Breadboarding different filter circuits help

Started by SpringbokUK, February 15, 2019, 09:19:02 AM

Previous topic - Next topic

SpringbokUK

Okay so I have a high pass rc filter that's 723hz cut off. If I connect it at the output of a transistor it does its job and cuts those frequencies below 723hz, great! But if I connect that little rc circuit to the middle lug of my volume it does the opposite and cuts everything above. I thought for that to happen I'd have to swap the resistor and cap round?

Unless when I connect it to the volume pot it's actually bringing the cutoff frequency down because lug 1 is to ground? Baring in mind lug 3 is connected to 2 resistors, 1 coming off a treble control and the other coming off bass control and they're coming off the transistor output. So really my HPF is 2nd order I think because of the treble control as well...But anyways yeah what's going on here?

PRR

> ...But anyways yeah what's going on here?

Schematics with values.
  • SUPPORTER

Rob Strand

Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

SpringbokUK

Quote from: PRR on February 15, 2019, 02:35:47 PM
> ...But anyways yeah what's going on here?

Schematics with values.

I'll draw one up later on. I was just being lazy trying to describe, ha.

Quote from: Rob Strand on February 15, 2019, 03:36:57 PM
You have a loading problem for sure.


Not quite sure I understand here mate but I'll post a schematic later when I'm free.

Keep in mind that my theory isn't great neither haha.

SpringbokUK

#4
Right so here's what I got.
Its the smooth jfet overdrive from wamplers book an  in place of the tone control I've added in AMZ's dual tone control.

At the bottom there ya see the high pass filter that I want to use and when the input is coming from C8 it works fine but I just tried it at lug 3 of the vol pot to see what happens and it seems to do the opposite

Sorry for how messy this is. Ignore the scribble after c7 that was just a mistake I made

If ya need any values let me know like I say its a hard to see



Rob Strand

#5
QuoteRight so here's what I got.
That's kind of what I thought.

When you look at each section they each make sense but when you put them together they don't behave as expected.  This is known as "interaction".  When you look at each section alone they don't interract.    To force the circuit to behave as if there is no interaction you would need a buffer between the volume pot and the filter.   Sometimes you do actually need the buffer, however, there's a way to choose the filter resistor to reduce the interaction.

You have a 100k pot and a 10k resistance in the filter.   The fact the 100k is large compared to the 10k is the source of the problem.   The 10k loads down the 100k making the interaction strong and the circuit doesn't behave how you think.

The pot actually looks like one quarter the pot value, ie. 25k (why requires more technical explanations).   A rough rule of thumb the load, in you case the filter, should be *more than* 10 times the impedance of the thing before it, ie the volume pot.   So 10 times 25k is about 250k.    In a squeeze you might get away with 5 times; 5 times 25k is 125k.   So to avoid interactions your filter needs to use a  resistor larger than 250k, maybe get away with down to 125k.    If you wanted to be heavy handed you could go 1MEG.  However, then you will get other problems the 470k to 1Meg amplifier input impedance starts to effect the filter.   So your filter is stuck between two points sources of interaction, one on the volume pot and one on the amp load.   The "best" value for the filter resistor is probably around 150k.

All you need to do is to adjust you cap so the filter works the same with 150k as it did with the 10k.   If you increase the resistance by a factor of 15 (150k/10k) then you need to decrease the capacitance by a factor of 15, so 100n/15 = 6.7n which means use 6.8n.

The story isn't quite complete because in the original circuit the JFET stage connecting to C8 has an output impedance depending on the trim pot setting.  You might find the filter is more like the filter resistor is 50k instead of 10k.   That means the cap needs to be only 3 times (3 = 150k/50k) smaller, so a cap value 100n/3 = 33n might sound more like you had before.

Instead of doing a lot of calculations, you can take a simpler approach to the whole thing:
- the filter resistor needs to be bigger than the pot value (but not too big)
- play with the cap so it sounds like it did before

EDIT:

A far better solution is to just put the cap between the 47k's and the volume pot.  The 100k pot now takes the place of the 10k filter resistor.

Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

SpringbokUK

Ah right I see. I'm guessing the 100k volume looks like 25k because of all the other sections of the circuit??

The 470k you're referring to is that R6 or R7? And the 1Meg amp input impedance you're referring to is the trim pot???
You'll have to forgive me because I'm kind of new to all this in-depth stuff so I'm trying to wrap my head around out without trying to seem hopeless. This whole thing actually came about when I was trying to add a mid scoop to the circuit for a more meatier overdrive. At the moment it's sort of like a fuzzy overdrive. That's probably because of the active pickups I'm using though.

