Jump to content
The Klipsch Audio Community

Recommended Posts

My bad, saw the "Austria" at the tail end of the previous post and your question about Oz; somehow missed you weren't addressing that post.

  • Haha 1

Share this post


Link to post
Share on other sites

Believe me, I have made the same kind of mistake. You made me double check what I saw...

Share this post


Link to post
Share on other sites

Guten Morgen.

Yes, i am from A. But speaking, i have this strange accent.


Gesendet von iPhone mit Tapatalk

Share this post


Link to post
Share on other sites

Georg,

I have an online acquaintance from Austria who uses the same music software I use, and he has online tutorials. He also has an accent, but that is all it is. His English is much, much better than my own Austrian German.

 

Welcome to the forums!

 

Bruce

Share this post


Link to post
Share on other sites

Thank you for the welcome.

Yes, we are only 8 million, but we are everywhere .


Gesendet von iPhone mit Tapatalk

  • Haha 1

Share this post


Link to post
Share on other sites

Dear all,

Another update on the KG journey.

 

In the long time since I've made a substantive update here I've been doing all the reading I can about horns, so compelling is the sound of these great speakers, to understand why this should be. This led me to the enormous and richly informative Econowave thread over on AudioKarma documenting the collaborative efforts of some real greats in the audio world, principally Zilch, who despite sadly passing some years ago, his endeavours are kept alive by the lively gang over there.

 

Over the two and a half years since the KG's came into my life, I've leant enough and purchased some proper measuring equipment (Omnimic and DATS), that now I'm able to empirically measure the KG's and be a bit more scientific about what I've done recently.

 

The Omnimic measurements I've been able to make show that in original, unmodified form:

  1. The KG's are hot on the top end, no secret there; and
  2. The 2 way crossover configuration with the woofers wired in parallel means that there is measurable lobbing, or destructive cancelation between the woofers that is entirely typical and expected.
     

Beginner updates quantified:

First I want to share is how the KG's measure with all of the mod's I've documented here so far, principal among these being the effect of the 4db L-pad modules.

 

The following frequency plots both show exactly the same data, that is the KG's measured at 1 meter with 5ms gating, on axis and all the way around to 90 degrees off-axis, in 10 degree increments. I have done this to ascertain both on and off axis horizontal response to get a feel for the totality of the output which is also called the sound power. 

 

871567522_KG5.5wCritesLpadandstuffing.png.a6c831ba1ddacfc12585d1e80b406406.png

 

In the graph above we can see that the on axis response is remarkably flat (other than the 1 - 2.5khz range) for what was essentially a stab in the dark effort with the 4db L-pad module. At this point I'm thinking my ears are great at picking a flat response. Nice one ears!

 

Taking the same curves we can easily display a polar map using the fabulous Omnimic software/microphone combo from Dayton Audio.

 

306550337_KGCritesLpadandStuffPolar.jpg.d6a46222578891ca55faa6069e477ed5.jpg

 

The main positive take away here for me is that the horn/waveguide of the KG's is constant directivity. This is a REALLY BIG DEAL! 😁 

 

At this point, if you are even vaguely interested in horn/waveguide loudspeakers (this is the KLIPSCH forum afterall, so believe me you are, but just may not know it yet! 😉) then you owe it to yourself to read Dr Geddes whitepaper on waveguides to understand why this polar graph is so important and how to read it!

 

Suffice to say once I saw this I knew why the KG's are so magic, but they can and should be better!  That dip at the crossover region is not good and is the subject of my attempt at improved crossover designs, called Stage 1 (a 2 way minimalist tweak) and Stage 2 (a 2.5 way complete makeover).

 

Unmodified KG 5.5 measurements:

Taking a step back first, if you have an unadulterated pair of KG's then this is the response of the system with the original crossovers simulated in XSim using the raw driver measurements in the KG enclosures.

 

1130789298_KGWoofersoriginalcrossover.png.833db8b6378b59d761f0f3a2458908b9.png

 

Yes, that treble is really HOT! The other important thing of note is the big woofer (green line) suck out at 1khz. That is occurring because this is a two way with two woofers playing the same frequencies. Not good.

 

The following graph shows the two woofers playing on their own (green and blue lines), no crossover measured in the KG enclosure at one metre with 5ms gating. The individual woofer responses look quite smooth. However the red line, with both woofers playing wired in parallel as they are in the standard crossover, (i.e. the green line in the previous graph) we can see the same big destructive suck out from roughly 1.5 - 2.2khz. 

 

1573775453_Woofercomparison.png.036d3de6c06ebab15d0befe7a6f5d0ee.png

 

That double bump after 2.2khz are the woofer cones loosing linearity and is called break up which I want to completely avoid if possible. 

