Help me design a high-pass filter to use with my Vandersteen Quatro speakers

[SoaDMTGguy]

My speakers are designed to be used with a 100 Hz first order high-pass filter. A pair of these RCA filters were included with the speakers: https://www.vandersteen.com/products/m5-hp

I want to design and build my own version of this filter, adding the ability to accept balanced and single-ended inputs/outputs, and maintaining the ability to adjust to match different amplifier input impedances.

The spec from the manual is a signal measuring 1v at the filter output at 1000 Hz should measure 0.707v at 100 Hz. The filter should be adjustable so that this attenuation can be achieved with a variety of amplifier input impedance values.

I know the basics of designing an RC filter, but I need help picking the resistance and capacitance values, and designing switchable loading.

 

[planet10]

The simplest ones are the simpliest of these. Series R followed by a shunt cap.

https://www.t-linespeakers.org/tech/filters/passiveHLxo.html

dave

 

[SoaDMTGguy]

So my filter resistance would be Series R + Input impedance, then the capacitor, right? Does output impedance of the preamp matter?

 

[rayma]

fhighpass = 1 / (2 π C x ( Rsource + ( Rshunt // Rload ) ) )

 

[Capt Grogg]

Its a highpass, so series c and shunt r including parallel input impedance value

 

[rayma]

The equation in post #4 is correct.

There also is a frequency independent attenuation factor of:

( Rshunt // Rload ) / ( Rsource + ( Rshunt // Rload ) )

 

[Capt Grogg]

Right, #4 posted while typing #5

 

[SoaDMTGguy]

What does the “//“ notation mean?

 

[rayma]

in parallel with

 

[SoaDMTGguy]

As long as the equation balances, can the values of C and Rshunt be as large or small as I want? Are there maxima or minima for either of this for any reason?

 

[rayma]

Rsource should much smaller than Rshunt // Rload

C should be much larger than Ccable // Cload

The source must be able to drive Rshunt // Rload with low distortion

For tube sources, Rshunt // Rload should be greater than 50k, the higher, the better

For ss sources, Rshunt // Rload should be greater than 2k, the higher, the better

Any source output capacitor should be much greater than C

 

[SoaDMTGguy]

Thank you rayma. I appreciate you. You answer so may of my and others questions so quickly and so well so often.

 

[planet10]

So my filter resistance would be Series R + Input impedance, then the capacitor, right? Does output impedance of the preamp matter?

In a Low Pass the R_in does not really affect things.

In a High Pass it does, i like using it as the R in the series C, shunt R circuit, it also needs to be quite high to aupport the (droopy) wnd order PLLXO.

dave

 

[SoaDMTGguy]

For my system, Rload is 110k and Rsource is 1k. Setting Rshunt to 422k produces a value for Rshunt // Rload of 87k, which seems good. Combing that with a 0.018 uF capacitor would set fHighPass at 100. Csource and Cload are unspecified for my equipment. Is it common for these values to significant? How do these values seem?

EDIT: Fixed values of Rshunt and C to standard part values

 

[rayma]

Total cable and load capacitance are likely to be on the order of 200pF or less,
so the 18nF would be ok, being about x100 larger.

 

[chrisng]

Based on the image of the M5-HPB balanced filter, I believe it uses a combination of capacitors and resistors to achieve 100Hz highpass for different amplifier input impedances, except the 100Kohm is connected directly to a capacitor without using a shunt resistor.
Since your amplifier input impedance is in 100kohm range, I think you can use a 15nf or 16nF capacitor and no shunt resistor is needed.

You can also add dip switches and arrays of capacitors and resistors to accommodate different amplifier input impedances.

 

[SoaDMTGguy]

If I measured the resistance and capacitance of my M5-HP when set properly for my amplifier, I could just copy those values to make a fixed filter, right?

 

[chrisng]

Yes! From my understanding of the M5-HPB balanced filter, for 100Kohm amplifier input impedance setting, no shunt resistor required, you can simply using your amp input impedance as R to calculate the C value.

https://www.digikey.com/en/resources/conversion-calculators/conversion-calculator-low-pass-and-high-pass-filter

I want to design and build my own version of this filter, adding the ability to accept balanced and single-ended inputs/outputs, and maintaining the ability to adjust to match different amplifier input impedances.

You can add dip switches and arrays of capacitors and resistors to accommodate different amplifier input impedances.
DIP switches 1,2 and 9,10 set 2 capacitors on each leg of the balanced output and DIP switches 3,4,5,6,7,8 set the shunt Resistors.

For unbalanced I/O, dip switches 9,10 and associated capacitors are not required.

 

[chrisng]

BTW, it's a first order high pass filter, you don't need a shunt resistor with any amplifier input impedance if you have the correct value capacitors.
For example, input impedance of 10Kohm, C=160nF.
20kohm, C=80nF
50Kohm, C=32nF

 

[SoaDMTGguy]

@chrisng Thank you, very helpful comments! Since a shunt resistor isn't necessary, why do you think Vandersteen used them? Is it easier to vary the amount of shunt resistance vs varying the amount of capacitance?