I learned how to use Akabak about a year ago. Not sure why I didn't think to check this before (Akabak makes it so quick and easy) but it looks like horn throat velocity can get pretty high.
First I checked the Labhorn at full power. Wiith it's ~2:1 compression ratio, throat velocity was up around 13m/s. (Obviously the Lab is an offset driver horn, so velocity at the throat is actually 0, max velocity happens a bit further down the horn). I only checked a couple of points near the throat so it's possible that it gets higher than 13m/s, I only spent a couple minutes on the Labhorn sim.
Next, I checked my last personal horn project. With it's ~5.5:1 compression ratio it's throat velocity is up over 25m/s at full power. (It's actually an offset driver as well, but for this I modelled it as end loaded to make it easy to see velocity right at the throat without having to search for the spot of highest velocity).
Usually we look at compression ratio only as far as force (potential driver destruction) but throat velocity looks like it might be a big issue too.
To make matters worse, usually horns with big compression ratios are built with very high aspect ratio throats (elongated rectangle as opposed to square), which is not necessarily the best shape to funnel high velocities through.
Anyone have any idea how much throat velocity might be safe before distortion goes way up? There's rules of thumb for ported boxes, is there one for horn throat velocity?
First I checked the Labhorn at full power. Wiith it's ~2:1 compression ratio, throat velocity was up around 13m/s. (Obviously the Lab is an offset driver horn, so velocity at the throat is actually 0, max velocity happens a bit further down the horn). I only checked a couple of points near the throat so it's possible that it gets higher than 13m/s, I only spent a couple minutes on the Labhorn sim.
Next, I checked my last personal horn project. With it's ~5.5:1 compression ratio it's throat velocity is up over 25m/s at full power. (It's actually an offset driver as well, but for this I modelled it as end loaded to make it easy to see velocity right at the throat without having to search for the spot of highest velocity).
Usually we look at compression ratio only as far as force (potential driver destruction) but throat velocity looks like it might be a big issue too.
To make matters worse, usually horns with big compression ratios are built with very high aspect ratio throats (elongated rectangle as opposed to square), which is not necessarily the best shape to funnel high velocities through.
Anyone have any idea how much throat velocity might be safe before distortion goes way up? There's rules of thumb for ported boxes, is there one for horn throat velocity?
Last edited:
My observation from distortion testing of various drivers in various horns with differing compression ratios leads me to believe the speaker construction and cone stiffness is more the determining factor for high distortion below Xmax than throat velocity.
Using my Keystone tapped horn using two Eminence Lab 12, or one Eminence 4015LF, or a BC18SW115 as an example, the 15, with the least compression ratio, and hence less throat velocity, had by far more distortion than the other two driver choices. It’s cone was not stiff enough for high power and low distortion.
I don’t think high throat velocity per se causes distortion, whether the speaker cone can take the pressure without deforming will determine distortion levels below Xmax.
Thanks, that's what I was hoping to hear.
BTW, what's the throat size of Keystone?
And do you think there's a line that just shouldn't be crossed, like maybe 30 m/s? 40? 50? I bet it's pretty easy to get up pretty high with a very high xmax driver and a horn with a very very small throat.
BTW, what's the throat size of Keystone?
And do you think there's a line that just shouldn't be crossed, like maybe 30 m/s? 40? 50? I bet it's pretty easy to get up pretty high with a very high xmax driver and a horn with a very very small throat.
Last edited:
In ported designs, too high a speed blows the air out of the port, chuff chuff, flap flap.
I have never really given m/s much thought in horn throat design.
The BC18SW115-4 loaded Keystone had more distortion at the same drive level as the BC18 in a ported cabinet, but also had 6 dB more output.
I'd suspect that reducing the drive level by 6 dB would result in less distortion from the Keystone than the ported cabinet.
At the narrow end, the Keystone throat is about 69 square inches, the compression ratio in the center of the expansion is 2.5/1 for an 18" IIRC.
Art Welter
Attachments
Based off 69 square inches with squared edges, the boundary layer thickening per velocity would start to detune the throat at ~30m/s and be completely turbulent at ~38m/s. (that's calculated with a very smooth surface, sheet PVC, if the wood is rough that number would head south a bit)
revboden,
My figure of 69 square inches at the narrow end of the throat was incorrect, it is actually 71.875 square inches there, 78.125 mid cone, and 84.375 at the bottom of the 18" speaker.
Could you please explain what you mean by "detune the throat", and what results you would expect from a "completely turbulent" throat.
Art
Hi Art,
It's basically about displacement thickness. As air moves over a surface the boundary layer gets thicker with velocity so the effective size of a port or horn shrinks in cross-section while the length stays the same. For a port this will tune the port lower because it has less area vs length. As the horn gets larger it would have a higher volume to surface ratio so it would be less and less effected as the sound moved down the length.
