SPICE Models for the 7199 Pentode-Triode

LT spice, Curve Captor, PSUDII and whatever other sims you can think of.

Post Reply
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

SPICE Models for the 7199 Pentode-Triode

Post by Ray »

For Stephie:

I am a newcomer to this forum and would like to say "thanks" for all the work you have put into making SPICE models available to the rest of us.

One model that I miss is the 7199 pentode-triode used in the Dynaco amplifiers. I've simulated the ST-70 driver circuit in LTspice using 7199 tube models from an expanded library originated by Norman Koren, but its pentode model in particular gives nonsense results (like 2 VDC on the plate of the pentode section). I've attached the library I'm referring to; a more recent version is available on the Yahoo LTspice forum but it shows similar results to this one.

I've checked the attached schematic for errors many times but have found none, and if I simply replace the 7199 sections with the EF86 and 6CG7 from the same library, the simulation results look correct (or at least reasonable).

If you have the time, would you consider adding 7199P and 7199T models to your tetrode and triode_nh libraries? Or can you spot what may be wrong with the published Koren library?

Thanks.

Ray
Attachments
Dynaco ST-70 Driver.asc
Dynaco ST-70 Driver
(3.36 KiB) Downloaded 636 times
Koren_Tubes.inc
Koren Tube Library
(45.46 KiB) Downloaded 2418 times
Last edited by Ray on Tue Oct 16, 2012 2:29 pm, edited 2 times in total.
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Norman Koren has done a wonderful thing with his models. A lot of the devices modeled in your included library I do not have in the drop down. However, the pentode model used for that part and most of the ones using that model have a fatal flaw (in my opinion) in that the screen grid current does not change with changing plate voltage, and that produces incorrect results, since, in a pentode the most constant current is the space current, which is the sum of anode and grid2. Thus, as the plate voltage increases, the screen current should decrease. In real circuit terms, that means that when you feed the screen only from a dropping resistor, particularly when you pre-bias the cathode (R6/R2 in your circuit), you'll get unrealistic results. It also makes things more difficult when the tube is characterized for 10mA, and you're using it around the 1mA level.

For instance, if you change R4 from 270k to 27k and R6 to 100k, the circuit produces reasonable output, although with a little less equivalent circuit gain.

Stephie <3

As time permits, I'll add the 7199P and 7199T into the library
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

Thanks for the quick reply, Stephie. At least I now understand why that model (particularly for the pentode) doesn't produce a useable result.

If you can get to the 7199 models at some point, that would be great.

Thanks again for all you've done.

Ray
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Thanks. I've currently got the triode models "open" since I am adding several devices into the library (Using build_triode). Also, the library as a hodge-podge of several sources had inconsistent set of functionality. I have added in grid current capability into all the devices in the drop down, and indicated those that really support A2/B class operation. (The simple resistor-diode models do not, although they are expedient for at least not kidding you during simulation... Too many people have fallen into the simulator trap of driving the grid positive and not realizing that THAT doesn't come for free in the real world. Also, some of the models produced current when driven into A2. That's not only bad behaviour, it's really misleading in the simulation. I've fixed the handful of devices that had that problem.

I'm expecting to be finished with that update sometime this week... I'll try to put the 7199T into that as well. Then I'll open tetrode.txt and make those consistent as well. I'll put the 7199P into that one.

-Stephie <3
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

Super -- thanks!
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

No, I haven't forgotten. I started to get into the "tetrode" model files and found SO many inconsistencies and limitations. (including the ones I added too). I am in the process of coming up with a reasonably universal model, and will put a "Build Tetrode" file similar to the "Build Triode", although I may modify "Build Triode" as a result of the tetrode investigations.

Some of the features I'm currently putting into the tetrode model are:
1. Support for sharp and remote cutoff devices.
2. Ability to best handle beam power vs pentode curves...they have a different "knee".
3. "Early" effect. (change in plate resistance vs current level).
4. Correctly modeled screen current.
5. Correctly modeled G1 current (for Class AB2, B and C operation). Most people on this forum use these in audio applications, but there are a few "RF freaks" who do use these in transmitters.
6. Saturation slope for high current low voltage "left line" modelling.
7. Kink effect for modeling tetrodes (vs pentodes). That's the wiggles in beam power tubes and in most true tetrodes at low plate voltage (lower than screen) and usually at low currents... although some tubes (6V6/6AQ5) have that effect show up at almost any bias.

