Parasit Baxandall tonestack debugging (SOLVED)

[mdcmdcmdc]

Hi folks,

I'm hoping that someone here might be able to help me figure out what's going on in a layout I built up a few nights ago (I posted a similar request for help over on the tagboard forum, so apologies if you've bumped into this same thread in two places).

The layout is Fredrik Lyxzén's active baxandall layout from his Parasit site:
https://www.parasitstudio.se/stripboard-layouts/active-baxandall-tonestack

Tthere's a discussion and a few variations in a thread here (the schematic for the design is also on this page):
http://guitar-fx-layouts.42897.x6.nabble.com/Bax-in-a-box-active-baxandall-tonestack-td7264.html

There are a couple of component changes between the version posted in the first link (parasit) and the versions from the forum discussion, and I built the version he posted on parasit with no substitutions or other mods. The circuit works for the most part and sounds really nice aside from one odd issue that I can't figure out.

- When the TREBLE knob gets to about 85-90% CCW (say, 7:30-ish) there's an audible 'pop' and the circuit cuts out completely from 85% CCW all the way to fully CCW (ie, treble completely cut). If I sweep up the range of the pot, it'll pass that same point and pop back on, signal returns.

- If I bypass the circuit and then click the footswitch again **with the treble pot in the problem range** the sound returns and i'm able to sweep up the full range of the pot. When I then sweep down and hit that same spot, the pop returns and the circuit cuts out.

- Just for the sake of repetition, that's treble knob at full CW= sound; sweep down to 85% CCW=pop and no sound; 85-100% CCW= no sound; cycle the 3PDT footswitch= sound; sweep up to 100% CW = sound; sweep down to 85% CCW=pop and no sound.

I've tried:
- replacing the treble pot (no change)
- checking for obvious cold solders or bridged stripboard tracks (nothing obvious)
- replacing the TL072 (tried 3 different ones, no change)
- my very sophisticated debugging technique of just randomly poking parts of the circuit with my finger to see if anything happens...

What I found is that if I bridge all three lugs of the treble pot with my finger, everything works perfectly fine. Would this suggest a grounding issue that might disappear when the circuit is properly boxed up? I'm currently trying it out through through my test box so the stripboard is sitting out in the open on some bubble wrap.

FWIW, the test box seems to work without issue in general, and I've used it without problems for a bunch of stripboard builds.

Voltages from the TL072 with the TREBLE fully CCW and the BASS fully CW:

DMM reads 9.30 at the source.

1 - 4.65
2 - 3.70
3 - 4.65
4 - 0.00
5 - 4.65
6 - 4.65
7 - 4.65
8 - 9.30

Here are some photos:






Normally I'd just start from scratch and see if the problem repeats itself with a fresh build, but since it's such a simple circuit I'd really like to understand what's actually going on.

Thanks very much in advance — apologies for the lengthy post, just want to provide as much info as possible and sort this out.

[mdcmdcmdc]

Well, I rebuilt the circuit on a new piece of vero and the problem repeated itself—connecting lugs 2 and 3 on the treble pot seems to solve the problem. If anyone might be able to help me understand why that's the case I'd be grateful!

[Vivek]

Could you please post a link to the actual schematic that you built (since there seems to be many changes proposed in the article that you linked to)

and also confirm that the Vero layout exactly implements the schematic without any errors.

[anotherjim]

The schematic is in the pcb version build doc...
https://www.parasitstudio.se/uploads/2/4/4/9/2449159/active_bax_eq_pcb.pdf
The stripboard layout around the treble pot connections look ok to me.

[Vivek]

Do the DC voltages on the IC change based on position of the treble knob ?

Also,

I would have expected "-"   pin 2 to have same voltage as  "output " 1 and "+"3 



Just to make sure there is no RF at work here, you could place 100pf caps across output and negative input of each Opamp

[Vivek]

Do the bass and treble controls work as expected when you do have sound ?

Does the circuit have no audible distortion ?

What are the DC voltages on the  treble pot for sound and no sound situations?

What are DC voltages on IC in the no sound position ?

If there is no sound, can you get sound to start just by touching the lugs of the treble pot ?

