Spice model with grid current 12ax7

[karsten21]

Hello,

I try to simulate a simple gain stage-> Cathode follower using a 12ax7. Taking some literature/schematics in account it's wise to bias the cathode follower less than 1v ( say 0.5 ) to let flow a small grid current from anode of gain stage. See example schematic.

If I simulate this, the Uk of the cathode follower is not what I assumed. It seemed, that the grid current which should flow if the bias less then 1v is not simulated.

Is this a "normal" issue on spice models, means most/all tube spice models do not simulate this? If I take a look to the model description a grid current simulation is implemented but it seemed to me that this do not work.

At least I would like to get the same behavior of the circuit but with lower HT ( 250v ).

Anny hints?

[TheGimp]

I've seen no reference in the tube specs for grid dissipaton so I think that might be a bad idea.

In addition, dropping the current in the input stage may cut your distortion from 4.7% to 2% (compare 820R vs 1K8) if I am reading this data sheet correctly.

http://www.pentalaboratories.com/pdfs/ECC83.pdf

[Robert McLean]

A quick answer for now, I will look for 12AX7 data and post more later.

It is fairly common for tube SPICE models to ignore grid current. Or if they do model it they just use a diode and a resistor in series with the grid etc. Often there is not any grid current data in a spec sheet, so you have to measure it yourself, this may be a reason why.

In any case, the model you are using does have grid current modeled, but perhaps it is not accurate enough. As written it has grid current starting when grid voltage is -.2 volts or higher. You may want to change this to -1 volts. And you may want to increase the magnitude of the current.
Also it has a grid current that is dependent on grid voltage only, that is not realistic. There is a better grid current model by Rydel, I will look for it and get back later.

but for now try changing the line shown :

from

bgg g k i=5e-6*(uramp(v(g,k)+0.2)^1.5)


to
bgg g k i=2e-5*(uramp(v(g,k)+1.0)^1.5)

These are just guesses on my part, try them and see.

[Robert McLean]

Here is a Rydel style grid current model for the 12AX7.

change from
bgg g k i=5e-6*(uramp(v(g,k)+0.2)^1.5)

to
bgg g k I=IF(V(G,K)>0, 5.10044246234457E-09 * ((( 4238.969348634 +V(A,K))/( 281.39443771702 +V(A,K)))**4)*V(G,K)**1.5,0)

This will gives grid current matching that I have measured in a 12AX7 some time ago.