In a completely turbulent state the particle flow is chaotic and the effects would be very hard to predict. my guess: It would sound muddled, like sound trying to get through randomly changing densities of fill.
rev.
It's basically about displacement thickness. As air moves over a surface the boundary layer gets thicker with velocity so the effective size of a port or horn shrinks in cross-section while the length stays the same. For a port this will tune the port lower because it has less area vs length. As the horn gets larger it would have a higher volume to surface ratio so it would be less and less effected as the sound moved down the length.
In a completely turbulent state the particle flow is chaotic and the effects would be very hard to predict. my guess: It would sound muddled, like sound trying to get through randomly changing densities of fill.
rev.
I suck at explaining things this complex... I'll try to find a good description. This may take a while. I'll also try to come up with some simplified equations.
these have some of the basics.
Stokes boundary layer - Wikipedia, the free encyclopedia
Boundary layer - Wikipedia, the free encyclopedia
Boundary-layer thickness - Wikipedia, the free encyclopedia
Reynolds number - Wikipedia, the free encyclopedia
http://www.engineeringtoolbox.com/reynolds-number-d_237.html
these have some of the basics.
Stokes boundary layer - Wikipedia, the free encyclopedia
Boundary layer - Wikipedia, the free encyclopedia
Boundary-layer thickness - Wikipedia, the free encyclopedia
Reynolds number - Wikipedia, the free encyclopedia
http://www.engineeringtoolbox.com/reynolds-number-d_237.html
Attachments
Last edited:
If I'm reading him right, and if I'm correctly extrapolating from my Labsub simulations, I think that for your sub the answer is only academic (although still very interesting), since I don't think you'll ever see 30 m/s in your horn throat.
Do you have an Akabak sim of your sub and if so could you check to see how fast it gets? I'm guessing it won't exceed 20 m/s at most.
Last edited:
I think if I understand his numbers, they are optimistic. They're just from a flow consideration. Keep in mind you're trying to force all this into a horn through through some rectangular opening from a 3D cone. That seems very turbulent.
I wonder what gedlee would say about this, or some of the pro guys. Certainly a very very interesting issue to bring up.
I wonder what gedlee would say about this, or some of the pro guys. Certainly a very very interesting issue to bring up.
Chaotic effects are very hard to predict indeed.
The overall margin of level difference between the BC18SW125-4 BR and the TH is 5.92 dB, about the same as the sensitivity difference was in the magnitude response.
However, the BR lost LF relative to HF in the high power sine wave tests compared to the Magnitude Response tests, perhaps due to port turbulence, AKA port compression.
The B&C TH is the inverse, the upper 90 Hz area peak reduced at high power sine wave testing.
OK, port compression is well known, the loss of upper frequencies, is kind of “different”.
Your guess, "It would sound muddled" would seem to apply, but the Lab 12s, however, retained their more prominent 90-100 Hz peak when run at high power, the opposite of "muddled".
My listening to the Keystones has been limited to the power I have available, somewhere a bit over 1800 watts per speaker, I'd like to have a bit over double that to take advantage of the BC18SW115 potential.
With the power available I notice little change in sound quality from soft to full tilt boogie, other than amp clipping.
P.S. neither the BC18SW115-4 loaded BR or TH show any change in magnitude response from around 20V peak when raised to 85V peak pink noise signal.
Art
Last edited:
I did the modeling of the using Hornresp, but it was not indicative of the final result, as the Keystone shaped exit influenced the response quite a bit.
Art
Yes, but that would require turning on the Dell PC and updating Hornresp, I have been on the PC only once since the Keystone design.
And after all the iterations I went through, the design looked just like the back of an envelope drawing I made weeks before...
If someone hires me to design a new sub I'll crank up the PC again, until that happens (not holding my breath :^)) I'll let those that like plugging numbers do the sims.
Art
I've been thinking about this for the last 24 hours...
There are too many variables to consider for a common DIY'er to even bother with. I may be able to sneak a design into the physics lab supercomputer but I doubt it's worth it. I'd get one answer for one set of variables, then you take the box out in the real world and temperature variations and humidity would probably have more effect on the sound than what the computations would predict. I modeled variations of between 20m/s and 40m/s over wood using a friction coefficient of .5 (smooth) and the layer grew from .25" at 20m/s to 2.8" at 40m/s. But as we are considering a pressure wave moving through the already somewhat turbulent flow in a horn throat....pressure, viscosity, bla bla bla (skip to end
) I would predict variations on the scale of +- 3-7db. in the HF range of the horn.I think a good rule o thumb is above 30m/s below 70"^2 strange things happen.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.