Turbo Tube Screamer ticking/tremolo issue

[Voltzy]

Hi all,

I have just finished building the TS9DX layout from here: https://tagboardeffects.blogspot.com/2013/07/ibanez-ts9dx-turbo-tube-screamer.html

For the most part the pedal works fine, however I get a strange tremolo effect when I turn the gain up while on the Turbo mode. You can hear it here: https://youtu.be/OCrpNYEmeVg in the video I already have it on turbo mode and increase the drive, then at the end I click it back to "Hot" mode (all the other modes work perfectly).

I noticed at the bottom of the tagboardeffects page someone else had the same issue as I, and fixed it by swapping out the 2.2uF cap with a 1uF cap, and this did fix the problem but it made the turbo mode incredibly dull and lacking anywhere near the power of the other settings, which imo is worse than the tremolo issue.

I measured the voltages on both max Turbo mode gain and also the most I could turn the gain up before the tremolo starts, this spot is about 186k ohms between pins 1 and 2 of the drive pot. Most voltages were moving around quite a lot (I was playing the sample in the vid through the pedal at the time) which I am guessing is corresponding to the loud/quite of the tremolo effect - I tried to find the upper and lower values that they were changing through as best as possible. Q1 is the transistor on the left of the layout and Q2 is on the right, I also triple checked the transistor orientation and there are no part substitutes:

Full Drive

Between signal gnd and circuit gnd - 0
Between signal gnd and circuit +9v - 9.02
Across power plug - 9.02

Q1
C - 8.77
B - 3.6 - 4.2
E - 2.8 - 3.9

Q2
C - 8.77
B - 3.2 - 4.7
E - 3 - 4.3

IC
Pin 1 - 1.5 - 7.5
Pin 2 - 3.8 - 5.2
Pin 3 - 3.8 - 5.2
Pin 4 - 0
Pin 5 - 1.7 - 7.4
Pin 6 - 1.6 - 7.6
Pin 7 - 0.8 - 7.7
Pin 8 - 8.77

D1
A - 1.3 - 7.6
C - 1.5 - 7.8

D2
A - 1.5 - 7.9
C - 1.3 - 7.9

D3
A - 1 - 7
C - 1.5 - 7.9

D4
A - 1.3 - 7.6
C - 1.1 - 7

D5
A - 1.5 - 7.9
C - 1.3 - 7.6

D6
A - 1 - 7.4
C - 1 - 7.1

D7
A - 0.9 - 7.6
C - 1.3 - 7.8

D8
A - 1.5 - 7.9
C - 1.5 - 7.7

D9
A - 9.02
C - 8.78

Half Drive (~186k ohms between pins 1 and 2 of drive pot)

Between signal gnd and circuit gnd - 0
Between signal gnd and circuit +9v - 9.02
Across power plug - 9.02

Q1
C - 8.77
B - 3.96
E - 3.53

Q2
C - 8.77
B - 3.94
E - 3.52

IC
Pin 1 - 4.36 - 4.41
Pin 2 - 4.43
Pin 3 - 4.37
Pin 4 - 0
Pin 5 - 4.36 - 4.40
Pin 6 - 4.36 - 4.41
Pin 7 - 4.36 - 4.41
Pin 8 - 8.77

D1
A - 4.1 - 4.15
C - 4.36 - 4.41

D2
A - 4.36 - 4.42
C - 4.1 - 4.15

D3
A - 3.16 - 3.7
C - 4.36 - 4.42

D4
A - 4.1 - 4.17
C - 3.7 - 3.72

D5
A - 4.17 - 4.2
C - 3.9 - 3.95

D6
A - 3.87 - 3.9
C - 3.65 - 3.7

D7
A - 4.17 - 4.19
C - 4.36 - 4.39

D8
A - 4.36 - 4.41
C - 4.17 - 4.22

D9
A - 9.02
C - 8.77

I noticed that I have some diodes with a higher voltage on the Cathode than the Anode, I double checked these and they are reading those voltages. I have also swapped out the 2.2uFcap for a fresh one to no avail. Here are some photos of the board (the uncut legs are the fresh 2.2uF cap). Since the photo I have re-flowed some of the dodgy looking solders and it has made no difference:






Sorry for the long post but I wanted to provide as much information as possible. Please let me know if there is any more information I can provide.

