And we haven't discussed how loading effects noise floor.
I kind of touched on that about over sizing grid resistors to isolate and prevent excessive grid current in DC coupled circuits.
Input circuits should be provisioned for all coupling methods that might be used.
They're also possibly the most intrinsically (before feedback) linear device (in an optimum setting) possible. Gm and perveance aren't very high, by modern standards, but operated in its ideal high impedance world, it can amplify with better than -50dB error components at output levels of 1/100 of B+ and monotonically down to single-billiard-ball-electron-charge level.
Semi-con field effect devices would seem to also be similarly monotonic, and Scott Wurcer has assured me that junction transistors are perfectly monotonic (below a single electron charge), but still have my doubts about the latter. Scott's dropped out of sight, so I can't ask him to elaborate. Hint, hint.
Much thanks and much appreciation, as always,
Chris
I think we lost him and transistor 3d sculpture guy that I can't remember his name from Analog Devices. His lab was always a mess, but very brilliant.
I'll save opening up the can on the transistor and why it was called a "trans-resistor" and why the newer transistor theory they teach now these days is incorrect.
Tubes are like mosfets in which they are trans-conductive devices. Because they are in always state of conduction. Some tubes require a heater for them to operate, but that isn't a bias, it is just there to establish the trans conductive state. The only thing is the resistance between plate and cathode is variable to plate voltage while the mosfet's resistance between the Source and Drain is constant.
The thing I don't understand with tubes is why you can get down to 4-5dB NF with the best. The cathode is very hot and you would think that this would determine the NF. If somebody could answer that I would be grateful.
Ok, I'm going to give you a rundown on the nature of a vacuum tube. This is going to open up another can btw...
A vacuum tube is an electron media crossing transmission line that transmits electrons from its cathode to plate from the 2 dimensional media of a wire to the 3 dimensional space of a vacuum. Current flow elements called grids are constructed between the cathode transmission antenna to the receiving plate antenna. The electron flow volume and flow rate are controlled by the grid elements. The noise of a vacuum tube is a product of the electrons resisting receiving from the plate back into the 2 dimensional media of wire.
Observation of the tube noise is on the plate it shows the lack of electron beam focus.
The electron beam focus is the steady flow of electron coming out of the cathode back into the plate.
In harmonic distortion is the product of the harmonic propagation effect of the cathode transmitting into 3 dimensional space
Last edited:
Limited values though from what I can see
https://www.mouser.ca/c/passive-com...ough-hole/?q=Dale CMF&series=CMF Non-Magnetic
Any other suggestions for brands/models?
Partial answer:
Due to the hot cathode, you get more white noise voltage than you would get with a FET with the same transconductance. On top of that, you normally get substantial 1/f-noise at audio frequencies. Still, the grid current and its associated noise current are very small, so if you transform up the source impedance enough, the noise figure can be low. At higher frequencies, things like induced grid noise increase the noise current and the obtainable noise figure.
From basic vacuum diode theory the shot noise may be interpreted as the Johnson noise of rp, where the space charge has an apparent cooling effect, making rp appear cooler than the cathode by a factor of 0.644.
Magnetic leads make absolutely no difference. The only time metal film resistors have a minuscule amount of inductance is when the value is low like under 10 ohms. The datasheets I have found show metal film to have no more inductance than a similar length of wire.
PRP model PR9372 is also nonmagnetic, see the blurb here. Vendors include Parts Connexion and Sonic Craft.
https://www.hificollective.co.uk/components/prp_pr9372_metal_film_resistors.html
maybe you need to look at it physically.
With a scope, looking at the plate in the RF region(tube noise), compare a grounded cathode stage with a metal film in the cathode and one with a non-inductive wirewound in the cathode circuit on another. Beyond hearing the difference.
Last edited:
Reading it does seem to be the space charge. When electrons detach from the cathode their energy is lost so they are as you say cold. There seems to be formula Rg = 2.5/Gm where Rg is the equivalent grid noise resistance.
The grid stopper resistor's role is to stop possible instabilities leading to distortion, overshoot or auto-oscillations at high level of gain or loudness.
This is a very common feature found on numerous tube guitar amps, but also on Hi-Fi amps.
Values can vary from 1K to 10K, but the 1.5K is probably the most encountered. Below a Fender Twin-Reverb :
T
This is a very common feature found on numerous tube guitar amps, but also on Hi-Fi amps.
Values can vary from 1K to 10K, but the 1.5K is probably the most encountered. Below a Fender Twin-Reverb :
T
Paralleled tubes (or FETs or bipolars) are prone to local RF oscillations between pairs or groups. This is the most common place you see them.
Fat chance.
Unless it's an "experiment" flawed on purpose to."prove" a point.
There is absolutely no physical mechanism by which a cathode metal film resistor should create visibly higher noise than a wirewound one.
Specially within the audible range as specified by thisusername
Unless it's an "experiment" flawed on purpose to."prove" a point.
There is absolutely no physical mechanism by which a cathode metal film resistor should create visibly higher noise than a wirewound one.
Specially within the audible range as specified by thisusername
it has to be an Ayrton–Perry wire wound non inductive resistor. There are several types.