flipper wrote:

As some of you no doubt think I sometimes do strange things mixing
transistors with tubes and I got another one for your amusement.

The motivation stemmed from wanted to slap something together with
parts at hand and, well, not having much at hand.

Any how, it boiled down to trying a 12u7; not normally very useful
with a decent B+ and not a tremendous amount of gain either.

But I did have an MPSA42 so I tried using it as a protective cascode
for the low voltage 12u7, an amusing roll reversal as I normally use
them to protect low voltage, low noise, transistors from the tube
circuits (like in the Looking Glass phase splitter mirror).

But, rather than just use the MPSA42 to fix a low B+ for the 12u7 I
put the load R above it. That fixes the plate, at 23.3V in this case
(24V minus a Vbe drop), and isolates it from the output signal.

The upshot is a gain of 130.6 (120k load resistor above the cascode
with AF in 8.262mV, AF out 1.079V). 2'nd harmonic is -63dB with the
rest under -110dB. (in spice)

I decided to do a comparison (all in spice) with a 'normal' anode R
configuration and figured same plate, same AF in, same load R and same
Rk (240Ohm) was fair. That meant, for a 'normal' R straight on the
anode, using whatever B+ would result in 23.3V on the 12u7 plate. That
turned out to be 201V (with ~1.47mA through it, matching the ~1.47mA
in the fixed plate configuration).

Gain was 15.08 with 2'nd down -71.9dB. That means my 'trick' increased
gain by 9.38dB while increasing 2'nd harmonics 8.5dB. And while the
MPSA14 was 'in the signal path" spice showed no detectable influence
on the harmonic distribution.

Second comparison was to set equal gains, selecting whatever load R in
the fixed plate configuration would result in 15.08 (and all else
equal) since there's no way to get the 'normal' configuration to
130.6. That turned out to be ~13.75k with 2'nd harmonics still at
-63dB, virtually equal to the high gain case.

So, little value to it, other than protecting a low voltage device,
unless one is going for high gain as it sacrifices 8.5dB of
distortion. On the other hand, try getting a gain of 130.5 out of a
12u7 the 'proper' way

What is the exact schematic for what you have tried?

I have never heard of a 12U7. Is it actually a 12AU7?

Patrick Turner.

I have never heard of a 12U7. Is it actually a 12AU7?

Patrick Turner.

http://www.nostalgiaair.org/tubes/12u7.htm

Peter Wieck
Wyncote, PA

Just for chuckles I tried spicing that one into the cascode and it
performed surprisingly well too, with a fixed 30V plate. So did a
6DJ8. If the spice models are good.

I have a spice 3f4 model of the 12U7 at
http://home.netcom.com/~wb2jia/tubes/spice4.txt

Copy and paste into your spice SW.

flipper wrote:

On 6 Jun 2006 13:25:44 -0700, " wrote:

I have never heard of a 12U7. Is it actually a 12AU7?

Patrick Turner.

http://www.nostalgiaair.org/tubes/12u7.htm

That's an interesting substitute suggestion they've got since the
12AE7 is double dissimilar triodes while the 12u7 is twin identicals.

Different AF, gm, bias, plate resistance, and heater current.

Same basing though.


Peter Wieck
Wyncote, PA

12U7 is like a 12AU7 but meant to operate with Ea =12.6V.
Gm = 1.6mA/V, Ra = 12.5k, µ = 20, all at Ia = 1mA.
Maybe they used them in car radios.

I have no idea of the cascode schematic you spoke at length about, and I
guess nobody else did either.

Patrick Turner.

flipper wrote:

On Wed, 07 Jun 2006 04:14:07 GMT, robert casey
wrote:



Just for chuckles I tried spicing that one into the cascode and it
performed surprisingly well too, with a fixed 30V plate. So did a
6DJ8. If the spice models are good.

I have a spice 3f4 model of the 12U7 at
http://home.netcom.com/~wb2jia/tubes/spice4.txt

Copy and paste into your spice SW.

Thanks!

I thought that looked familiar Yeah, I got that set, plus the other
three, from your site and that's the 12u7 model I'm currently using in
circuitmaker

Btw, In case you're curious I put up a quick copy of two approaches
I'm tinkering with under the title "The Strange Case of Mixing Low
Voltage Tubes and Transistors in High Voltage Circuits or, what's that
12u7 doing in there?"

http://flipperhome.dyndns.org/

Ah, now I see what you are doing at

http://flipperhome.dyndns.org/LowHighMix.htm

What on earth is the advantage of using a 12u7 low gm triode to power an
emitter circuit of an MPSA42?
If the bjt voltage gain was 100, then RL for the triode is only 1.2k, so
the triode gain = 1.8 approx
so all that you achieve is a triode buffer with hardly any triode gain.
But it may sound well so all is not lost.
The normal way to cascode is to have the high gm device powering a cathode
of a triode with its grid taken to a fixed bias.

Perhaps there is an advantage I cannot see...

The mpsa42 has a fairly high amount of source resistance in series with
the base, so since the base
current = 1/hfe x Ic, a considerable signal voltage will exists at
the base as well as the emitter, and in
fact the gains are somewhat different....

BTW, where you have a mosfet of j-fet driving a tube cathode with fixed
grid bias
the overall gain is simply gm of the fet device x the above triode anode
RL, and this is true regardless
of whether you have a 12AT7 or 12AU7 as the triode.

Patrick Turner.

12U7 is a low voltage (design center max 30V plate) twin medium mu
triode originally intended for 12V car radio service. I had picked up
a couple, along with the 6GM8s, for the low voltage wall wart stuff I
was playing with (guitar to stereo preamp and the PC Speaker amp) but
ended up using the 6GM8s.

Here's a spice 3f4 model of the 6GM8:

* 6gm8 low voltage twin triode low mu Spice 3F4 model
..SUBCKT X6GM8 P G K

*the following 3 lines should be one very long line
Bp P K
I=((0.0652280007e-3)+(0.02616312706e-3)*V(G,K))*uramp((12.19320993)*V(G,K)+V(P,K)+(5.22 5386744))
^ 1.5 * V(P,K)/(V(P,K)+(0.599401572))

Cgk G K 1.7P
Cgp G P 1.3P
Cpk P K 0.5

..ends X6GM8

Also you can find this at the bottom of
http://home.netcom.com/~wb2jia/tubes/spice4.txt

"Patrick Turner" wrote in message
...


I have no idea of the cascode schematic you spoke at length about, and I
guess nobody else did either.



RDHB page 533.xi

Iain

Iain Churches wrote:
"Patrick Turner" wrote in message
...


I have no idea of the cascode schematic you spoke at length about, and I
guess nobody else did either.



RDHB page 533.xi

Iain

Eh? No illustration xi on p533 of the RDH4. Is the RDHB different from
the RDH4?

Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review

flipper wrote:

On Wed, 07 Jun 2006 15:18:27 GMT, Patrick Turner
wrote:



flipper wrote:

On Wed, 07 Jun 2006 04:14:07 GMT, robert casey
wrote:



Just for chuckles I tried spicing that one into the cascode and it
performed surprisingly well too, with a fixed 30V plate. So did a
6DJ8. If the spice models are good.

I have a spice 3f4 model of the 12U7 at
http://home.netcom.com/~wb2jia/tubes/spice4.txt

Copy and paste into your spice SW.

Thanks!

I thought that looked familiar Yeah, I got that set, plus the other
three, from your site and that's the 12u7 model I'm currently using in
circuitmaker

Btw, In case you're curious I put up a quick copy of two approaches
I'm tinkering with under the title "The Strange Case of Mixing Low
Voltage Tubes and Transistors in High Voltage Circuits or, what's that
12u7 doing in there?"

http://flipperhome.dyndns.org/

Ah, now I see what you are doing at

http://flipperhome.dyndns.org/LowHighMix.htm

What on earth is the advantage of using a 12u7 low gm triode to power an
emitter circuit of an MPSA42?

Hehe. Well, the 12u7 wasn't picked for it's gm, it was picked because
I have a couple. The 6GM8 would be 'better' in that regard, however,
after I did some modeling I decided I didn't need that much gain as I
only wanted 'just enough' NFB to lower pentode output impedance.

Btw, I originally mentioned a gain of 300, and it was, but I knocked
it down a bit in the one posted to 166 (increased the Rk).

It's not as easy as it might seem to get a lot more gm as the voltages
available, and my wanting to use 1 gain stage, force a rather high
load R, which means rather low idle currents, and that cuts way into
the gm of high gm tubes.

The 6GM8 does give me more gain, though, and I also have one of those
left, but, as I mentioned above, I decided I didn't need that much
gain (could always lower the load R though, for less, but I'm not sure
that buys me anything) and I would kind of like to keep it for a
future low voltage project where it's high gm could be put to better
use.

If the bjt voltage gain was 100, then RL for the triode is only 1.2k, so
the triode gain = 1.8 approx

But that isn't where the stage voltage gain takes place and, ideally,
it would be great if the effective RL there was 0.

so all that you achieve is a triode buffer with hardly any triode gain.
But it may sound well so all is not lost.
The normal way to cascode is to have the high gm device powering a cathode
of a triode with its grid taken to a fixed bias.

Well, yes. I realize that's the normal tube cascode.

The MPSA14 is doing essentially the same thing.

Setting aside, for the moment, real world issues of internal impedance
and such, the bjt fixes the 12u7 plate at 23 volts and passes signal
current on up to the 120k load R. Which is the same thing a tube
cascode does except transistors are quite happy working with no more
than a volt across them so I get more V available to drop idle
current across the load R.

I did spice it up with a tube cascode but didn't see any advantage to
it.

The MPSA14 isn't doing any gain but it is in the signal path so I
suppose that makes it bad audiophile JuJu.

Whoa, the majority of the gain is due to the mpsa14? doncha mean mpsa42?

Although, rather moot if I
use the current mirror phase splitter since they're 'in the path' too.

Not really. The mirror tries to keep the dc currents the same but essentially
any
sonic effect they have is well bypassed by the 330uF cathode C.


On the other hand, the current mirror phase splitter should be a near
perfect power supply noise eliminator since the exact same noise would
be imposed on both (I think, anyway, as I haven't analyzed the effect
of the MPSA92 cascode on that one) and then negated in the PP output
stage. The remaining 'noise point' would be the MPSA14 cascode Vref.

I managed to eliminate all the caps in the gain stage, though

Perhaps there is an advantage I cannot see...

The gain, mainly. Although the origins of it was to keep the 12u7
plate under 30V. Things 'evolved'

The mpsa42 has a fairly high amount of source resistance in series with
the base, so since the base
current = 1/hfe x Ic, a considerable signal voltage will exists at
the base as well as the emitter, and in
fact the gains are somewhat different....

Yes, there is some signal 'loss' to the base-emitter drive. Spice
shows that to be about 121nA out of a 9.01uA signal, about 1.3% .

A FET could be used to eliminate that but, overall, the loss is
imperceptible and akin to having a 1% load R tolerance, or the gain
being 164 vs 166, or 166 vs 168. Just doesn't seem to matter all that
much.

BTW, where you have a mosfet of j-fet driving a tube cathode with fixed
grid bias
the overall gain is simply gm of the fet device x the above triode anode
RL, and this is true regardless
of whether you have a 12AT7 or 12AU7 as the triode.

What would be the purpose of doing that?

A j-fet such as 2sk369 has gm = 40mA/V so if the RL at the triode anode is 20k,

the gain is 800 regardless of what triode you use. The voltage for the fet to
operate need only be so Ek of the triode is +8V, and all the other voltage is
across the triode.
Linearity depends on the transfer curcve for the fet since the triode
effectively has the Rd of the fet as its
Rk so its own thd is subject to reduction by current NFB.
A mosfet with gm higher than 2sk369 would give even more gain and this gain can
all be reduced
by a source resistor to tailor the gain and linearize the whole stage.
But a pair of mpsa42 connected as a darlington pair driving a triode would also
take advanatge of the
highgm needed for current drive into the cathode and the tube provides most of
the voltage gain...

In your circuit you need more than the output stage grid drive voltage because
you only have a
concertina driver stage.

Your circuit reminds me of so many from the old tube era where a couple of 6BM8

were used similarly. Dynaco ST70 went further with a triode-pentode which has
the pentode anode load bootstrapped
to give heaps more open loop gain.



Patrick Turner.

Got a question, though. The design center max plate for the 12u7 is
30V but do you know what that's based on? I mean, I can't believe the
thing would arc at, say, 100V so I imagine it's because of the
specially treated cathode.

I have never been interested in tubes meant for car battery power supplies.
I don't exactly know what they did to the cathode; something about the space
charge was different.



The question is, what's 'acceptable' prior to conduction, I.E. prior
to warmup? Say I did a tube cascode using the 12u7. The operating
points are within 30V but before they come on there would be, say, 280
on the top triode, which would then swing down through warmup to 54V
(27 on each), or so.

Use a neon tube to limit the high turn on voltage if you are worried.

if you are only experimenting, and nothing blows up, by the time you change it
to another circuit
maybe the tubes will be undamaged....

Patrick Turner.

Andre Jute wrote:

Iain Churches wrote:
"Patrick Turner" wrote in message
...


I have no idea of the cascode schematic you spoke at length about, and I
guess nobody else did either.



RDHB page 533.xi

Iain

Eh? No illustration xi on p533 of the RDH4. Is the RDHB different from
the RDH4?

I think our Iain meant RDH4 page 533 where at the page bottom a dissertation
(xi) starts on cascode
which continues into page 534 where there is a sketch of an incorrectly biased
dodgy cascode circuit
which isn't ever actually used like cascodes actually are used.
( the top tube has its grid drive directly connected to the bottom tube grid.)
Probably the author is trying to dodge trouble around patents or some other
reason.
But otherwise considerable math is presentented about cascode.

But cascode gain is easy to work out as well as Effective Rout from the top
anode.

The load on the bottom tube anode = anode load of the top tube / gain of the top
tube.
This is where top tube has a fixed grid voltage.

Patrick Turner.



Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review

Patrick Turner wrote:
Andre Jute wrote:

Iain Churches wrote:
"Patrick Turner" wrote in message
...


I have no idea of the cascode schematic you spoke at length about, and I
guess nobody else did either.



RDHB page 533.xi

Iain

Eh? No illustration xi on p533 of the RDH4. Is the RDHB different from
the RDH4?

I think our Iain meant RDH4 page 533 where at the page bottom a dissertation
(xi) starts on cascode
which continues into page 534 where there is a sketch of an incorrectly biased
dodgy cascode circuit
which isn't ever actually used like cascodes actually are used.
( the top tube has its grid drive directly connected to the bottom tube grid.)
Probably the author is trying to dodge trouble around patents or some other
reason.
But otherwise considerable math is presentented about cascode.

But cascode gain is easy to work out as well as Effective Rout from the top
anode.

The load on the bottom tube anode = anode load of the top tube / gain of the top
tube.
This is where top tube has a fixed grid voltage.

Patrick Turner.

No, no, no, Patrick, this won't do at all. I (and all the other
illiterates on RAT) demand a piccie. How can we be expected to
understand anything without a piccie? This Iain Churches is
*oppressing* by his elite learning when he demands we turn a page to
discover further information. Contrast, for instance, you own patience,
again demonstrated above, when we innumerates of RAT, by far the
majority (and therefore in a democracy by definition always *in the
right*), demand you do our math for us. I think, and of course everyone
who knows what is good for him will agree, that Iain should draw us a
picture. QED.

Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review

....with plenty of piccies of course!

PS I think the illustration in the RDH is one of those thingies that
says loud and clear to those who have burned their fingers a few times
"here's the principle but if you build it too bad it won't work until
you read a few pages further and instead build the actual working
schematic". The authors of a book like the RDH, and particularly in its
period, before the socialization of education, when even engineers were
usually correctly presumed to be gentlemen, had absolutely no worries
about frivolous suits for patent infringement.

Andre Jute wrote:

Patrick Turner wrote:
Andre Jute wrote:

Iain Churches wrote:
"Patrick Turner" wrote in message
...


I have no idea of the cascode schematic you spoke at length about, and I
guess nobody else did either.



RDHB page 533.xi

Iain

Eh? No illustration xi on p533 of the RDH4. Is the RDHB different from
the RDH4?

I think our Iain meant RDH4 page 533 where at the page bottom a dissertation
(xi) starts on cascode
which continues into page 534 where there is a sketch of an incorrectly biased
dodgy cascode circuit
which isn't ever actually used like cascodes actually are used.
( the top tube has its grid drive directly connected to the bottom tube grid.)
Probably the author is trying to dodge trouble around patents or some other
reason.
But otherwise considerable math is presentented about cascode.

But cascode gain is easy to work out as well as Effective Rout from the top
anode.

The load on the bottom tube anode = anode load of the top tube / gain of the top
tube.
This is where top tube has a fixed grid voltage.

Patrick Turner.

No, no, no, Patrick, this won't do at all. I (and all the other
illiterates on RAT) demand a piccie. How can we be expected to
understand anything without a piccie? This Iain Churches is
*oppressing* by his elite learning when he demands we turn a page to
discover further information. Contrast, for instance, you own patience,
again demonstrated above, when we innumerates of RAT, by far the
majority (and therefore in a democracy by definition always *in the
right*), demand you do our math for us. I think, and of course everyone
who knows what is good for him will agree, that Iain should draw us a
picture. QED.

LOL. I'll have to do. Time and patience is not someting blokes who are tree kwaters
irish
always have an abundance of.

I don't have time to exactly check out the math for cascode in RDH 4 for a pair of
triodes.

But to get the numbers for a cascode where the top triode grid is grounded, bottom Rk
bypassed,
then as I said above top tube gain = µ x RL / (RL + Ra ) or simply Va / ( Vg to Vk ) .

RL of the bottom tube anode looking into the top tube cathode = top tube RL / top
tube gain.

Gain of the bottom tube = µ x Rkin of top tube / Ra + Rkin of top tube.

Gain of total overall = gain of top tube x gain of bottom tube.

The above first principles logic flow allows the use of dissimilar triodes, say 12AU7
on the top and 6DJ8 on the bottom, or vice versa.


For ECC88/6922/6DJ8, with typical total RL = 40k load for top triode, gain top = 25
approx,
Rkin top = 40 / 25 = 1.375k, gain of the bottom = approx 7.
Total gain = 25 x 7 = 175.

Rout at the top tube anode = RL of top anode in parallel with effective Ra of top
tube.
Ra of top tube + [( µ + 1) x Ra of bottom tube.]
In this case with 6DJ8 its 40k // 160k approx = 32k approx, in other words
marginally lower than the RL, but very much above Ra for the top tube.
The cascode thus works like a poor man's pentoad.
The miller effect is reduced because bottom tube gain is lower than if you had the
tubes
cascAded rather than casOded.

Where the cascode is good is in an RF circuit where the top tube RL is only maybe 3k,
and
total gain is maybe only 25, but with no miller effect since the bottom tube gain is
very low.

The cascode circuit makes an excellent input stage for an MM amp.
And with a j-fet on the bottom its excellent for MC.

Allen Wright made a big deal of cascode in line stages in FVP and in some secret input
LTPs for
PP amps in a "Forced Symmetry" idea,
( to which I muttered "why force anything Allen, why not just let it be?")

The Hedge circuit of the 1960s also used cascoded pairs of triodes in a cross coupled
LTP gain block.

But I don't like them in power amps; too slow and not enough voltage headroom, and the
Rout is
high like a pentode, and I like driving output tubes with the low Ra of triodes.

Patrick Turner.






Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review

...with plenty of piccies of course!

PS I think the illustration in the RDH is one of those thingies that
says loud and clear to those who have burned their fingers a few times
"here's the principle but if you build it too bad it won't work until
you read a few pages further and instead build the actual working
schematic". The authors of a book like the RDH, and particularly in its
period, before the socialization of education, when even engineers were
usually correctly presumed to be gentlemen, had absolutely no worries
about frivolous suits for patent infringement.

AH, AH. I found it. Cpk P K 0.5 is missing the p for picofarad

Now it works. Thanks.

corrected my spice4.txt file per above, thanks

Don't happen to have a 6GU7 model, do you?


Try this:

* 6GU7 Spice 3F4 model
..SUBCKT X6GU7 P G K


Bp P K
I=(0.06810278442e-3)*uramp(V(P,K)*ln(1.0+(-0.07171045201)+exp((2.250885686)+(2.250885686)*((2 2.31757643)+(-193.4922211e-3)*V(G,K))*V(G,K)/sqrt((48.95241165)^2+(V(P,K)-(24.3417218))^2)))/(2.250885686))^(1.260156564)

Cgk G K 3.4P
Cgp G P 3P
Cpk P K 0.44p


..ends X6GU7

http://home.netcom.com/~wb2jia/tubes/spice4.txt
at the bottom

AH, AH. I found it. Cpk P K 0.5 is missing the p for picofarad

Now it works. Thanks.

Thanks, corrected it in
http://home.netcom.com/~wb2jia/tubes/spice4.txt

Don't happen to have a 6GU7 model, do you?

Added it to the bottom of
http://home.netcom.com/~wb2jia/tubes/spice4.txt

* 6GU7 Spice 3F4 model
..SUBCKT X6GU7 P G K

* next few lines below should be one line
Bp P K
I=(0.06810278442e-3)*uramp(V(P,K)*ln(1.0+(-0.07171045201)+exp((2.250885686)+(2.250885686)*((2 2.31757643)+(-193.4922211e-3)*V(G,K))*V(G,K)/sqrt((48.95241165)
^2+(V(P,K)-(24.3417218)) ^2)))/(2.250885686)) ^(1.260156564)
*the few lines above should be one line

Cgk G K 3.4P
Cgp G P 3P
Cpk P K 0.44p


..ends X6GU7

flipper wrote:



Super! Thanks

I've got 4 of those but didn't have a model.

Do you have a quick and simple way of creating new models? One that I
could use to do the same thing? And pentodes?

Reason I ask is I get oddball tubes from time to time and it would be
great to have a simple means to whip up a spice model for them.


I've been using curvecaptor for triodes and diodes. But no support for
pentodes unless they are triode connected. See
http://sourceforge.net/projects/curvecaptor/
and
http://sourceforge.net/project/showf...roup_id=138442

This program is a bit hard to learn. What you need as source data is a
gif image file (jpg I think works too) of the plate curves. From the
program you pick out which gif file, then with the mouse you start
plotting data points onto the image file. First are voltages (Vp axis),
then mouse up to a pull down menu to get into the plate current mode Ip
axis), and after that "plate chasterictics curve". You can change the
numbers in the little white boxes to the right of the menu selections.
Vp and Ip increment, but Vg does not. To start on Vg set it to 0v and
plant crosses (with clicking the mouse) at short intervals along the 0v
grid curve. Then for say grid voltage at -2v you have to edit the Vg
voltage in the little white box up top. The increment number doesn't do
anything here. Then start clicking on the -2V grid curve. "Save
Markers" fairly often, as I haven't figured out how to correct a
misplanted + yet. When I screw it up I revert to teh last batch of
saved markers. Eventually you get all the curves done, and then save
markers and then at the bottom hit "capture curves" and then you see a
list of numbers. Then hit "build model" and after about a half minute
you'll see the program's best fit to one of about 8 template models.
Then hit "see spice model" to get the model. On this screen, down below
is a selection of several varieties of spice model code (3f4, ORCAD,
Circuitmaker, and such). At this point just cut and paste (ctrl C and
ctrl V) into your spice simulator. There are a few syntax sillyness
you'll have to edit (like adding the X in front of the tube name).

flipper wrote:

On Fri, 09 Jun 2006 22:11:36 GMT, robert casey
wrote:


AH, AH. I found it. Cpk P K 0.5 is missing the p for picofarad

Now it works. Thanks.

Thanks, corrected it in
http://home.netcom.com/~wb2jia/tubes/spice4.txt

Don't happen to have a 6GU7 model, do you?

Added it to the bottom of
http://home.netcom.com/~wb2jia/tubes/spice4.txt


Super! Thanks

I've got 4 of those but didn't have a model.

Do you have a quick and simple way of creating new models? One that I
could use to do the same thing? And pentodes?

Reason I ask is I get oddball tubes from time to time and it would be
great to have a simple means to whip up a spice model for them.

Also read http://www.diyaudio.com/forums/showthread/t-56327.html
Mentions that you also need Tcl/Tk to run the GUI - get it from
ActiveState
http://www.activestate.com/Products/...x?id=ActiveTcl
Then run the curvecaptor.tcl (if you did associations properly
when installing Tcl/Tk, you'll just need to double-click it). Read this
thread for more info.

'make' is a build tool of Unix origins. It is used for automatic
compilation of source code. You don't have to do anything, though, as
you already have the binary distribution.

But is there some requirement about the graph dimensions? My first one
was apparently too big (I figured the bigger the better to pick
points) and there doesn't seem to be any 'fit' option.

Guess you have to edit the size of the image file using your favorite
image editor, like Photoshop. The setting of the size (resolution, ie,
800x600) on your VGA display

That also made me wonder about alignment as some of mine are
apparently manual scans of a book page and are either rotated a bit

I figured the Vp and Ip axises point plots are used by the program to
compensate out any rotation errors.

or, even worse, distorted from a good orthogonal.

DOn't know if the program will remove non linear image distortions

Mucho thanks.

Too bad there isn't a similar one for pentodes.

But this will help a LOT since, as I mentioned, I often end up with
oddball tubes.

Thanks.

Be sure to post the models you make. Thanks again

OK, but NEWBIE ALERT, NEWBIE ALERT

This is my first shot at it after carefully reading the instructions
that don't exist grin

All I claim is they 'do something' in my circuitmaker Which, in
case anyone missed that, means they're circuitmaker models.


My models made with curvecaptor

************************************************** *******************************

After adjusting for some cut and paste stupidity between my newsreader
to notepad and then to circuitmaker, your models look quite good. I
have a simple curve tracer simulation I use to show plate curves that I
can compare to published curves.