[Voltzy]

When I first built the circuit I wired up each of the pots and the rotary switch and then just plugged the input, output, 9v+, and ground into my breadboard to test that the circuit works and everything worked fine. It is only since boxing it up that the issue has arisen.

The only things that I changed between the breadboard and the box is the footswitch (no LED), the input and output jacks (I have hard-wired ones on my breadboard), the power jack, and the addition of an enclosure ground (soldered to the back of the volume pot). All of my grounds are starred to the output jack ground (input jack, circuit ground, DC plug ground, and enclosure).

My breadboard has a small power filter circuit between its DC plug and the breadboard consisting of a 220uF and a 100nF capacitor, so I originally thought that it needed this in addition to the 100uF power filter cap already in the ts9dx circuit. I have tried powering the boxed pedal through my breadboard power filter which didnt fix the problem. I have also since added a 100nF power filter cap to the circuit board with no change.

I tried removing the additional enclosure ground and there is also no change.

I don't see how the footswitch or the input and output jacks could cause this issue... If they were not working then I would not get any sound in any mode which isn't the case.

Could it be some small capacitances or resistances inherit in the breadboard caused it to filter or cancel out whatever is causing the issue? I am really clutching at straws.

[GibsonGM]

Some sort of self-oscillation, IMHO...very low freq "motorboating", maybe.  Wires too close to each other maybe; the extreme gain present there means that ANY questionable build practices then get magnified.  The signal is probably coupling back upon itself in a positive feedback loop.   It's slow, I'll give you that! 

Maybe replacing in/out wires with shielded types would help...switch wires too...hopefully it isn't the layout that's causing this...a carefully placed cap or 2 in the right place might help....

[Voltzy]

Some sort of self-oscillation, IMHO...very low freq "motorboating", maybe.  Wires too close to each other maybe; the extreme gain present there means that ANY questionable build practices then get magnified.  The signal is probably coupling back upon itself in a positive feedback loop.   It's slow, I'll give you that! 

Maybe replacing in/out wires with shielded types would help...switch wires too...hopefully it isn't the layout that's causing this...a carefully placed cap or 2 in the right place might help....

OK thank you, I will try shortening the wires going to the pots and the switches, the in/out jack wires are already as short as possible. If this doesnt work I will also try replacing them with sheilded wires.

Where would you suggest placing caps to limit the feedback? Maybe a power cap across the power pins of the IC?

[GibsonGM]

I'm not sure if you talked about this (I didn't have time to read thru the detail of your 1st post!), but sometimes bypassing IC pins with a pair of "uF" and "nF" level caps helps...so, something like 10uF and 100nF, in parallel, between the IC's + and - pins.

My suspicion is that maybe going to shielded wire, grounded at 1 end only, would be good to do here for input/output wires, right off.  And be sure signal wires cross at right angles only, like a " + ", and do not run in parallel with each other, like in a cable run.   Keep power and signal wires separate if you can!

Little things can have a big effect...you might even open it up, and try moving wires around with a pencil or something while it's doing its "tremolo", to see if you can affect it...("Chop sticking")

[Voltzy]

OK so no luck so far...

I have replaced the wires going from the input/output jacks to the switch, and the input/output wires going from the switch to the circuit with shielded wires, with only one end of each wire's sheild attached to ground. I have also shortened all of the wires for the pots and the selector switch as much as possible and ensured that they don't run together and are only perpendicular to each other. I also used a platic pen to move around those wires inside of the enclosure while the effect was running and I didn't have any change. I will try adding 10uF and 100nF caps accross the power pins if I have time tonight but I am not confident that will help.

Any other ideas?

[ElectricDruid]

If it only does it in the mode where there are no diodes doing any clipping (I don't know if that is "turbo" mode or not), my guess would be that the op-amp is latching up. Some op-amps react badly to being pushed beyond their rail limits and the voltage will hang at one extreme or the other for a little while until the situation settles down. Others cope fine and will just hard-clip at the rails, which is what you'd want here.

What op-amp did you use, and have you got any others you could try instead?

HTH,
Tom

[Voltzy]

What op-amp did you use, and have you got any others you could try instead?