 

Here is the original crossover schematic copied from here on the forum into the wonderful free XSim design software. Technically what we have here is a third order highpass on the tweeter and a first order lowpass on the woofer with a impedance correcting shunt or Zobel (C3 & R1) on the woofer:

 

1115925327_KGOrigcrossoverschematic.png.3c9708fc3d5c9a2e851ff19ec1ba0a08.png 


 

 

Crossover upgrade time:

So I would like to offer two crossover upgrade options to deal with these problems:

 

Stage 1 - a minimalist 2-way tweak: 

This crossover aims to do the minimum and alters the:

  • tweeter (S1) circuit a little (at L1) and adds an L-pad (R3 and R4); and
  • woofer (S2m) by adding an additional inductor (L3) and nudging the C3 and R1 values.

 

1264745852_KG5.5XSimnewcrossoverschematic.jpg.e21e5de88b2325df82cf69d9514a0d08.jpg

 

The simulated response is a big improvement (if I may say so myself), but the woofer (green line) breakup bump is only 5db down and affecting the system response (blue line):

 

1696775236_KG5.5XSimnewcrossovergraph.jpg.b0fed7ca8370ad117f30ffe4f007d650.jpg

 

And finally the system impedance:

 

752562667_KG5.5XSimnewcrossoverImp.jpg.437ca925b78d03024a6668b0fc1d8a95.jpg

 

 

Stage 2 - a complete 2.5-way redesign:

Here everything is a new value and this design adds a new additional inductor to the lower woofer (S3m) for the .5 in the 2.5. Personally I think this more complete redesign will reap improved off axis (polar) response which will reward the listener with a more consistent sound stage.  

 

1837099773_KGnew2.5waycrossoverschematic.png.5eed77797f4641e0c4bcada2af4dc6e1.png 

 

This is a slightly more complex design that completely removes the nasty breakup lump at 2.2khz and flattens the response through the crossover and overall.

71634410_KGnew2.5waycrossoversimgraph.png.41fbed8dac3d7ad2ca0d6c73c54fcec2.png

 

And the system impedance:

1568456526_KG5.5XSim2.5waycrossoverImp.jpg.f40b16ea43e25570af2c227fea7d537d.jpg

 

 

From here I need to build up these new designs and measure the KG's with them installed. This may take a while because I have three kids and a full time job but I'm on a mission so I'm excited to see how much improvement I can get from the old girls yet! 

 

Rest assured when I do I'll be sure to post up the results and would welcome any and all feedback. It's all a journey afterall.

 

1662239602_KGOmniMeasure.thumb.jpg.2497a9fc84b72fa8aaedcf1bacb77d67.jpg

 

  • Like 1

Share this post


Link to post
Share on other sites

Maz4bz is back and i can feeling the wind of change;-)

Have you seen my other posting regarding the plans i have to rebuild the crossovers of my little KG family incl 3x 2.5, 2x 3.5 and 2x 5.5? Wouldn‘t it be great to improve them even further by adding the improvements you described above!!?

Georg


Gesendet von iPhone mit Tapatalk

  • Like 1

Share this post


Link to post
Share on other sites

So you've flattened the response graph In the (original) crossover region with, in part, the addition of L3.   Looks good on paper but what undesirable sonic qualities are thus introduced with the woofers being now slightly out of phase with each other everywhere?

  • Like 1

Share this post


Link to post
Share on other sites
On 7/30/2019 at 4:01 AM, oldman4u said:

Maz4bz is back and i can feeling the wind of change;-)

Have you seen my other posting regarding the plans i have to rebuild the crossovers of my little KG family incl 3x 2.5, 2x 3.5 and 2x 5.5? Wouldn‘t it be great to improve them even further by adding the improvements you described above!!?

Georg


Gesendet von iPhone mit Tapatalk

 

Hi Georg,

Thanks for the tip off to your thread and I'm so glad you're having some fun with the KG's too.

 

Certainly my work on the 5.5's here is directly applicable to your 5.5's. Also the cabint bracing, acoustic stuffing and other minor tweeks will most likely work on the rest of the KG family too.

 

With regards to the crossover work, each speaker is a unique design unto itself and would require individual measurement to know what works. Therefore I would not think that the crossover changes I've proposed for the 5.5's will work with the rest of the KG family. If you have a look at Troels Graveson's site you will see there what he has to say about this matter. 

 

I tried to use an existing SEOS design but substituted in a different tweeter compression driver that purportedly was interchangeable because the original unit specified is no longer available. Suffice to say that did not work out as well as I'd hoped and a complete crossover redesign was required.

