mic pre HPF

[Morning_Star]

How it is done in the Trident S20?  That one seems to work really well.

 

[mitsos]

Simon, I thought the green had an HPF already... at least the last version.. not that I built one or anything.

no idea about the S20... but

as far as this goes, I was thinking of putting it after the opamp..  I don't know how this would mess with the output trafo (since it would mess with the input side...). Since I wanted to keep it passive (but still want a 12dB/o filter) I figured I could cascade 2 filters, and to separate them could put one before the opamp and the other after.. again, don't want to do a relay, but it seems inevitable so this is very close to being axe'd. or at least to be continued later...

 

[simonsez]

Simon, I thought the green had an HPF already... at least the last version.

When i planned to bulid green pre, i have 3 different schemo in front of me. I always love the original sounds like, so i decide to bulid from the original schemo thus no hpf in the preamp stage. I think the green pre is balanced transformerless circuit (?) so if i want to put hpf and keep the circuit balanced it should be done at both side (+/-) and need precisely match both cap and R like Chae said...matching R is easy, but Cap?..may be i will try it..  

 

[abbey road d enfer]

There's a number of line input stages that use adjustable caps to trim CMRR, but these are limited to small values (<200pF) so are not suitable. You have to adjust CMRR by putting additional low-value caps in parallels with the one that's lacking.

 

[mitsos]

ok, you guys are losing me.  Can we do some numbers to help?  I've been plugging stuff into excel since reading CJs post about reactance but no idea what the number should mean, since reactance = (-)resistance at the -3dB point and beyod that goes to all kinda of numbers.

to make sure I have everything correct in my mind:

So let's say we're using a 1uF cap and 2K resistor, that would make the F = 79.6 Hz and X=-2000 right? But if you start plugging in different F to the Reactance formula you get all kinds of numbers.. Where is reactance(X) high enough or not high enough? As F goes up, X goes down, but not sure where it becomes a problem....

Keeping in mind a 1:7 to 1:10 input trafo (I am using Cinemag 75101, 1:10, with 150K RL, for ~1K5 input Z)... would these numbers work, or would it be better to use a higher R and lower C? If I can get the C under .033uF I can get 1% Wima caps, but I'd have to use 33K to get it down to 146Hz...so not sure if it's feasible.. Also, for a second order, you can't just cascade 2 filters can you?

Could this be put on the output and would it have the same/similar problems? Chae's point about headroom is valid, but if we did not take that into account? 

Going back to Chae's point on putting it between the I/T and opamp... This would go after the zobel, RL etc. right? Wouldn't the RL effect the filter since it is going to ground? Or the other way, wouldn't the RC filter affect loading on the trafo since it has an R to ground?


SIMON: this green schemo has the HPF built in... it's made of the opamp with C10,C20, R12 and R14 around it.. Fc=87 Hz, if I calc'ed it correctly.
http://www.mendelt.nl/greenpre_V14/_Green_trim_HPF%20CCT.gif
Fc=1/(2pi*sqrt(r12*r14*c10*c20))

 

[simonsez]

Thanks Abbey...
Mitsos, I build based on the original schemo (like supergreen).....

 

[abbey road d enfer]

ok, you guys are losing me.  Can we do some numbers to help?  I've been plugging stuff into excel since reading CJs post about reactance but no idea what the number should mean, since reactance = (-)resistance at the -3dB point and beyod that goes to all kinda of numbers.

Not exactly; reactance is not the easiest entity in this matter. Just consider the impedance of the cap. For a 1st order filter, the -3dB point happens when the impedance of the cap is equal to the resistance.

So let's say we're using a 1uF cap and 2K resistor, that would make the F = 79.6 Hz and X=-2000 right? But if you start plugging in different F to the Reactance formula you get all kinds of numbers.. Where is reactance(X) high enough or not high enough? As F goes up, X goes down, but not sure where it becomes a problem...

There is no specific problem with the actual value of the impedance, only when it increases, there may be associated effects that must be analysed.

Keeping in mind a 1:7 to 1:10 input trafo (I am using Cinemag 75101, 1:10, with 150K RL, for ~1K5 input Z)... would these numbers work, or would it be better to use a higher R and lower C?

The problem is the inductance of the fransformer. If you under load the secondary (or the source impedance is too high) the inductance becomes dominant and the filter becomes second-order.

If I can get the C under .033uF I can get 1% Wima caps, but I'd have to use 33K to get it down to 146Hz...so not sure if it's feasible..

For a 1.5k input Z and a 80Hz turnpoint, you need 1.5 uF; in fact since you have two caps in series, you need two 3uF caps. Unless you have a split primary with the cap in between.

