Quick question, have I inadvertently created a variable low pass filter by placing an RC network after the volume control potentiometer? I know its best to place the low pass filter right at the input jack but this was an afterthought to prevent one of my first ever amps from picking up a strong local radio station.
Does the variable resistance of the top half of the pot alter the cutoff frequency or does it not matter when configured as potential divider.
The attached drawing shows what I currently have going on.
Thanks.
Does the variable resistance of the top half of the pot alter the cutoff frequency or does it not matter when configured as potential divider.
The attached drawing shows what I currently have going on.
Thanks.
Attachments
> have I inadvertently created a variable low pass filter
Yes. Worst-case, 3,500r against 470pFd, is 96KHz. Few-tenths-dB at 20KHz.
Why can't it go before the volume pot?
Yes. Worst-case, 3,500r against 470pFd, is 96KHz. Few-tenths-dB at 20KHz.
Why can't it go before the volume pot?
> Where does the 3.5kohm value come from?
I assume the "PC/Phone" is zero impedance (near enough).
I assume the 1uFd is also "zero impedance" at least above the bottom bass.
If the pot is full-up, your added 1K is driven from zero (Phone+1uFd).
If the pot is full-down, your added 1K is driven from zero (ground return).
The worst-case is with pot at half-way. Then a 10K pot is 5K to the top in parallel with 5K to the bottom. 2.5K. Plus your added 1K, is 3.5K.
Check: the "Phone" is probably not much over 32 ohms output (since it has to drive 32r with good efficiency). This is 30X smaller than any resistor in pot or added for the high-cut. Assuming you do not need 3% precision in high-cut frequency (we probably never do), it is "zero enough".
I assume the "PC/Phone" is zero impedance (near enough).
I assume the 1uFd is also "zero impedance" at least above the bottom bass.
If the pot is full-up, your added 1K is driven from zero (Phone+1uFd).
If the pot is full-down, your added 1K is driven from zero (ground return).
The worst-case is with pot at half-way. Then a 10K pot is 5K to the top in parallel with 5K to the bottom. 2.5K. Plus your added 1K, is 3.5K.
Check: the "Phone" is probably not much over 32 ohms output (since it has to drive 32r with good efficiency). This is 30X smaller than any resistor in pot or added for the high-cut. Assuming you do not need 3% precision in high-cut frequency (we probably never do), it is "zero enough".
Last edited:
you have choices.
add a Buffer stage after the vol pot. It drives the cable and any subsequent capacitance.
and/or reduce the vol pot impedance.
and/or reduce the attenuating capacitor.
move the capacitor to the input socket.
If some, or all, of these are not attractive enough, then start again.
add a Buffer stage after the vol pot. It drives the cable and any subsequent capacitance.
and/or reduce the vol pot impedance.
and/or reduce the attenuating capacitor.
move the capacitor to the input socket.
If some, or all, of these are not attractive enough, then start again.
You have omitted the DC blocking capacitor between the chipamp input and the vol pot wiper.
When you add that you will need a bias resistor from chipamp input to audio ground.
The 1uF and 10k vol pot are also acting as an accidental high pass filter.
Try a pair of back to back 22uF electrolytics to replace the 1uF. That restores the low bass performance of the input stage.
When you add that you will need a bias resistor from chipamp input to audio ground.
The 1uF and 10k vol pot are also acting as an accidental high pass filter.
Try a pair of back to back 22uF electrolytics to replace the 1uF. That restores the low bass performance of the input stage.
I'm really struggling to understand what's going on 🙁, if I turn the volume pot down it gets quieter whilst near the top it gets louder. But the way you describe makes it sound like it should get louder again when turned down. When the pot is fully down the output of the voltage divider should be near 0v.
At this point I'm just tying to understand what is happening electrically because its not what I originally thought was going on. Also when you say "add a Buffer stage after the vol pot. It drives the cable and any subsequent capacitance" how does a later stage in the chain drive a cable capacitance that's feeding it? Wouldn't that be down the audio source?
But doesn't the 1uF at the aux input take care of that? Is it better to add the DC blocking cap after the pot at the chips input pin?
Doesn't that stay at a fixed value though because its just 1uf + 10k to ground or does that also depend on the wiper position?
This stuff is more confusing than you'd think lol.😱
Did you read this "sticky" Thread?
http://www.diyaudio.com/forums/anal...iometer-passive-preamplifier.html#post3157081
Did you read ESP's
Potentiometers (Beginners' Guide to Pots)
Designing Our First Filter
http://sound.whsites.net/dwopa2.htm
http://www.diyaudio.com/forums/anal...iometer-passive-preamplifier.html#post3157081
Did you read ESP's
Potentiometers (Beginners' Guide to Pots)
Designing Our First Filter
http://sound.whsites.net/dwopa2.htm
Last edited:
I've already read those but they don't really give you a solid solution that I can just get on and solder in place. I know that everything basically acts like one big RC filter, but without getting into op amps I'm just looking for what values to use since I can't understand the explanations given ie why the worst case scenario for the pot is the halfway point. I've been thinking about that all day now but I just can't seem to get my head around it. 🙁
My configuration is this: PC/phone headphone output>1uF DC blocking capacitor>10k volume pot>1k 300pF RC filter to stop rf>amp input. Lets just say for example I was using the LM386 and wanted the above chain, what would be the best values to choose so all the frequency filtering going on wouldn't eat into the audio spectrum too much.
