Bringing 828 cabs up to snuff...

[Altec Best]

IME, impedance of the drivers certainly affects passive XO design, but doesn't have much of an attenuation or SPL impact.

Maybe not as much attenuation, but what I mean is that with the lower impedance, the 4Ω drivers will play louder than the 8Ω drivers (They Will Receive More Power) The 288's have a sensitivity of about 108 dB and the 416's 98 dB or thereabouts..And that's the reason why they (288's) sound so hot. I haven't measured my drivers in awhile and I should be doing that very soon to finish up my project.

[alancohen]

Been reading up on Mr. Hiraga's A5s. At first when I saw his crossover I was concerned that there was no L-pad incorporated. That gave me the willies since the 288s are so hot. Reading on it seems the HF attenuation is built into the design to balance the 515s and 288s. I just ordered the parts and will build then next week. Time will be tighter as I finally go back to work Monday.

BTW, I sold the Behringer active XO. If I end up going active I will be using the miniDSP system.

Hiraga's crossover for Altec A5x

[fpitas]

BTW, I sold the Behringer active XO. If I end up going active I will be using the miniDSP system.

A friend was looking at the miniDSP and the Steinberg UR824, and went with the Steinberg because it has a bit better noise floor etc:

UR824*:**|*http://www.steinberg.net/

The Steinberg is a few bucks more, but it sounds great.

I should add, he performs all the crossover and EQ functions in his PC using J. River and some plug-ins, then sends USB to the Steinberg.

[alancohen]

The steinberg looks nice, but its $750 on amazon. 2 miniDSPs including plugins are $240. And I really don't want a rack mounted unit. I want something I can bury.

[alancohen]

I'm back to building a passive crossover for my A5s. I am going with Mr. Hiraga's concept.

As I mentioned before, I have 16 Ohm 515Bs and 8 Ohm 288-8Ks. My idea was to use Hiraga's 16 Ohm values for the LF and the 8 Ohm values for the high.



According to the Sound Practices article cited here:

Hiraga's crossover for Altec A5x

Reply #8 / Page 10 / bottom of center column it is R5 @ 5 Ohms that is the matching resistor.

Given that I am using an 8 Ohm 288, do I need to make any adjustments to the R5 resistor? Any one have the skills to calculate what that value should be if it needs to be changed? I'm thinking it needs to be higher since it's set at 5 to match an 8 Ohm HF driver to an 8 Ohm LF driver. Does it need to be 10 Ohms as in the 16 Ohm crossover?

[aditya]

I'm back to building a passive crossover for my A5s. I am going with Mr. Hiraga's concept.

As I mentioned before, I have 16 Ohm 515Bs and 8 Ohm 288-8Ks. My idea was to use Hiraga's 16 Ohm values for the LF and the 8 Ohm values for the high.



According to the Sound Practices article cited here:

Hiraga's crossover for Altec A5x

Reply #8 / Page 10 / bottom of center column it is R5 @ 5 Ohms that is the matching resistor.

Given that I am using an 8 Ohm 288, do I need to make any adjustments to the R5 resistor? Any one have the skills to calculate what that value should be if it needs to be changed? I'm thinking it needs to be higher since it's set at 5 to match an 8 Ohm HF driver to an 8 Ohm LF driver. Does it need to be 10 Ohms as in the 16 Ohm crossover?

Hey Alan you are doing great there !

Now to answer your question.... that R5 resistor has nothing to do with any impedance/matching activity of the HF/LF drivers. It is purely for the attenuation of the HF section alone. Actually it should have been a L-Pad, but as they are the noisier components, Mr. Hiraga (after performance evaluation) has got it replaced with a fixed value resistor. In your case you should also do that. I mean, wire up a WW-pot of 50 ohms as a variable resistor, and find the value at which it gives you the proper balance, and then replace it finally as Mr. Hiraga did.

I hope you are building the 16 ohms version, and not a mix of 16 & 8 ohms filters for your 16 ohms 515 and 8 ohms 288 respectively ?

[alancohen]

Hey Alan you are doing great there !

