The basics were all that I was asking for. I'm sure all that information is available on the net somewhere, but it isn't that easy to find. Feeding me bits n' pieces doesn't help.
I spent most of last night looking at LOW pass filters like you suggested, thinking I needed to pick some really high frequency like 50 kHz and design a filter to pass everything below that, not 2 Hz like Mooly suggested.
Yes it does...thank you!
A few questions:
1) So if I want a gain of only 2, then I would use 10K resistors for Rf, Rg and from the input to ground, correct?
2) How did you calculate what size cap that goes in series with Rg for a -3dB point of 2Hz? Can you please share the formula?
3) Is the 100 ohm resistor still necessary on the + input for stability?
4) When calculating the input cap's value, which resistance would I use? The 10K going to gnd, the 100 ohm or the value of the volume pot?
Attachments
DC blocking capacitors on both the input and the negative feedback leg.
Gain set by the ratio of Rf : [Rg+Xc of cap]
one input filter set by cap & 10k
That's a good start.
As said earlier determine the midband gain (I like to call it passband gain).
Initially ignore all the capacitors and determine the passband gain
Then use F-3dB = 1 / [2PiRC] for either a high pass filter (the LF limit of the passband), or for a low pass filter (the HF limit of the passband).
The low pass filter is missing. Add one to the input.
For lowest output offset, check that the "resistances" seen by the +IN pin match the "resistances" seen by the -IN pin.
These may need a very small adjustment to compensate for input offset current inequalities.
Gain set by the ratio of Rf : [Rg+Xc of cap]
one input filter set by cap & 10k
That's a good start.
As said earlier determine the midband gain (I like to call it passband gain).
Initially ignore all the capacitors and determine the passband gain
Then use F-3dB = 1 / [2PiRC] for either a high pass filter (the LF limit of the passband), or for a low pass filter (the HF limit of the passband).
The low pass filter is missing. Add one to the input.
For lowest output offset, check that the "resistances" seen by the +IN pin match the "resistances" seen by the -IN pin.
These may need a very small adjustment to compensate for input offset current inequalities.
Last edited:
I'm confused now...
I thought the image in post #15 is a low pass filter, or is it a high pass? If not, what does a low pass filter look like?
Okay, I understand that I have a high pass filter in the circuit as I have it drawn.
What does a low pass filter look like?
There are two reasons you might.
1: the load you are driving might be low. so it might be forming a low pass filter that will mess with the sound. So you use a larger capacitor to assure that the corner frequency remains very low.
2: the load you are driving might also be capacitor coupled. series capacitors halve their value. see above wrt frequency response.
I really prefer to design out the offset and DC couple my outputs. This isn't always realistic, though, and in some designs would require hand-matching resistors to fractions of an ohm and other annoying things.
No worries folks...
1/ Yes, correct.
2/ I cheat and use a nifty circuit calculator pinned to Internet Explorer
calculators for electronic circuit design
To work it out long hand you take the formula Fc=1/2 π rc and rearrange for C which is C=1/2πrf. You know R because you have worked it out for the gain you need and you know f because that is your chosen frequency where you want the output to have rolled off by 3db.
3/ For 99.9% of opamps, no its not necessary.
4/ Use the 10k.
Mooly, thank you for your replies and the link to the nifty calculators.
Can you please shed some light on what a low pass filter would look like on the input of my AD825 circuit? Is it just the opposite of a high pass where you have a resistor in series and a capacitor to ground?
I thought placing a small cap across Rf limited the bandwidth of the op amp?
Can you please shed some light on what a low pass filter would look like on the input of my AD825 circuit? Is it just the opposite of a high pass where you have a resistor in series and a capacitor to ground?
I thought placing a small cap across Rf limited the bandwidth of the op amp?
The top one's not a low pass, rather a high pass. A low pass on the input would have the capacitor to 0V, you have the resistor to 0V. So in that schematic you have two high passes.
<edit> To address your earlier question, yes a low pass has a resistor passing the signal (this is called the series element) and a capacitor to GND (the shunt element). So it is indeed the opposite of a high pass - the two components (R and C) are swapped. Yes to the last question, a feedback cap has the effect of rolling off the gain of the circuit (not strictly that of the opamp, that remains the same). The small cap across Rf is another example of a low pass filter.
<edit> To address your earlier question, yes a low pass has a resistor passing the signal (this is called the series element) and a capacitor to GND (the shunt element). So it is indeed the opposite of a high pass - the two components (R and C) are swapped. Yes to the last question, a feedback cap has the effect of rolling off the gain of the circuit (not strictly that of the opamp, that remains the same). The small cap across Rf is another example of a low pass filter.
Last edited:
Initially I was told I needed a cap in series with the + input to stop DC from reaching the volume pot's wiper.
If the capacitor is going to ground as in a low pass filter, won't DC go through the input resistor and right into the volume pot's wiper resulting in the ole' "swishing" sound when rotating the pot.
So in the end, which filter is really more important in my posted circuit as I'm getting tired of dwelling on RC filters.
Because I don't like caps in the signal path...especially two.
If I knew which filter was more important, I wouldn't be asking now would I??
The 'in the signal path' thing's just a myth because current flows around loops, not along 'paths'. If you're really serious about avoiding caps then any decouplers you've got on the supply rails (including reservoir caps you may have after a rectifier) will have to come out because the current flows through them too.
The low pass at the input's for taking out RF (but not over a very wide band of frequencies due to various circuit parasitics) - if you're not concerned about losing dynamics then you can omit it.
The low pass at the input's for taking out RF (but not over a very wide band of frequencies due to various circuit parasitics) - if you're not concerned about losing dynamics then you can omit it.
Do you have any idea how many capacitors are in the signal path before the recording you are trying to listen to?
Hundreds. Possibly thousands.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.