[anotherjim]

I would have expected "-"   pin 2 to have same voltage as  "output " 1 and "+"3 

Me too. The opamp should auto-correct pin2 to equal pin3 voltage. It looks like the meter load is dropping the voltage on pin2 and the opamp can't drive the output voltage higher to correct it (if you have 2 test meters you could check that). It  suggest there is something wrong in the feedback path from the opamp output back to the inverting input.

[mdcmdcmdc]

Thanks everyone for chiming in - it's greatly appreciated. I'll do my best to answer your Qs. Also I'll add the caveat that my DMM is very low-quality, so take the readings with a grain of salt. They're reading consistently, but they may not be accurate.

For reference, with the circuit disengaged, I'm getting the following voltages from the IC:
1 - 4.65
2 - 4.65
3 - 4.65
4 - 0.00
5 - 4.65
6 - 4.65
7 - 4.65
8 - 9.30

Could you please post a link to the actual schematic that you built (since there seems to be many changes proposed in the article that you linked to)

and also confirm that the Vero layout exactly implements the schematic without any errors.

This is the layout I followed:


...which is based on this schematic:


The layout changes the values of C3 and C4, which I believe alters the centre-point of the treble/bass split.

Do the DC voltages on the IC change based on position of the treble knob ?
Also, I would have expected "-"   pin 2 to have same voltage as  "output " 1 and "+"3 

Just to make sure there is no RF at work here, you could place 100pf caps across output and negative input of each Opamp

The voltages do change as I sweep the treble put - at full CW pin 2 reads 4.69, at full CCW it reads 3.78.

To the point about 100pF caps - forgive my inexperience here, but that would mean connecting pin 1 to pin 2 with a 100pF cap, and then pin 6 to pin 7 with a 100pF cap? Would you suggest doing that at the socket/chip or finding somewhere to bridge those together on the stripboard?

Do the bass and treble controls work as expected when you do have sound ?


Yes, there don't seem to be any weird phase issues or anything unexpected. Bass knob boosts or cuts bass, treble knob boosts or cuts treble.
I didn't notice this previously, but I just double checked the sweep of the treble knob and at about 95% CW there's now an audible pop and a lot of RF-sounding noise that intensifies when I touch pin 2 with the DMM.

(NOTE: I just realized that I didn't have an instrument plugged in, so connecting a guitar to the input of the test box cleared up the RF at the high end of the treble pot sweep. Also, I was testing the circuit previously on my wood floor, and it's now sitting on a metal table...)

Does the circuit have no audible distortion ?

None that I've noticed. As per the descriptions I've read of Baxandall EQs, it's a very natural/"hi-fi" sounding EQ.

What are the DC voltages on the  treble pot for sound and no sound situations?


treble full CCW (no sound, guitar unplugged):
1 - 3.95
2 - 3.95
3 - 4.00

treble full CCW (no sound, guitar plugged in):
1 - 3.94
2 - 3.94
3 - 3.83

treble full CW (sound, guitar unplugged):
1 - 4.48
2 - 4.48
3 - 4.70

treble full CW (sound, guitar plugged in):
1 - 4.44
2 - 4.44
3 - 4.65

treble at signal loss (85-90% CCW, guitar unplugged):
1 - 3.97
2 - 3.98
4 - 3.89

treble at signal loss (85-90% CCW, guitar unplugged):
1 - 3.95
2 - 3.96
4 - 3.77

treble at RF noise (95% CW, guitar unplugged):
1 - 4.70
2 - 4.48
3 - 4.48

**RF noise does not occur at 95% CW with guitar plugged in)**

What are DC voltages on IC in the no sound position ?


Treble 85-90% CCW (guitar plugged in)
1 - 3.90
2 - 3.80
3 - 4.63
4 - 0.00
5 - 4.65
6 - 4.65
7 - 4.65
8 - 9.30

If there is no sound, can you get sound to start just by touching the lugs of the treble pot ?


No, I have to sweep the pot past the pop-point (there's a pop in both directions) and then sound returns.

If I bridge the lugs with my finger I get an audible ground hum, but the signal remains from full CW to full CCW. The RF noise at the very top of the sweep is greatly reduced and the pop/silence at the bottom disappears.

As far as I can tell the only way to get sound to return in that bottom part of the treble pot sweep is by cycling the footswitch.


