How do I change this to a single supply filter?

[maor]

Hello everyone.

Thanks to one of the contributers here, I used the TI filter program to make a good filter for a friend, but now, I'm trying to avoid using a dual supply so it can work on an effects pedal power (9v adapter) But the only problem Is I can't manage to figure out how to bias it correctly.
I tried changing all grounds to the bias voltage and changing the negative supply to a ground but that didn't work, I tried putting a 1M ohm resistor before biasing the amp at the v+ supply but didn't work either.
I could really use some help.
Thanks in advance,
Maor.

 

[indulis]

Everything that was "GND" should be biased to 4.5V and pin 4 of the opamp should now tie to GND instead of -9V

 

[maor]

But I already did that and it didn't work!

 

[indulis]

And exactly how did you bias things up to 4.5V? You can't just use a resistor divider.

 

[indulis]

Sorry... with the 14 pin package, it's pin 11 that ties to GND

OH... also don't forget the input source return has to be biased up to 4.5V as well.

 

[Gary B]

I’ve done this kind of thing successfully using two incandescent lamps as a voltage divider. Produced an awful lot of heat so I went to two zeners and two resisters to build a symmetrical reference. The junction between the zeners is the ground reference point. The top and bottom of the zeners are the + and – power rails. For this to work, the input voltage has to be a little more than the chip needs. Since you need +9 and -9 volts for the chip to work with the values given, you need a 20 to 24 volt source.

 

[maor]

Hey thanks, the source return bias fixed it.

 

[Nigel Goodwin]

And exactly how did you bias things up to 4.5V? You can't just use a resistor divider.

You can most of the time.

 

[indulis]

Only if you don't sink/source too much current from the divider node.

 

[Nigel Goodwin]

Only if you don't sink/source too much current from the divider node.

Which you would only rarely be doing.

 

[indulis]

Like in this case where your biasing up 3 resistor dividers

 

[Nigel Goodwin]

Like in this case where your biasing up 3 resistor dividers

I see no particular problem with that circuit? - although the pointless dividers could do with removing.

 

[audioguru]

Why not simplify it and make it like this:

 

[Hero999]

Well I agree that the potential dividers can be removed and the gain reduced.

However, for educational purposes, here's how to convert it to a single supply with the potential dividers left in.

It's pretty simple, double the value of the resistor and use two, forming another potential divider across the power supply. This works because the impedance looking into a potential divider, is equal to both the divider impedances in parallel. I've rounded all the values to the nearest E24 value, for example for the 4.7k resistors, double that gives 9.4k, the nearest E24 value is 9.1k.

Another potential divider and a capacitor is added to all the other points of the circuit which are normally connected to 0V.

 

[schmitt trigger]

Something that I've used in the past with great success for exactly this type of an application is a "virtual ground" IC, like the following device from TI:

https://www.electro-tech-online.com/custompdfs/2010/05/tle2425.pdf

The voltage is split very accurately, with a low output impedance and low noise, while drawing a minimal amount of current.

 

[audioguru]

The opamps have Fet inputs so the bias voltage does not need to be a low resistance. The bypass capacitor I used has a low impedance.

 

[Hero999]

I forgot to add AC coupling capacitors to the input and output - I've edited the schematic.

 

[maor]

Uhm, Trying to bias without connecting input return to bias doesn't work.
It seems that changing all the grounds to bias only works properly.
Also, the circuit changed, the whole point is to cut the 400Hz and the 2.5Khz points.

 

[audioguru]

The "half the supply voltage" DC bias voltage is needed only at the (+) inputs of each opamp. R3, R5 and R8 must have a very low impedance connection to ground which is best if they connect directly to ground. The value of your filter capacitor C13 is much too small to make the bias voltage without a signal so everything in your circuit that connects to the bias voltage have signals all mixed up and the spec's for the filters are also messed up. Maybe that is why your circuit works only when the input return is also connected to your high impedance bias voltage.

What feeds this circuit? It must have a very low impedance like the output of an opamp.
Why are you using bandpass filters when you actually want notch filters?

 

[maor]

I used the TI filter pro to make the design of this filter, I set it to notch, why do you claim that its a bandpass?
also, what do you mean by feed, if your referring to what powers the filters its a 9v at 200mA ac adapter, the common one for effects pedals, else I have no idea what you mean.

So what should I do in this situation?