram should have looked like the one attached, showing both the chamber and port, as specified by you. Also, for the sake of completeness I should have perhaps mentioned that the presence of the chamber and port can alternatively be confirmed by selecting the Schematic and Chamber options in the Loudspeaker Wizard, as shown in Attachment 2.

Kind regards,

David
 

Attachments

  • Attach_1.png
    Attach_1.png
    49.9 KB · Views: 209
  • Attach_2.png
    Attach_2.png
    45.9 KB · Views: 206
I took your suggestion and searched this thread. I searched for the term "velocity" and got over 200 hits, but only one of those was by Bjorn and it didn't pertain to what I looking for. I would really like to find any info Bjorn has on matching up simulated particle velocity with real world measurements.

I haven't measured port velocity. I did measure diaphragm displacement of a tapped horn once, at the time I was helping David develop the tapped horn model and we wanted to verify some results.

As for why the results are different, it may be that viscous losses and turbulence play a part. These factors are not included in my own SW at the moment. I suspect they are not included in Hornresp or AkAbak either, since the results match.

Are you able to find out if WinISD and BassBoxPro includes corrections for these effects, and if so, what models they use?

Would you be able to find out if your hot wire anemomenter mainly responds to the DC component of the flow, or if you can actually measure the rms or peak velocity of a sinusoidal bidirectional flow?
 
Just a quick status on the Semi-Inductance calculator spreadsheet...
I've just about got all of the comments/corrections/suggestions from the Beta testers rolled in. Will post spreadsheet here soon for Hornresp users to take for a spin.

All that is needed is the free-air impedance data. Currently it can import ASCII impedance files that are comma, space, or TAB delimited as well as ARTA/LIMP binary *.lim format. The *.zma files that Parts Express posts which include both TAB & spaces between columns also work fine. I'll include a step-by-step example for one of the Dayton drivers.

I guess I missed this. I suppose I'm too late to be a beta tester but I'm super excited to get my hands on a copy of this. I hope data-bass does an article on this.
 
Hi David,
I have a few questions again.

1. What means "acoustic path length" parameter? Is it distanse between axises of driver and port measured on front panel? Or is it depth between surfaces?

2. How to simulate cases when driver and port are in different surfaces? As shown by Martin J. King, the parameter "driver position ratio" provide different responce. Look please Fig. 7.1, 7.13, 7.15, 7.17 on this http://www.quarter-wave.com/Horns/Back_Horn.pdf

3. And about power responce again. I already asked this question: http://www.diyaudio.com/forums/subwoofers/119854-hornresp-784.html#post5180117 In sheets by Martin J. King summary SPL is calculated for driver and port both. This feature would be more usefull. Because SPL more familiar.
 
Its certainly possible my test instrument isn't giving me entirely accurate measurements. The manufacturer specs say it should be + or - 5%. Also I'm not using it exactly as intended since its in a rapidly changing airstream so there could be some error there as well.

I have tried different placements within the port and I do get about a 10% change in measurement depending on where the probe is placed.

I took your suggestion and searched this thread. I searched for the term "velocity" and got over 200 hits, but only one of those was by Bjorn and it didn't pertain to what I looking for. I would really like to find any info Bjorn has on matching up simulated particle velocity with real world measurements.

thejessman,

I've also spent some time looking at port air velocities, I design and build a lot of pro-sound boxes these days and compression/efficiency/driver cooling are a pretty critical parameter. I have a couple of comments on your experiment:

1. A hot wire anemometers are great for turbulent flows to tell you the average speed, but I've found units like the testo with the protective thingy (I can't come up with a better term) are pretty sensitive to probe orientation.

2. I too have that anemometer and was trying to use it to validate some port speeds on a 360L double 18" PA sub box. I was getting numbers a bit low, but not near as far off as you were.

3. I suspect (and David if I'm wrong, please correct me) that David assumes a round port...rectangular ports have higher losses and different end corrections I believe.

4. Your port has a high aspect ratio of ~5:1...the higher the aspect ratio the more divergent it will be from the velocities of a round port.

5. I believe the best way to do this is through a high SPL microphone not an anemometer....but it will have to withstand between 140 and 170dB SPL depending on velocity.
 
Ports are difficult... As you mention a round port behaves different than a rectangular one. It even makes a difference if the port is at the edge of a cabinet or in the middle of the baffle, it´s distance inside the box to the rear wall can become relevant if the distance is small and other effects can lead to great differences between simulation and reality.

Most of the times the port is something I tune after building the prototype.
 
I believe the best way to do this is through a high SPL microphone not an anemometer....but it will have to withstand between 140 and 170dB SPL depending on velocity.

Not necessarily. Just run the sim at a lower voltage and check to see if the SPL out the vent matches the sim. Note that the measurement will be influenced by the output from the driver, so it might only be valid at Fb, when the contribution from the driver is at a minimum. Hornresp assumes that the driver will behave linearly, and that might only be true at lower SPL levels anyway.