Author Topic: DC bias shift due to charge buildup on coupling cap driving Asymmetrical clipper  (Read 1185 times)

Vivek

From http://www.runoffgroove.com/thunderbird.html




Output varies with time. Initially there is more amplitude



Waveshape in first few ms (sine wave 1000Hz, 3V peak fed in)
Soft clip on positive cycle due to series resistor near positive clipping diode. Even more compliance in the bottom cycle



Waveshape after 800ms
The cap has built up a charge, shifting the signal waveform more positive. The top becomes broader since the diode clips at a fatter part of the sine wave. The bottom becomes smaller since there is less signal being presented below 0V.


 


anotherjim

Is that a real circuit or modelled? And if modelled, are there components subject to thermal drift and/or leakage?
An apparent DC drift may only be the wave shape "centering". The centre of an assymetric waves' amplitude is not 50% of peak to peak.
Croeso i Diystompboxes.

If they didn't hear you then you didn't say it.

merlinb

An apparent DC drift may only be the wave shape "centering".
That's the same thing as a DC drift.

A dynamic shift is more important to a good overdrive tone than almost anything else, so I have to applaud the OP for pursuing this route instead of wasting time with different coloured LEDs. However, for a true tube-like distortion (which most people seem to like best) you want the dynamic shift to have the opposite effect -initially less amplitude, then a little more after 50ms or so. Better yet, clipping that is initially asymmetrical, then grows more symmetrical with time. Keep it up!  :icon_biggrin:
« Last Edit: October 30, 2020, 06:25:22 AM by merlinb »

Vivek

Is that a real circuit or modelled? And if modelled, are there components subject to thermal drift and/or leakage?
An apparent DC drift may only be the wave shape "centering". The centre of an assymetric waves' amplitude is not 50% of peak to peak.


That bit of circuit is from real circuit from THUNDERBIRD at ROG.

but I analysed it with LTSPICE.

Its not just wave centering because

The peak to peak amplitude changed with time
The wave shape changed with time.



Vivek



A dynamic shift is more important to a good overdrive tone than almost anything else, so I have to applaud the OP for pursuing this route instead of wasting time with different coloured LEDs. However, for a true tube-like distortion (which most people seem to like best) you want the dynamic shift to have the opposite effect -initially less amplitude, then a little more after 50ms or so. Better yet, clipping that is initially asymmetrical, then grows more symmetrical with time. Keep it up!  :icon_biggrin:


I understood
Sag = cleaner, higher amplitude signal initially and then dirtier and less amplitude signal

Bloom, Swell etc may be the opposite

But I cannot seem to find raw data of waveshapes or oscilloscope snapshots or WAV files on the internet that show these characteristics of a tube amp. Hence it is so hard to study.

So far I only saw very few DIY circuits, some Crate Amps, some Levin amps that try to emulate change in waveshape over time. and they do it is somewhat similar fashion of a cap collecting charge over time. They dont have that cap reducing charge over time.

While it appears that almost no AMP IN A BOX or Distortion pedal has implemented this feature.


About initially asymmetrical ie pleasing even harmonics, progressing to symmetrical clipping , I feel this could be related to instantaneous signal level rather than integration over time. I need to read to old patents more carefully, but I was thinking of implementing this in a second clipping stage which clips earlier asymmetrical clipped waveform to be symmetrical if above a particular value.


Or maybe this circuit already does this, if I input sine waves with envelope approximately same like a guitar input. Right now I posted input = Sine wave 3V peak, 1000Hz fixed


 

EBK

I actually have a pedal circuit in my unfinished pile that does this, based on the ideas in the expired SLM (Crate) patent:

Jack Orman also explored the main part of the SLM circuit a bit on his page (I'll try to dig up the link).

Edit:  Here it is.  About 2/3 of the way down on this page:
http://www.muzique.com/schem/projects.htm
"Tubetype Distortion"

My circuit is different from Jack's, but the main bits are very similar.  Ignore Jack's last bit about the patent, by the way.  The knowledge belongs to everyone, as the patent is long expired.
« Last Edit: October 30, 2020, 08:15:58 AM by EBK »
  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss

Vivek

Thanks for pointing that out!!




Yes I had earlier input that circuit into Spice and analysed it too. It has somewhat similar buildup of charge on the Cap. So the overall amplitude graph was similar as what I posted earlier. It does not do "Bloom", "Swell"

The interesting thing is the DC gain resistor 42 that can be converted into a pot, to change the effect of the DC shift

But the circuit needs  pre-dist EQ and post-dist EQ since the patent only tried to show the way to get a DC shift


https://patents.google.com/patent/US5032796A/en

It would be nice to see your schematic


PS: Crate and Levin Amps have 2 stages with this DC shift, to increase the effect.
« Last Edit: October 30, 2020, 08:27:10 AM by Vivek »

Vivek

Levin schematic with 2 stages that show DC shift

I would feel that the second stage integrates the signal that already has been clipped by the first stage and hence shows less dynamic range. It might have been interesting to feed the second Opamp the DC control signal generated by the first IC.

