Type A Crossover Question

[dhsettim]

I am currently using an ALK Universal Xover on my 1977 Khorns, but with a Coral M-100 alnico 1 inch squawker driver.

It is 105dB and 8 ohms, so I have found the 4-1 taps to sound best with my system.

Some time back I made up a Type A Xover with good quality parts (as I had read that both Bruce Edgar and PK both preferred it) but I found some harshness which I assumed was the 9 Khz peak that has been described, and went back to the ALK. I'm not sure if I was using the Coral at that stage.

I want to try again so can I successfully use a 8ohm driver with the Type A, and would the autoformer taps be the same? (Eventually I want to crossover higher as the Coral is said to go up to 18 KHz).

I know, I know... try it and see...but I don't always trust my ears.......

thanks y'all

tim

[Daddy Dee]

Tim,

I don't think I can be helpful with your question, but had thought that spike issue was a K55/K401 issue. If it's a horn reflection back on itself, I suppose building a P Trap would either eliminate the problem or demonstrate that the harshness is coming from something else.

I used both the 'Type A and ALK universal Type A replacement in my LaScalas and Khorns. In my system, I don't think I picked up on harshness with one and not the other. The happiest I ever was with the sound was using the type A w/ 4500 hz mod and a CT-125 tweeter. Of course, it was always using the K55.

What's your take on the sound of the Coral compared to the K55?

[JohnA]

Hi Tim,

What type terminals does your K-55-V have? Not all of them have the 9k bounce.

If
the Coral driver will go from 350 to 18k, you don't need a
tweeter. If it is 105 dB/w/m and 8 ohms, you will need to design
a new crossover, with a 2 dB L-pad and caps sized for the crossover
point and new impedance.

[Erik Mandaville]

As John said, you can redesign the crossover for the new impedance. The autoformer is just one way to attenuate the driver, however a more common approach, which is also used in the new Heritage models, is to use an L-pad. I've compared both types at crossover points of 4kHz, 4.5kHz, and 6kHz. As far as my own experience, a two resistor L-pad does not compromise the sound in any way. It is also easy to calculate the desired attenuation.

Good luck,

Erik

edit: You can also take the same approach many well-known companies have used in the past: An adjustable L-pad that will maintain the impedance of the driver regardless of the position. You can have the advantage of much better control of the attenuation until the drivers blend, and then just leave it there. Altec Lansing was one company that used these in their networks, but they are common and work.

[Erik Mandaville]

Tim:

This may be a case where an impedance equalization network would be helpful. If you decide to make another network and you have the specs for the Re (resistance) and Le (inductance) of the voice coil for the driver you're using, the calcuations are straightforward. If you have access to an impedance curve for the driver, it may indicate a rise in impedance due to inductance of the voice coil. The RC filter will help flatten that rise.

Erik

[dhsettim]

Thanks John and Everyone,

My K55 has the push terminals. Are they the problem drivers?

My thought was the any 9 KHz peak would probably not be present with a different driver.

I'm afraid I don't have the abilty to design a new crossover for an 8 ohm driver so likey to just stick with the ALK.......until I can go active....

It is a while now since I listened to the K55 vs the Coral but I remembeer the Coral being sweeter and more detailed. I'm not good at describing HiFi changes....they were just better. The details I have on the driver a very sketchy but from a frquency response graph that was fuzzily posted on the net I think it is more like 400 to 18K. Showed it to be smooth to 18K with rapid drop ater that but I have no way of knowing how acurate it is.

I did try TAD2001 briefly and even tho' they are supposed to go to 22K I still felt a tweeter was necessary.

In the end I may just try the Type A with the 8 ohm Coral and justsee what happens.

Anyone got any recommendations for active crossovers?

thanks, tim

[JohnA]

The push-terminal drivers are the ones that sometimes had the 9k bounce.

There
are several of us that could design a crossover if we had impedance and
response curves on the horn you want to use (K-400?)

[djk]

A Coral M-100 is a Japanese copy of the JBL 2410. As such it is rated at 10W program when crossed at 500hz at 12dB/oct. I really can't recommend 6dB at 400hz (type A network). The diaphragms are stupid expensive (about $250 each), and if they don't hit the phase plug and fracture, they will still work harden quite fast and lose their 'sweet' sound.

They are 110dB/W on a horn like a K400, so 4-1 is about right (6dB of attenuation). I would use about a 10µF cap to the input tap #5 and add about a 5mH coil in parallel with the driver. This will give roughly a 12dB/oct slope at about 350hz.Then I would add a 68 ohm 2W resistor in parallel with the driver. The resistor will dampen the Q of the network a bit and raise the crossover up to about 400hz.

While these type of drivers will measure flat on an old style horn, they will always sound rolled-off in the high end. On a modern CD style horn the response will start to roll off about 6dB/oct above about 4Khz . This requires about 12dB of boost in the 16Khz region to make the horn and driver flat, or a woofer no more than about 98dB and do the EQ in the network (I've done this before and it can sound great).

[dhsettim]

Thanks for your very kind offer John, but looks like DJK has put the stops on that idea.

Wow, what an amazing mine of information thank you DJK.

I might just stick with the ALK universal as I know and like it until I can actively crossover.

But then again the temptation to tweak is strong........

thanks y'all

tim

[Erik Mandaville]

But then again the temptation to tweak is strong........

tim

You can build the same the same network Djk suggests, if that's what you think you would like to do, yet make some value changes of parts to compliment a variable L-pad rather than an autoformer. Autoformers can be awkward in terms of making attenuation changes, and a variable L-pad, which will match the impedance of the driver regardless of the amount of reduction or gain, can be much, much easier in that respect. If you're planning on going two-way with this, it might be of interest to study, if available, an impedance curve of the driver; and, if it seems like it would be prudent, install the RC network that would give you a more linear response. Who knows, you may prefer how it sounds without that network, which is something that can help put to use the tweaking urge you mention: take it out or leave in as sounds best to you.

The variable L-pad works just like a volume control.

Two-way certainly simplifies the design.

Erik