902-8Bs on 511Bs

[Colt45]

I have a pair of 902-8Bs on 511Bs I bought to make 2 way La scalas. I built a pair of 1st order crossovers at 650 Hz and hooked them up. They sound good if you keep the power down. At loud conversation level, bass notes make the 902s buzz like the wires of a snare drum. One is worse than the other.

What is the recommended crossover frequency and slope for the 902-8B on a 511B?

What else might cause the noise? How do I identify non-Altec diaphragms?

It's time to get serious with them. i've picked up a pair of Peavey FH1s (La Scala bass horn clones) and will be using them as sound reinforcement at church and outdoor events. I want these things to be nearly stupid-proof.

Obviously, I have work to do.

[scott fitlin]

Change from the 6db per octave slopes to at least 12db slopes. Crossing over at 650hz on such a shallow slope allows the horn to get down way too low.

I would use an 18db butterworth filter and a xover point of 800hz, especially since you will be using the speakers for various kinds of sound reinforcement work!!

[Colt45]

I had figured on 18 db slopes, but I'd rather not go up to 800. The FH1 is losing output by then.

Did Altec not specify a frequency and slope?

[Hofmannhp]

Originally posted by Colt45
I had figured on 18 db slopes, but I'd rather not go up to 800. The FH1 is losing output by then.
Did Altec not specify a frequency and slope?
Hi Colt,
I agree with scott and recomend 800Hz with min12 better 18db/oct.
Please think of the fact that the 902-8B is only good for 15 watts between 500 and 20kHz. A lot of low freq energie can easy destroy the expensive diaphragma.

HP

here are the specs:

[scott fitlin]

Originally posted by Colt45
I had figured on 18 db slopes, but I'd rather not go up to 800. The FH1 is losing output by then.

Did Altec not specify a frequency and slope? Colt you are actually correct about the FH1, but the 902 definitely should not be used with a 1st order filter at 650hz!

You could, if you wanted to spend additional funds, use a 10 or 12 in low mid driver between the FH1 and 902. This would allow you to operate each cabinet/driver/horn in their optimum frequency ranges!

Or you could consider changing from the FH1 to a front loaded horn similar to an Altec A7, or possibly a direct radiator.

Altec did specify using 12db or higher slopes with their compression drivers. Actual xover point depends on what horn your using, but 800hz is a good xover point, more musical sounding than 500hz. And if you cross the 902 over at 800hz with a slope of 18db per octave you will have no problems with the horn drivers.

[GordonW]

OK, here's my take... the real issue here, is excess low-frequency energy getting to the driver. I've attached a crude Bode plot (log/log frequency response diagram), showing the difference between the curves, for a 12 dB/oct. 800 Hz highpass crossover, an 18 dB/oct. 650 Hz highpass crossover, and a 24 dB/oct. 650 Hz highpass crossover.

Note the colored shaded areas in blue- these represent areas where the driver is getting MORE low frequency energy, with a 650 Hz crossover, vs. an 800 Hz crossover. The light blue is the difference between 800 and 650 Hz, specifically, and the dark blue, is the DIFFERENCE in energy for an 18 dB/oct. crossover and a 24 dB/oct. crossover.

From the diagram, it's simple to see, that the 650Hz, 18 dB/oct. crossover is passing more energy to the driver, all the way down to 410 Hz. This is BELOW the cutoff frequency of the horn, which is a DEFINITE no-no. However, note the 24 dB/oct, 650Hz curve... it's already passing LESS energy to the driver, by 520 Hz and down. This NARROWLY is higher than the horn cutoff.

My take? The 24 dB/oct, 650 Hz crossover MIGHT (I would almost say SHOULD, but it depends on the fragility of the parts condition as to how certain it would be) be OK on the driver/horn combo at hand. If you can be sure of the operator not getting too "out of hand" driving the system, I would probably be willing to give this combo a shot. However, the 18 dB/oct 650 Hz crossover would be a dangerous combo with this horn/driver, and would be something I WOULD NOT even consider doing, IME and IMHO.

[GordonW]

One other thing- the nice thing about a 24db/oct crossover, is that you can change the the Q (the "sharpness" of the transistion at f0, ie whether it starts to "lazily roll off" slightly before f0 with a low Q filter, or drives right into close to f0 before starting to rolloff with a higher-Q filter. This allows you to shape the response. Basically, it's tunable... if you run into a situation where you need to drop the output just a tad at f0, just lower the high-pass filter Q. In fact, the filters I modelled, were Butterworth 4th order filters; for a L-R, you'd be down another 3 dB at f0.

I would be inclined to try a Bessel 4th order highpass (Q=.577) on that one... it's kinda in-between. A good compromise between trying to meet the folded horn and protecting the driver, I'd think.

Finally, in the end, a couple dB dip at 500-700 Hz, really isn't going to be THAT bad in a PA speaker. I would seriously doubt, that anyone would even notice, realistically. Also, if you needed, you could just use a 31 band EQ, and run up the 630 Hz slider just a tad (1-3 dB)... problem solved...

Regards,
Gordon.

[Colt45]

Here is the *published* curve for the FH1. I think I'll raise the crossover point to 800 at 6 dB on the bass horn and 800 at 18 dB for the tweeter and see what happens.

If I add a polyswitch in parallel with 100R and in series with the 902s, downstream of my L-pad, the crossover point will drop to about 425 Hz when the polyswitch opens, but at much reduced power to the driver. That should do it.

BTW, isn't the real cut-off frequency of the 511B about 250 to 260 Hz?

[boputnam]

Originally posted by GordonW
...the nice thing about a 24db/oct crossover, is that you can change the the Q (the "sharpness" of the transistion at f0, ie whether it starts to "lazily roll off" slightly before f0 with a low Q filter, or drives right into close to f0 before starting to rolloff with a higher-Q filter. This allows you to shape the response. Basically, it's tunable... There are some great commercially available xovers with this "response" adjustment - makes for some very interesting fine tuning, just as you say. Enables some very pleasing results. :yes:

[Hofmannhp]

Originally posted by Colt45

........... I want these things to be nearly stupid-proof........
thats why I think you should take the near 800Hz crossover. You get a lot of "stupid proofness".
The FH1 SPL curve seems to be ideal for this I mean.
cut the FH1 lazily above 800Hz and the 902 with 12db under 800Hz. The result maybe like this (see the colored curve) with a lot of headroom for the whole system.

HP

[Colt45]

O.K., This looks like the final design. It is a first order crossover, rolling off the FH1 at 800 and rollong in the 902/511B at 1800 Hz with a "Solen" first order slope (equal to a Butterworth at about 3000). With this, I don't need an L-pad, but will add a polyswitch and resistor to limit tweeter power.

The graph was generated with X-over Pro using the published FH1 response and measured 902/511B response. These smoothed response curves are +2/-3 from 95 to 20k Hz adn -10 at 63 Hz in half-space.

They're not built, yet. What do you think?

[GordonW]

The problem I see, is that it's still 6 dB electrical slope on the horn. Drivers like that, need at least a 12 dB electrical high-pass, to limit diaprhagm excursion. It takes at least a 12 dB slope, to NOT have increasing driver excursion with decreasing frequency, even below resonance! With a 12 dB/oct. slope, the driver excursion will be CONSTANT with frequency below resonance, so if it's safe at 1800 Hz, it'll be safe everywhere.

What I'd try? Converting the 6 dB crossover at 1800 Hz, to a 12 dB/oct. slope with a REALLY low Q, at a slightly lower frequency. Try adding a 2 to 4 mH inductor, to ground after the cap in the highpass, and adjusting its value and/or the cap value, to make it as flat a blend as possible with the woofer cabinet. You might have to fiddle with the inductor on the woofer horn too... but I'd bet, you can probably get it pretty flat again. With the second order highpass on the 902 @ above 1000 Hz, I'd think it should be pretty rugged...

Regards,
Gordon.

[Colt45]

"I'd bet, you can probably get it pretty flat again."

That is the very reason I picked the high 1st order crossover. I was not able to get 2nd order to be as flat. I agree it would be better.

I am discussing it with the software author via e-mail. He wrote the program to use T-S parameters for all drivers and to use the acoustic response as a minor tweak. This results in odd interactions with the filters and unrealistic response graphs ( a 2 nd order Butterworth filter with a 6 dB peak just above the crossover and 11 dB down at 20k). I'm not sure I trust the graph above as accurate.

He is essentially unfamilar with horns. The program is written for direct radiators.

I was not able to tell that before paying for it. I need the #$%^ computer to draw the graphs so I don't spend days of tedium doing it by hand since I'm in the frequency range where both horns are behaving badly.

[jkc]

Here is my 5 cents worth on crossovers for 511Bs.
I use 24 dB/Octave Linkwitz-Riley Electronic Crossovers and 500 Hz sounds fine.
The advantage is you are 6 dB down at cross over and phase Coherent.
I them drive them with a 100 Watt amp which means they are unlikely to be clipped as they suffer from throat distortion first.
I have had the same 802 diaphragms for 20 years + in some 808 drivers and they have seen plenty of use.
They may sound a little better crossed over a little higher but mainly it is a power handling thing or what you can get out of whats on the bottom.
Look here for a filter design
http://sound.westhost.com/project09.htm