Shading/filtering on a 5 coil planar tweeter (group delay ?)

[WrineX]

Hi there, i have build a small planar driver 5.8cm wide x 22cm high, with 5 coils on them the middle being the tweeter, and flanked above and under by the same coils in parallel (next time io do them in series) and above and under those 2 are another 2 coils also in parallel.
i used caps in parallel with the coils i want to have a roll off, so it adds to the lower end but does not hinder the dispersion of the high frequencies to much.
here a schematic of the thing made up in Xsim. and how the roll off looks like and the group delay. (since i wonder if i need to concern myself about that)

They are crossed at the moment around 600 hz (should be higher but at my volume it works fine)
it greatly increased vertical dispersion, but to my ear it sounds great but maybe a bit to spacious >? i have trouble getting a rock solid phantom center in my room for ages, so i am not sure if i am hearing now a problem with phase because of this setup ? or delay ? or just my room. (normally i have more lines source like tweeters, so copuld also be this more point source style gives much more room troubles) does anyone know if the group delay as shown above is problematic ? i did measure this contraption vs a bare panel without any filtering and they both measure in rew gated almost exactly the same group delay even without crossover below 800 hz or so it rises. (my back wall starting to play up >?)BY the way how can the tweeter be a negative delay ?.....

here a rather badly taken picture of one of the shitty looking things, its just an illustration 🙂 by now the bass drivers are under and above the tweeter. sounds better!

 

[WrineX]

I thought if this group delay it might be possible to dump flanking and put tweeter at the top of a panel then filter from top to bottom. then tilt the panel around 1-1.3cm at the top delaying the tweeter around 38 uSec. and pulling the bottom driver 38 usec forward in time 🙂 making the overall goup delay flatter? in the part thats does most of the stearing (left right)
not sure if going from the top

tweet
upper mid
lower mid,

gives other problems though 🙂 instead of the

lower mid
upper mid
tweet
upper mid
lower mid

 

[neodymium]

Please, more photos!

 

[TNT]

Anything less then 100 us can be ignored I think. In these day and ages its better to be equipped with a DSP system so that one can polish off the whole system at the end if needed. Personally I wouldn't worry about that - it seem to behave perfectly as a minimal phase system which means it can be easily "corrected" by e.g. CamillaDSP etc. Don't you change loob directivity when you change the timing like you describe - i.e. make the speaker "point" into an other direction (up/down). MTMs has its quirks - like not outputting an even power curve (i.e. the energy emitted from the speaker in total, is not linear i.e. missing some frequency bands).

Nice builds as usual!! These start to resample that Swiss company's combined tweet/mid planar. Which cost a pile of SFRs.

//

 

[WrineX]

Anything less then 100 us can be ignored I think. In these day and ages its better to be equipped with a DSP system so that one can polish off the whole system at the end if needed. Personally I wouldn't worry about that - it seem to behave perfectly as a minimal phase system which means it can be easily "corrected" by e.g. CamillaDSP etc. Don't you change loob directivity when you change the timing like you describe - i.e. make the speaker "point" into an other direction (up/down). MTMs has its quirks - like not outputting an even power curve (i.e. the energy emitted from the speaker in total, is not linear i.e. missing some frequency bands).

Nice builds as usual!! These start to resample that Swiss company's combined tweet/mid planar. Which cost a pile of SFRs.

//

thanks, yeah i wonder if MTM or (so i dont have to write tweeter upper mid lower mid all the time 🙂 ) is better better then a TMM.... last M being filtered lowered. haha this contraption is a hell to write about 🙂
yeah it might change the lobe. not sure if that matters ? almost all speakers function this way woofer mid tweet. hmmm i might just make one panel to see if tilting it does anything to begin with. if so i could make 2 and make it a MTM again but tilt them separately (moving the tweeter part slightly further away from the listener) i do have to make the coil of the tweeter part halve the height. since both panels will hold one halve. (hope i can get them close enough together)
any one by the way know why all CBT's are in an arc tweeter being without delay. while in my measurement the opposite problem is there ? tweeter is to much forward compared to the rest in my Xsim

 

[WrineX]

well i changed it up, going back to separate drivers. 3 way... i think there is some phasing issues in such a setup. it becomes insanely wide soudn stage, witch indicates to me some phase issues. i cant seem to measure them but they are there. i might measure the wrong thing.
splitting the panel up and putting the part that has the rolloff 2 cm more forward compared to the tweeter at the top (tilting), adds +- 2.5 dB to the 2K region. so apparently there is a problem. i could try double split panel moving the tweeter part backwards..... but a true 3 way might be simpler and stable when you move your head or are 10 cm longer 🙂 . bit more costly on the crossovers though. 🙁

 

[TNT]

To see that I suppose one has to do directivity measurements - say every 5 deg... ?

//