![]() |
Unity horn build |
Post Reply ![]() |
Page <1 6789> |
Author | |
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
old crossover left, new version right (designed with more accurate REW data). HPF shifted up to 1.5kHz and changed to second order rather than third (plus polarity reversal on tweeter). got rid of bulb and high-shelf bypass cap - rather implement these in LMS. series resistor and PTC moved to amp-side of filter inductor to protect whole circuit from over-powering. large cap is DC blocking and signal for both 6" and 1.4" now goes through it. same resistor network used as previously posted. ![]() |
|
![]() |
|
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
is coil too close to the tweeter? it is 3.3mH (though measures only 2.9mH). it's a neo tweeter. ![]() |
|
![]() |
|
odc04r ![]() Old Croc ![]() ![]() Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
![]() ![]() ![]() ![]() ![]() |
I wouldn't have thought that the coil's field would significantly affect the presumably much more intense field in the tweeters gap. What average current do you estimate the coil is seeing during typical use? The majority of a coil's magnetic field is inside it along the central axis, outside field strength is weak in comparison.
|
|
![]() |
|
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
without digging out old Akabak model (and old computer to run it) I am not sure. would the peak current through the coil be at the crossover frequency? and equal to the current through the driver? so about 1.5A max and much less at other frequencies.
Edited by snowflake - 12 August 2019 at 12:25pm |
|
![]() |
|
studio45 ![]() Old Croc ![]() ![]() Joined: 16 October 2007 Location: United Kingdom Status: Offline Points: 3855 |
![]() ![]() ![]() ![]() ![]() |
As I see it, with the two components in the orientation shown, the tweeter's external field lines will be parallel to the mounting board, while the coil's will be perpendicular. So they are not particularly well set up to couple.
There may be a very small modulation of the tweeter's magnetic field, and the coil will experience some mechanical forces as current passes through it - this will probably introduce a minor amount of distortion, but likely at a level far below other sources of distortion.
|
|
Studio45 - Box Builder Commotion Soundsystem -Mobile PA
|
|
![]() |
|
odc04r ![]() Old Croc ![]() ![]() Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
![]() ![]() ![]() ![]() ![]() |
Yup I agree with that assessment. I usually model xovers in Spice as well as Akabak if I'm trying to integrate them fully into a design. Spice is so much better for estimates of power dissipation as a function of frequency etc and I find helps when sanity checking. Peak current through the coil is going to depend on a few factors including musical content and the crossover design. Hard to say.
|
|
![]() |
|
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
measurements of old vs new crossover. on axis and 25 degrees off axis. old: green and blue, new: red and orange. smoothed 1/3 octave. ignore below 400Hz as measured in small room due to rain.
much improved. mid-range lobbing isn't too extreme and expected high-frequency fingering above 3kHz is acceptable. I was having to do quite a lot of corrective eq with the old crossover (obviously too much vocal). with the new crossover it sounds pretty nice with no eq at all. next will try and play around with splay angles. see below for response plot Edited by snowflake - 19 August 2019 at 6:31pm |
|
![]() |
|
odc04r ![]() Old Croc ![]() ![]() Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
![]() ![]() ![]() ![]() ![]() |
Yep, nicely damped the peak at 1KHz, much smoother roll off now and aintegrates well with the top. Hows' the phase looking? I'm still extremely keen to build a pair of Unitys but being realistic it's going to be years before I get to the final and finished items. But I can start with the mid/high portion of the design. I'm considering 3D printing the horn/mid assembly and then integrating that with a wooden cabinet with the low mids.
|
|
![]() |
|
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
if you can 3d print the tweeter and mid mountings it will save a lot of woodwork and allow you to play around with chamber volumes and port shapes that just weren't practical with a router and rail saw. PM me if you want the CAD drawings of what I built as a starting point.
|
|
![]() |
|
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
forgot to switch on the active filter - 3rd order LP and HP @350Hz. mid/top output inverted and delayed 33cm
![]() phase (old green, new red) ![]() group delay ![]() Edited by snowflake - 19 August 2019 at 2:34pm |
|
![]() |
|
odc04r ![]() Old Croc ![]() ![]() Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
![]() ![]() ![]() ![]() ![]() |
That was my thinking, good for prototyping. I was also wondering if extending the port length to the mid chambers such that they were time aligned to the HF diaphragm could be of any practical value. Cheers for the CAD offer, maybe sometime in the future. I've already spent a while trying to flatten 3D CAD models of Unity horns into sheets for cutting - it's surprisingly non trivial. Not a typical use of CAD but there are ways. The plan in the meantime is to upgrade the CNC to a 4x2 ft bed and then I can just run the sheets off and finish them with a table saw to give near perfect results with minimum effort. It's not strictly needed but I'm going to do it anyway so why not use it. No obvious discontinuities in the phase with either xover but the newer does appear to have a more linear slope up to 2k so hopefully that leads to better mid-tweeter integration |
|
![]() |
|
snowflake ![]() Old Croc ![]() Joined: 29 December 2004 Location: Bristol Status: Offline Points: 3086 |
![]() ![]() ![]() ![]() ![]() |
I now realise that using a DC blocking cap is a really bad idea because it removes all electrical damping at low frequencies. fortunately the HF and mid drivers are also protected by the cutoff frequency of the conical horn which increases the nearer to the throat you get, and by the rear chambers of the drivers. so with that removed the crossover is just a second order filter on the compression driver, an L-pad to match the levels and a protection fuse. Everything else is done by the horn geometry, front chambers, and tap sizes. In this case the midrange roll-off turned out very close to target without any electrical components, if not a series inductor can be used to modify the response of a FLH as explained in Kolbrek (though it is probably as quick to find the optimal value by trial and error than go through the theory). The acoustic LPF of the unity mid is one order higher than a FLH because of the cancellation of the reflection to the horn throat - so fourth order which matches the 2nd order acoustic and 2nd order electrical high pass filters applied to the compression driver. Edited by snowflake - 05 April 2021 at 2:42pm |
|
![]() |
Post Reply ![]() |
Page <1 6789> |
Tweet |
Forum Jump | Forum Permissions ![]() You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |