12x4 spice model

[Mahela]

Could someone with insight in spice modeling enlighten me in this matter:

I want to model a vacuum tube rectifier power supply and tried the 5Y3GT (available in multisim library) but somehow that spice model produced a very strange output (thousands of volts, -and my input AC is only 160 V) - that spice model looks like

.SUBCKT 5Y3GT A A K

*             .----- Anode
*             | .--- Cathode
*             | |
*SUBCKT 5Y3GT A K
BGP  A  K  i=2.69E-4*(V(A,K)^1.5+V(A,K)^1.5)/2
.ENDS 5Y3GT
============= Model template =================
x%p %tP1 %tP2 %tK_H1 %m

I then modified the model to a more accurate version by Stefano Perugini  at http://digilander.libero.it/paeng/vacuum_diode_models.htm and ended up with this model:

.SUBCKT 12X4 P P K
+ PARAMS Ka= 0.0012211748   Kb=   0.0004528104   
+ A=  0.550030857    Eps= 0.1
E1     1   0    VALUE = {Ka + Kb * V(P,K)}
RE1    1   0    1G
E2     2   0    VALUE = {V(P,K) + EPS}  
RE2    2   0    1G  
G1     P   K    VALUE = {V(1)/2 * (PWR(V(2), A) + PWRS(V(2), A))}
RPK    P   K    1MEG  

* CPK  P   K    .5n   

.ENDS

============= Model template =================
x%p %tP1 %tP2 %tH2 %tK_H1 %m

Now- obviously I have made a nonworking model, but I fail to see how, and - why is the EWB native model not working?

My circuit is more or less identical to mr Peruginis.

Thanks!

[lacy]

I tried the circuit in the picture and it seems to work correctly with the 5Y3GT from the Mater Database. I do not know why you are getting the results that you are seeing. What I am seeing is an output voltage of about 205VDC at the junction of the 220ohm resitor and the 100K resistor and one of the 220uF capacitors.

Maybe you need to check that your grounds are in the right place and that you are measuring from ground with your instrument. This is all I can tell you without examining your actual MS10 file.

As far as the other model is concerned I have not tried that yet and I am really no model expert. I will try to see if I can get the last model you posted to work, but I can make no promises on that.

I hope this helps somehow.

[OStep]

Mahela,

It appears you have something wrong with your symbol to model pin mapping

.SUBCKT 12X4 P P K
+ PARAMS Ka= 0.0012211748   Kb=   0.0004528104   
+ A=  0.550030857    Eps= 0.1
E1     1   0    VALUE = {Ka + Kb * V(P,K)}
RE1    1   0    1G
E2     2   0    VALUE = {V(P,K) + EPS}  
RE2    2   0    1G  
G1     P   K    VALUE = {V(1)/2 * (PWR(V(2), A) + PWRS(V(2), A))}
RPK    P   K    1MEG  

* CPK  P   K    .5n   

.ENDS

============= Model template =================
x%p %tP1 %tP2 %tH2 %tK_H1 %m

The template tells me that you're calling a subcircuit with 4 pins, but in fact your subcircuit only has 3 pins.  Double click on your component, go to 'edit component in DB' and go to model tab. Make sure you only map the 3 symbol pins to the model pins in the order they appear on .subckt line. That 4th symbol pin should be 'nc', for 'no connection"

Regards,

[Mahela]

Thank you for your comments.

The simulation of native EWB 5Y3GT produces an error log that says "timestep too small" and "problem in node

6" which is the junction on heater R12, C11 and R12. ?? The scope is attached in junction pin8 and R12 and

produces an output climbing to 2.5 kV prior to abortion.
Is it necessary to somehow save the new model to make it accessible to EWB?

I attach the error log of 12X4 and the EWW model 5Y3GT simulations if that makes any sense.
I modified the part 12X4, and right now it is made up of with an A and B contained in one package (template

is the same but with 1P1 and 1P2) both with (p)plate and (k)cathode in simulation model.

##################  Model Data Report  ##################

.SUBCKT 12X4     P K
+ PARAMS Ka= 0.0012211748   Kb=   0.0004528104   
+ A=  0.550030857    Eps= 0.1
E1     1   0    VALUE = {Ka + Kb * V(P,K)}
RE1    1   0    1G
E2     2   0    VALUE = {V(P,K) + EPS}  
RE2    2   0    1G  
G1     P   K    VALUE = {V(1)/2 * (PWR(V(2), A) + PWRS(V(2), A))}
RPK    P   K    1MEG  
* CPK  P   K    .5n   
.ENDS

============= Model template =================
x%p_b %t2P1 %t2P2 %m

Regards and thanks all.

[lacy]

Is there any way you could post your original Multisim file with the 5Y3GT tube so that it may be examined? hooked everything up according to your picture and got no error messages and this makes me interested in what may be occurring here.

[Mahela]

Yes of course.

I get (on 12X4 circuit) a netlist error prior to running the simulation. Did you managed to run the 12X4 model successfully?

Your  help is most valuable.

M

[OStep]

confirm your pin mappings...

right now the tube's A and K model pins are each connected to a sinusoidal voltage sources, so I am not sure what you planning to accomplish

Also, I suggest you use the expression

BGP  A  K  i=2.69E-4*(pwr(V(A,K),1.5)+pwrs(V(A,K),1.5))/2

in your model

[Mahela]

Thanks MaxNI for your comments. I am not sure about your suggestions, would you please clarify. About the schematics: (5Y3GT) the AC source should be connected to the both plates and the cathode to R12, well as I see it here. I changed the model as to your suggestions, and still no luck. I am confused as this is a EWB native model. I fail to see that the plate and cathode should be both connected to the AC source? Is that visiable in the model? The aim is to make a fullwave rectifier with this vacuum tube diode.

I enclose a screenshot of the desktop.

Regards all.