Dipole tweeter Xover

[bjarnetv]

as the thread title says, i need som crossover design help for my speakers!

the speaker currently consists of L.cao 8" fullrange speakers mounted in an open baffle on top of 2x15" woofers.
i currently use a minidsp 2x4; the L.cao's are driven by some Quad II monoblocks and the woofers by a rotel rb1510 icepower amp.

the only problem is that the l.cao needs a lot of EQ above 4000hz, and even then it lacks some top end sparkle.

i therefore plan to throw some tweeters into the mix, and as a planar lover, the tweeters on top of my wishlist is a pair of ESS AMT units

now the big issue here is that the tweeters need to crossed over with the L.caos, but the Quad II doesnt like loads under 8ohm, so as far as i understand, i need to make a series crossover.

i dont have much experience making passive crossovers, so please forgive my ignorance if i do something stupid

here is my first go at the crossover, as i want to make the AMT play as low as possible to evade the issues around 4khz.

i figure the sensitivity is close enough between the two speakes that i wont need an l-pad on the L.cao.

so, is this the worst plan ever, or something we could work with?

[IMGDEAD]http://img651.imageshack.us/img651/4048/img20120902172422.jpg[/IMGDEAD]

 

[VaNarn]

The capacitor value should be 10mFd and it would be useful to place a resistor of 3.9 Ohm(10 W) in series with the AMT to adjust the sensitivity and impedance for a closer initial match.There are other refinements that could be implemented later,such as a impedance correction circuit for the Lcao and possibly a change of component values to suit a Quasi second order alignment( C= 7mFd and L=.9mH for zeta .707).If there is another xover network for the 15'' and the 8'' some changes may be needed to avoid interaction.

 

[bjarnetv]

if i add a 3.9ohm resistor as you recommend, then 10uf would indeed be correct.
regarding the 3,9 ohm resistor; how much would it affect the sensitivity of the tweeter? wont it eat up a lot of power from the amp?
seeing as the 8" is rated at 94db and the AMT 92db, adding a resistor doesnt really sound optimal.

also, if i understand series XO correctly, the amp would see a 8ohm load from the 8" at lower frequencies, around 16ohm at the crossover point and 8ohm in the high frequency, as opposed to 4 ohm if i didnt include the resistor?

the 15" and 8" is crossed over at 150hz using a miniDSP active XO.

 

[VaNarn]

There is some variation in the AMT1's sensitivity,which is dependent on the production date.The very earliest type did not have the gauze screens and were 96-97 dB SPL for a 1W/1m (2V) input. What tends to be overlooked is that the AMT should be fitted with a baffle and I credit Mark Gulbransen with pointing this out.With your proposed cct and taking the Lcao as being 8 Ohm at 2kHz the impedance would be around 4 Ohm at the xover freq.The 150 Hz L.F. xover is well away from having any interaction problem. Measurements will allow a more refined approach but if you are satisfied with the results after prolonged listening then you may not need to go further.

 

[bjarnetv]

There is some variation in the AMT1's sensitivity,which is dependent on the production date

i was thinking of some new production units from parts express, rated at 92db

With your proposed cct and taking the Lcao as being 8 Ohm at 2kHz the impedance would be around 4 Ohm at the xover freq.

so i guess i must have misunderstood series crossovers then, as i thought having the tweet and mid in series would end up with 16ohm, as opposed to the 4ohm you get from parallel crossovers


also, because of the impedance mismatch, would it be better to just save up some more and go for a 8ohm AMT like RAAL 140-15D or beyma tpl150, as i wouldnt need the power eating resistor in series with the tweeter?

 

[VaNarn]

While I cannot comment on the RAAL or Beyma units as I have not heard them,I do regard the ESS HEIL AMT as representing a very high standard . As far as I know the resurrected production of the AMT has the same D.C. resistance of 3.9 Ohm and sensitivity figures of the previous American series(when used with identical diaphragms).Since speaker sensitivity varies with freq. I would not place too much reliance on precise matching initially; as there are methods that can be used to overcome minor deviations.One of them is to apply an impedance correction cct across the 8'' speaker to bring it's impedance down to 4 Ohm at 2kHz.this effects impedance only ,not sensitivity but it does absorb power.You could also vary the suggested value of the resistor in series with the AMT, down to as low as 2.7 Ohm ,as the series network is not an overly critical cct.The quasi second order network usually has a slight dip at the x.o. freq. which helps to achieve a smoother blend ,although it still has a lobing problem.This can be dealt with by suitable tilting or setback of the AMT.

 

[bjarnetv]

the reason i am being so difficult with all these questons is because the amps only have 15 watts to play with and works best between 8-16ohm.
i am just worried i am turning the speakers into something that wont work to well with them

 

[VaNarn]

There is nothing wrong with being cautious before proceeding with a project.I am surprised that no one else has offered advise,especially on the sensitivity of the AMT's.What a shy lot at diyAudio ! It is worthwhile to emphasise that the majority of your listening will be at levels below 1 W and since your spkr.system would be considered of high sensitivity by current standards,that 1W of sustained level is a very loud sound even for one channel. If the power spectrum requirement is evaluated for civilised music even of the most dynamic nature,then the upper frequency region which is of concern to you,should not be a problem for the Acoustical Quad II.This amplifier was capable of driving that co's electrostatic speaker,which is far from an ideal resistive load.

 

[bjarnetv]

yeah, dont be so shy people; join the discussion!

after a bit more googling, i think it might be a good idea to raise the XO point to 3k, as it seems most people find them a bit nasal when crossed too low.

 

[puppet]

OK ... I'll chime in. Old vintage motor w/new diaphragms I'd guess the sens. to be @95-96db. I'm using them in an active system ... same amp for mid/woof which is 93db sens. Heils are at half volume on their channel(s).

A third order (bessel) sounds nice with them. I'd try to keep the HP above 1400hz or so. With the L.cao driver you've got a good starting point would be 2k. ... but you might need to notch that 3k spike a bit, even with a third order LP. (shrug) try it.

FWIW, I use a 1600hz HP on my heils and they don't sound nasal to me at all.