Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. The sunn still shines online!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  

Planned distortion. Design for Saturation or Cutoff?

7/22/2005 5:06 AM
BK Planned distortion. Design for Saturation or Cutoff?
I've been studying load lines for a preamp I'm working on and started to consider how to force the signal to clip by adjusting the bias point of the tube. So, I wonder which operating point is a better place to explore distortion? Should you, say, use a pair of stages in cutoff yielding symmetric clipping of the final signal? Alternate two stages in cutoff/saturation?  
Any thoughts/comments?  
7/22/2005 6:58 AM

You could rig up one gain stage with a spdt switch that switches between two cathode resistors, one that biases it close to cutoff and the other biasing it close to saturation. Then you can compare how the clipping sounds either way.  
Someone here posted a link to a paper that examined the distortion generated by ]a gain stage direct coupled to a cathode follower. Two points from that paper that stuck out in my mind were: (a) clipping at saturation sounds smoother and is generally preferred to clipping at cutoff, and (b) the distortion produced by the combination of the gain stage and the cathode follower is fairly close to symmetrical.  
So, consider that option. You might try tweaking the value of the plate resistor and scoping the output to see whether & how the value of the plate resistor affects the shape & symmetry of the distortion. In my tweed twin clone, I accidentally used a plate resistor on V2 that was much lower than it was supposed to be -- either 10k or 1k, whereas it was supposed to be 100k -- and the distortion was nice and smooth. I just wasn't producing enough p-p voltage swing for that amp.  
7/22/2005 7:49 AM
Ray Ivers
12AX7 Stage w/Symmetrical Clipping
I'm of the firm belief that truly good-sounding clipping does not occur at cutoff, and the distortion I find the most pleasing occurs at the other end of the spectrum, whether or not that represents actual saturation or not. IME it's not easy to design a tube stage for what I consider true 'saturation' (i.e., absolute-maximum cathode current) and it's probably not too good for the tube either, although it's quite easy to adjust pentode load lines for plate 'bottoming'.  
In every tube SE RC-coupled preamp distortion circuit I've ever seen, the vast majority of clipping occurs by the cutoff mechanism only. The onset of grid current can give the illusion of bottom-of-plate-waveform 'clipping', but it's actually a voltage-dividing effect between grid-circuit and driving-circuit resistances. As Shea stated, the GS/CF combo clips semi-symmetrically and is by far my favorite preamp-distortion circuit, at least so far...  
After wanting to try this for - I don't know, ten years? :D - last week I finally got off my ass and at least simulated a symmetrical-clipping 12AX7 circuit, which I hope to get around to putting in a customer's amp in the next few days. The circuit component values I ended up with were 330R cathode R and 120k plate R @ 375V B+ - but the grid was low-impedance-driven (this is critical for true symmetrical clipping) from the output of an IRF820 MOSFET source follower through a .022uF cap and 1k series grid resistance, w/100k grid R to ground. A tube CF, op-amp output, JFET, BJT, xfmr, etc. could also be used for drive, and direct-coupling would be cool if feasible.  
This circuit appears to be absolutely brutal on the 12AX7 clipping stage, indicating plate-circuit 'saturation' at 3.5mA, a lot of current draw for a 2.5mA-max tube. The waveform looks exactly like I wanted it to, though; just like a push-pull triode output stage without a trace of bias-shift or crossover garbage.  
7/22/2005 8:49 AM
Dai Hirokawa

I remember reading something about that which is hopefully useful:  
[QUOTE]Tube DIY Asylum  
I designed just such an animal for a guitar amp for a friend a few years ago.  
[ Follow Ups ] [ Post Followup ] [ Thread: Display All Email ] [ Tube DIY Asylum ]  
Posted by Mark Kelly (A) on March 01, 2005 at 01:54:01  
In Reply to: DIY tube distortion effect. posted by kupfer_m on February 28, 2005 at 23:37:58:  
There is no great trick to it. Just put in an ordinary gain stage feeding through a pot into a remote cutoff pentode with switchable "triode mode". This feeds another pot which is reverse geared off the first pot then an output buffer. The voltage output from the first stage must be higher than the second stages max input for best effect.  
The total attenuation through the two pots stays more or less constant but the voltage seen by the "distortion stage" varies with pot position. You can switch the distortion stage between triode and pentode mode for different sounds, IIRC I also included adjustable idle current so that you could sit in the high current part (away from cutoff, less distortion) or the low current part (nearer cutoff, more distortion) of the triode characteristic.[/QUOTE]  
7/22/2005 1:57 PM

Interesting, Ray. Can it be made to operate within the 12AX7's max ratings if you don't mind a little bit of asymmetry or bias shifing?  
Also, is your gain stage direct coupled to a cathode follower? I assumed that it was when I read your post, but I'm not 100% sure.  
7/22/2005 3:51 PM
BK Throwing Away Gain
It occured to me, upon further consideration of the load-line, that the useable headroom(?) at the grid is about 3 or 4 volts p-p. If I hit the grid with a source of 4V p-p the stage is sure to clip symmetrically (assumming a bias point of -2V).  
Is it common to "throw away" most of a stage's output (say 45V p-p) down to something like 2V p-p in a cascaded gain design? I would guess if I didn't the think wouldn't work as tghe grid goes way positive and into conduction.  
7/22/2005 6:24 PM

"It occured to me, upon further consideration of the load-line, that the useable headroom(?) at the grid is about 3 or 4 volts p-p. If I hit the grid with a source of 4V p-p the stage is sure to clip symmetrically (assumming a bias point of -2V)."
Not really. The shape of the clipping is different on each half of the waveform - rounder on one side, harder on the other. Also, one side usually clips before the other, and if grid current starts flowing then the bias point is likely to shift from the blocking cap charging up.  
One way you can be sure it isn't all that simple is that Ray spent 10 years daydreaming about designing a preamp circuit that clips symmetrically, then put the effort into simulating the circuit to get it all worked out correctly. Nobody is as passionate about making tube distortion symmetrical as Ray is, I reckon.  

  Page 1 of 6 Next> Last Page>>