Again i cannot see how this differs fundamentally to any power supply with an adequately sized last filter cap.
Shoog
Shoog, the problem with getting rid of the cathode by-pass is that when the signal swings too much, the tube stops conducting (which is why Brits call them valves). When this happens the cathode drops to zero (or whatever bias point is below that cathode resistor).
When the signal allows the valve to conduct again, this situation causes it not to conduct for a short time (a form of distortion). Morgan Jones and others describe this nicely.
The by-pass capacitor quickly restores the charge (which is highly desirable) but as you may know also can create a feedback loop (and a different form of distortion).
Using a CCS or a choke (not both) is nice for other reasons, but it won't skin this cat.
Ian
You don't seem to grasp the distinction between a voltage source and a current source. They are opposites/duals/inverses of each other.Shoog said:
A CCS and the output of a good power supply are about as opposite to each other as it is possible to get in electronics. You seem to think of them as somehow similar!
Is this not the principle of fixed bias as drawn on my original schema ??
CCS anode load and fixed bias for the cathode.
Shoog
An adequately spec'd power supply operates at both fixed voltage and current - otherwise we would be having power supply sag - which we design to avoid. It is the last cap which allows the AC current variation.
Shoog
No. PSU is almost always fixed voltage and variable current. Not perfect, but that is the aim. CCS gives you the opposite: fixed current and variable voltage. That means that you don't know what the quiescent anode voltage will be, unless you also provide variable bias (which defeats the point of a CCS!). The CCS tries to set the current by adjusting how much voltage is wasted - but in an output stage you don't want any waste.
This thread does not seem to be making progress. You asked what is wrong with your circuit. We have told you.
This thread does not seem to be making progress. You asked what is wrong with your circuit. We have told you.
There are two objective which it achieves - it fixes the standing current (essential) and eliminates the cathode bypass cap (desirable).
The fact that it consumes some voltage is unfortunate - but given the other requirements (discussed) it only replaces wasted cathode voltage with wasted anode voltage - outcome been the same overall wastage. The degree of wasted voltage can be finely tuned by adjusting the fixed bias applied by the negative supply.
I am willing to accept a critique, but so far they are not of sufficient magnitude to negate the desired advantages.
Shoog
The wasted voltage is multiplied by mu, roughly. This could be several thousand for a pentode. So no, not the same overall wastage.
Wavebourn and I have tried to explain. I am out.
You don't recognise the magnitude of the critique. You asked for comments. You have received comments. You have failed to understand the comments.Shoog said:
Wavebourn and I have tried to explain. I am out.
i am also entertaining this somewhat crazy idea, i plan on using a CCS on the anode of a triode and fixed biasing on the grid....
my reasoning is that the CCS ensures that i get my desired plate current, the fixed biasing on the grid is to ensure that i get my desired plate voltage....is this practical?
For a pentode, maybe you could put the CCS on the screen grid to set the plate current, since the screen grid intercepts a fixed (at a given plate V) fraction of the plate current. Then a cap from the screen grid to ground to take care of the AC signal current variation. Then adjust the grid 1 bias to get the plate V where you want it. With an inductor/xfmr type load though, you will have to measure the plate current while adjusting the grid 1 bias.
For a P-P case, a single CCS for both screens might produce some interesting results. Soft saturation for one.
For a P-P case, a single CCS for both screens might produce some interesting results. Soft saturation for one.
Last edited:
thanks......i am going to use this circuit to drive the gate of a mosfet in a source follower configuration....
I use a CCS to set the quiescent DC operating conditions in the IT coupled driver stages of my newer SE amplifier designs. This does exactly what you need and is scalable. How well it would work with a UL or pentode connection is unknown (works fine in triode) and for PP you need two in order to keep the primaries separate at DC, and must provide an AC path back to ground (good quality film bypass cap) for the cold end of each primary. Lots of heatsinking will be required for any appreciable current, try to keep the voltage drop across the device to the minimum required for good performance at low line.
See below...
See below...
Attachments
Perhaps I shouldn't, but I'll take a shot.
A transformer's secondary changes voltage when there is a change in current in the primary. If the current in the primary stays constant, ...
So it seems my original intention has been demonstrated to work if the CCS is bypassed to ground by large enough capacitors. Thank you Kevinkr for your useful input of practical experience.
I will lash it up and give it a test to see if it can be scaled to the output stage. I envisage that due to the transformer a relatively simple CCS should suffice.
Shoog
I will lash it up and give it a test to see if it can be scaled to the output stage. I envisage that due to the transformer a relatively simple CCS should suffice.
Shoog
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.