Below is a cut and paste of my triodecurv.ckt file, and hopefully you
can create a copy of this in your circuits directory and hopefully
circuitmaker can make sense of it. Start to cut and paste at the line
"CircuitMaker Test" and end at the last zero at the bottom.

triodecurv.ckt:

CircuitMaker Text
5.6
Probes: 5
v1#branch
AC Analysis
0 130 22 65280
v1#branch
Operating Point
0 130 22 65280
v1#branch
Transient Analysis
0 130 22 65280
v1#branch
Fourier Analysis
0 130 22 65280
v1#branch
DC Sweep
0 131 22 65280
41 35 62 4 97 108 86 46 82 16 73 113 113 59 105 107
0 5 0 1e+09 0.001 2
33
13 10 13 12 18 10 20 20 12 20
10 13 13 10 20 13 46 20 14 20
18 17 14 16 20 20 20 20 10 13
20 18 11
0 0 30 100 10
176 75 1020 719
7 5.000 V
7 5.000 V
3 GND
8.75e+06 10
24 100 0 1 0
20 Package,Description,
30 C:\PROGRAM FILES\CM60S\BOM.DAT
0 7
2 4 0.500000 0.500000
344 171 1188 493
144179216 0
0
0
0
0
0
0
8
7 Triode~
219 129 114 0 3 7
0 4 3 2
0
0 0 336 0
4 6GU7
30 -5 58 3
2 V2
44 -26 58 -18
0
0
14 %D %1 %2 %3 %S
0
0
0
10

0 1 2 3 1 2 3 6 7 8
27928028
88 0 0 0 0 0 0 0
1 V
5130 0 0
2
38877.7 0
0
7 Ground~
168 62 126 0 1 3
0 2
0
0 0 53360 0
0
4 GND2
-14 -26 14 -18
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
391 0 0
2
38877.7 1
0
5 SAVE-
218 130 22 0 10 11
0 0 0 0 0 0 0 0 0 0
1
0
0 0 57552 0
1 A
3 -26 10 -18
0
0
0
14 *DC 6.65 153 0
0
0
0
1

0 0
0 0 0 0 1 0 0 0
4 SAVE
3124 0 0
2
38877.7 2
0
7 Ground~
168 239 141 0 1 3
0 2
0
0 0 53360 0
0
0
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
3421 0 0
2
38877.7 3
0
9 V Source~
197 130 40 0 2 5
0 5 4
0
0 0 16752 0
2 0V
15 -2 29 6
2 V1
15 -12 29 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
8157 0 0
2
38877.7 4
0
9 V Source~
197 239 95 0 2 5
0 5 2
0
0 0 17008 0
4 300V
9 -2 37 6
3 VDS
12 -12 33 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
5572 0 0
2
38877.7 5
0
9 V Source~
197 64 72 0 2 5
0 3 2
0
0 0 17008 0
3 10V
12 -2 33 6
3 VGS
12 -12 33 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
8901 0 0
2
38877.7 6
0
7 Ground~
168 116 164 0 1 3
0 2
0
0 0 53360 0
0
4 GND1
-14 -26 14 -18
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
7361 0 0
2
38877.7 7
0
6
2 1 3 0 0 4224 0 1 7 0 0 3
103 114
103 51
64 51
3 1 2 0 0 4096 0 1 8 0 0 3
118 137
118 158
116 158
2 1 4 0 0 12416 0 5 1 0 0 4
130 61
130 74
129 74
129 88
2 1 2 0 0 8320 0 7 2 0 0 3
64 93
62 93
62 120
2 1 2 0 0 0 0 6 4 0 0 2
239 116
239 135
1 1 5 0 0 8320 0 5 6 0 0 4
130 19
130 12
239 12
239 74
0
0
4 0 0
0
0
3 VDS
0 500 2
3 VGS
0 -30 -2
3 0 1 4
0 8e-07 1e-08 1e-08
0
0
0 0 0
0 0 0
0
0
0 0 0 0
0
0 0 0 0 0
0 0 0 0
0
0 0 0
0
0 0 0
5 -1 10 10 10 0 10 10 0
3368 2259520 100 100 0 0
77 66 617 396
0 97 640 568
617 66
77 66
617 66
617 396
0 0
0 0 0 0 0 0
12401 0
4 100 200
0
0 0 100 100 0 0
77 66 767 186
0 0 0 0
767 66
77 66
767 66
767 186
0 0
0 0 0 0 0 0
12401 0
0 1 100
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0

robert casey wrote:


OK, but NEWBIE ALERT, NEWBIE ALERT

This is my first shot at it after carefully reading the instructions
that don't exist grin

All I claim is they 'do something' in my circuitmaker Which, in
case anyone missed that, means they're circuitmaker models.


My models made with curvecaptor

************************************************** *******************************



After adjusting for some cut and paste stupidity between my newsreader
to notepad and then to circuitmaker, your models look quite good. I
have a simple curve tracer simulation I use to show plate curves that I
can compare to published curves.

Below is a cut and paste of my triodecurv.ckt file, and hopefully you
can create a copy of this in your circuits directory and hopefully
circuitmaker can make sense of it. Start to cut and paste at the line
"CircuitMaker Test" and end at the last zero at the bottom.

triodecurv.ckt:

CircuitMaker Text
5.6
Probes: 5
snip

Looks like it works, I posted it and then cut and pasted from my post
back to circuitmaker, so nothing got munged. However, to get it to run
right, go into "Simulation" on the top toolbar, then pick "Analysis
setup". Enable the checkbox to the left of the DC button, and then
click the DC button. Unckeck all other buttons. When you do click the
DC button you'll see a window with primary and secondary (enable both).
You want VDS as primary and VGS as secondary. (now I remember what I
did to make this, I modified the FET curvetracer cir file, which I
cannot find right now). VDS is the plate B+ supply that circuitmaker
will vary, and VGS is the grid voltage
that also gets varied. Start, stop and step size I think are self
explanatory, remember to make the grid voltages negative.

"Andre Jute" wrote in message
oups.com...


No, no, no, Patrick, this won't do at all. I (and all the other
illiterates on RAT) demand a piccie. How can we be expected to
understand anything without a piccie? This Iain Churches is
*oppressing* by his elite learning when he demands we turn a page to
discover further information. Contrast, for instance, you own patience,
again demonstrated above, when we innumerates of RAT, by far the
majority (and therefore in a democracy by definition always *in the
right*), demand you do our math for us. I think, and of course everyone
who knows what is good for him will agree, that Iain should draw us a
picture. QED.

Hi Andre. I drew it out, with a big stick in the sand. Then, while I
was away looking for my digicamera, a 6ft wave came in........:-((

regards to all
Iain

Hi Andre. I drew it out, with a big stick in the sand. Then, while I
was away looking for my digicamera, a 6ft wave came in........:-((

regards to all
Iain

Are you suggesting that Mr. McCoy go "pound sand"?

Peter Wieck
Wyncote, PA

"Andre Jute" wrote in message
oups.com...

Iain Churches wrote:
"Patrick Turner" wrote in message
...


I have no idea of the cascode schematic you spoke at length about, and
I
guess nobody else did either.



RDHB page 533.xi

Iain

Eh? No illustration xi on p533 of the RDH4. Is the RDHB different from
the RDH4?



Hello Andre

I am referring to RDH4 Classic ed. The text re cascode begins page533.
There is a fundamental circuit fig 12.51B on page 544.


Morgan Jones "Valve Amplifiers" has a good description of the cascode
page 90-96 with a full circuit diagram on page 91 (fig 2.17)

Strangely enough, as Patrick points out, comparison of the RDH4
diagram with schematics found elsewhere shows up some remarkable
discrepancies.

People seem to have favoured the cascode for its low noise floor.
It was said to have the advantages of the pentode with none of the
drawbacks.

Regards to all
Iain

Yeah, it cut and pasted just fine and the DC was already set with the
appropriate parameters.

The only 'bleep' was the 6GU7 model it couldn't find because your name
for it was a bit different.

In short, it works.


Great. Below is a ckt file for diode curves:

CircuitMaker Text
5.6
Probes: 5
v1#branch
AC Analysis
0 130 22 65280
v1#branch
DC Sweep
0 130 22 65280
v1#branch
Operating Point
0 130 22 65280
v1#branch
Transient Analysis
0 130 22 65280
v1#branch
Fourier Analysis
0 130 22 65280
41 35 62 4 97 108 86 46 82 16 73 113 113 59 105 107
0 5 0 1e+09 0.001 2
33
13 10 13 12 18 10 20 20 12 20
10 13 13 10 20 13 46 20 14 20
18 17 14 16 20 20 20 20 10 13
20 18 11
0 0 30 100 10
176 75 1020 719
7 5.000 V
7 5.000 V
3 GND
8.75e+06 10
24 100 0 1 0
20 Package,Description,
30 C:\PROGRAM FILES\CM60S\BOM.DAT
0 7
2 4 0.500000 0.500000
344 171 1188 493
144179216 0
0
0
0
0
0
0
6
10 Vac Diode~
219 138 114 0 2 5
0 3 2
0
0 0 336 0
3 5U4
42 -5 63 3
2 V2
51 -26 65 -18
0
0
11 %D %1 %2 %M
0
0
0
5

0 1 2 1 2 -1610612676
88 0 0 0 0 0 0 0
1 V
5130 0 0
2
5.89246e-315 0
0
7 Ground~
168 129 175 0 1 3
0 2
0
0 0 53360 0
0
4 GND1
-14 -26 14 -18
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
391 0 0
2
5.89246e-315 5.26354e-315
0
5 SAVE-
218 130 22 0 10 11
0 0 0 0 0 0 0 0 0 0
1
0
0 0 57552 0
1 A
3 -26 10 -18
0
0
0
14 *DC 6.65 153 0
0
0
0
1