Currently still working out the details, although at present, the parameters you would enter are a whole lot more straightforward in that the "starting point" can be entered almost directly from the data sheet rather than trying to guess.

At this point, "wish list" can be entertained.

Stephie <3
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

Sounds great, Stephie. I'll be interested in finally modeling the original Dynaco ST-70 with its 7199 tube, something I haven't been able to do with the Koren models.

Your hard work on these libraries is really appreciated!

Ray
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Hi Ray and all. I've completed the model and added a 7199 as its first instantiation. I reran your Dynaco file with the new tetrode.txt and it seems to bias properly now. The anode voltage was 150 volts and transient analysis produces a reasonable results, so the model is behaving properly. I also increased the 1mV signal you were using as a source up in steps to 10 volts, and the system produced expected results. (Yes, it clipped severely at 10 volts input, but that's expected.

I will publish the tetrode.txt file and a new tool, BuildTetrode.asc in the next few days. In the mean time, attached is a "temporary" tetrode.txt file, with the 7199 added. Let me know if this seems to work out for you.

Warmest Regards,
Stephie <3
Last edited by sbench on Mon Nov 19, 2012 2:48 pm, edited 1 time in total.
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

Thanks, Stephie. My original ST-70 schematic had an incorrect B+ supply but the corrected version I posted shortly thereafter seems to produce reasonable results with your new model. The plate currents for the pentode and triode sections look about right, based on analysis of the original power supply and its documented voltages at a no-signal condition. This is great!

May I suggest that you name the subcircuit 7199P (rather than just 7199) to distinguish it from its triode cousin?

Ray
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Thanks. Actually it's pretty easy to create any tube using that model. It has one limitation that I will correct before releasing it, having to do with triode mode and A2 operation... need to add one parameter, but I may defer it... want to work on triode mode using pentodes more, but I need to implement the devices that have both sets of curves published in both -grid and +grid regions...

I will change the tube to 7199P in the tetrode.txt file... good suggestion.

Note to all..... I deleted the attached file on my previous post. Looking at the results with different tubes I found the following:

1. The method I used to change the cutoff of the tube is flawed. I need to do more work on it.
2. The params ARE easy to use, so it's pretty easy to add a new tube into the fray. (that was a +)
3. Operating the tube in triode mode with positive bias grid is severely flawed. I need to do work on that as well.

Ray: The 7199 as shown in the file you're using DOES work as intended, so you can continue your sims, but the distortion characteristics will not be entirely accurate... I'll, of course, include an updated model when I release the "official" files.




Stephie <3
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

Stephie,

I've been playing with the 7199 model and it is giving reasonable looking results. Thanks again for sharing your models.

Ray
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

Stephie,

I'm simulating the stock Dynaco ST-70 in stages and ran into some warnings and errors when using the 6CA7_EL34 model in the tetrode.txt library. I've attached the LTspice schematic for the output stage only, along with the log file from a transient simulation. (The warning about the single connection to Eyelet_13 is expected, since this feedback connection goes back to the driver input stage.)

I don't believe there is anything wrong with the circuit as drawn, so I thought I should bring this to your attention as you work on the tetrode library. If you do spot an error in my schematic, please let me know.

Thanks.

Ray

UPDATE:

I examined the tube model and noticed that the equation for Gg had curly braces for an interior expression instead of normal parentheses. The 6BQ5_EL84 model had the same error. I have corrected both subcircuits in the library and have attached that file here. The lines in question have been commented.

Problem solved!
Attachments
tetrode.txt
Revised tetrode.txt file
(24.14 KiB) Downloaded 584 times
ST-70 Output Stage.asc
ST-70 Output Stage
(6.16 KiB) Downloaded 576 times
ST-70 Output Stage.txt
LTspice Error Log
(1.77 KiB) Downloaded 451 times
terry123
Posts: 3
Joined: Tue May 20, 2014 6:42 am

Post by terry123 »

Too many people have fallen into the simulator trap of driving the grid positive and not realizing that THAT doesn't come for free in the real world.
soundguy
Posts: 1
Joined: Sat Nov 14, 2020 6:27 pm

7199 subcircuit for LTSpice

Post by soundguy »

I'm new to this forum, and am looking for the 7199 subcircuit mentioned in this thread. I downloaded tetrode.txt file from the link I found in the thread, but it does not include the 7199. Where can I download this subcircuit?

Thanks.
Post Reply