 

Notwithstanding, any changes you make are reversible so I say why not have a try and see if you like what you hear. You may learn something along the way!

 

On 7/30/2019 at 6:13 AM, glens said:

So you've flattened the response graph In the (original) crossover region with, in part, the addition of L3.   Looks good on paper but what undesirable sonic qualities are thus introduced with the woofers being now slightly out of phase with each other everywhere?

 

Hi Glens,

Because I have measured on axis with the tweeter I have designed the crossover to set the phase correctly in that direction .  In my case when I sit on my lounge my ears are at the same level as the tweeter so this is how I have measured and designed the crossover. If you have a play with XSim you can see there you to adjust the relative acoustic offset of the drivers which sets the phase and forward lobe.

 

What this means is that the response should be uniform off axis in the horizontal plane. However in the 2way configuration the vertical axis will have a very narrow region where the drivers are in phase because (as I understand the physics of the matter) the distance between the tweeter and lower woofer dictates that the forward lobe will be very narrow when that lower woofer is called to play up to the crossover frequency.  There is a great video by Wayne Parnam the designer of the famed Pi Speakers which demonstrates the concept of the vertical forward lobe and its importance to power response. Also Audioholics has done a thorough write up on the problems with the two woofer design many centre channel speakers have (same problem, different orientation) and how this affects response.

 

These nulls are in fact (if I'm not mistaken!) all phase related which are effectively dealt with via the 2.5 way approach that JBL and other manufacturers commonly use. 

Share this post


Link to post
Share on other sites

All that you said jibes well with the general understanding I've had for quite some time.  But that, particularly the xover to the tweeter, isn't what I was talking about, except as a point of reference.

 

So you've got two woofers covering the same ground and you want to instead fade one of them out before it gets up to the tweeter crossover range.  Simple, you insert an inductor for a 6 dB / octave low-pass filter before one of them.  Now that woofer will bow out of the picture before the other, which continues upward to meet the tweeter.  As was (also) done in the JBL you referenced.

 

Not a separate low-pass network for each woofer, but an extra coil for one of them.  That's what I was talking about.  That extra inductor in front of the one woofer.  The woofer which now has a frequency-dependant time-delay compared to the other.  That'll make a lobe...

Share this post


Link to post
Share on other sites
On ‎7‎/‎30‎/‎2019 at 9:18 PM, glens said:

That's what I was talking about.  That extra inductor in front of the one woofer.  The woofer which now has a frequency-dependant time-delay compared to the other.  That'll make a lobe...

 

Hi Glens,

Actually it wont affect the phase much at all, on the measured axis.  Have a look at my sim here with phase added. 

 

Original crossover response graph now with phase curve switched on:

361177181_KG5.5XSim2.5waycrossovergraphincphase.jpg.fbab99d483f5a903e8c68c044020f115.jpg

 

Now I can remove the effect of L3 (the .5 part) using the short function in Xsim (denoted by the thick line through L3) on the bottom woofer to bring this crossover back to a 2way:

 

319386736_KG5.5XSim2.5waycrossoverschematicincphaseindshorted.jpg.efbd71de2c22dcb86cf1eb787f4c53e8.jpg

 

Here's the new frequency response and phase:

930921713_KG5.5XSim2.5waycrossovergraphincphaseindshorted.jpg.5336d0151a8bfe13e03965003248e3c9.jpg

 

Overlaying these for a side by side makes the difference easier to see:

Green = 2way; and

Grey = 2.5way.

458838143_KG5.5XSim2.5waycrossovergraphoriginalincphaseindshorted.jpg.1b1404c357cf3990719e39c0f8409342.jpg

 

 

What I'm aiming to do here is described far better in the literature for the JBL Everest DD66000 where they describe the ".5" augmentation approach of that magnificent design on page 11:

 

"A second 1501AL operates in the bass frequency range from below 30Hz to around 150Hz, where it is rolled off at a gradual 6dB/octave. The first-order slope ensures proper amplitude and phase summing between the two woofers over their total operating range. Both woofers operate below 150Hz, but only one of them extends up to the 700Hz crossover point. This is done to achieve proper directivity control throughout the entire woofer operating range, while delivering powerful and extended low-frequency performance."

 

Now this graph looks familiar, also from page 11 the .5 is applied to LF1:

 

DD.jpg.c0c2fa179ede712b5d9192b8001f3982.jpg

 

On page 18 they say:

"The design intent is to use both woofers in the bass frequencies and slowly transition to a single woofer in the midrange. This technique allows a primary crossover point between just two drivers (i.e. LF2 and HF) and permits proper control of the directivity pattern of the system, while providing tremendous power and air movement capabilities at the lower frequencies. "

 

So how to roll off the KG's second woofer at 6db per octave? Add a single coil to create a first order (6db) transition out. See here the JBL service manual for the DD66000 which sets out the crossover design in complete detail. If its good enough for the Everest, it should be good enough for the KG's?