0 0
0 0 0 0 1 0 0 0
4 SAVE
3124 0 0
2
5.89246e-315 5.30499e-315
0
7 Ground~
168 239 141 0 1 3
0 2
0
0 0 53360 0
0
0
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
3421 0 0
2
5.89246e-315 5.32571e-315
0
9 V Source~
197 130 40 0 2 5
0 4 3
0
0 0 16752 0
2 0V
15 -2 29 6
2 V1
15 -12 29 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
8157 0 0
2
5.89246e-315 5.34643e-315
0
9 V Source~
197 239 95 0 2 5
0 4 2
0
0 0 17008 0
4 300V
9 -2 37 6
3 VDS
12 -12 33 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
5572 0 0
2
5.89246e-315 5.3568e-315
0
4
2 1 2 0 0 8320 0 1 2 0 0 3
127 137
129 137
129 169
2 1 3 0 0 4224 0 5 1 0 0 3
130 61
130 88
138 88
2 1 2 0 0 0 0 6 4 0 0 2
239 116
239 135
1 1 4 0 0 8320 0 5 6 0 0 4
130 19
130 12
239 12
239 74
0
0
4 0 0
0
0
3 VDS
0 30 0.1
0
0 0 0
3 0 1 4
0 8e-07 1e-08 1e-08
0
0
0 0 0
0 0 0
0
0
0 0 0 0
0
0 0 0 0 0
0 0 0 0
0
0 0 0
0
0 0 0
5 -1 10 10 10 0 10 10 0
2908 2259520 100 100 0 0
77 66 617 396
0 97 640 568
617 66
77 66
617 66
617 396
0 0
0 0 0 0 0 0
12401 0
4 10 20
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0

The only 'bleep' was the 6GU7 model it couldn't find because your name
for it was a bit different.

In short, it works.


And this one is for pentodes, one pentode connected, and another triode
connected. Note that this ckt assumes that the G3 is tied to the
cathode, thus the use of the tetrode symbol.

CircuitMaker Text
5.6
Probes: 5
v1#branch
AC Analysis
0 130 22 65280
v1#branch
Operating Point
0 130 22 65280
v1#branch
Transient Analysis
0 130 22 65280
v1#branch
Fourier Analysis
0 130 22 65280
v1#branch
DC Sweep
0 132 23 65280
41 35 62 4 97 108 86 46 82 16 73 113 113 59 105 107
0 5 0 1e+09 0.001 2
33
13 10 13 12 18 10 20 20 12 20
10 13 13 10 20 13 46 20 14 20
18 17 14 16 20 20 20 20 10 13
20 18 11
0 0 30 100 10
176 75 1020 719
7 5.000 V
7 5.000 V
3 GND
0 0
24 100 0 1 0
20 Package,Description,
30 C:\PROGRAM FILES\CM60S\BOM.DAT
0 7
2 4 0.500000 0.500000
344 171 1188 493
144179218 0
0
0
0
0
0
0
13
7 Ground~
168 95 269 0 1 3
0 2
0
0 0 53360 0
0
4 GND3
-14 -26 14 -18
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
8901 0 0
2
5.89289e-315 0
0
8 Tetrode~
219 107 232 0 4 9
0 4 4 3 2
0
0 0 336 0
4 6AU6
26 -30 54 -22
2 V6
47 -26 61 -18
0
0
17 %D %1 %2 %3 %4 %S
0
0
0
9

0 1 2 3 4 1 2 3 4 0
88 0 0 0 0 0 0 0
1 V
7361 0 0
2
5.89289e-315 5.26354e-315
0
9 V Source~
197 200 33 0 2 5
0 6 4
0
0 0 16752 0
2 0V
15 -2 29 6
2 V5
15 -12 29 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
4747 0 0
2
5.89289e-315 5.30499e-315
0
9 V Source~
197 351 102 0 2 5
0 7 5
0
0 0 16752 0
2 0V
15 -2 29 6
2 V4
15 -12 29 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
972 0 0
2
5.89289e-315 5.32571e-315
0
2 +V
167 447 177 0 1 3
0 7
0
0 0 53744 0
4 100V
-14 -22 14 -14
2 V3
15 -12 29 -4
0
0
13 %D %1 0 DC %V
0
0
0
3

0 1 1 0
86 0 0 0 1 0 0 0
1 V
3472 0 0
2
5.89289e-315 5.34643e-315
0
8 Tetrode~
219 140 119 0 4 9
0 8 5 3 2
0
0 0 336 0
4 6AU6
26 -30 54 -22
2 V2
47 -26 61 -18
0
0
17 %D %1 %2 %3 %4 %S
0
0
0
9

0 1 2 3 4 1 2 3 4 0
88 0 0 0 0 0 0 0
1 V
9998 0 0
2
5.89289e-315 5.3568e-315
0
7 Ground~
168 62 126 0 1 3
0 2
0
0 0 53360 0
0
4 GND2
-14 -26 14 -18
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
3536 0 0
2
5.89289e-315 5.36716e-315
0
5 SAVE-
218 130 22 0 10 11
0 0 0 0 0 0 0 0 0 0
1
0
0 0 57552 0
1 A
3 -26 10 -18
0
0
0
14 *DC 6.65 153 0
0
0
0
1

0 0
0 0 0 0 1 0 0 0
4 SAVE
4597 0 0
2
5.89289e-315 5.37752e-315
0
7 Ground~
168 239 141 0 1 3
0 2
0
0 0 53360 0
0
0
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
3835 0 0
2
5.89289e-315 5.38788e-315
0
9 V Source~
197 130 40 0 2 5
0 6 8
0
0 0 16752 0
2 0V
15 -2 29 6
2 V1
15 -12 29 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
3670 0 0
2
5.89289e-315 5.39306e-315
0
9 V Source~
197 239 95 0 2 5
0 6 2
0
0 0 17008 0
4 300V
9 -2 37 6
3 VDS
12 -12 33 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
5616 0 0
2
5.89289e-315 5.39824e-315
0
9 V Source~
197 64 72 0 2 5
0 3 2
0
0 0 17008 0
3 10V
12 -2 33 6
3 VGS
12 -12 33 -4
0
0
11 %D %1 %2 %V
0
0
0
5

0 1 2 1 2 0
86 0 0 0 1 0 0 0
2 Vs
9323 0 0
2
5.89289e-315 5.40342e-315
0
7 Ground~
168 128 167 0 1 3
0 2
0
0 0 53360 0
0
4 GND1
-14 -26 14 -18
0
4 GND;
0
0
0
0
3

0 1 1 0
0 0 0 0 1 0 0 0
3 GND
317 0 0
2
5.89289e-315 5.4086e-315
0
13
4 1 2 0 0 8192 0 2 1 0 0 3
96 255
95 255
95 263
0 3 3 0 0 16384 0 0 2 8 0 6
102 122
86 122
86 177
74 177
74 235
81 235
1 0 4 0 0 8192 0 2 0 0 4 3
107 206
107 191
200 191
2 2 4 0 0 4224 0 3 2 0 0 3
200 54
200 229
133 229
2 2 5 0 0 12416 0 6 4 0 0 5
166 116
184 116
184 198
351 198
351 123
1 0 6 0 0 0 0 3 0 0 13 2
200 12
200 12
1 1 7 0 0 8320 0 4 5 0 0 4
351 81
418 81
418 186
447 186
3 1 3 0 0 8320 0 6 12 0 0 4
114 122
102 122
102 51
64 51
4 1 2 0 0 8192 0 6 13 0 0 3
129 142
128 142
128 161
2 1 8 0 0 4224 0 10 6 0 0 4
130 61
130 79
140 79
140 93
2 1 2 0 0 8320 0 12 7 0 0 3
64 93
62 93
62 120
2 1 2 0 0 0 0 11 9 0 0 2
239 116
239 135
1 1 6 0 0 8320 0 10 11 0 0 4
130 19
130 12
239 12
239 74
0
0
4 0 0
0
0
3 VDS
0 300 1
3 VGS
0 -10 -1
3 0 1 4
0 5e-05 2.5e-08 2.5e-08
0
0
0 0 0
0 0 0
0
0
0 0 0 0
0
0 0 0 0 0
0 0 0 0
0
0 0 0
0
0 0 0
5 -1 10 10 10 0 10 10 0
2992 2259520 100 229 0 0
77 66 617 396
0 97 640 568
617 66
77 66
617 66
617 396
0 0
0 0 0 0 0 0
12401 0
4 50 200
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0
0 0 0 0 0 0
0 0 0 0
0 0 0 0
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0
0 0 0
0

Looks like it works but I have no diode models and I've never been
able to figure out how to add a model when circuitmaker grays out the
Model Data button.


Long time ago I faked it by editing a solid state diode model to become
a tube diode. I've since have a tube diode model, but I forgot how I
did that.

I managed to make a new sub class "tube diode" but, in keeping with
all the other screwball things in circuitmaker, it won't let me name
it vacdiode, because that exists (was hoping it would overwrite,
substitute, or default to adding a model since none are in it), won't
let me delete vacdiode, which it swears I must do in order to name
mine that, and it won't let me add anything to the existing vacdiode
nor will it macro copy anything to it either.

So I don't know how you got yours into vacdiode.sub.

