Help with Low Pass filter

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

k brown

Well-known member
Joined
Jul 7, 2021
Messages
1,250
Location
California
I tried replicating this Hi Cut filter from the Sony ECM-969 to use with other too-bright capsules, but when used with a different FET (and biasing values), it behaved like a drastic Hi Pass instead. I tried isolating it from the FET biasing circuit with a cap, but same results.

I'm guessing the behavior of the circuit depends on a specific source impedance from the preceding FET circuit (or I may have just bolloxed up the filter circuit itself); I.e. - the FET's 75k resistor is functioning also as a part of the filter?

Any suggestions how to modify this to a 'universal' circuit that can be used to experiment with taming bright capsules? The filter values would need to tinkered with to get the roll-off point to match the lift of a specific capsule, but need to get it to work first !

Thanks
 

Attachments

  • Screen Shot 2021-09-22 at 10.03.34 AM.png
    Screen Shot 2021-09-22 at 10.03.34 AM.png
    303.5 KB · Views: 38
Last edited:
That JFET there is just a source follower / unity-gain buffer, so the output impedance should be pretty low. Have you tried simulating that piece of circuit, perhaps?

And/or are you 100000% sure you made the connections right (and/or didn't make connections where there should be none)? Like C1 connecting (only) to the transistor's emitter, not its base?
 
The JFET is biasing the bipolar transistor so you can't change the bias and you can't just put a cap in there. So you'll need to find a JFET that biases vaguely around the one in the schem and then adjust R3 to land on 4V or whatever the transistor needs to bias properly.

That circuit is highly sensitive to bias point, filter resistor and capacitor values, supply voltage, probably temperature and more. You would really need to study it intently and preferably get FFTs to make sure it's working as designed.

But your immediate problem is that the second transistor is not biased properly at all. Replace R3 temporarily with a potentiometer, try different JFETs (ones of the same part number as well because individual JFETs have different bias points) and adjust the pot until you see the source around 4V and maybe listen to hear it start working.
 
I'm not concerned with the FET biasing in this circuit; I don't even have that mic. I'm interested in using this Hi-Cut with other capsules/FETs.

You're saying that because the 75k is a wrong value for the FET, it's also mis-biasing Q2 - that it wouldn't result in Q2 having 3.3v on it's output as shown on the schem?

Could Q2 simply be biased with a voltage divider on it's base, like Q3?
 
Last edited:
Right. The input at R4 of that filter stage is what biases the transistor to be on. So whatever it's connected to needs to be ~4V (3.3V + 0.6 diode drop = 3.9V) so that the emitter is ~3.3V. Assuming the supply is 9V. If it's not, then you would probably use all sorts of different resistor values and capacitors because everything depends on everything else in this circuit. It's not like you can just cut and paste this part of a circuit into another circuit. You would have to understand things a lot better to do that. If you just want to see it work for educational purposes, then you could probably just use a voltage divider on the input of like two 100K resistors and a coupling cap before that. Then you would be able to feed in some signal and see the low cut work.
 
That is a 12dB an octave filter are you sure you want that? It also needs a DC coupled low resistance drive or the output resistance of the stage before needs to be subtracted from the first resistor calculated value. There are other ways to do lowpass filters
 
You might want to look up "sallen key npn" on Google. That's a discrete implementation of a 2nd-order Sallen-Key filter using an NPN transistor. I've used the (switchable) high-pass version in some of my mic circuits.
 
Thanks I'll try that - I Googled second order low pass, simple second order LP, etc. - didn't get much - I'll try including 'Sallen Key'. First order might be better for rolling off the top of a mic, though.
 
Or in case you might want a notch rather than a low-pass or a high-shelf attenuation, you might want to look into npn or pnp gyrators.
 
Back
Top