I 100% need the RC filter in there as it picks up radio without it plus my laptop likes to output a 250khz sinewave for 30 seconds after playing sound.
I have the thing built already and do notice changes in dynamics if the volume control (10k) is changed, I can even see it with an oscilloscope if I feed it a 1hz-20khz sweep and put it in roll mode. Its not much but its definitely there so if I can learn how to avoid it on smaller amps like this then I can make the changes on bigger amps in future (its more noticeable on the LA4705 amp I built).
The output impedance of a pot is equal to the upper resistance in parallel with the lower resistance. This reaches a maximum at electrical mid-position, when the output impedance is equal to one quarter of the total track resistance. If you use a log pot, which is normal for a volume control, then electrical mid-position is just below maximum mechanical position.
Hi, this is the part that I don't understand.
Why is this so and why does the lower half of the pot behaves that way. Wouldn't the RC low pass filter have the most effect when the pot is turned all the way down presenting 10k in addition to the RC low pass filter?😕
Attachments
Last edited:
When the pot is turned all the way down you get zero voltage from a zero impedance, then fed into a 1k and 470pF filter. The same when the pot is at maximum (assuming zero source impedance - an approximation but often a valid approximation).
When the pot is halfway (electrically) you have a 5k in series with the signal and 5k in shunt with the signal. This gives you half the signal from a 2.5k impedance. The reason it is 2.5k is that both the 5k resistors connect to zero impedance, so looking back in from the pot output you see two 5k in parallel.
When the pot is halfway (electrically) you have a 5k in series with the signal and 5k in shunt with the signal. This gives you half the signal from a 2.5k impedance. The reason it is 2.5k is that both the 5k resistors connect to zero impedance, so looking back in from the pot output you see two 5k in parallel.
But wouldn't only the top half of the pot have an effect as that is the path the signal is taking to the input of the chip? How does it appear in parallel when its always presents an adjustable 10k in series with the audio input pin?
Still very confused sorry 🙁 Maybe some kind of laymans term analogy would help me understand better, because I would really like to get this thing sounding good.
Last edited:
> the way you describe makes it sound like it should get louder again when turned down.
"If the pot is full-down, your added 1K is driven from zero (ground return)."
Turn-down feeds "ground" (silence) to the wiper.
As for the resistance around a pot at several settings: if Andrew's reading-list, and DF96's comments, are not helping, then get an ohmmeter and some clipleads and measure. For wiper resistance, clip the two ends together (low impedance source across the ends) and measure the resistance from wiper to one end. Turn. This is the resistance "seen" by your 1K resistor.
___________________________
> add a Buffer stage after the vol pot. It drives the cable and any subsequent capacitance.
Aside complexity-- if his amplifier rectifies radio, won't the buffer also rectify radio?
Yes, he may have a poor amplifier and could build a radio-resistant buffer.
He's clearly not yet ready to design that.
His R-C solution apparently does work. And should. The variable frequency is not a real problem because the worst-case hardly shaves the top of the audio band (better than catching Peruvian radio when you want disco). His issue is that he's not sure how to calculate the frequency. It's tough the first time you venture into Network Analysis.
"If the pot is full-down, your added 1K is driven from zero (ground return)."
Turn-down feeds "ground" (silence) to the wiper.
As for the resistance around a pot at several settings: if Andrew's reading-list, and DF96's comments, are not helping, then get an ohmmeter and some clipleads and measure. For wiper resistance, clip the two ends together (low impedance source across the ends) and measure the resistance from wiper to one end. Turn. This is the resistance "seen" by your 1K resistor.
___________________________
> add a Buffer stage after the vol pot. It drives the cable and any subsequent capacitance.
Aside complexity-- if his amplifier rectifies radio, won't the buffer also rectify radio?
Yes, he may have a poor amplifier and could build a radio-resistant buffer.
He's clearly not yet ready to design that.
His R-C solution apparently does work. And should. The variable frequency is not a real problem because the worst-case hardly shaves the top of the audio band (better than catching Peruvian radio when you want disco). His issue is that he's not sure how to calculate the frequency. It's tough the first time you venture into Network Analysis.
Last edited:
Ignore almost everything you may have read about 'signal path'. Both parts of the pot affect the signal. Instead, make sure you really understand voltage and current and Ohm's Law.Signaling said:
The 10k is not an adjustable resistor. It is an adjustable potential divider. Very different. Learn about potential dividers. Only then can you understand a volume pot.
- Status
- Not open for further replies.