Rename your old vacdiode.sub file to something else, and try using the
below to create a new vacdiode.sub file

*tube rectifier
..SUBCKT X38HK7R A K

b1 A K i=5.3e-3*uramp(V(A,K))^(1.51)
Cpk A K 11P
..ends X38HK7R
*tube rectifier (heater not done)
..SUBCKT X35W4 A K

b1 A K i=3.4e-3*uramp(V(A,K))^(1.44)
Cpk A K 2.4P
..ends X35W4

*tube rectifier
..SUBCKT X35Z5 A K

b1 A K i=3.4e-3*uramp(V(A,K))^(1.44)
Cpk A K 2.4P
..ends X35Z5

*tube diode detector
..SUBCKT X6AL5 A K

b1 A K i=2.2e-3*uramp(V(A,K))^(1.44)
Cpk A K 2.46P
..ends X6AL5

*tube rectifier (heater not done)
..SUBCKT X5641 A K

b1 A K i=1.04e-3*uramp(V(A,K))^(1.44)
Cpk A K 2.4P
..ends X5641

*tube diode detector
..SUBCKT X6AL5-4vh A K
*reduced heater voltage, crude limited plate current above 8V
b1 A K i=2.4e-3*(uramp(V(A,K))-uramp(V(A,K)-8))^(1.24)
Cpk A K 2.46P
..ends X6AL5-4vh

* 3DG4 - DH full-wave power rectifier (350mA) one section
..subckt X3DG4 P K
* Child-Langmuir law: mean fit error 3.52662 mA
* Bp P K I=(2.698962402m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 2.22284 mA
Bp P K I=(2.733421052e-3)*uramp(V(P,K)+(-0.3709046687))^(1.5)
* Perugini model: mean fit error 2.21657 mA
* Bp P K
I=((2.604459133m)+(-0.0005396471359m)*V(P,K))*uramp(V(P,K)+(-0.2400294538))^(1.514047938)
..ends X3DG4

* 5AR4 - IDH full-wave power rectifier (250mA) one section
..subckt X5AR4 P K
* Child-Langmuir law: mean fit error 3.57654 mA
* Bp P K I=(3.241174316m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 3.48688 mA
* Bp P K I=(3.257047182e-3)*uramp(V(P,K)+(-0.1046690061))^(1.5)
* Perugini model: mean fit error 0.685736 mA
Bp P K
I=((1.801056955e-3)+(-0.009848868441e-3)*V(P,K))*uramp(V(P,K)+(0.5667044198))^(1.7184223 81)
..ends X5AR4

* GZ34 - IDH full-wave power rectifier (250mA)
..subckt XGZ34 P K
* Child-Langmuir law: mean fit error 5.32372 mA
* Bp P K I=(3.968200684m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 4.64243 mA
Bp P K I=(4.050621191e-3)*uramp(V(P,K)+(-0.3522660485))^(1.5)
* Perugini model: mean fit error 1.81874 mA
* Bp P K
I=((1.770910894m)+(-0.002635208504m)*V(P,K))*uramp(V(P,K)+(1.90184266) )^(1.721344738)
..ends XGZ34

* 5AU4 - DH full-wave power rectifier (325mA) one section
..SUBCKT X5AU4 P K
* Child-Langmuir law: mean fit error 2.91988 mA
* Bp P K I=(0.9904121399m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 2.03865 mA
Bp P K I=(1.001318345e-3)*uramp(V(P,K)+(-0.5914399553))^(1.5)
* Perugini model: mean fit error 1.39243 mA
* Bp P K
I=((1.243495938m)+(0.001247697537m)*V(P,K))*uramp( V(P,K)+(-0.8353102945))^(1.43333432)
..ends X5AU4

* 5V3 - DH full-wave power rectifier (350mA) one section
..subckt X5V3 P K
* Child-Langmuir law: mean fit error 4.98363 mA
* Bp P K I=(1.085098267m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 4.57773 mA
Bp P K I=(1.093537823e-3)*uramp(V(P,K)+(-0.4353512857))^(1.5)
* Perugini model: mean fit error 2.33562 mA
* Bp P K
I=((0.9707914136m)+(-0.0009299376857m)*V(P,K))*uramp(V(P,K)+(-1.077144669))^(1.549894427)
..ends X5V3

* 5U4 - DH full-wave power rectifier (275mA) one section
..subckt X5U4 P K
* Child-Langmuir law: mean fit error 2.37547 mA
* Bp P K I=(0.7784729004m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 0.949764 mA
Bp P K I=(0.787802982e-3)*uramp(V(P,K)+(-0.6644060839))^(1.5)
* Perugini model: mean fit error 0.844211 mA
* Bp P K
I=((0.6623814366m)+(-0.0002767670064m)*V(P,K))*uramp(V(P,K)+(-0.07623613033))^(1.545073678)
..ends X5U4

* 5V4 - IDH full-wave power rectifier (175mA)
* Data from GE 5V4GA datasheet - traced on 09-09-01 by AF
..SUBCKT X5V4 P K
* Child-Langmuir law: mean fit error 1.39637 mA
* Bp P K I=(1.393901825e-3)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 1.20084 mA
Bp P K I=(1.404180087e-3)*uramp(V(P,K)+(-0.1877759095))^(1.5)
* Perugini model: mean fit error 0.948085 mA
* Bp P K
I=((1.324187528e-3)+(0.0007522446486e-3)*V(P,K))*uramp(V(P,K)+(0.1960102559))^(1.5054931 63)
..ends X5V4

* 5Y3 - DH full-wave power rectifier (125mA)

* Data from GE 5Y3GT datasheet - traced on 09-09-01 by AF
..SUBCKT X5Y3GT P K
* Child-Langmuir law: mean fit error 1.15597 mA
* Bp P K I=(0.3533050537m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 0.958542 mA
Bp P K I=(0.3560945008e-3)*uramp(V(P,K)+(-0.4420163009))^(1.5)
* Perugini model: mean fit error 0.656689 mA
* Bp P K
I=((0.3169037014m)+(5.675621737e-05m)*V(P,K))*uramp(V(P,K)+(0.6363978284))^(1.51803 242)
..ends X5Y3GT

* 274B - DH full-wave power rectifier (225mA)
..subckt X274B P K
* Child-Langmuir law: mean fit error 1.17132 mA
* Bp P K I=(0.3503318812m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 0.983373 mA
Bp P K I=(0.3458749385e-3)*uramp(V(P,K)+(0.539276053))^(1.5)
* Perugini model: mean fit error 0.393065 mA
* Bp P K
I=((0.368828156m)+(-0.0003153165025m)*V(P,K))*uramp(V(P,K)+(-0.6317505224))^(1.505633947)
..ends X274B

* 6X5 - IDH full-wave power rectifier (70mA)
..subckt X6X5 P K
* Child-Langmuir law: mean fit error 0.829063 mA
* Bp P K I=(0.6876983643m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 0.666578 mA
Bp P K I=(0.6965588691e-3)*uramp(V(P,K)+(-0.2713646197))^(1.5)
* Perugini model: mean fit error 0.350721 mA
* Bp P K
I=((0.3658726552m)+(-0.0008687929442m)*V(P,K))*uramp(V(P,K)+(1.38423595 7))^(1.683868589)
..ends X6X5

* 6X4 - IDH full-wave miniature power rectifier (70mA) (this is
essentially 6X5 in 7-pin base)
..subckt X6X4 P K
* Child-Langmuir law: mean fit error 1.10646 mA
* Bp P K I=(0.6787719727m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 0.603786 mA
Bp P K I=(0.6955865558e-3)*uramp(V(P,K)+(-0.5138012527))^1.5
* Perugini model: mean fit error 0.372965 mA
* Bp P K
I=((0.4819803969m)+(-0.001701085314m)*V(P,K))*uramp(V(P,K)+(-0.06370839623))^(1.636154216)
..ends X6X4

* 6CA4 - IDH full-wave miniature power rectifier (150mA)
..subckt X6CA4 P K
* Child-Langmuir law: mean fit error 0.585402 mA
* Bp P K I=(1.672480011m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 0.352346 mA
Bp P K I=(1.68085724e-3)*uramp(V(P,K)+(-0.1132675647))^(1.5)
* Perugini model: mean fit error 0.349954 mA
* Bp P K
I=((1.673471733m)+(-0.0001489801373m)*V(P,K))*uramp(V(P,K)+(-0.1167055168))^(1.502190772)
..ends X6CA4

* EZ81 - IDH full-wave power rectifier (150mA)
..subckt XEZ81 P K
* Child-Langmuir law: mean fit error 3.98557 mA
* Bp P K I=(1.703564453m)*uramp(V(P,K))^(1.5)
* Child-Langmuir law with contact potential: mean fit error 3.06551 mA
Bp P K I=(1.616914135e-3)*uramp(V(P,K)+(0.8508860213))^(1.5)
* Perugini model: mean fit error 1.36344 mA
* Bp P K
I=((3.069306816m)+(0.1206292787m)*V(P,K))*uramp(V( P,K)+(1.116202179))^(1.087711051)
..ends XEZ81