 

One of the many great things about Xsim is the crossover design file contains all of the measured driver data. You can pick up the latest release of Xsim here over at diyaudio.com and have a play with my crossover design .dxo file now attached.

 

Anyone think they can improve on this? Go on, I dare you to have a try! 

 

If you can improve it we will all benefit. :D

 

Cheers,

Maz.

2.5 Way Crossover.dxo

Share this post


Link to post
Share on other sites

Please try a simulation with completely separate (and appropriately different) low-pass filters with the same number of elements each, so both woofers have the same phase delays.

Share this post


Link to post
Share on other sites

Another update on the journey.

 

Stage 1 Crossover Upgrade - design, build and measurements.

 

I played with my crossover design for an improved 2way and made a few tweaks (new Xsim .dxo attached):

 

1661339408_2wayschematic.jpg.71c05c907a0b809aa112597a0c2493f6.jpg

 

Simulated response looks good, woofer (green) breakup is now 10db down:

 

188363156_2wayrespgraph.jpg.4d17bce9fe869c94dc7e40c7f3dc87f0.jpg

 

Impedance never drops below 4ohmns...

 

 

1542544163_2wayimpgraph.jpg.83e93f08ac87853de8bcda9c3268e264.jpg

 

For this modified 2way design I simply cobbled together some extra caps, inductors and changed the resistor. Here is the crossover.

 

The little green cap is added to a 15uF to make a 16uF (parallel caps sum values) and the original 2ohmn resistor is swapped out for a 1ohmn.....

 

233005639_2wayCrossover.thumb.jpg.7e5cea8ee014d646e0e24f3214f716d1.jpg

 

The little brown cap is soldered across the original 8.2uF (which is on the other side) to make it 9.1uf....

 

2092708347_2wayCrossoveraddedcap1.jpg.763f78aa2e60c0b12464eb202965142b.jpg

 

 

And the 1.1mH coil is soldered into the positive woofer lead hole (+W)....

 

506486870_2wayCrossoveraddedinductor.jpg.8c7157deb0f5fd9ef8ab0b9ab9f7adb5.jpg

 

whilst the woofer lead is simply soldered to the other end of the coil....

 

258023292_2wayCrossoveraddedinductor1.thumb.jpg.ce4850af48e6eaacc728e243b78bb330.jpg

 

On the other side of the crossover board I swapped in a 1mH inductor in place of the original....

 

1643376169_Inductorhighlight.jpg.2b4f0f72f99e61c8e2998f5581a78ad4.jpg

 

 

And finally, where would we be without an L-pad for good measure....

 

1448717655_2wayCrossoverLpad.jpg.376e25ae07c3ec33a4a1b704f6a33d35.jpg

 

Simple, quick and cheap, perhaps $10 per speaker in parts seen here, sky's the limit if you like high end crossover parts that wont measure any differently!  😉 

 

If I end up with this design in my KG's permanently I'd stick everything down with hot melt glue and cable tie the inductor to the board. Done!

 

The moment of truth. Here is an overlay of the modelled and measured response on axis. The measured Red curve is without smoothing, Xsim applies 24db/octave smoothing by default to the modelled driver (Blue) curves. I'm blown away by how accurate Xsim is:

 

522358387_2wayrespgraphsimvsmeasured.jpg.6483092697dbe66d16849b2360e40293.jpg

 

Here next is a comparison of original crossover (Red), original crossover with 4db L-pad in yellow and new crossover in green.  I'm calling this a win for the new Stage One 2way design. :D

 

1843754331_Original_originalwL-pad_New2way.png.d05453feab5137507659906926e6bffd.png

 

Here is the polar from 0-90 degrees off axis:

 

591732466_horizontalpolar.png.a87fae609423ae1c309f372e811ff6db.png

 

Here is the Dr Geddes style polar, not perfect but a big improvement over the original through the crossover region. Easily tweaked using the .dxo file attached:

 

531268454_horizontalpolarheat.png.cc073df8c753b3306b1f9b35e7e713cb.png

 

For context. Horizontal polars are measured by rotating this turntable:

 

485019213_Polarturntable.jpg.d459e3a4b512b6a2e3cf609461fedea0.jpg

 

 

But! Lets look now at the vertical off axis. Here is the vertical in the upward direction:

 

1029718053_Verticalpolarup.png.de3d7e5b0b24a6cdbd98f87570ed46a6.png

 

Here is the vertical in the downward direction:

 

818455217_Verticalpolardown.png.f3e6b3d29a2b3736577b9af3241b3600.png

 

So I believe this shows that the KG's in a 2way configuration are not too tolerant of being vertically off axis with the horn. Those big suck outs in the crossover show the effect of the forward vertical lobe. If you stand/sit in front of them off axis vertically, depending on how close you are, you may not be hearing these lovely ladies at their best. And I want them to be dressed in their best!