Thanks again for taking the time to respond - it's greatly appreciated!

[mdcmdcmdc]

I stuck a 100pF cap across pins 1 and 2 of the IC and it solved the problem:



So, per your note above, this would indicate an RF issue? Should I expect this to clear up when the circuit is boxed, or should I add the 100pF cap to the circuit permanently?

[Vivek]

I would suggest 100pf between output and inverting input permanently, on both opamps.
These RF protection caps work best if they are as close physically to the Opamp as possible.



And across C5 also

[Vivek]

The input impedance of this circuit is R2 100K ohms.

That not good if a guitar is plugged in.

That's OK if another pedal feeds to this circuit.

[mdcmdcmdc]

Thanks again for the help on this - it's greatly appreciated!

[Vivek]

Next possible steps in the learning process

A) redesign front end to increase input impedance to at least 500k so that a guitar could be plugged in without tone suck.

B) calculate the frequencies used by the Baxandall and check they are suitable for guitar (and not just a hifi Baxandall hastily plugged into the circuit)

C) try to redesign as a slight boost so that you get at least 3vpp  signals at output.

D) add a Mid control ? Presence control ?

If you are interested, we could do this together.

[mdcmdcmdc]

I'd love to - sounds like a great project, and I'm hear to learn.

As far as the frequency response of the circuit, I did take a look at the TSC at http://www.guitarscience.net/tsc/james.htm to see the difference between the values in the first version of the circuit that Fredrik posted in the forum and the revised version that appears on the Parasit site.

However, it looks like his circuit/schematic is not a 1:1 implementation of the schematic in TSC, so I had to do a little bit of guesswork as far as what-goes-where. I'm not sure if this is simply due to the TSC being based on a passive tone stack and Fredrik's version being active.

For what it's worth, the mid-point of Fredrik's implementation actually sounds pretty good on guitar. I think the inherent problem with the bax tone stack is that it boosts EVERYTHING over the middle point, so there's the potential for a lot of high frequencies that don't add anything useful to an electric guitar. Maybe running the output through a HPF and LPF or a bandpass would clear it up a bit.

I'd love to hear your thoughts on adjusting the input impedance of the circuit.

[Vivek]

Does this help to calculate the corner frequencies ?

https://web.archive.org/web/20141227031258/http://www.geocities.jp/dgb_studio/bax_calc_e.htm

[Vivek]

I entered the schematic into LTSPICE

Do you feel the Treble control changes very little in the middle, and has major activity only towards its ends ?


If we place a 1nf cap between wiper of Treble cap and R9, it will prevent the highs from getting boosted so much. Like a low pass somewhat.



Treble control is not so sensitive in the middle.

[anotherjim]

I think the Tech21 bass amp pedals have some good ideas to pick at.
The input amp is properly protected and gives some initial gain. The input impedance is high for guitar/bass use.
The output has a protected opamp buffer/line driver.
The low and high bax filters before the output buffer are preceded by a mid control.
The volume control is before the EQ so can also be used to keep any EQ in boost from clipping.

[mdcmdcmdc]

I entered the schematic into LTSPICE

Do you feel the Treble control changes very little in the middle, and has major activity only towards its ends ?


If we place a 1nf cap between wiper of Treble cap and R9, it will prevent the highs from getting boosted so much. Like a low pass somewhat.



Treble control is not so sensitive in the middle.

I wouldn't say the treble does very little in the middle, but there is definitely a pronounced boost in treble at the very top of the pots sweep.

That calculator was very helpful - thanks, my only issue will be to be able to understand the numbers it outputs versus the relatively easy visual feedback of a graph.

[Vivek]

Here are the calculations and a graph that explains the various frequencies





https://freecircuitdiagram.com/wp-content/uploads/2008/09/baxandal_tone_control_and_formula.gif

It uses almost the same values as in your circuit, but with a slight schematic change near Bass Pot. However that is easy to analyse if you consider how the Bass circuit looks at the extreme travels of the pot.

[PRR]

...Treble...

9dB cut/boost at 1kHz is mighty heavy-handed.

In other fields we aim for 3dB or less @1kHz.

Here, that would mean trying half or 1/3rd the treble cap(s) values.