* EZ80 - IDH full-wave power rectifier (90mA)
..subckt XEZ80 P K
* Child-Langmuir law: mean fit error 1.07583 mA
* Bp P K I=(0.7265075684m)*uramp(V(P,K))^1.5
* Child-Langmuir law with contact potential: mean fit error 0.805197 mA
Bp P K I=(0.7527537983e-3)*uramp(V(P,K)+(-0.553916594))^(1.5)
* Perugini model: mean fit error 0.279904 mA
* Bp P K
I=((0.08550573077m)+(-0.000848630625m)*V(P,K))*uramp(V(P,K)+(3.586198141 ))^(2.164324675)
..ends XEZ80


* 5896 4v on 6.3 heater not valid over 30ma; with contact potential
..SUBCKT X5896-4hv P K
Bp P K I=(1.259102866e-3)*uramp(V(P,K)+(0.6788214279))^(1.5)
..ends X5896-4hv

* 6B10 diodes (common cathode) one diode section
..SUBCKT X6B10diode P K
Bp P K I=(1.437120118e-3)*uramp(V(P,K)+(0.6846418964))^(1.5
Cpk P K 1.85P
..ends X6B10diode

* 5903-13vh half heater voltage, crude plate current upper limit 8ma
..SUBCKT X5903-13vh P K
Bp P K
I=0.008-(uramp(0.008-((1.298177535/1.0e3)*uramp(V(P,K)+0.07-(0.09430728624))^(1.11))))
..ends X5903-13vh

* 5903-13vh CircuitMaker model
..subckt x5903vh13 P K
Bp P K I=(0.538177535/1.0e3)*uramp(V(P,K)+(0.9430728624))^(1.5)
..ends x5903vh13

Hi Robert,

why different models for 5AR4 and GZ34? So far
I thought this just to be the same tube within
American & European naming systems.

Tom

--
Anyone who considers protocol unimportant has
never dealt with a cat. - R. Heinlein

flipper wrote:

On Tue, 13 Jun 2006 02:33:42 GMT, robert casey
wrote:



I managed to make a new sub class "tube diode" but, in keeping with
all the other screwball things in circuitmaker, it won't let me name
it vacdiode, because that exists (was hoping it would overwrite,
substitute, or default to adding a model since none are in it), won't
let me delete vacdiode, which it swears I must do in order to name
mine that, and it won't let me add anything to the existing vacdiode
nor will it macro copy anything to it either.

So I don't know how you got yours into vacdiode.sub.

Rename your old vacdiode.sub file to something else, and try using the
below to create a new vacdiode.sub file
snip

Now that I've got things sorted out I went back to the pentode curve
tracer to check some in triode mode and, assuming your tracer is
accurate and I see no reason why it wouldn't be, there's no WONDER I
had biasing problems when I went to build the 6V6 Plain Jane amp. The
curves are way off, so I looked at the source and all the models of
the 6V6, 6BQ5 and 6AQ5 I've got are exactly the same, apparently
copies of each other except for the name, of course,... but the
published plate curves ain't.

6GK6 - 6BQ5... ok (sorta, it's off but not nearly as much) but the
6AQ5-6V6 are quite different.

6BQ5/EL84 to 6V6 shouldn't be close and their difference is like EL34 and
6L6GC.

The old data manuals I have don't seem to show a 6V6 = EL84.
Anyway, about the only curves I really need for any tube is the Ra line for
where Eg1 = 0V, and for pentodes and tetrodes these Ra lines and knees in
curves can vary widely for screen voltages.....

I have never placed any serious value on simulated testing of tube amps in
the PC because I'd be wasting a lot of time
and the real world breadboard tests and carefuly intelligent tweaking and
measuring/listening
is to me a shirtload more value than running airy fairy spice tests. Garbage
in = garbage out,
and so there is no substitute for real world testing with a hot soldering
iron and away from the keyboard.

Your comments about someone's tube curve tracer are amusing because anyone
that can make a tube tracer
which is accurate to within 3% of the real curves that exist is a true wizard
indeed. You should assume that unless you have proven the tube tracer to be
correct, it probably isn't. Never ever assume anything.
The testers used in 1957 were not always very accurate either; bloody liars
they were, actually.
And the curves shown in manuals are not always right; maybe drawn to look
pretty and interpolated by visual
reckoning between a series of dots on a page and made to look pretty. Often
the tubes can measure better than
the curves indicate especially when power triodes are involved which may give
less distortion that
indicated when using a ruler during load line analysis.
So entering curves from manuals into spice simulation programs could give a
dodgy outcome
especially with regard to distortions, where only slight line changes could
be interpreted by the PC algorithms
to produce very different relative harmonic spectra % compared to what ya get
when ya solda somink and measure the bugger.

Get real I say.

Patrick Turner.

Now that I've got things sorted out I went back to the pentode curve
tracer to check some in triode mode and, assuming your tracer is
accurate and I see no reason why it wouldn't be, there's no WONDER I
had biasing problems when I went to build the 6V6 Plain Jane amp. The
curves are way off, so I looked at the source and all the models of
the 6V6, 6BQ5 and 6AQ5 I've got are exactly the same, apparently
copies of each other except for the name, of course,... but the
published plate curves ain't.

6GK6 - 6BQ5... ok (sorta, it's off but not nearly as much) but the
6AQ5-6V6 are quite different.

The RCA tube manual says that, within its limitations, the 6AQ5 is the
same as the 6V6.

6BQ5/EL84 to 6V6 shouldn't be close and their difference is like EL34 and
6L6GC.

The old data manuals I have don't seem to show a 6V6 = EL84.

It's 6V6 = 6AQ5
Anyway, about the only curves I really need for any tube is the Ra line for
where Eg1 = 0V, and for pentodes and tetrodes these Ra lines and knees in
curves can vary widely for screen voltages.....

Be sure to set the screen voltage in my circuit to match that quoted in
the published curves. I've made that error embarrassingly often... :-)

I have never placed any serious value on simulated testing of tube amps in
the PC because I'd be wasting a lot of time
and the real world breadboard tests and carefuly intelligent tweaking and
measuring/listening
is to me a shirtload more value than running airy fairy spice tests. Garbage
in = garbage out,
and so there is no substitute for real world testing with a hot soldering
iron and away from the keyboard.

That's the next step after the simulation work is done. I wouldn't go
into production only on simulations. Even at the big name semiconductor
makers test "first silicon" on breadboards before going into production.

Your comments about someone's tube curve tracer are amusing because anyone
that can make a tube tracer
which is accurate to within 3% of the real curves that exist is a true wizard
indeed. You should assume that unless you have proven the tube tracer to be
correct, it probably isn't. Never ever assume anything.
The testers used in 1957 were not always very accurate either; bloody liars
they were, actually.
And the curves shown in manuals are not always right; maybe drawn to look
pretty and interpolated by visual
reckoning between a series of dots on a page and made to look pretty. Often
the tubes can measure better than
the curves indicate especially when power triodes are involved which may give
less distortion that
indicated when using a ruler during load line analysis.
So entering curves from manuals into spice simulation programs could give a
dodgy outcome
especially with regard to distortions, where only slight line changes could
be interpreted by the PC algorithms
to produce very different relative harmonic spectra % compared to what ya get
when ya solda somink and measure the bugger.

Flipper is doing what one needs to do with simulations. You see
something that looks crazy, and then you start digging in to find why.

Presumably the published curves were extracted from several tests of
nominal tubes. I've done at work characterization of various
semiconductor IC chips, and you always do lots of chips.

I hope you read this right after the other because I worked it out. I
decoded the .dat files and found where the user defined symbols and
standard ones are listed and edited them back. Takes more than one
place too, which is confusing as hell. Change it here and nothing
appears to change till you change it in the other 3 or 4 places.
sometimes twice in the same dern file.



I vaguely remember agony of this sort with trying to create my own
symbols and getting the models to go with them.


So now I have two identical copies in the menu: vacdiode and
tubediode. But I'll delete one of those out.

The built in "Vac Diode" is also back and just as useless as it was to
begin with. I don't know if it's needed as some kind of 'class'
reference to something else or what but I don't think I'll try
removing it again. Your default ckt still pulls that one up when the
macro utilities are invoked but it doesn't really matter.

I haven't messed with macros. I'll have to look to see what they do.
I know some pieces of this SW pretty well, and other pieces not at
all... :-)

The next step would be to breadboard it and see if the improvement or,
conversely, added distortion (in the common anode feedback comparator
case) 'mattered'. For example, in the SRPP case, tube variation
mismatch might negate a lion's share of the benefit, or maybe not.


Unfortunately, I don't have the equipment to do a 'real world' test of
that nature, so that will have to be left to others (or trust
Broskie), but the spice model indicates it might be worth looking at.

I've used my PC's soundcard and CoolEdit software to make some
distortion measurements. I've also used Cooledit to create various test
tones that I burned to a CDR to play from a decent CD player. That then
goes into the device under test, and the output of that via attenuation
feeds the soundcard. Cooledit captures this audio, and then I can then
use Cooledit's frequency spectrum plot. But first I connected the CD
player directly to the soundcard to verify that it doesn't add its own
distortions. Noise floor is about 95dB. 50dB harmonic levels show up
nicely. There are limitations, of course (my soundcard samples at
44.1KHz, newer ones can do 96KHz and thus would allow you to see any
ultrasonic crud).

