• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Stability problems, help please?

Hi,
Just built a riaa-amplifier using a pentode input and dc-coupled CSS loaded output stage. Correction is in the pentodes plate-load.
For some reason the output stage can not be adjusted by the CCS to a stable Vp operating point. Plate voltage goes up and down whilst Ip remains stable at idle operation.
When I replace the CCS for a resistor it becomes stable..
CCS is a cascode DN2540..
Is someone willing to shine a light on this for me please? Can it be made stable using a CCS?
 

Attachments

  • riaa_e80f_d3a_ccs_fet_supply_step_down.gif
    riaa_e80f_d3a_ccs_fet_supply_step_down.gif
    26.4 KB · Views: 197
Last edited:

rayma

Member
2011-04-29 8:37 pm
Plate voltage goes up and down whilst Ip remains stable at idle operation.

Try increasing the input stage's decoupling capacitor from 2.2uF to at least 20uF, and see
if the low frequency oscillation is reduced, or at a lower frequency.

If that doesn't work, instead try increasing the input stage's screen decoupling capacitor to at least 20uF.
 
Last edited:

euro21

Member
Paid Member
2004-09-17 12:16 pm
Budapest
Are you sure, that D3a operating point (150V, 20mA, -1.5V) is correct?
In the datasheet the closest point 160V, 20mA, -1.5V.

1.) Is the 325V is rock stable, or change? One RC filter stage maybe necessary.
2.)Try to degenerate CCS with parallel resistor (1-2mA current).
3.) Try to add fix resistive load to tube - for example 250k-.
 
Are you sure, that D3a operating point (150V, 20mA, -1.5V) is correct?
You are right, I am using 16mA @ 150V..
Is the 325V is rock stable, or change? One RC filter stage maybe necessary.
Yes, it is rock solid.
Try to degenerate CCS with parallel resistor (1-2mA current).
Try to add fix resistive load to tube - for example 250k-.
I take it that the 250K is the parallel resistor?

Thanks!
 
Is it the output stage which is oscillating at a subsonic frequency?
Yes it is the output stage but I am not sure if it is oscillation since I do not see something alike on a scope.
I just replaced for both channels the CCS for a 4K7 resistor and everything is very stable now. Vp is stable at 250Volts @ 16mA.

When the CCS is in the circuit I can adjust Vp to 250~270V by varying Ip at the CCS. However when left for a few seconds it climbs to 290~300V. Decreasing Ip leads to a sudden fall of Vp to 160~180V.
I replaced one channel CCS for a resistor which resulted in stable behaviour but the other channel still on CCS is not stable (same behaviour as described before).
So I have a fixed Vg1 and Ia and a tube that is settling Vp all over the place..
 
You do understand that you can't set current at cathode and anode separately? The cathode current (and hence the anode current) is set by the grid voltage (from the previous stage anode) and the cathode resistor. You can't separately set the anode current. What you are seeing is not oscillation but the effect of the poor valve trying to satisfy your conflicting requirements; all it can do is adjust its anode voltage to get as close as possible to your impossible request. With a resistor at the anode it can adjust the anode current to match the cathode current; with a CCS it can't do this so has to do the best it can, which will normally mean the anode voltage going as low or high as possible.

So, you need to decide: do you want to set the valve current at the anode or cathode?
 
Yes, exactly, that is what my gut feeling said: You might be regulating too much.. I thought by adjusting the CCS to the demand of the cathode things might be okay, but it is not.
So, the design is my bad. But.. learned from it.
Things sound very much okay with the plate resistor, channel separation is my next focus.
Thanks for the advice and considerations!
 
Last edited:

euro21

Member
Paid Member
2004-09-17 12:16 pm
Budapest
In my opinion:
If you drive with CCS loaded tube anode the (low C) parafeed transformer, and the trafo another leg bound to cathode block capacitor, the PSRR has sharp edge at low frequency (primarily depends of the parafeed C and L), even tube conduct or not.

Before the real operating point will be stable (this stage is DC coupled from the large R-large C powered previous stage), this "signal" would be modulate the operating point.

If you destroy this sharp edge (charging time change, damping resonant Q etc.) the situation would be improves.

IMHO the easiest way to increase parafeed C value, to 20-47uF (depends of the transformer L).
 
Nope, it must be a DC-biasing problem. Leaving the parafeed transformer out of the circuit results in the same problem.
But thanks for the suggestions!

The problem is what the wideband D3A anode sees. It's nothing with biasing; it won't solve it....The SS CCS probably has an ft equal to that tube, and presents the anode as an extremely high undamped impedance in series with the output transformer leakage parasitics. Since the tube is screened, it's nearest companion to create Miller effect will be the wiring around the tube socket, the placement of g1 and anode wiring are opposite as shown. For audio work no-one needs these problems.
I bet if one replaced the CCS with an equivalent resistor, the problem will vanish, as confirmed in the forum. I had the same issue using EF184's, and replaced the CCS in the anode by a resistor, and there it stays. The damping effect of the resistor did the trick. Morgan Jones does mention that using a SS CCS will oscillate when any opportunity. One has been warned !
If one wants to keep the SS CCS, then it will have to be snubbed; but this might conflict with the value of the o/p transformer leakage resonant frequency into a complex oscillatory mode as the output loading will also effect.
Tough....might just have to put up with the resistor method.

richy