Vertical measurements inspired by Wayne Parham, note the compass taped to the side of the enclosure at the level of the horn:

 

1715462441_Polarmeasurements.jpg.7a5235b49ca1c6c9b19349002e350578.jpg

 

And for completeness here's the verticals of my tweaked 2way design and the original crossover, very similar off axis performance suggests the suck out is acoustic rather than electrical:

 

1454690986_Verticalpolarupold_new.png.b83f2e04131ae0f718b672657f1840fe.png

 

968533254_Verticalpolardownold_new.png.1a2cd33252fcf737c55b8685a75fcf3c.png

 

It will be interesting if a 2.5way improves the range of angles before the big cancelation suck outs occur - watch this space.  :emotion-22:

 

2way crossover.dxo

  • Like 2

Share this post


Link to post
Share on other sites

Another minor update. :P

 

Refining the Stage 2 crossover 2.5-way upgrade design.

 

Looking closely at the JBL DD66000 crossover schematic mentioned in a previous post opens another approach to the implementation of the 0.5 circuit leg of the Stage Two 2.5way design.

 

Tweaking my previous Xsim design (I'm calling the "2.5way V0.1") I have come up with a modified (V0.2) design that realises about 2db more efficiency from the woofers by connecting L3 directly to positive. C4 and R4 are optional as they have no affect on the frequency response, but can be added to smooth the impedance. This additional efficiency from the lower woofer is easily matched by adjusting the value of R3 which acts like a volume control on the the tweeter level. 

 

456321324_2.5wayV0.2.jpg.75f4962ca6165930d7b41b36d5cd3c80.jpg 

Here is a comparison of the difference between the V0.1 (blue) design and V0.2 (grey)....

 

252872090_2.5wayfreqrespgraphV0.1vsV0.2.jpg.2ee7431a03f5d24d6fa86f35fd40ffcc.jpg

 

The left-hand axis of this graph is zoomed in here to 2db increments.  This shows the other benefit of this V0.2 design is a slightly smoother response, with the KG's now effectively no more that +/-2db from 200hz to 13khz, or 18khz if we are able to forgive that 1khz wide notch . Also, I have not included phase above because removing all of the low pass section from the lower woofers signal path has zero effect on its acoustic phase. 

 

And here is the impedance graph. Electrical phase is included here, this shows 2.5way crossover V0.1 (blue) with the impedance peaks up around 30ohmns effectively addressed by the optional C4 and R4 on the lower woofer circuit (grey). Tune these to your taste or omit.

 

998617854_2.5wayimpgraphV0.1vsV0.2.jpg.40cad6fb236215e3372a024bb7da4e17.jpg

 

And for comparison here's the measured Stage One 2way design (red) vs the simmed Stage Two 2.5way V0.2 (blue) both without smoothing....

 

599873392_2waymeasuredvs2.5wayV0.2respgraphsim.jpg.aa7c1a60a92c5b1fe40a8d33f8fad038.jpg

 

Frankly, I'm really excited about this result and am very much looking forward to building up this design! New Xsim V0.2 .dxo attached.

 

2.5way crossover V0.2 (.5 to amp).dxo

Share this post


Link to post
Share on other sites
Posted (edited)

I have 4 pairs of kg5.5's which I swap in and out of my listening room. I have always felt they should sound better than they do. it is wonderful the work you are doing and I am watching intently as you progress and enjoying it very much. I have only tried 3 things over the years with these. one pair I added padding and some bracing and changed the diaphragms. I wasn't overly impressed, however it was a little better. I stopped at that point because I have no measuring gear except for the 1st version of wt3 from parts express.I should get serious one day and buy a microphone and learn to use it.seeing what you are able to do with Xsim is impressive. I haven't tried it before.Keep up the great work and thank you!

Edited by Pjstone
spelling error
  • Thanks 1

Share this post


Link to post
Share on other sites
8 hours ago, Maz4bz said:

Looking closely at the JBL DD66000 crossover schematic mentioned in a previous post opens another approach to the implementation of the 0.5 circuit leg of the Stage Two 2.5way design.

 

Tweaking my previous Xsim design ...

 

Well, you took it half way to where I asked you to go with it.

 

8 hours ago, Maz4bz said:

... realises about 2db more efficiency from the woofers by connecting L3 directly to positive.