I wasn't sure my sound cards were low enough distortion themselves, in
either output or capture, to be of much use but if yours works well
enough maybe mine would too.

That's why I did a test without the circuit under test first. Your
soundcard is probably fine for this, nothing special about my soundcard.
Just a 16 bit job.

I seem to be a bit short in the software audio tools department but I
need to look around and take stock of what I've got as I may have
something lurking in the archives.


I just made up a 6AQ5 spice model, it is very different from a 6BQ5.
I used the 5AQ5A designation (it's an audio tube made for series string
TV sets) so I can recognize it as one I made up. It seems to compare
well with the pentode and triode connected curves, except the slight
tetrode kink isn't really modeled.

*Vacuum Tube Tetrode (Audio freq.)
..SUBCKT X5AQ5A 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)/6)
*the URAMP(V(S,K)/# mostly determines peak plate current, grid line
spacing nearly constant, the sqrt and ^ adds some nonlinearity
*the number at the right end determines slope of grid lines
Bgs gs 0 V=URAMP(V(S,K)*0.19-(sqrt(V(G,K)*2.2)^(1.86))+V(A,K)/130)
*the exponent sets the linearity of grid line spacing, and big impact on
peak plate currrent
Bgs2 gs2 0 V=V(gs)^(1.55)
Bcath cc 0 V=V(gs2)*V(at)
*
* Calculate anode current, grid line spacing adjust and peak plate current
*
Ba A K I=0.26E-3*V(cc)
*
* Calculate screen current
*
Bscrn sc 0 V=V(gs2)*(1.1-V(at))
Bs S K I=0.46E-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 8p
Cak A K 8p

Cg1a G A 0.4p

..ENDS X5AQ5A

Also can be found at the bottom of
http://home.netcom.com/~wb2jia/tubes/spice.txt

Now that I've got things sorted out I went back to the pentode curve
tracer to check some in triode mode and, assuming your tracer is
accurate and I see no reason why it wouldn't be, there's no WONDER I
had biasing problems when I went to build the 6V6 Plain Jane amp. The
curves are way off, so I looked at the source and all the models of
the 6V6, 6BQ5 and 6AQ5 I've got are exactly the same, apparently
copies of each other except for the name, of course,... but the
published plate curves ain't.

6GK6 - 6BQ5... ok (sorta, it's off but not nearly as much) but the
6AQ5-6V6 are quite different.

As noted elsewhere in this thread, I just created a 6AQ5 circuitmaker
model. No magic software here, it was a lot of trial and error using
that pentodecurv circuit and adjusting various parameters in the models
to get it to look close to the published curves. I named it 5AQ5A, as a
way to remind myself that I made it up, and a 5AQ5A is a series string
TV tube identical to a 6AQ5 other than the heater. Find it at the
bottom of my page:
http://home.netcom.com/~wb2jia/tubes/spice.txt

This model ought to simulate a lot closer to your 6V6 circuit than a
6BQ5 model would. A 6BQ5 model doesn't come anywhere close, no wonder
it didn't work for you above.

With the problem finding circuitmaker models I thought I'd pull out
Orcad again and see if I could make heads or tails of it.

Well, I can made heads and tails but little else.

I don't suppose you have any idea how to make that blasted thing run a
simulation, do you?

Sorry, never worked with Orcad. Though I remember it as just being a
schematic capture tool. Though I've seen mention of it being a
simulator. Maybe you need additional software for the simulations?

Oh BROTHER!

One thing that we have to credit to Mr. McCoy is an active, shameless
and limitless imagination.

It reminds me in a perverse way of the attorney who died and went to
the Pearly Gates to meet St. Peter.

St. Peter went to his great book and, looked it over for a few moments
and said: I see by the records that you are 143 years old. The attorney
thought for a moment and said: Huh? I dropped dead at 47 with a massive
heart-attack. Whereupon St. Peter replied: I only added up your
billable hours.

Mr. McCoy will face similar questions when his many, varied and
oh-so-world-changing careers are added together.

Peter Wieck
Wyncote, PA

Follow up. Since I had some luck with the Orcad model I decided to try converting it
to circuitmaker, because I ran across some 'tips'. Like most 'tips' they ended up
being a bit short but, anyway, I do have something that at least runs. Maybe you'd
like a copy and maybe you could double check the conversion.

Duncan's, the author of the model, tips were...

snip

Thanks, I saved your post to a file. I'll try to convert some other
PSpice models. Your conversion looks pretty close to the 5AQ5 model I
made up. I can see someone using your 6V6 model and my 5AQ5A model in a
P-P amp simulation to see what happens with a pair of slightly
mismatched tubes. Harmonic distortions, bias mismatching if both share
a cathode resistor or the grids are fed a minus bias voltage.

On Thu, 15 Jun 2006 20:16:43 -0500, flipper wrote:

On Thu, 15 Jun 2006 19:16:41 GMT, robert casey
wrote:



With the problem finding circuitmaker models I thought I'd pull out
Orcad again and see if I could make heads or tails of it.

Well, I can made heads and tails but little else.

I don't suppose you have any idea how to make that blasted thing run a
simulation, do you?

Sorry, never worked with Orcad. Though I remember it as just being a
schematic capture tool. Though I've seen mention of it being a
simulator. Maybe you need additional software for the simulations?

I've got the 'Unison Suit', albeit an old version.

I was particularly frustrated when I posted as I *had* been able to
make it at least do 'something' before and I still don't know what's
so screwed up with the test I was trying at the time but I sorted it
out in general later on and even managed to get a 6V6 model to work in
it. Mainly by finding something that 'worked' and using that as a
guide to expand from.

Orcad makes Circuitmaker's obtuseness seem crystal clear by comparison
Which is one reason why I ended up using circuitmaker so much. I
was more interested in getting something done than in becoming an
Orcad guru.

Circuitmaker is simpler to use for small projects and has a much
better simulation interface but it's harder to find models for and
schematic capture has a nasty tendency to get confused and crash.

I'll probably stick with Circuitmaker when I've got good models for
it.

Have you tried LTSpice? Get it off www.linear.com for free. It has a
very big following and is continuously updated.

- YD.

--
Remove HAT if replying by mail.

flipper wrote:


Oh, wait, I just ran another curve on the 5aq5 with grid 15 to 0, step
-5 and I get 4 curves. Which is the same as I get with 0 to -15, step
-5, or 15 to -15, step 5. I'll bet it's reflecting around 0, rather
than just 'stopping', because the curve is there.. Or in math terms,
it appears there's something making it an absolute value.


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)..

Try the below, it gives reasonable positive grid curves except for less
than about 100V on the plate in tridoe mode:


*Vacuum Tube Tetrode (Audio freq.)
..SUBCKT X5AQ5A 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)/6)
*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)*0.19+V(G,K)*1.73+V(A,K)/130)
*the exponent sets the linearity of grid line spacing, and big impact on
peak plate currrent
Bgs2 gs2 0 V=V(gs)^(1.5)
Bcath cc 0 V=V(gs2)*V(at)
*
* Calculate anode current, grid line spacing adjust and peak plate current
*
Ba A K I=0.32E-3*V(cc)
*
* Calculate screen current
*
Bscrn sc 0 V=V(gs2)*(1.1-V(at))
Bs S K I=0.55E-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 8p
Cak A K 8p

Cg1a G A 0.4p

..ENDS X5AQ5A

On Sat, 17 Jun 2006 06:28:36 -0500, flipper wrote:

On Sat, 17 Jun 2006 01:36:21 -0300, YD wrote:

On Thu, 15 Jun 2006 20:16:43 -0500, flipper wrote:

On Thu, 15 Jun 2006 19:16:41 GMT, robert casey
wrote:



With the problem finding circuitmaker models I thought I'd pull out
Orcad again and see if I could make heads or tails of it.

Well, I can made heads and tails but little else.

I don't suppose you have any idea how to make that blasted thing run a
simulation, do you?

Sorry, never worked with Orcad. Though I remember it as just being a
schematic capture tool. Though I've seen mention of it being a
simulator. Maybe you need additional software for the simulations?

I've got the 'Unison Suit', albeit an old version.

I was particularly frustrated when I posted as I *had* been able to
make it at least do 'something' before and I still don't know what's
so screwed up with the test I was trying at the time but I sorted it
out in general later on and even managed to get a 6V6 model to work in
it. Mainly by finding something that 'worked' and using that as a
guide to expand from.

Orcad makes Circuitmaker's obtuseness seem crystal clear by comparison
Which is one reason why I ended up using circuitmaker so much. I
was more interested in getting something done than in becoming an
Orcad guru.

Circuitmaker is simpler to use for small projects and has a much
better simulation interface but it's harder to find models for and
schematic capture has a nasty tendency to get confused and crash.

I'll probably stick with Circuitmaker when I've got good models for
it.

Have you tried LTSpice? Get it off www.linear.com for free. It has a
very big following and is continuously updated.

Hadn't tried it but I downloaded it and will take a look. Thanks.


Judging by your gripes about Circuitmaker you might find it
friendlier. And a friendly large user base at Yahoo Groups.

- YD.

--
Remove HAT if replying by mail.