I got the chance to have my speakers measured at BD-Design, needless to say, I grabbed that chance. Here are the resulting Frequency response graphs, on axis:

I chose a crossover point at around 2500Hz. Low enough for the woofer to avoid the troubles at 4k, and high enough for the tweeter to be safe. After a few hours of tweaking, designing and simulating in LspCad, this is the first filter and the filter I'm currently testing by listening to it:

As you can see, I did a baffle step correction for the woofer and did some serious padding on the tweeter, since it's almost 6db more efficient than the woofer. It gives me this on axis frequency response:

There's a high and narrow peak at around 600Hz. I suppose it's from standing waves inside the box (it was tested without dampening material). I stuffed the box lightly with wool, and a new measurement will have to show if this worked. The dip in the tweeter's response is there, but I don't know why. I did not correct for it in the crossover. The phase response is promising, around 2500Hz, the tweeters phase tracks the woofer's phase almost exactly. (visible in the indiviual phase response plots, not shown here)

I first listened to it in mono, to do a first quick check if there were big noticable mistakes.

The mono test revealed no big errors, so I purchased the parts to build a second crossover, so I could listen in stereo. I confiscated my sister's bed, to make a quick stereo setup, hooked up an amp and a cdplayer and started listening.

Now that I could listen in stereo, I immedately found two things wrong with the speakers, luckily they were quite positive.
First, the speakers sound really good. They shouldn't. They are my first own built stereo pair of speakers, containing my first self made crossover. They can't sound this good. I'm not really qualified to give an objective opinion, so I hope that in the near future, I can have some friends over to give their opinions on the speaker. Also, I want to have them measured again, with the crossover in place, so I can check if the peaks and dips in the response are gone, and if the FR looks anywhere near the simulation.
Second, they produce bass. The Linkwitz isn't even in place yet, and already some of my bass test tracks sound just fine. But these are 12L closed boxes, with an F3 at around 60Hz. This must be becouse of the small room they're placed in, or just my imagination playing tricks on me. I'm not saying that these speakers produce chest-thumping bass, just that the low notes in my favourite music can be heard just as good as on the larger speakers I have listened to. I will still finish the Linkwitz transform, but I'm not convinced anymore that I'm going to use it. (remember:these speakers are to be placed on a desk, with me 80cm in front of them. They're not meant to fill a livingroom with music)

They past this test with flying colours, so I placed them on the desk and will try them out for the next few weeks, to check if I keep liking them.

In the next few weeks I will keep updating this page, with updated crossovers, and new measurements. Some tests of my own, showed that the measurements on these pages are not fully correct. And I have also designed some new crossovers, which I plan to try out some time soon. Or maybe it's time to start saving up for LspCad pro!