Now to answer your question.... that R5 resistor has nothing to do with any impedance/matching activity of the HF/LF drivers. It is purely for the attenuation of the HF section alone. Actually it should have been a L-Pad, but as they are the noisier components, Mr. Hiraga (after performance evaluation) has got it replaced with a fixed value resistor. In your case you should also do that. I mean, wire up a WW-pot of 50 ohms as a variable resistor, and find the value at which it gives you the proper balance, and then replace it finally as Mr. Hiraga did.

I hope you are building the 16 ohms version, and not a mix of 16 & 8 ohms filters for your 16 ohms 515 and 8 ohms 288 respectively ?

That's funny, I was laying in bed last night and thinking that I should just get a few different value resistors and swap them around until I get the best SPL match. Then I thought about just putting an L-pad in there. Then I went back to the resistors since they're reasonably cheap and I can always reuse them for something else.

But I did get components to mix the two. I have 3.3mH inductors for the 288s and a 6.6 for the 515, along with the appropriate caps and resistors. It appeared to me that the two circuits were parallel and completely independent .

[aditya]

It appeared to me that the two circuits were parallel and completely independent.

No it is not. Both actually depend on each other to create the decided x-over point & slope, as we are dealing with transducers, which are not resistive but reactive loads, meaning frequency dependent devices. In the absence of symmetry, either the x-over frequency or the x-over slope-overlap (dip or peak) will change, depending on the type of design whether parallel or series. You should go for the 16 ohms type for both, and hook up the 288 with a series 8 ohms resistor to make it behave like a 16 ohms device, which though will introduce a 3db loss factor here, still will not be an issue as the 288s are quite hot. Then use the variable pot to find the correct attenuation. Finally bypass the series 8 ohms resistor with a suitable small value cap in order to bring up the very top end of the hf, as the resistor will put in some losses here in that area.

[Audio_by_Goodwill]

No it is not. Both actually depend on each other to create the decided x-over point & slope, as we are dealing with transducers, which are not resistive but reactive loads, meaning frequency dependent devices. In the absence of symmetry, either the x-over frequency or the x-over slope-overlap (dip or peak) will change, depending on the type of design whether parallel or series. You should go for the 16 ohms type for both, and hook up the 288 with a series 8 ohms resistor to make it behave like a 16 ohms device, which though will introduce a 3db loss factor here, still will not be an issue as the 288s are quite hot. Then use the variable pot to find the correct attenuation. Finally bypass the series 8 ohms resistor with a suitable small value cap in order to bring up the very top end of the hf, as the resistor will put in some losses here in that area.

Those are parallel circuits. They are independent from each other. I would agree that the loads are not resistive, and will change. However, what's to guarantee that the two sides will change together any way? You're recommending to replace a resistive load into the 16ohm design to simulate a 16 ohm driver that's not purely resistive..... how is that better?

[aditya]

Those are parallel circuits. They are independent from each other. I would agree that the loads are not resistive, and will change. However, what's to guarantee that the two sides will change together any way? You're recommending to replace a resistive load into the 16ohm design to simulate a 16 ohm driver that's not purely resistive..... how is that better?

If you are agreeing to the fact that the loads are not resistive, then this itself is proof enough of the fact that both the sides are not independent of each other, and this is the sole reason why text book calculations for x-overs never work alright, which we all are aware of. Actually speaking, in this particular case of 16 ohms 515 & 8 ohms 288, if one goes for a 16 ohms filter-setup for the LF & a 8 ohms one for the HF in the parallel mode, then the o/p will have a slight bump around 250Hz and a suck out at 1KHz. If one switches to the series mode, then the x-over 'f' will shift to 750 Hz.roughly. This is nothing new, and you surely can visualize it loking at the ckt.

Regarding my suggestion of adding a series resistor to the 288 to make it behave like a 16 ohms driver, this is definitely better than the other choice (while keeping Hiraga's x-over in view). Actually it will help flatten out the dips and peaks of the driver to some extent. Ofcourse we have to bypass the inductive effect suitably. As GM has pointed it out many times....many actually run their A7s with the woofer (416x only as its cone cry helps here) directly hot, and the HF driver hooked via a series resistor (direct again) till the levels are okay, and then the series resistor is bypassed suitably with a capacitor.

And also not to forget the fact that the A7 is a very thrusty/forward sounding setup, and to make the sound flat & smooth is not an easy task. It will just shout back with the smallest error.