King of Tone fading sound with both clippers on

[lars-musik]

I built (another) King of Tone and experience some weirdness in the clipping behaviour. When the soft clipping is switched in, the following hard clipping does some "slow gear"-alike fade in. If I strum the strings hard it goes away, when I stop playing shortly it comes back. Almost like there is some capacitor re-charging somewhere. If I select soft OR hard clipping everything works fine.


Now I think it doesn't make make much sense to cut the soft-clipped signal by the hard clipping because it cuts the waveform earlier anyways. But I'd really like to understand what is causing that phenomenon.
Thanks!

PS: Here's the complete schematic. The supply voltage is selectable by a header with jumpers. The phenomenon occurs with both voltages.

[anotherjim]

Running on 9 or 18v?
My first thought is "is VB drifting?". If the hard clippers pump enough current into VB, then that affects every amp referenced from it.

[Gus]

As anotherjim posted does it happen at 9VDC or 18VDC?
I would also watch the VB with a voltmeter as you play the guitar into the effect.
What brand of caps are you using for C400 and C401? Do you have a cap tester to check value and ESR?

First stage clipping can have 3 diodes removed using the switch you show in the schematic the switch would need to be wired different. Use two and two diodes, short one diode for the 2nd setting

[lars-musik]

Hi! Thanks for chiming in!

Running on 9 or 18v?

Had it in my first post: Both...

The phenomenon occurs with both voltages.

What brand of caps are you using for C400 and C401? Do you have a cap tester to check value and ESR?

The caps are "AISHI"s and another one I've got here measures at 105µF, VLoss=1,6%, ESR=0.28 Ohms.

First stage clipping can have 3 diodes removed using the switch you show in the schematic the switch would need to be wired different. Use two and two diodes, short one diode for the 2nd setting

I know, I just wanted to have the possibility to plug a completely different set of clipping diodes in there. The symmetrical/asymmetrical just sounded best to me. The switsch in onboard, so not really rewireable.

[lars-musik]

That is super weird!
At 17,3V VB is about 8.25V idling, when strumming the guitar it rises to 11V when the FIRSTof the two channels is hard-clipping. On the OTHER channel VB drops to 5.4V when activated and I strum the guitar.

Something is seriously wrong here! But I have no idea what.
Here is the whole schematic. What the heck!

[dschwartz]

Check your diodes..if only one side is clipping at the opamp feedback, it will mess the bias and react to envelope..
This is what i use on the "broken mode". Of my drive maker design..it can be useful if controlled..

[lars-musik]

Does that mean that asymmetry in the opamp feedback is not good?

Gesendet von meinem Lenovo K33a48 mit Tapatalk

[rezzonics]

Vb bias voltage is generated by a resistor bridge 2x 47k, that works OK if there is no current on Vb, but there is a 1k pulldown in series with a forward diode on the 2nd clipping stage and another forward diode on the first clipping stage.
2nd clipping stage should not be connected to ground instead Vb?

[Gus]

First is this vero or built on a PCB? Check your VB connections to start.

Two things to try for a band-ad fix

Make the VB divider from two 10Ks instead of 47ks VB increase the 100uf value if you want, VB will "move" less.

AND/OR

Use a cap between the opamp output and R203 and ground the VBs after R203. Try 10uf and higher values, + side to opmap

[anotherjim]

It might be possible, depends on how things are laid out for you, run only at 18v. Disconnect the VB supply voltage divider (lift the 47k's) and link VB to the 9v supply.

[lars-musik]

2nd clipping stage should not be connected to ground instead Vb?

Is it so? The schematic seems to be verified about a million times. This is what really puzzles me: Besides the asymmetrical clipping I changed nothing!

First is this vero or built on a PCB? Check your VB connections to start.

It is a fabbed PCB derived from the schematic I posted earlier.



Diptrace gives me no differences between the schematic and the PCB.

Make the VB divider from two 10Ks instead of 47ks VB increase the 100uf value if you want, VB will "move" less.

What makes a voltage divider from lower resistances  better?

ground the VBs after R203

How do I ground VBs?

[Gus]

Have you built this circuit with the same PCB design before? If not check the PCB by hand with a meter set to OHMs

Have you checked if you used two 470k resistors instead of 47Ks for the VB divider.

I finally looked at the Marshall Bluesbreaker based schematics I found looking up this effect, there is stuff I would change.

[rezzonics]

A 47k divider from 9v gives you only 4.5v in the mid point Vb if there is a current on Vb that is much smaller than the 100uA that the resistor divider requires. It should be the case since opamp input current is very low, specially on FET opamps. Something is probably wrong in your circuit and is asking for more current than it should. Bad opamp, bad or reversed diode.
AC changes on Vb are filtered by the 100uF cap C401, maybe it's dead.
Your schematics look OK and second diode clipping stage has to be connected to Vb.
If you use 10k instead of 47k on resistor divider, current on it is 450uA instead of 100uA. So a 50uA on Vb generates only 10% imbalance compared to 50%, that can give you a clue but it's probably not the root cause, since original schematics use 47k

[rezzonics]

What does it mean  ALT_DRV_DIO value in the input cap C200?

[lars-musik]

AC changes on Vb are filtered by the 100uF cap C401, maybe it's dead.
Your schematics look OK and second diode clipping stage has to be connected to Vb.

Thanks a lot for the explanation. I changed the voltage divider to 10K and upped the capacitor to 220µF and now it seems to work fine.


I had the PCB fabbed and cannot image they messed up something. As ususal I checked all parts with my DMM before soldering them in, so I at least tried to minimize errors in that field. I even had a second board populated last night. Strange enough, one channel worked fine and the other had the very same problems. After changing the VB resistors to 10K there also, both channels now work.

What does it mean  ALT_DRV_DIO value in the input cap C200?

It seems I had both parts marked when I renamed one of the diodes in the "Alternative Drive Diode"Section. Should be 10n (and that's what I soldered in).

Check your diodes..if only one side is clipping at the opamp feedback, it will mess the bias and react to envelope..
This is what i use on the "broken mode". Of my drive maker design..it can be useful if controlled..

I matched the diodes in the standard slot now (both have a Fv of about 1.34V now). But obviously that's not an option for the asymmetrical clipping option. Or do I get this wrong?

[lars-musik]

Check your diodes..if only one side is clipping at the opamp feedback, it will mess the bias and react to envelope..
This is what i use on the "broken mode". Of my drive maker design..it can be useful if controlled..

Could you link to the circuit you are referring to? I did a search here on the forum but came up empty.

Thanks!

[anotherjim]

I'm staring to think, if you have a lot of things using VB, it might be better to have several local dividers instead of trying to reference everything with only one. Might make board layout easier since one less bus to run around. Where a section has only +inputs to reference, these local VB resistors can be high, 100k = less useless current drain.
Seem the problem here was having x2 the circuit using a reference divider intended for x1 circuit? Most amps non-inverting so any drag on VB voltage gets reinforced.

[lars-musik]

Seem the problem here was having x2 the circuit using a reference divider intended for x1 circuit? Most amps non-inverting so any drag on VB voltage gets reinforced.

Hi Jim, thanks for the suggestion. I am puzzled, because this King of Tone schematic is out there since years and has possibly been built a million times.
But maybe I'll try a second voltage divider sometimes. A pity that I have these boards already fabbed. 

[pinkjimiphoton]

i'd make r23 bigger, like 10k-47k or so. you may be saturating the next stage into cut off with just the 1k resistance shown.