Interesting little trick I'm working on for a jfet drive 'clipping' control
A long time ago, working with a Sallen-Key filter, I realized that you can take the output signal from the junction of the 2 gain feedback resistors instead of the op amp output, taking advantage of the stage gain to shape the filter, but with the output being unity gain (at the expense of headroom). Obviously volume pots exist, but it's a neat trick, I thought
Anyhoo- I wanted to have a sort of 'drive' or 'clipping' or 'dirt' control, but without it meaning a huge signal on the input of the next stage- I also didn't want a second control just to attenuate the signal (what would we name the knobs? The first one is gain, the second is ungain?), and a fixed attenuator leaves me at the mercy of the input signal.. so I came up with this
(https://i.postimg.cc/1gVn69MN/IMG-5438.jpg) (https://postimg.cc/1gVn69MN)
When you turn the pot, the gain goes up, but by taking the output from that junction, it's in a sweet spot where the RMS volume is pretty close to constant, but the signal is more and more clipped
(https://i.postimg.cc/ZCh5s8Hn/IMG-5439.jpg) (https://postimg.cc/ZCh5s8Hn)
Compared to the signal at the collector, which is the typical output of this block
(https://i.postimg.cc/vgRJ8zc6/IMG-5440.jpg) (https://postimg.cc/vgRJ8zc6)
It's dialed in mostly by ear- the sim graph is a little arbitrary, I found values that made it look about right, but the 330R and the 56R are both trimpots on my board
Now- one may ask 'why don't you just want the gain, who hates gain? You can just attenuate it'
For the wider circuit I'm working on, I just don't want the next stage to be hammered and clipped at the input, nor do I want to attenuate it- it's either a clunky double knob deal, or a fixed attenuator, but I just want the clipping.. also it's just a neat problem to solve
It's also just a little more common on the bass side to have pedals that utilize more headroom-based clipping and 'dirt' controls with no/little volume add.. perhaps the combo of more solid state amps and DI ubiquity
I'll also show the typical transistor block I based that on, and the 'op amp analog' .. maybe it's unnecessary, but I doubt I explained anything well in the OP lol
(https://i.postimg.cc/zb7jBsNM/IMG-5442.jpg) (https://postimg.cc/zb7jBsNM)
(https://i.postimg.cc/ykkP1pbQ/IMG-5441.jpg) (https://postimg.cc/ykkP1pbQ)
As above, 'taking the signal from the junction of the feedback resistors' would be the 680 and 1k.. I'll often add coupling caps and a pot there as a gain control.. I use it as one of a couple 'discrete op amp' blocks
(https://i.postimg.cc/cgvbp1wh/IMG-5443.jpg) (https://postimg.cc/cgvbp1wh)
Edit: **credit goes to talkbass user fDeck / Francis Deck for the above three images.. he designed a popular HPF, which he thoroughly documented; the original version used this discrete 'op amp' block. Him and a variety of his circuits were actually extremely formative for me; He designed a quick and dirty jfet buffer that was my first active circuit build ever, his HPF was my second build, and a rich subject of study as I tried to learn electronics.. that gain block has been an ever present theme in a lot of my designs over the decade and a half (wow that hurts to think about)
Now .. at the same time I guess my solution *is* just kind of attenuating.. but the constant volume thing is very useful when the wave shape is changing.. the hard dial in is the 330R in the first post.. because the clipping actually makes this signal smaller at that point, so I have to get it a little bit above the source
Not exactly sure of the implications on the loading of having the signal so close to ground there, but for my goals, it will be input into another jfet, and I don't seem to be suffering much for highs
> what would we name the knobs? The first one is gain, the second is ungain?
"Gain" and "Loss"?
Input and Output?
If you like ungain, then the other is unloss.
A very interesting concept! I can think of lots of applications for this, like 'one knob' effects. I love the aesthetic simplicity of one knob effects but they're hard to acheive. A constant-volume distortion would be a great project.
Quote from: Eddododo on March 05, 2025, 01:04:49 AMI'll also show the typical transistor block I based that on
FWIW, I don't think 2N4416 is a good choice for particular configuratuion..
(its wide V
GS(off) spread might take Q2 Collector close to either saturation or cut-off..)
"Pre Gain" "Post Gain" ? I have this on a tube preamp I built...input gain and 'drive', I just call them pre and post.
Neat.
Have you actually built it? How does it sound when changing clipping?
I'd think the changing in waveform would change the perceived loudness, even if the signal amplitude is more or less the same.
Thanks!
Quote from: marcelomd on March 05, 2025, 08:47:51 AMNeat.
Have you actually built it? How does it sound when changing clipping?
I'd think the changing in waveform would change the perceived loudness, even if the signal amplitude is more or less the same.
Thanks!
I think it is pretty likely that you are correct- The circuit on my breadboard was tuned by ear for the 'constant volume' effect, meaning that If I took images from a scope, that the spread off the waveform shapes may have a wider spread than the sim.. Luckily, tuning this by ear is kind of the point anyway. The clipping control itself can probably use some tweaking and tapering- A lot of the clipping is bunched up at the top of the sweep, despite using C-taper pots and/or Tapering resistors.. That being said, I'm hearing a fairly smooth, increase in the audible Drive throughout this sweep, and just the hardest clipping seems to skyrocket at the top.. But for my personal purposes, I actually don't know if I even really desire the most extreme clipping, so it may be as simple as increasing the 56R a bit, and removing aggressive range.
I do wonder how finicky it's going to be for different instruments at the input - I wouldn't be shocked if it needs to be retuned for different signal levels or tonal characters..
Quote from: merlinb on March 05, 2025, 04:35:45 AMA very interesting concept! I can think of lots of applications for this, like 'one knob' effects. I love the aesthetic simplicity of one knob effects but they're hard to acheive. A constant-volume distortion would be a great project.
Ironically, this circuit is tied to a very convoluted effort: a do-it-all bass preamp.. I've been playing with Fetzer valves BUT MORESO JC Maillet's triode simulator (the class-A exaggerator!) for a little
Je ne sais quoi.. i've really fallen in love with the sound of the Exaggerator (enhances the transfer curve such that the negative peaks are elongated and the positive peaks are compressed), And I wanted to add a little dirt BEFORE it so that the clipped signal can be enhanced the same way- Which is why I don't want to slam the input of the next stage because it 'ruins' the peak stretching
Quote from: antonis on March 05, 2025, 04:53:02 AMQuote from: Eddododo on March 05, 2025, 01:04:49 AMI'll also show the typical transistor block I based that on
FWIW, I don't think 2N4416 is a good choice for particular configuratuion..
(its wide VGS(off) spread might take Q2 Collector close to either saturation or cut-off..)
I'm actually using a 2n5485 for it- the pictures in the second post come from a talkbass user named fDeck, Who designed a very popular HPF, For which he published a ton of documentation and notes, etc.. I'm not sure how he chose that Fet in the original circuit, but I know in later iterations and other circuits, he tends to use 2n5486..
The basic circuit with the Bjt current source came from a Jensen application note, but fDeck is responsible for the interesting gain control grafted onto it
Personally, I don't worry a whole lot about choosing FETs- I have a bin each of J201, 2n5485/2n5486, and 2n5457, and I generally just try to work within those and tune by ear, because I get depressed when I try to carefully choose Fets and then find them somewhere for $2.50 each
I posted it on Reddit as a 'guess what this knob does' puzzle.. perhaps that would have been more fun here