67.169.53.166
In Reply to: RE: Unbypassed cathode resistor and feedback posted by danlaudionut on October 26, 2009 at 17:32:46
Henry and I see this one differently and I know both are correct in how you see the models.
First think of the pentode and electron beam theory. In fact think of the beam tetrode if you will. In this physical layout, it is specially designed to have all the attraction head straight toward the screen grid. The attraction is designed to be maximum there and it is "screening" out the the impact of the plate as much as can be. The electrons fly by if aligned well with the input grid 1 and suppressor grid 3. Now why is there something called a suppressor grid that is usually just connected to the cathode inside or set to cathode potential in the beam tetrode? This is at a lower voltage and it picks up secondary emissions lost from anywhere: the screen grid's secondary "multiplier electrons" and the plate's "impact splatter secondary emissions" of high energy electrons. The screen's primary job is to hold a high accelerating potential close to grid1 to start sending out a beam of the electrons past it to the plate. The plate and screen are too far apart for the plate to impact input grid1. The electrons are already accelerated toward the plate and increasing plate voltage does nothing to accelerate more electrons toward it. Again the screen is screening off the plate and its voltage matters for current acceleration.
So we have flat plate curves for the pentode as the plate voltage goes up, the Ip remains flat and hence operating in the constant current region. So what? That means that plate secondary emissions are swept up and never sent back to the input grid and there is no contact for any real NFB inside there. Try NFB by adding a large plate resistor and it will not change its essential gm gain, and this is a real gm device. And what is the Rp of a good pentode? Essentially infinite. It's a current source, not a voltage source.
Now take off the screen and suppressor and it's a triode now. What do we lose and maybe gain here? We lost plate isolation due to the loss of the screen, and we lost secondary emission clean-up by losing the suppressor. What does all this new dirty electrons running amok do for us now? Well now the plate is in close contact to the grid and cathode for attracting and accelerating electrons. And when we add more voltage to the plate, like we did with the screen, it will now increase electron current flow straight to the plate output. But this time a powerful electron bombardment does create a splash of plate electron secondary emission, and those negative ion clouds push back on more electrons running in as it isn't so positively charged anymore. So there is a new force slowing down an attempt to accelerate electrons to the output, and it operates in a way to negatively slow down more with ever more voltage and current hitting the plate. This is what I call the physics of a natural NFB inside. And the device is no longer a constant current, it now has internal plate resistance, Rp.
But like most people see, the force is not pushing hard against the grid voltage, yet it counters your effect at the grid to push foward what was easy on the pentode, and hence has lower gain too. So now the plate voltage changes as well as current changes. This shifts the plate curves radically upward and no constant current region can be found. It's all triode region. If you increase plate current on the triode's plate by raising just plate voltage, it will rise fast and high at low Vgk and almost quite linear. In fact the plate current rises faster than the plate voltage. When you go past most published curves at very high plate currents that starts in on secondary emission and high red glowing plate electron bombardment you will start seeing the bend over where the plate can take no more current and is blasting off its plate a lot. This is increasing plate resistance fast. Some tubes are designed to handle it in that region.
But then watch as you see the high gm tube work Ip vs Vp with higher Vgk levels. The grid is meant to control Ip as it is still a transconductance amp by Vgk. But at higher Vp it is having less and less control over this valve. The currents flatten out and Rp increases again. With lower allowed current by Vgk input, Vp itself is more hampered against getting Ip to move up. And that's what you want. But there will be no constant current region and there will be real plate resistance to see.
Now set the load line for a vertical load, or a plate resistance real low. The lowering of Vgk from cutoff to all on will send it through only a few Vgk changes before it will burn up Ip as Vgk nears 0. Near cutoff it moves rapidly in changing Ip as it turns on, all completely nonlinear. The pentode would be fine with it and amplify well. To drop distortion rapidly, make the loadline horizontal, so that little plate current is altered as Vgk moves and gently moves Vpk. It's now much more linear. But we accomplish that at the plate side only! For the pentode, little difference in linearity for given plate loads. Built-in plate-grid NFB not there in a pentode, yet is in a triode.
So with a variable Rp device called the triode, a non-touching NFB arises by increasing the plate load, and you can take it out to a full CCS. On a near infinite non-varying Rp device called the pentode, a CCS will cause disaster as two fighting CCS devices try to set the ONE current.
So is there a PURE NFB inside a triode type device, or is there just the characteristics of a three terminal device? It hardly matters, but the triode is the only device made by man that can do that single trick, and is unique and not yet SS replaceable. BJT's might look like them, but their characteristics are current amps, not transimpedance, and the harmonic structure naturally is horrible. So only MASSIVE NFB makes them get linear looking.
Many special devices were invented for unique characteristics wanted, but lost as they got replaced by mass production for alternative applications.
Nothing replaces the mechanical line stretcher for microwave use, but it is a rarity item today as more electronic and more expensive equipment built uses alternative methods to encompass more measurement flexibility. But for one thing, that old line stretcher is what we look for sometimes and hope no one threw away. Triodes are on the list. It was the simplest and best for pure sound, but not for modern computers and most applications today where poor substitutes for cheaper solutions came into being. So it being first is an accident, and would not even be seen if the technology for SS was there in 1900.
-kurt
Follow Ups:
Post a Followup:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2009, Audio Asylum, all rights reserved ]
[ Go Green ]
[ Take5 ]