KEF Q500 new crossover

[wtnsmk]

At least finally I could measure with 4.2ms gate:

freq3 - Google Drive

The issue was the speaker was too close to the ground. With this measurement the speaker is from 120cm from the ground.

 

[Zvu]

Try to set this offsets just for test:

For tweeter all offsets are 0
For midrange there is only Z offset and it is -11mm
For woofer we have Y offset of -300mm and Z offset at -50mm

All of this is just for test purposes and aren't real offsets.


Offsets are entered with microphone as a reference point.

Now, since you haven't used ARTA, you have to derive Z offsets yourself. If microphone was pointed to the tweeter, all offsets for it are 0. Since you have concentric driver for mid-hi you only have to derive the Z offset for midrange and since i have LS50 which is almost the same as your concentric it is safe to say that Z offset for your midrange is around -11mm (as is with mine concentric driver). For woofer, based on your cabinet and driver placement sketch, woofer is 30cm below microphone position so we have X offset 0 and Y offset is -300mm.

To derive acoustic Z offsets for drivers you have to place the mic on axis to the tweeter at 1m distance. Don't touch the volume of the amplifier (or any other volume control in chain for that matter) and set start frequency of the sweep at 300Hz. Measure the tweeter and name it say Q500TW. Then you connect the midrange in parallel with the tweeter and make another sweep. Name it Q500MID.TW. Then you set start frequency of the sweep at 20Hz. Disconnect tweeter and measure just midrange. Name it Q500MID. Then connect the woofer in parallel to the mid and make another sweep, you get it, name it Q500MID.WF. Then disconnect the midrange and measure the woofer, name it Q500WF.

Now you have everything that is necessary to derive Z acoustic offsets yourself.

Set listening distance in VituixCAD at 1m. Make the tweeter in sim and upload the tweeter Q500TW and impedance. Then make a midrange in sim, and import Q500MID file. Import Q500MID.TW file as Reference (that's what it's called in LspCAD that i use) and compare it to the combined curve. Add Z offset in mm until they overlap. Then disconnect the tweeter and make a woofer in sim and upload Q500WF and impedance. Now upload as a Reference Q500MID.WF and compare it to combined curve. Set X offset at 300mm and add Z offset (usually negative value) until the curves overlap. Now connect the tweeter again and you can begin to simulate your crossover.

 

[wtnsmk]

Thanks I will do these measurements.

The weird thing, in Xsim, I'm already really close to the measured response (brown measured, black simulated), if I DON'T invert the mid speaker...:

 

[Zvu]

I'd, again, advise you to use Holmimpulse or Arta. I'm not sure that rew is good for modeling frequency response. It has few different types of smoothing that all give slightly different results. I'm not saying that's bad but it takes knowledge to know when and what to use. With Holm and Arta it is much simpler.

As for differences in Xsim, maybe it is optimized to work with rew measurements like Vituix is for ARTA.

 

[wtnsmk]

I'm doing the measurement in HOLMImpulse and I can't change the freq range. There is a start freq in device & signal, but if I put that up to 3-400Hz my measurement is just completely useless. I see some crazy curve. Also which sign I should use? Log or chirp? They look different in the low range

 

[wtnsmk]

Looks like this:

 

[Zvu]

Red Spade Audio: How to measure your crossover with Holm Impulse

 

[wtnsmk]

Red Spade Audio: How to measure your crossover with Holm Impulse

Thanks for the link, it was really useful, but unfortunately doesn't understand my questions.

These are my tweeter responses if I start the measurement from 400Hz:

 

[wtnsmk]

Could you please have a look on the file? Maybe it just a problem with my visualisation...

holmimpulse - Google Drive

Thank you very much!

 

[wtnsmk]

Ok I think I found the issue, I tried to change the start freq on the first (device & signal) tab rather than the highpass on the measurement tab... really confusing

Here are the measurements:

correct.zip: holmimpulse - Google Drive

you will see these measurements:
LF - woofer
LF MF - woofer with midrange (normal polarity) - 20-20
LF MF 2 - woofer with midrange (reversed polarity) - 20-20
MF - midrange - 20-20
MF HF - midrange + tweeter (reversed polarity) - 300-20
MF HF 2 - midrange + tweeter (normal polarity) - 300 - 20
HF - tweeter - 300-20

 

[wtnsmk]

I think I'm getting really close:

Yellow is the reference, with the real crossover.

 

[Zvu]

Your reference measurement must be made under same conditions and with same gate lenght as your frd files of you want to achieve complete overlap.

I'll download holm files as soon as i get back from work.

 

[wtnsmk]

Your reference measurement must be made under same conditions and with same gate lenght as your frd files of you want to achieve complete overlap.

I'll download holm files as soon as i get back from work.

Thanks man, I really appreciate it, I think my GF is even more if I can finally move back everything and she can start move around in the flat again haha....

 

[Zvu]

While measurements for midrange and tweeter are spot on:




Woofer and midrange for some reason aren't:



It looks like the volume was moved during the measurements because the woofer is about 15dB below midrange in last set of measurmeents, unlike the ones you've did yesterday.

 

[Zvu]

Just to see what a good job you've done with midrange and tweeter, look how crossover from factory cabinet looks now:



Not too shabby ha

 

[wtnsmk]

I haven't change the volume. The woofer is lower than on the full measuremnt because the back part wasn't screwed back so the air could go out. I have another measurement only from the woofer where I covered the hole. But when I measured the midrange and woofer together the hole wasn't covered either.

 

[wtnsmk]

All the volumes are on the same level now, what should I re-measure?

 

[Zvu]

Unfortunately, everything.

You really don't need to measure reversed polarity. So, using nomenclature you've established, you need to measure:

HF 300-20
MF HF 2 - midrange + tweeter (normal polarity) - 300 - 20
MF - midrange - 20-20
LF MF - woofer with midrange (normal polarity) - 20-20
LF - woofer

No holes this time, everything should look like it will when we finish.

 

[wtnsmk]

Unfortunately, everything.

You really don't need to measure reversed polarity. So, using nomenclature you've established, you need to measure:

HF 300-20
MF HF 2 - midrange + tweeter (normal polarity) - 300 - 20
MF - midrange - 20-20
LF MF - woofer with midrange (normal polarity) - 20-20
LF - woofer

No holes this time, everything should look like it will when we finish.

I understand, except the normal polarity. Are we only going to use that to measure distance? Or we're going to use it with normal pol?

Thanks!

 

[Zvu]

I just quoted you from last post on page 7.

Bottom line is that we don't need MF HF reversed polarity or LF MF reversed polarity when taking measurements used for deriving sim files.