 

Sure, you dropped two series inductors from the "sub" path.  While you were 1 order's worth of phase angle separation (displacement) between the woofers before, you're now 2 orders' worth.  What you gained in signal level getting through to the "sub" you lost a nibble back with the phase mismatch, depending to a larger extent on wherever your head (or mike) is at in relation to the box.

 

8 hours ago, Maz4bz said:

I have not included phase above because removing all of the low pass section from the lower woofers signal path has zero effect on its acoustic phase. 

 

That is utterly impossible.  The crossover elements cause electrical phase changes to the signal and since that's what's driving the drivers, their (acoustic) output has to reflect that fact.

 

I think I fetched a copy of that software the other day, to use in a virtual Windows box in order to see what's going on with it, and possibly to show you what I'm talking about, but haven't got to it yet.  If it's simulating no difference in acoustic phase with/without crossover elements I feel it's incomplete.

Share this post


Link to post
Share on other sites
9 hours ago, Pjstone said:

one day and buy a microphone and learn to use it

 

Hi Pjstone, thanks for the encouragement!

 

I have the Omnimic. I was going to get something cheaper until I watched this by 123Toid. I also got DATS V2. Once you have these two tools then XSim is the glue that enables it all to come together. The XSim thread has a lot of how-to info and Ryan at Impulse Audio, designer of many DIY Sound Group systems has this nice tutorial (and a more advanced Part 2) and many design walk throughs where he shows how he uses XSim which is very insightful for its operation and the subtler aspects of crossover design. 123TOID also has some great XSim tutorials on various facets of crossover design.

 

Stage 2 Crossover - 2.5way V0.1 (blue) vs V0.2 (grey) frequency and phase. Only showing the effect of changing the 0.5 config (I didn't bother to change the L-pad on the tweeter to bring it down)....

 

502606857_2.5wayfreqphasegraphV0.1vsV0.2.jpg.9b672aec72e6adf9ec71ef25a2fa0599.jpg

 

 

The author of Xsim is Bill Waslo. Same guy that developed Omnimic for Dayton Audio, SynergyCalc etc. The work he has done for us DIYérs is amazing. Getting to know his work a little better has fed my addiction. :P.  Would love to build myself some synergy horns (Danley style) one day for my ultimate HT build!

 

Share this post


Link to post
Share on other sites

Hello maz.

I believe sometimed decisions are made without having the best technical possible solution in mind - also at Klipsch.

The time you invest at the moment could help us all to rise our speakers to a new level and i am more than willing to give it a try. So i will wait with the rebuild of my Crossover until you finished your researches. In the meantime i will improve my 3D printing skills and do some additional research related to the maximum temperature the elements of a crossover can reach and how to avoid temperature related issues.

Thank you

Georg


Gesendet von iPhone mit Tapatalk

  • Thanks 1

Share this post


Link to post
Share on other sites

Gents, thanks for your encouragement. Really appreciated.

 

I have learnt most of what I know about audio from great forums like this one. If it weren't for all the generosity out there I'd still be tweaking blindly. I post this work as a way to give back. If I can save just one pair of KG's going into the garbage I'll feel rewarded. On that note please do let us all know if you try any of these tweaks. What I do here suits my ears and maybe not yours. Let us know either way.

 

Outdoor polar measurements - original crossovers Vs original + 4db L-pad mod Vs Stage 1 upgrade:

Just a small update - I did some full horizontal polars of the KG's with original crossovers and with Lpads outside last weekend at the same time as I measured the new Stage 1 two way design. These were done with the speakers hoisted up about 1.2 metres off the ground which helps give cleaner measurements than inside due to less room reflections.  

 

The Lpad tweaked measurements that were provided previously were done indoors, these should be a bit more accurate. 

 

I think these are good reference points for changes going forward.  🧐

 

Curves:

Original KG crossovers...

1771067509_0-90DegreesFreqPlot.png.18ce8a95ccb543f661b2003ef2eebc20.png

 

Original KG crossovers with 4db Lpads….

608873982_Lpad0-90DegreesFreqPlot.png.87539656742bdeba5aaac947b4ae7bed.png

 

Stage 1 Upgrade Crossover....

1248674388_Stage10-90FreqGraph.png.058e6b435273fe27923ceae5ab1a6083.png

 

Polars:

Original KG crossovers...

1295765511_0-90degreesPolar.jpg.5159d2dbc849decc2714f3c76d282d22.jpg

 

Original KG crossovers with 4db Lpads….

323085785_Lpad0-90degreesPolar.jpg.0489673388ae164a1ba7a1e7f8e99b91.jpg

 

Stage 1 Upgrade Crossover ....