« Last Edit: October 30, 2020, 08:30:57 AM by Vivek »

EBK

It would be nice to see your schematic
This is still unfinished, but here you go.  Feel free to suggest improvements.  The JFET portion is a rather crudely cobbled bit, so go easy judging that.   :icon_razz:



  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss

Vivek

Wow, Interesting, thanks for sharing !!!

That patent DC shifter needs fairly high input signal voltages to be able to do its stuff.

What is the peak voltage of signal that reaches R5 ?

I found in the simulations that one way to reduce the requirement of big input signals is to scale everything down ie 2 or 3 diodes in series for the clipping instead of zenners, and a schotty for the DC integration diode.

Please help me to understand the subcircuit A,B,C,D. Thanks
« Last Edit: October 30, 2020, 09:09:08 AM by Vivek »

EBK

Please help me to understand the subcircuit A,B,C,D. Thanks
The subcircuit is lovingly ripped out of the Klon and scaled a bit differently.  I wanted to feed some additional bass signal forward around the distortion portions and mix it into the output to have more bass content without more mud.  I thought it would be fun to just see what would happen dropping that existing bit in.  It sounded interesting enough in my initial testing that I haven't bothered changing it yet.  Another crudely cobbled bit that looks more sophisticated than it is.  :icon_wink:
  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss

antonis

Couldn't IC1B & IC2B be incorporated in a single stage, Eric..??


P.S.
"Honey Butter Biscuit".. ??? !!!  :o
(you didn't give your best, did you..??)  :icon_lol: :icon_lol:
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

EBK

Couldn't IC1B & IC2B be incorporated in a single stage, Eric..??
No.  You see, this one has two stages.  It's one louder.  :icon_lol:

Quote
P.S.
"Honey Butter Biscuit".. ??? !!!  :o
(you didn't give your best, did you..??)  :icon_lol: :icon_lol:
The very first prototype of the circuit, which was a lot different at the output stage, sounded sweet, rich, and warm to me.  That's what inspired the name.  This iteration is (sadly) less warm, but I've kept the name.
  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss

11-90-an

Quote
P.S.
"Honey Butter Biscuit".. ??? !!!  :o
(you didn't give your best, did you..??)  :icon_lol: :icon_lol:
The very first prototype of the circuit, which was a lot different at the output stage, sounded sweet, rich, and warm to me.  That's what inspired the name.  This iteration is (sadly) less warm, but I've kept the name.
[/quote]

I do like the name, though.. :icon_mrgreen:

I need to know what all these adjectives for tone are. I now feel like creating a pedal named “food” or something like that and having a “sweet” knob, “bitter” knob, “spicy” knob, etc.
flip flop flip flop flip

Vivek

Please help me to understand the subcircuit A,B,C,D. Thanks
The subcircuit is lovingly ripped out of the Klon and scaled a bit differently.   :icon_wink:



Does one end of R25 need to go to an AC ground like VBias ???

Vivek

Couldn't IC1B & IC2B be incorporated in a single stage, Eric..??


I think IC1B clips asymmetrically since the zeners are not the same value

Then the output of IC1B is inverted

and 1C2B clips the inverted signal asymmetrically ie clips the other side of the signal

All this cannot be done with one opamp



« Last Edit: October 30, 2020, 09:54:07 AM by Vivek »

EBK

Please help me to understand the subcircuit A,B,C,D. Thanks
The subcircuit is lovingly ripped out of the Klon and scaled a bit differently.   :icon_wink:



Does one end of R25 need to go to an AC ground like VBias ???
Yes.  I corrected that, but didn't take a new pic after I drew it in.  I'm still looking for my latest copy.  I believe I scaled some resistors and caps differently too....
  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss

Vivek

EBK, it would be a good idea to estimate the peak signal value at various stages in the circuit

I feel that in case Vcc was infinity and there were no zeners, IC1B should have output maybe 200 times the bigger zener voltage. Then the zeners clip the output at around 1/200 of peak at max gain.

PS : Input DC decoupling cap needed ?

PS : Cutting bass before clipping needed ?

EBK

I'd also recommend taking a look at this paper for another take on emulating triode DC shifting:
https://www.researchgate.net/publication/290534843_Triode_emulator_-_Part_2

I think Peavey may have done something similar in some of their amps.
  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss

EBK

EBK, it would be a good idea to estimate the peak signal value at various stages in the circuit
Still looking for my notes.  I'll post whatever I find.
  • SUPPORTER
"I want to go back to being weird. I like being weird. Weird's all I've got. That, and my sweet style." --Maurice Moss