Filter for buffer

[carlmart]

In spite of having proposed a simpler way to two-way active amps, using boards like the NIGC sold by BrianGT, doing a first order high-pass filter on mid-high amp and low-pass filter on bass amp, the aproach has some limitations.

But there might be a more interesting application of the idea on an IGC by doing part of the filter (or all of it) in the buffer itself.

My first proposal for a two stage 1st order buffer plus IGC would be this:

[carlmart]

The second option is a bit more complicated in the filter stage, as it will take more parts and you can choose between Butterworth and Bessel second order filters.

This proposal is for a two-way active IGC, for mid/high and bass stages. Mid and high should have their own passive xover.

The IGC should be the standard one already published in this Forum.


Carlos

[carlmart]

The second option is a bit more complicated in the filter stage, as it will take more parts and you can choose between Butterworth and Bessel second order filters.

This proposal is for a two-way active IGC, for mid/high and bass stages. Mid and high should have their own passive xover.

The IGC should be the standard one already published in this Forum.

The filters were taken out from Walt Jung's "Audio IC Op-Amp Applications", unfortunately out of print.


Carlos

[Pedja]

Hello Carlos,

Generally, both options you posted look good to me. One notice though.

In the first case the impedance of the source could vary too much to consider the filter before the first stage as an option for crossover. Actually, it would not be that strange if the additional input buffer would be found necessary.

Btw, there was the earlier thread discussing similar things: Line level XO and gainclones
(Ignore notice about the thump and the diamond in my post. The guess was based on some previous experience with it, but when I tried the diamond, it did had the thump. Opamp buffers, as shown by Carlos, are better candidates to make you free of it though and it won’t be bad to check some open loop monolithic buffers for this, I won’t be surprised if some have solved thus issue.)

Pedja

[carlmart]

Originally posted by Pedja

In the first case the impedance of the source could vary too much to consider the filter before the first stage as an option for crossover. Actually, it would not be that strange if the additional input buffer would be found necessary.

By first case you mean the one on the drawing, buffer + IGC?

In fact there was something missing on my drawing, which is the pot before the buffer. So there might be an impedance variation for different pot points. One trick might be to invert the pot, putting the variable arm toward the source, but that could be a problem for some sources.

Perhaps a buffer before the filter is in order. Implementing an OPA627 for the buffer and a BUF634 on the filter might do it. That should work for both filter versions I proposed.

[QUOTE][i]Btw, there was the earlier thread discussing similar things: Line level XO and gainclones

My fault no to have read that thread. Some of my proposals are there, though the buffer, filter and IGC would have to be completely separate because levels will have to be set.

[i]
Opamp buffers, as shown by Carlos, are better candidates to make you free of it though and it won’t be bad to check some open loop monolithic buffers for this, I won’t be surprised if some have solved thus issue.

Nuuk's circuit, perhaps physically assembled as Kuei Yang suggests to avoid caps, might be an option. But thumps of any type should be out if possible.

Second order Bessel filters, set around 150Hz might be an idea. AD's Filter Design Tool provides the right values.


Carlos

[carlmart]

Originally posted by Pedja

In the first case the impedance of the source could vary too much to consider the filter before the first stage as an option for crossover. Actually, it would not be that strange if the additional input buffer would be found necessary.

By first case you mean the one on the drawing, buffer + IGC?

In fact there was something missing on my drawing, which is the pot before the buffer. So there might be an impedance variation for different pot points. One trick might be to invert the pot, putting the variable arm toward the source, but that could be a problem for some sources.

Perhaps a buffer before the filter is in order. Implementing an OPA627 for the buffer and a BUF634 on the filter might do it. That should work for both filter versions I proposed.

[i]Btw, there was the earlier thread discussing similar things: Line level XO and gainclones

My fault not to have read that thread. Some of my proposals are there, though the buffer, filter and IGC would have to be completely separate because levels will have to be set.

[i]
Opamp buffers, as shown by Carlos, are better candidates to make you free of it though and it won’t be bad to check some open loop monolithic buffers for this, I won’t be surprised if some have solved thus issue.

Nuuk's circuit, perhaps physically assembled as Kuei Yang suggests to avoid caps, might be an option. But thumps of any type should be out if possible.

Second order Bessel filters, set around 150Hz might be an idea. AD's Filter Design Tool provides the right values.


Carlos

[carlosfm]

Carlos,

I think the easy way would be to use a double op-amp for each channel.
That way you have buffer-filter-buffer.
An OPA2132 always gives very good results.

[Nuuk]

Nuuk's circuit, perhaps physically assembled as Kuei Yang suggests to avoid caps, might be an option. But thumps of any type should be out if possible.

When I get time I want to try that active crossover with a couple of GC stereo amps. I already have the speakers here and could run each pair of GC's from the regulated PSU's that I am currently building into another buffered amp.

And then I could modify the crossovers like KYW suggests...........

[theChris]

2 points:
1.) damping -- i'm pretty sure you can't get an underdamped response with that. you don't have any reactive feedback for the amplification sections, nor any resonant networks. i don't see how you can get a butterworth response. or are you setting the 1st cap large enough to get a 1st order approximation? last time i tried something similar i found that i had to use negative resistances.
2.) sourcing -- the buffer is to ensure that a low-impedance source is driving the crossover. If this is true then you can do that.