692951012_Stage10-90Freqpolar.png.b5d5026bcff36af3ea31f74082118a4f.png

 

Looking at the 4bd Lpad polar it was interesting to reflect on my earlier comments in the beginners mods section of this thread about how the Lpads seemed to bring up the bass. Here you can literally see the heat (red) in the 2khz and below coming into the response picture.

 

The take away for me is to trust my ears, they seem like an accurate gauge of my preference for a flatter response!

 

Cheers. :emotion-22:

Share this post


Link to post
Share on other sites

Dear all,

Another update on my KG journey.

 

Stage 2 crossover upgrade:  V0.22 2.5-way design, build and measurements.

So getting to a workable 2.5way design has been a bit of a process. Measuring speakers is definitely a skill, with the results affected by a lot of variables. To get to this point I have measured each of the KG's individual drivers on no less than seven occasions, each with a little variation to try to ensure I was going to land on a reliable result.

 

My big learning here has been that being able to set the relative acoustic offset of the drivers is key to getting the forward lobe pointed in the direction that I actually listen. My first attempt at the 2.5way crossover resulted in a downward tilted lobe which would be a great design if you like to sit on the floor when listening! 😉  

 

The version presented here however does indeed present a forward lobe that is directly in front of the tweeter horn which is the level my ears are at when seated on my lounge. Here is the new 2.5way V0.22 crossover schematic:

 

361626537_2.5waycrossoverschematicV0_22.jpg.49205ed067c923c39b3efa6b33b1d06f.jpg

 

One thing to note here is that I dropped the series resistor on the tweeter circuit, it wasn't necessary. Now you can very easily vary the tweeter output by adjusting the R3. I recommend only going up in value here if you'd like some added sparkle on top. Going lower will put the impedance above 10khz below 4ohms which may be a challenge for some amps or receivers.

 

Here's the predicted response:

1899946601_2.5waycrossovergraphV0_22.jpg.f02162e52e712a9954cee0483f3955b8.jpg

 

 

And predicted impedance:

 

466495788_2.5waycrossoverimpgraphV0_22.jpg.b1019ceed59a25fe6131df046c2becc0.jpg

 

 

I also wanted to show how I go about building a crossover, as I'm prototyping and most definitely not a professional, I'm trying to keep costs down.

 

A while back I was lucky to snag a big box of brand new-old stock Energy Connoisseur and assorted other crossovers and terminal plates for a bargain on eBay from a local repairer. I deconstructed these for their LCR parts and ended up with a nice variety of components to enable me to start building up my XSim designs.....

 

580199282_Suppliesmore.jpg.6b7083a4125a43072c49d679b675caeb.jpg

 

 

Separating out my parts...

 

1953000225_Capsandresistors.jpg.cddc3a7c147311ae0e0fd83fa2640f72.jpg

 

 

Supplies.jpg.e96ddfe50f7d39fdf71cd78c3a0834e0.jpg

 

 

For this crossover I didn't have a 2mH inductor on hand so I had to unwind one of my larger value units. I chose a 3.4mH as I had 3 units of this value on hand and I'm aiming to have a pair of each value to enable me to build up prototypes for pairs of speakers so that I can listen to the results of each design in my lounge room.

 

options.jpg.91bef0bb654ce18749829a7d351c0a7d.jpg

 

An essential tool for anyone keen on crossovers is a multimeter with measurement for capacitance and inductance. Here measuring the inductor to be unwound....

 

original.thumb.jpg.1de236f4652c0a1a122824f3db0665eb.jpg

 

 

And now measuring the unwound inductance. I keep the excess wire as it's insulated and great to use as jumper wire between components on the crossover, I just burn off the insulation with my soldering iron by melting some solder onto the end of the wire...

 

unwound.thumb.jpg.878f00d992873b3b8336836d6f9fc30d.jpg

 

 

Some pics of the design build up, we're not fancy but we are cheap! 😁

 

Spread out here to keep the inductors away from each other, those cross braces added before are great landing pads for the new crossovers....

 

1250625007_Crossover2waysection.jpg.0985bab67c957acb5124a39b1c0650d8.jpg

 

A little bolt is great for tying in all the negative side of the circuit. This is an image of the first version of my 2.5way design (with inductors too close together) ....

 

19375022_negativepost.jpg.f8429eea21c484baf819e31dc9377ed7.jpg

 

The 0.5 section for the lower woofer installed in the enclosure, trying to keep those two big inductors away from each other...

 

177250610_2waysection.thumb.jpg.e9ba4005aff4f956f1cc8c1130f07f4d.jpg

 

 

 

Horizontal Measurements

All measurements are outside off the ground, the KG measured at 1 meter with 5ms gating, on axis and all the way around to 90 degrees off-axis, in 10 degree increments horizontal, no smoothing....

