Problem in HF class-A amplifier. Delayed response to signal

[neazoi]

Hello, I have made a small amplifier for HF (1-30MHz), operating in class-A.
The amplifier works ok if I continuously fed an input signal.
But when I try to switch on/off the input signal, the amplifier fails to respond quickly to these changes and it slowly increases it's output signal up to the maximum. This change takes as long as 1 sec to happen from min to max output signal.
What could the problem be?

 

[BradtheRad]

Do you have a DC blocking capacitor somewhere? If its value is very large, then it will take a while to adapt after large abrupt changes. (Such changes as might come from switching something on or off.)

The bias voltage is thrown off, while the capacitor is gradually charging. Eventually it reaches the state where you get normal operation again.

 

[mtwieg]

Anything on the timescale of seconds if going to be thermal drift, I bet. Unless you are repeatedly exceeding the FET's SOA and you're watching it die slowly.

 

[neazoi]

Do you have a DC blocking capacitor somewhere? If its value is very large, then it will take a while to adapt after large abrupt changes. (Such changes as might come from switching something on or off.)

The bias voltage is thrown off, while the capacitor is gradually charging. Eventually it reaches the state where you get normal operation again.

I have DC blocking/coupling caps at the input and the output of the amplifier.
100nF. Are they too large? I am interested in 1-30MHz range

 

[tony_lth]

What is the power output of your PA?

 

[BradtheRad]

I have DC blocking/coupling caps at the input and the output of the amplifier.
100nF. Are they too large? I am interested in 1-30MHz range

The RC time constant comes into play here.

Your input resistance is probably high. Perhaps it goes low when the input is shut off, then it goes high when the input is turned on? And the capacitor must acquire it's charge all over again, through a very high resistance?

Or, perhaps you switch in the input, causing the capacitor to pull the bias very high, or very low?... Until it acquires a charge, then the amplifier can again operate in its normal region.

Anyway the problem might be solved if you reduce the capacitor value.

 

[albbg]

I suppose your system is a 50 ohms, I don't think the size of the capacitors can play a role (we are speaking of seconds). It seems something related to the thermal behavior or to a self-biasing.
What kind of technology are you using (LDMOS, GaN, etc) ? Is the transistor properly biased ?

 

[neazoi]

The RC time constant comes into play here.

Your input resistance is probably high. Perhaps it goes low when the input is shut off, then it goes high when the input is turned on? And the capacitor must acquire it's charge all over again, through a very high resistance?

Or, perhaps you switch in the input, causing the capacitor to pull the bias very high, or very low?... Until it acquires a charge, then the amplifier can again operate in its normal region.

Anyway the problem might be solved if you reduce the capacitor value.

It is an HF BJT class-A amplifier 1W output, 100mW or less input.
I have done all measurements on 50R scope, but I do not think the input and output of the amplifier are matched to 50R, because there is just a 100nf cap in the input to the base, and a choke-capacitor in the collector DC/output.
I do not know if that helps but I have noticed that if the RF drive level is low, this behaviour happens. When I increase the input RF lebel things are much better (no lag). But I need the low drive level so that the harmonics are kept to minimum.
Shall I play with the base bias of the amplifier transistor?

 

[albbg]

It is an HF BJT class-A amplifier 1W output, 100mW or less input.
I have done all measurements on 50R scope, but I do not think the input and output of the amplifier are matched to 50R, because there is just a 100nf cap in the input to the base, and a choke-capacitor in the collector DC/output.
I do not know if that helps but I have noticed that if the RF drive level is low, this behaviour happens. When I increase the input RF lebel things are much better (no lag). But I need the low drive level so that the harmonics are kept to minimum.
Shall I play with the base bias of the amplifier transistor?

I think you are experiencing biasing problems. Are you sure the transistor is driven a class A ?

 

[Dan Mills]

Bipolar power stages are notorious for requiring that the bias supply be really stiff, especially if they are not actually running in class A (A QRO push/pull bipolar amp can demand AMPS of bias current under some conditions).

How are you keying the thing?
If your keying disconnects the input, then upon it being reconnected, the input cap (Which will have self discharged) will take some time to recharge via the bias network, during which time the device will be under biased.
The cure if this is the case is to stick a resistor across the input so the cap has a path to bleed off leakage current when key up.

Regards, Dan.

 

[neazoi]

Bipolar power stages are notorious for requiring that the bias supply be really stiff, especially if they are not actually running in class A (A QRO push/pull bipolar amp can demand AMPS of bias current under some conditions).


How are you keying the thing?
If your keying disconnects the input, then upon it being reconnected, the input cap (Which will have self discharged) will take some time to recharge via the bias network, during which time the device will be under biased.
The cure if this is the case is to stick a resistor across the input so the cap has a path to bleed off leakage current when key up.

Regards, Dan.

You are right, the amp is keyed by switching the input RF on and off (from previous stages.
You mean to connect a resistor at the input of the amplifier to the ground?
If I scale the capacitor accordingly, this could work into benefit, to soften the keying at the rising edge?

 

[D.A.(Tony)Stewart]

One time constant is associated with the risetime or inverse BW of the resonator (LC?)

The other is the negative feedback circuit for DC biasing the current, which is more likely the problem affecting output power regulation.

Pls. Post your circuit and work done with waveforms.