I found what I was doing to cause that error. Looks like when I took
V(G,K) to the 1.8something power and then the square root of that it
also did an absolute value as well... Took that out (I was trying to
impose some nonlinearity but that's the wrong place for that)..


I also corrected the 34DG5 and the 32ET5A models (audio pentodes used in
AM radios). Updates can be found at
http://home.netcom.com/~wb2jia/tubes/spice.txt
Don't know how many people were waiting on these, but here they are :-)

flipper wrote:

The 35B5s are
so dern cute I'm thinking of making a mini toy amp. Like maybe with a
pair of those junker 5 buck 'surplus' OPTs from AES so it looks kind
of like a 'real amp' shrunk down.


Use some sub-mini twin triodes with the 35B5s to complete the shrunk
down toy amp look. Like a 6111 low mu twin triode or a 6112 higher mu
twin triode.


:-)

flipper wrote:



Idle for the 35B5 is about 4.5W so, forget winding losses, R droppers,
and gain stages, just the two 35B5s come to about 20 watts for 3 watts
of audio (1.5W per channel). That's not a 'to scale' transformer

Maybe I can find a way to 'hide' it, or call David Copperfield.

So much for it being a 'simple' project


I just made up a 35C5 spice model (35B5 = 35C5 except for the pin out)
so maybe a simulation would be enough.... :-)

*Vacuum Tube Tetrode (Audio freq.)
..SUBCKT X35C5 A S G K
*
* Calculate contribution to cathode current
*
*the number at the right end determines sharpness of knee
Bat at 0 V=0.636*ATAN(V(A,K)/29)
*the URAMP(V(S,K)/# mostly determines peak plate current, grid line
spacing nearly constant
*the number at the right end determines slope of grid lines
Bgs gs 0 V=URAMP(V(S,K)/6.2+V(G,K)+V(A,K)/220)
*the exponent sets the linearity of grid line spacing, and big impact on
peak plate currrent
Bgs2 gs2 0 V=V(gs)^(1.58)
Bcath cc 0 V=V(gs2)*V(at)
*
* Calculate anode current, grid line spacing adjust and peak plate current
*
Ba A K I=1.20E-3*V(cc)
*
* Calculate screen current
*
Bscrn sc 0 V=V(gs2)*(1.1-V(at))
Bs S K I=1.1E-3*V(sc)
*
* Grid current (approximation - does not model low va/vs)
*
Bg G K I=(URAMP(V(G,K)+1)^(1.5))*50E-6
*
* Capacitances
*
Cg1 G K 12p
Cak A K 9p

Cg1a G A 0.6p

..ENDS X35C5

And at the usual web page:
http://home.netcom.com/~wb2jia/tubes/spice.txt

flipper wrote:

On Mon, 19 Jun 2006 19:22:14 GMT, robert casey
wrote:


flipper wrote:


The 35B5s are
so dern cute I'm thinking of making a mini toy amp. Like maybe with a
pair of those junker 5 buck 'surplus' OPTs from AES so it looks kind
of like a 'real amp' shrunk down.


Use some sub-mini twin triodes with the 35B5s to complete the shrunk
down toy amp look. Like a 6111 low mu twin triode or a 6112 higher mu
twin triode.


:-)


Hehe. Yeah, I thought about those but, wouldn't ya know it, none in
the 'free' bin They're surprisingly pricey too, for a toy amp
anyway, but I can get 6021s on the cheap.

I've got some 7 pin triodes and pentodes but they're a tad too tall
for a 'scaled' 12A'x'7, but maybe not for a 6CG7. Well, except you
can't get two in 7 pins so you end up with two bottles.

There's the 6J6 but the cathode is shared between the two triodes.
COmmon TV tuner tube in older sets from the early 1950's.

I also thought about doing a gasp solid state front end and just
hide it.

But the bigger problem is power iron. It doesn't scale and actually
works opposite because with low plate voltages, like these el-cheapo
xB5 tubes were intended for, you need more bias current to get even a
watt or two and that increases the needed filament power to burn all
them electrons off. That pip squeak 35B5 filament sucks 5.25 watts but
a 6V6/6AQ5, for over twice the output power (at twice B+), draws only
2.83W.

Maybe forget the 35B5s and use 6AQ5s.... :-)

flipper wrote:


The plate curves show screen, with 0 grid, dropping below 5mA
somewhere around 230, or so, but yours is up in the 20mA range. And
really dramatic under 50V, up to 110mA, or so, while theirs doesn't
make 25mA till 15V where it ends.


Spotted it: the published curves I used had a separate scale for screen
current (on the right). DUH! Anyway, I corrected the model, change
this line:
Bs S K I=0.23E-3*V(sc)

And at the bottom of http://home.netcom.com/~wb2jia/tubes/spice.txt is
the entire corrected model.

flipper wrote:



Oh, wait, I think I see it. If you look at the low voltage end of the
plate curves the model drops off differently than my datasheet.Take 0
grid and 20V B+. Published curve shows something around 68mA but your
model is doing around 44mA. The knee is too early and not as sharp.

I changed a few things, particularly the number that impacts the knee.
The line with the ATAN in it. But I can't recreate the tetrode kink
that well, so it will be off some at very low plate voltages.

*Vacuum Tube Tetrode (Audio freq.)
..SUBCKT X35C5 A S G K
*
* Calculate contribution to cathode current
*
*the number at the right end determines sharpness of knee
Bat at 0 V=0.636*ATAN(V(A,K)/13)
*the URAMP(V(S,K)/# mostly determines peak plate current, grid line
spacing nearly constant
*the number at the right end determines slope of grid lines
Bgs gs 0 V=URAMP(V(S,K)/6.2+(V(G,K)*1)+V(A,K)/270)
*the exponent sets the linearity of grid line spacing, and big impact on
peak plate currrent
Bgs2 gs2 0 V=V(gs)^(1.58)
Bcath cc 0 V=V(gs2)*V(at)
*
* Calculate anode current, grid line spacing adjust and peak plate current
*
Ba A K I=1.150E-3*V(cc)
*
* Calculate screen current
*
Bscrn sc 0 V=V(gs2)*(1.1-V(at))
Bs S K I=0.23E-3*V(sc)
*
* Grid current (approximation - does not model low va/vs)
*
Bg G K I=(URAMP(V(G,K)+1)^(1.5))*50E-6
*
* Capacitances
*
Cg1 G K 12p
Cak A K 9p

Cg1a G A 0.6p

..ENDS X35C5

http://home.netcom.com/~wb2jia/tubes/spice.txt