As far as biasing goes the original schematic from the book said to use 100k trims but for some reason 100k was too large for Q1 and too small for Q2. I guess it's fine now.

Times like this I wish I had an EQ spectrum analyser.

If you wouldn't mind, could you perhaps link me to a website or give an explanation of why the 100k is actually kind of 25k?

And if I was to put that 100nf/10k combination at the output cap C8, does it act as a first order HPF and the treble control acts as a 2nd order HPF??

I'm really trying to understand all this because ultimately I want to glide through when I start designing properly. Don't want to be stuck in the clone zone forever, ya know?

Thanks for your response!

SpringbokUK

Quote from: Rob Strand on February 16, 2019, 05:33:24 PM
EDIT:

A far better solution is to just put the cap between the 47k's and the volume pot.  The 100k pot now takes the place of the 10k filter resistor.

Also would this not have interaction as well? With the 47k's in parallel making the overall resistance smaller and less than the value of the pot? Sorry if I'm missing a point here!

Rob Strand

#8
QuoteAh right I see. I'm guessing the 100k volume looks like 25k because of all the other sections of the circuit??

The 470k you're referring to is that R6 or R7? And the 1Meg amp input impedance you're referring to is the trim pot???
You'll have to forgive me because I'm kind of new to all this in-depth stuff so I'm trying to wrap my head around out without trying to seem hopeless.
Sometimes when you start out it's best not to think of this stuff too much.  You are best understanding the basic blocks like voltage dividers, high and low-pass filters, biasing.    The fine details can really bog you down.   However, now and then you will get caught out and it's the fine details like interactions which stuff everything up for you.  Over time you learn when you can ignore it and when you have to take it into account.

For the 25k thing.  Imagine the pot set midway so it forms a divider made from two 50k resistors.  The impedance looking into the output terminal is actually the two 50k's in parallel.  Two 50k's in parallel is 25k.
There's some notes here,

https://www.reddit.com/r/electronics/comments/1dx35x/thevenin_equivalent_of_a_simple_voltage_divider/

I kind of simplified things at the 25k level.  In reality we need to account for the two 47ks connected to the pot and also the tones control.  If you really want to be picky the resistance depends on where the volume pot is set.  It goes on and on with details.  Too many details to deal with when you start out.   The important thing is the pot looks like some sort of high "equivalent" resistance and that was the cause of your original problem. 

For the 470k to 1MEG thing.  An amp might have an input impedance of say 470k to 1MEG; it will vary from amp to amp.   When you ultimately connect this pedal to an amp, the amp input impedance will load down the pot.   Now if you just have the pot there the amp impedance will just drop the level and you naturally crank the volume up to what ever you want, so you don't care.  However, if you have a filter sitting at the output you don't want amps with different input impedance loading down the filter, which then shifts the filter cut-off and changes the tone.   It's just and extra variable you can do without.

QuoteAlso would this not have interaction as well? With the 47k's in parallel making the overall resistance smaller and less than the value of the pot? Sorry if I'm missing a point here!
The main point is the pot already provides a resistance to ground (more or less taking the place of the 10k in you filter).  So instead of adding the filter which adds one more level of interaction, we just use the pot for the "10k" (except it's 100k).   The 47k's do have an effect but we can allow for that.    The two 47k's and the bass and treble pots will end-up looking like about 33k (that's the Thevenin impedance look into the point where the two 47k's meet).   For the HPF cut-off the 33k *adds* to the 100k, so the 10k in your filter now looks like 100k + 33k = 133k.   Those last two sentences might start bending your mind.  Again it's more techy details you probably shouldn't be thinking about; I could explain it but it might not mean anything to you.

Here's the steps which might help you see why you add the resistors on the high-pass filter,

Click twice for detailed image,

Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

SpringbokUK

Haha cool. I think I'm sorta there for understanding the bigger picture of it all.

The fine details absolutely bog me down. I'm a guy that can't settle until Im satisfied with my results and I can produce those results with conviction. Unfortunately it's hard to do that with those fine details that get in the way and wreck my head.

Its easy enough to follow a schematic and produce a clone but to understand it all seems a far venture. Like you finish a clone, get to know components, values, don't question the "how" and just do it but there doesn't seem to be a stepping stone from that, just a big massive leap to the advanced fine detail stuff that I can't make haha.

The original wampler circuit had an additional cap and resistor in series before the gate of Q2 but for some reason that sent the bias wild so I removed them. And apparently you can make a bass control out of the resistor and cap in parralel before that. Instead of resistor, use a pot and I did that and nothing happened. Controlled nothing.

For now I'm just going to really on my ears and bookmark every thread on here that has useful info.

Thanks for the help man :)