Filtering the Control Signal of a Fairchild 660: Passive is ok?

[niccolo gallio]

hi there,
I'm one of the crazy persons that built a 660 circuit using Drip Electronics' boards.
It works as it should and i'd like to add an high pass filter before the control section. Do you think it can be done?
Do you think it's an "easy" enough thing to do?
I would like to keep it passive, a simple RC first order network with switchable values for three frequencies points to be selected, or should i use some kind of buffer?
Where would be the best place to hook this filter up? i guess I'm not good enough to see in the schematics where is the control return path to the 6386 coming from..
Please excuse me if this is too much of a request or if it's too stupid of a question, trying to learn something from errors and blushing..

thanks for looking

niccolo

 

[Kingston]

A great place for the high pass filter is right before the sidechain input transformer. I don't know the Drip 660 specifically, but in these kind of fairchild topologies there is usually:

output transformer (from the vari-mu amp) -> pad - > sidechain input transformer

With a DPDT switch you can have a balanced high pass filter:

output transformer (from the vari-mu amp) -> pad -> high pass filter ->  sidechain input transformer

You need two caps and a shunt resistor.

two 2.2uF caps and a 2k resistor is a good starting point. I seem to recall I used those values in a similar configuration and that it worked well. A 6dB/octave (1st order) high pass filter with a cutoff somewhere around 100-150hz.

For three frequency points it gets more complicated, and perhaps that is unnecessary anyway. on/off with a well selected high pass frequency should suffice most compression tasks. You will find the effect quite subtle, unless you cut above maybe 1khz, but then it sounds crap anyway.

 

[niccolo gallio]

Thanks a lot for the kind answer,

I agree on keeping it simple, I'll just start with one freq as you suggest.

I'm not sure if I understand the role of the "Pad" object in your diagram, are you suggesting I should pad the signal at that point or are you indicating a component that "usually" is there?

Also, four paths come from the output trafo, should I apply the filter to both pairs?

I am attaching a snip of the part between the output trafo (T102) and the control input trafo (T103), is the section after R149 and R150 good for the filter circuit?



Is this correct as a balanced filter network?



Thanks a lot again

nicco

 

[Kingston]

[edit]

The schematic above is Fairdhild 670 stereo version, not 660 (mono?) as I understand.

There is mid-side mixing going on with the transformers above, and that complicates the high pass filter tremendously.

Your drawing is correct for normal balanced L + R signal. Your 660 should have a pad like R145-R150 above, but without the sidechain tricks. You can stick your drawing right after that.

Find the correct schematic first.

If the unit you have actually is the stereo version, then a good place for the high pass filter would actually be right after the sidechain output transformer, before the bridge rectifier and the 4K7 resistor. Similar values should apply there as well. Your drawing is a good starting point there as well.

 

[niccolo gallio]

Oops,
I'm sorry but I said 660 because I thought would make things simpler, but instead it is a 670, even though the matrix switch is not fully operational but yeah, it is a 670.
Sorry for the confusion, I should have been more clear.

So you say that after the control am section is a good place right? I can see that because we are coming straight balanced from that and not four ganged..

So.. before the bridge rectifier.. would it be better before or after R136?



thanks a lot again

 

[Kingston]

I would put it before R136. It's easier to calculate the exact RC filter that way.

You can calculate the cutoff with this:

http://www.muzique.com/schem/filter.htm

(but even now 2X 2.2uF with 2k resistor is a good starting point)

remember that the capacitance you get from that calculation will need to be doubled for each side.

also remember the filter R will be parallel with the 3K9 (R136) load resistor, so take that into account with your calculation:

http://www.sengpielaudio.com/calculator-paralresist.htm

 

[niccolo gallio]

Thanks a lot Kingston for the infos and the links!
If I can abuse some more of your patience and knowledge please allow me to pose another couple of questions:

1): When you say

remember that the capacitance you get from that calculation will need to be doubled for each side.

it seems pretty clear to me that if I find that 2uF is ok for a unbalanced RC filter, then if I want to turn it balanced I'll need two 4uF caps, right?
Then I went looking for some info on wikipedia and found this:



The first image on the left can represent the unbal filter we're talkin about, where Z is the resistance and Y is the capacitance, and then there is this image:



that seems to represent the balanced filter in question, but here it seems implied that the Z value should be half of what used to be when unbalanced, this is the exact opposite of what I thought I understood.
Both images are taken from this article:
http://en.wikipedia.org/wiki/Electronic_filter_topology

2): When you say

also remember the filter R will be parallel with the 3K9 (R136) load resistor, so take that into account with your calculation

you mean that this is more important for the filter of for R139? I mean, should I choose the R of my filter as not to change too much the total parallel resistance seen by the output transformer (formerly R139)? or should I not worry about this and choose an R that, in parallel with R139, gives me the R needed for m filter?
And what about using R139 as the R of the filter, adding only the two capacitors?

Thanks a lot again, your posts are very informative.

nicco

 

[Kingston]

You're complicating things far too much already.

http://www.gearslutz.com/board/geekslutz-forum/356482-easy-diy-passive-filter.html

 

[niccolo gallio]

My apologies, I've crossed your threshold of patience and, rightfully, I've been given the GS tratment..

 

[emrr]

Since the SC amp input is also the audio output path, insertion of passive EQ components may subtly equalize the audio output.  I would definitely compare measurements to be sure. 

If you have a problem there, I would consider using smaller coupling caps in the SC amplifier.

 

[Kingston]

Since the SC amp input is also the audio output path, insertion of passive EQ components may subtly equalize the audio output.

Too subtle to measure at least on poorman, which has a reasonably similar load at that point. The balanced pad takes care of that.

signal chain:

audio output -> load isolation (balanced pad) -> balanced high pass -> SC input

 

[emrr]

Good to know.  I've seen it be a problem elsewhere. 

 

[niccolo gallio]

Thanks a lot,
now I get that in a C section the capacitors are in series, thus their value must be doubled!

thanks again