 

853445095_0-90FreqGraph.png.9d3a45273a0cfa1cc97c3ba575ec6a9b.png

 

 

Dr Geddes style polar plot of the same data. Controlled directivity! This is our best result yet and brings a big smile to my dial....  :D

 

Polar.jpg.3f7453b40f84398f5e9a739b5d0211e9.jpg

 

 

 

Now for verticals.

Upwards....

 

32699446_VerticalUpwardsFreqGraph.png.ae5ff4b4453a4d553680e65956b7c0a9.png

 

Downwards....

 

2083107481_VerticalDownwardsFreqGraph.png.63d64bd79018b544d0a12662585bfbc6.png

 

 

 

The moment of truth - would we bother with a 2.5way design when the original was a simpler/lower cost 2way? 

 

Let's allow the data to answer that question.  :emotion-14:

 

Upwards - 2.5way (black) @ 25 degrees compared to 2way (red) @ 20 degrees ….

 

899705929_20degrees2.5wayvs2way.png.a0298227bc8df15ccd6033b2addbbad8.png

 

 

Downwards 20 degrees - 2.5way (black) compared to 2way (red)….

 

122885579_-20degrees2.5wayvs2way.png.e60b69cbdae788f3a1db42f4c1328645.png

 

 

That ought to be enough to answer the 2.5way vs 2way question. Whichever way you go, both will be a big improvement.  :emotion-51:

 

The 2.5way will just sound better (I believe) in a wider range of settings, like when I'm standing up to iron my work shirts! It will also mean that the "sound power" or totality of direct and reflected sounds will be flatter. Dr Floyd O'Toole of Harman (JBL) investigates this in detail and concludes that this is the defining element of what trained listeners find most appealing in double blind tests. I've linked his paper here to this post as it is the basis for my efforts at finding the flattest response both on and off axis for the lovely KG's. 

 

This off axis performance data also kind of begs the question as to why did Klipsch choose the 2way knowing (I assume) the performance implications this would have? I suspect that this may have been a marketing decision to go with the higher SPL of the original crossover as a way to stand out on a crowded HIFI shop floor full of speakers from the opposition where first impressions count for everything. If I'm correct in this assumption then I understand their reasoning - we all want Klipsch to succeed! As I understand the psycho-acoustics we are drawn to the loudest sounds, initially. Perhaps we all appreciate the efficiency at first and then acclimate to thinking "these could/should be better" whatever that means to each of us, with our own unique tastes? 

 

So how does she sound? Notwithstanding a healthy dose of confirmation bias - AWESOME!

 

Yamaha AX-570 my new amp!

I've only built up one crossover (I'm still short another 2mH coil!) but last night I hooked her up to my new purchase, a lovely Yamaha AX-570, the bigger brother of my beloved AX-450. As lady luck would have it this old workhorse popped up on our local Gumtree for sale over the last weekend in less than stellar condition. She's missing a few knobs and her pre-main jumper bars. There were quite a few kids scampering around the house I picked her up from. She didn't emerge from that place unscathed!

 

I negotiated the price to $30 and got her home for a test run. 

 

AX-570a.jpg.233cc0daf452d8ba704f1f08e348bd36.jpg

 

So everything works great, a pair of uber short RCA leads plugged into the pre-main sockets and all functions are online. My Harmony remote operates all the remote functions too! More HIFI happiness! 

 

With a real 100 watts under the bonnet of the AX-570 I connected up my android phone to the CD input, engage Pure Direct and let rip on the upgraded KG with some FLAC encoded RUFUS Solace which has some THUNDERING bass.

 

And oh my, the KG's soaked it up and put out some super smooth sounds. We definitely lose some SPL with the 2.5way design compared to the 2way by taking out the lower woofers output early. That's the point after all. But we lose nothing in the bass department. Relatively we might say we gain in the bass department due to the reduction in output through the treble region necessary to balance out the reduced midrange output from just the upper woofer - if that makes sense. This means the tweeter is really loafing along now and I think results in lower overall distortion. I'll need to measure that and include in a future update. 🧐

 

In any case the old gal immediately sounded GREAT with this new crossover design, even though she was just plonked on the floor in my work shed. Huge bass, which I love about these speakers, but now with oh-so-smooth mids and highs. I'm ecstatic about the results. Can't wait to get the second crossover built up and the lovely KG's back into my main system. The journey continues.  :P

 

Cheers!  :emotion-22:

 

1688623871_Soundinggood.thumb.jpg.bd7bbac65406a6f7a25698a9d4887b36.jpg

 

 

  • Thanks 2

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...