Input tube bias

[eeyore]

Trying to understand the workings of a tube input stage. If we assume that we have a maximum of 2Vrms at the input of a power amp, and assume that we don't want to drive the input tube into positive current (A2), then would it be fair to assume that the driver stage bias must be less then -2.8V given the peak to peak voltage of 2Vrms is 5.65Vpp? I understand that the full signal at 2Vrms is not common, but still, we are trying to avoid A2 operation on the driver stage. Am I making sense or do I have it completely wrong?

 

[llwhtt]

In a cathode biased tube stage you don't want to exceed the voltage on the cathode. With your example you would need 2.828VDC, maybe a little more, on the cathode to avoid any problems.

Craig

 

[eeyore]

Sorry, I forgot to mention I am talking about a grounded cathode with either grid or cathode bias. I assume that the tube specific cathode to grid voltage limitations must be respected. So assuming a 2Vrms input, the cathode bias should be say 3V to avoid A2 operation?

 

[Gordy]

eeyore, you have it right.

In addition, some tubes begin to draw grid current before reaching the 0 point, so if you expect to receive the full 2V rms (2.828V peak) input then I would suggest staying away from the -1 V area, hence a bias point around -4V (or more).

Regards,
Gordon

 

[eeyore]

Yeah, thanks for the replies, this can cleared this up for me. In short, biasing the input stage is no different to the power stage, to avoid A2, never let the grid swing to positive voltage and where possible, allow a bit of headroom. Cheers!

 

[kevinkr]

Once you have determined the minimum input bias you then need to determine the operating point you want and hence the plate voltage required to achieve the desired quiescent current. Of course you could just use a CCS and set the current to your target value - making sure that the B+ is sufficient to not run into voltage compliance issues at the most positive portion of the plate swing..

 

[eeyore]

So for argument sake using a ECC99 with a CCS. Bias the voltage of gid at -6V and setting the current at around 20mA would mean a standing voltage of 200V or so. Assuming a 2Vrms signal, the grid would experience say -3V to -9V. Meaning the voltage across the tube be between 130V and 270V. Therefore, the B+ supply for the ECC99 with a CCS should be say 300V? Or have a something amiss?

If it was choke/interstage loaded instead of a CCS, the B+ can be closer to the operating point voltage of a little over 200V (accounting for losses in the transformer)

 

[kevinkr]

So for argument sake using a ECC99 with a CCS. Bias the voltage of gid at -6V and setting the current at around 20mA would mean a standing voltage of 200V or so. Assuming a 2Vrms signal, the grid would experience say -3V to -9V. Meaning the voltage across the tube be between 130V and 270V. Therefore, the B+ supply for the ECC99 with a CCS should be say 300V? Or have a something amiss?

If it was choke/interstage loaded instead of a CCS, the B+ can be closer to the operating point voltage of a little over 200V (accounting for losses in the transformer)

In this case a choke sounds like the right way to go. Theoretically you can swing twice the supply voltage on the plate, in practice it will be significantly less but within the range you need. One of my designs swings more than 300Vpk on a 200V supply into a 50K load with excellent linearity..

 

[eeyore]

But if it were using a CCS, the B+ would need to be 300V then?

 

[kevinkr]

But if it were using a CCS, the B+ would need to be 300V then?

Even more actually for good linearity at the target amplitude, 350V - 400V depending on the tube used, that's one of the benefits of the choke.

 

[DualTriode]

Hello eeyore,
Checking assumptions. It sounds like you are looking at the workings of a grounded cathode triode input stage. You are talking of 2+- volts bias on the grid and wanting to avoid A2 operation. The discussion so far sounds like fixed bias possibly with the cathode resistor bypassed with a capacitor?
Another option is to let the bias float, get rid of the bybass capacitor. If you do this, the grounded cathode triode input stage can idle with a grid voltage closer to -1 volt and still not go into A2 with an input of 2 volts. As the input voltage goes in one direction the bias floats the other way. AKA local negative feedback or one less capacitor in the signal path.
DT
All just for fun.

 

[Gordy]

So for argument sake using a ECC99 with a CCS. Bias the voltage of gid at -6V and setting the current at around 20mA would mean a standing voltage of 200V or so. Assuming a 2Vrms signal, the grid would experience say -3V to -9V. Meaning the voltage across the tube be between 130V and 270V. Therefore, the B+ supply for the ECC99 with a CCS should be say 300V? Or have a something amiss?

That's right. Allows a few volts across the CCS even at maximum signal swing.


(In reality a bias point of -4V would most likely be fine. 160V on the anode, about 250V supply, 20mA CCS load. Plonk a pot and grid-stopper in front and cap couple the anode to a Hammond 119DA and you have a headphone amplifier!!!).

 

[eeyore]

How then would one determine the maximum dissipation of the CCS? Taking the ECC99 example above, the maximum current/voltage would be at the minimum voltage on the tube. Assuming 20mA, and B+ at 300V. Lowest swing is around 130V. Would then the CCS have to dissipate p=vi = 0.02 x 170 = 3.4W?

 

[Gordy]

Yes, that's right.