Also, for a second order, you can't just cascade 2 filters can you?

Well, you may, but using only R's and C's, you will achieve only overdamped filters, that means the slope will not be as sharp as LC's or active, but it has been done with success in many HiFi preamps.

Could this be put on the output and would it have the same/similar problems?

Do you mean passive between the preamp and the xformer? Not a good idea... Will be dependant on output load and the xfmr won't like it; big infrasonic hump, very bad for THD...

Chae's point about headroom is valid, but if we did not take that into account? 

This must be put in perspective with the actual conditions; this may be a problem if you wanted to do recordings in the middle of an earthquake, but generally VLF components are sufficiently low level to not impair the overall headroom.

Going back to Chae's point on putting it between the I/T and opamp... This would go after the zobel, RL etc. right? Wouldn't the RL effect the filter since it is going to ground? Or the other way, wouldn't the RC filter affect loading on the trafo since it has an R to ground?

The input resistance is part of the filter. You have to take it into account for you calculation of the cap.

 

[mitsos]

wow I wish there was a headspin emoticon!

For a 1.5k input Z and a 80Hz turnpoint, you need 1.5 uF; in fact since you have two caps in series, you need two 3uF caps. Unless you have a split primary with the cap in between.

Is this one 3uF cap on each input with a 1K5 to ground?  Or would this be for the cascaded filter? I admit I don't get the split primary with cap in between part.

Well, you may, but using only R's and C's, you will achieve only overdamped filters, that means the slope will not be as sharp as LC's or active, but it has been done with success in many HiFi preamps.

So this can be done... I assume the slope would be somewhere between 6 and 12 dB for such a second order filter? 

Any suggestions for where I can read about this stuff more in depth?

 

[abbey road d enfer]

wow I wish there was a headspin emoticon!

For a 1.5k input Z and a 80Hz turnpoint, you need 1.5 uF; in fact since you have two caps in series, you need two 3uF caps. Unless you have a split primary with the cap in between.

Is this one 3uF cap on each input with a 1K5 to ground? 

No, that would be with 1.5k differential impedance, no additional R's.

I admit I don't get the split primary with cap in between part.

Check the attached schematics

Well, you may, but using only R's and C's, you will achieve only overdamped filters, that means the slope will not be as sharp as LC's or active, but it has been done with success in many HiFi preamps.

So this can be done... I assume the slope would be somewhere between 6 and 12 dB for such a second order filter? 

Exactly

Any suggestions for where I can read about this stuff more in depth?

You have to read and analyse as many schematics as you can, and ask here if you nedd an explanation. Genuine audio designers are not really secretive, but they want you to find out by yourself what they have invented.

 

[mitsos]

thanks for the visual aid! helps a bunch... to my untrained eye the first (2 caps + trafo) looks like the most straightforward solution. Gonna have to try that one out.

I'm still unclear about the split primary one.. how does that work? Shouldn't the cap be in series?

I've been checking all the schematics I can find, most new ones are active, or are channelstrips and HPF is in the EQ section.. One interesting one is the old BBC AM9, not sure how that one works. It's got an R and C in parallel, in the middle of the circuit.. doesn't make sense, does it?

 

[okgb]

i recall the peavey tube pre's block diagram showing a [ two pole ?]
one stage before the tube input and another stage inbetween
the tubepreamp stages

 

[abbey road d enfer]

I'm still unclear about the split primary one.. how does that work? Shouldn't the cap be in series?

It IS actually in series: signal goes to 1st winding, then to cap and ultimately through 2nd winding.

I've been checking all the schematics I can find, most new ones are active, or are channelstrips and HPF is in the EQ section..

I remeber Sennheiser had a location mixer with that type of arrangement, but it's not really frequent.

One interesting one is the old BBC AM9, not sure how that one works. It's got an R and C in parallel, in the middle of the circuit.. doesn't make sense, does it?

Yes, interesting. But the Bass Cut is not the cleverest arrangement, IMO. Slope is a meager 3dB/oct, with only 8dB attenuation of infrasonics. Simulation is not in accordance with the comment written on the dgm. I see a 430Hz turnoff frequency instead of 200. Maybe the intent was not to filter out pops and rumble;maybe for compensation of proximity effect????

 

[mitsos]

ok, yeah I see now it's in series, but it wasn't really making sense. Would the trafo impedance ratio be the same in this scenario?

...gonna keep looking at this stuff.. I'm gonna get some 312 boards done, and will try to test some of this stuff there..

thanks all!!

 

[Crusty2]

Going all the way back to an idea near the beginning of the thread - how about switching the cap in the shunt leg for your roll-off, but alter your gain in the feedback loop with a good quality pot or switch...?