The IC that I used has KA4558 EE13BF written on it and was bought from Jaycar. This is the only 4558 they have on their website so I am assuming it must be equivalent to this: https://www.jaycar.com.au/lm4558-dual-high-performance-op-amplinear-ic/p/ZL3958 and this is their datasheet for it: https://www.jaycar.com.au/medias/sys_master/images/9117340172318/ZL3958-dataSheetMain.pdf. The only op-amps that I have that fit this pinout are some TL072s so I could give them a go.

Looking at the circuit layout it looks like the turbo mode does indeed use diodes to clip, but one thing I have noticed is I couldn't find a voltage divider to supply a bias voltage to the op-amp input, it passes 9v through two 10k resistors in series to drop the voltage at the input pin3 of the op-amp. Is this just a lazy way of achieving the same goal? Does the 4558 need a biased input?

[thermionix]

KA4558 datasheet, specifies no latch up:

https://www.mouser.com/ds/2/149/KA4558-889307.pdf


The two 10K resistors in series make a voltage divider.

[Voltzy]

The two 10K resistors in series make a voltage divider.

OK I see it now, there is a third resistor to ground that forms the voltage divider and the second of the two 10ks mentioned above taps off the middle of that.

[thermionix]

I was really just guessing without a schematic.  I don't get paid enough here to follow vero layouts.  Regular Tube Screamers connect Vref at a number different points, even the "ground" end of the volume pot.  I don't know how the Turbo compares.

[Voltzy]

Alright fair enough. Would it help if I traced out a schematic from the vero layout?

[ElectricDruid]

KA4558 datasheet, specifies no latch up:

https://www.mouser.com/ds/2/149/KA4558-889307.pdf

Makes sense. 4558's are usually safe for clipping.

Oh well...time for another theory, I guess!

[Voltzy]

OK so I have put together a schematic based on that board layout (hopefully it is correct):



Here is the board layout again for convenience:

https://tagboardeffects.blogspot.com/2013/07/ibanez-ts9dx-turbo-tube-screamer.html

When turbo mode is on, the "SWA & B" connection coming off pin 2 of the op-amp is connected to "SWB4" on the right (2.2uF and 4.7k to ground). SWA1, SWA2, SWA3, SWB1, SWB2, and SWB3 will not connected to anything. I could redraw the schematic with it locked into turbo mode if that will help.

A couple of things I have noticed:

There seems to be no output DC blocking cap, there is an output cap on Q2 but the 9v is essentially wired directly to the circuit output. Should this even be working like this? I would have expected a DC hum coming through or am I missing something?

When pin2 of the op-amp is connected to SWB4, do the cap and resistor create a high pass filter?

When the mode switch is in + mode, pin2 of the op-amp is connected to SWB2 on the right and is also connected straight to ground. This mode works fine I am just curious as to how.

[thermionix]

C7 should not be connected to +9V.  On the tagboard layout there's a cut, hidden under the 10K resistor.

[Voltzy]

Ah that makes more sense. Looks like I missed a bunch of cuts, I will have to revist the schematic

[Voltzy]

OK this one looks a lot better:



So the first three different modes just change what set of clipping diodes are used, and what cap and resistor to ground is attached to the first op-amp feedback loop. Turbo mode is not using any clipping diodes at all.

Could it be the first op-amp's gain is increasing so much that it is overloading the second?

[Voltzy]

So I am still struggling with this. Do the different capacitors to ground from the first op-amps gain loop effect the amount of gain in the first op-amp?

I am interested in possibly changing out the 51pF compensation capacitor to see if that helps, unfortunately I rarely get much time to tinker... If this capacitor was faulty or the wrong value, would it cause this tremolo issue?

[thermionix]

Do the different capacitors to ground from the first op-amps gain loop effect the amount of gain in the first op-amp?

If you're talking about C11-C14, those effect the frequency response, the resistor in series (R18) is what basically sets the overall gain.  Though I guess you could say changing the frequency responce changes the gain of some frequencies.

51pF compensation capacitor... If this capacitor was faulty or the wrong value, would it cause this tremolo issue?

No I don't think so.

I wonder if you're experiencing motorboating?  That's like a low frequency instability.  I've only seen it in amps, but I guess it happens in pedals too.

[Voltzy]

I wonder if you're experiencing motorboating?  That's like a low frequency instability...

If this was the case, would it be more prevalent with a bass guitar as compared to high on the fretboard on a normal guitar? The guitar I have been using for testing is in baritone tuning so it's somewhere in between.