Different 3636 autotransformer tap settings for Type A

[Kudret]

I will be building a new Type A with 12 ga 2.4 mH inductor and 12 uF capacitor. With the above components the crossover points are already moved, but probably this is not too significant.

I would like to try different attenuation levels for the squawker by changing the tap settings on 3636. I believe I need to either use a swamping resistor across 0-5 or reduce the size of the 13 uF capacitor as I increase the attenuation to maintain the crossover points.

I want to try -3.8 dB and -5.3 dB and -6.0dB attenuation levels. How do you calculate the size of the required resistor or capacitor for these settings?

4-0 (-3dB default setting)

5-2 (-3.8 dB) or Y-1 (-3.9dB)

X-1 (-5.3 dB) or Y-2 (-5.4 dB)

3-0 (-6.0 dB)

Is there any sonic difference to these two approaches (changing the size of the resistor vs capacitor)?

Kudret

[BEC]

Kudret,

The 3636 is built to give values of attenuation between 1 and 12 dB in 1 dB increments.

See PDF

Bob Crites

B&K_AT-1-1.pdf

[Kudret]

Bob, Thanks for pointing that out. I should have used the table on the right for Type A in the following figure:

In this case, say for -4dB attenuation (input on X, output on 3) where does the swamping resistor go? Also, what size of resistor/capacitor do I use?

Kudret

[BEC]

Kudret,

You will have to discuss the resistor with Dean or Al. I try to avoid even talking about resistors in crossovers.

Bob

[Al Klappenberger]

Kudret,

If you are building a Klipsch "A" you need to use the 3 dB setting and NO swamping resistor. You also need to stick with the cap value the Klipsch used. The swamping resistor is NOT aplicable to the "A" network without major changes.

Al K.

[Deang]

You're mixing apples and oranges Kudret.

[Kudret]

Dean, I found this thread.

http://forums.klipsch.com/forums/permalink/758463/758558/ShowThread.aspx#758558

Discussion is about AA and I thought it applies to A as well.

Kudret

[djk]

Assume the K55 is about 15.3 ohms, this will give a 400hz crossover with a 13µF cap and the -3dB tap on an autoformer.

To keep the same nominal impedance and crossover point when changing taps, a resistor may be added in parallel with the midrange driver. The impedance multiplication factor of the autoformer on various taps (and the calculated resistor value for nominal impedance and crossover point):

-3dB 2:1 (none)

-4dB 2.51:1 (59.8) ((150))

-5dB 3.16:1 (26.4) ((82))

-6dB 4:1 (15.3) ((62))

-7dB 5:1 (10.2) ((51))

-8dB 6.3:1 (7.1) ((45))

-9dB 8:1 (5.1) ((41))

-10dB 10:1 (3.83) ((38))

-11dB 12.6:1 (2.89) ((36.4))

-12dB 16:1 (2.18) ((35))

A 16 ohm L-pad may be connected as a 60 ohm variable resistor between its terminals #2 and #1 for test purposes. Measure and adjust for the nominal value, then you can tweak it a bit and listen to what's happening. Replace with a fixed resistor when you're satisfied with the results. An 8 ohms L-pad will go up to as much as 40 ohms between terminals #2 and #1, and could probably be used as well.

The set of values in double ((brackets)) is for use on the input side of the autoformer. These are standard 5% resistor values down to the -7dB tap, below there they will need to be a combination of resistors to get the right value.

It is important to note that these are parallel resistors, not series (that some object to).

[djk]

Bump.

When you edit an old post it doesn't bump up, so hence the bump.

[Kudret]

Thank you, djk.

I believe one can also reduce the value of the 13 uF cap to keep the crossover points the same. What would be the appropriate cap values to achieve the same? and is one approach preferable to the other?

Kudret

[DrWho]

I think changing the cap value will also change the slope too...no?

[Deang]

Not really -- it's still a first order high pass.

I believe one can also reduce the value of the 13 uF cap to keep the crossover points the same.

No, you stay with 13uF regardless. The resistor values Dennis gave are so you can attenuate without changing the cap value.

Keep in mind that this is a different way of doing it than the way Al does it -- in other words -- you don't use the 10 swamping resistor between taps 5 and 0, and no longer float the ground for the squawker.

[djk]

There are really too many ways to do it all.

You can add the resistor and keep the cap the same, or you can change the cap, or you can do some of both.

I prefer to have some resistance in parallel with the driver to swamp the impedance peaks in the lower end of the horn. If you're only using 3dB of attenuation, the only way you can do it is to increase the cap values. High quality big caps cost big bucks. A good resistor costs only a buck or two.

If you know about how much attenuation you are going to need, hook the driver to those taps, add a 15 ohm resistor in parallel, and then calculate the value of the cap you need.

An example of this would be using a 26µF cap with a 4-0 tap and a 15 ohm resistor (A or AA type networks) .

The table I gave in the other post gives two alternate resistors, one for across the driver, the other for across the input taps. Pick one or the other, you don't need both. If you feel adventuresome you could see if one method is better than the other. I have no opinion on this, and have done it either way depending on the material at hand.

If using no resistor, the cap valuse must change as you attenuate. For a 400hz network the values would be:

-3dB 2:1 13µF (stock)

-4dB 2.51:1 10.36µF

-5dB 3.16:1 8.23µF

-6dB 4:1 6.5µF

-7dB 5:1 5.2µF

-8dB 6.3:1 4.13µF

-9dB 8:1 3.25µF

-10dB 10:1 2.6µF

-11dB 12.6:1 2.06µF

-12dB 16:1 1.625µF

The only use I can see for the -12dB values would be with a Heresy woofer and a BMS driver, but you get the general idea.

[Kudret]

Thank you so much again, djk. I certainly appreciate having all the possibilities laid out before I start experimenting.

Kudret

[freddyi]

Thanks DJK - that's a really helpful chart and explanation of how to keep one cap value plus get some swamping of Z peaks.