I traced it from the Mazda Belvu datasheet.This is the best 12AX7 model I have, made by Wayne Clay using the Adrei Frolov curve captor :
snip...
It almost fit like a glove on a hand, specially in the most useful area, the bottom left corner ...
I always do that when I use a new model for my simulations.
Here's the LTSpice variant:
Code:
* 12AX7A-mz LTSpice model
* Rydel model (5 parameters): mean fit error 0.0364781mA
* Traced by Wayne Clay on 10/01/2006 using Curve Captor v0.9.1
* from Mazda Belvu data sheet
.subckt 12AX7A-mz P G K
Bp P K I=
+ ((0.001162961073m)+(0.0001179305949m)*V(G,K))*uramp((89.27317344)*V(G,K)+
+ V(P,K)+(50.06656412))**1.5 * V(P,K)/(V(P,K)+(1.191219264))
Cgk G K 2.3p ; 0.7p added (1.6p)
Cgp G P 2.3p ; 0.7p added (1.6p)
Cpk P K 0.53p ; 0.2p added (0.33p)
Rpk P K 1G ; to avoid floating nodes
d3 G K dx1
.model dx1 d(is=1n rs=2k cjo=1pf N=1.5 tt=1n)
.ends 12AX7A-mz
Code:
* 12AX7A-mz PSpice model
* Rydel model (5 parameters): mean fit error 0.0364781mA
* Traced by Wayne Clay on 10/01/2006 using Curve Captor v0.9.1
* from Mazda Belvu data sheet
.subckt 12AX7A-mz P G K
Gp P K VALUE=
+ {((0.001162961073m)+(0.0001179305949m)*V(G,K))*limit((89.27317344)*V(G,K)+
+ V(P,K)+(50.06656412),0.0,1.0e16)^1.5 * V(P,K)/(V(P,K)+(1.191219264))}
Cgk G K 2.3p ; 0.7p added (1.6p)
Cgp G P 2.3p ; 0.7p added (1.6p)
Cpk P K 0.53p ; 0.2p added (0.33p)
Rpk P K 1G ; to avoid floating nodes
d3 G K dx1
.model dx1 d(is=1n rs=2k cjo=1pf N=1.5 tt=1n)
.ends 12AX7A-mz
Last edited:
6J6 3f4 Spice:I also look for a 6J6A model ?
Another strange tube, a twin medium-mu triode with a common cathode ... This tube can be use in a phase splitter, a two channel mixer or both triode in parallel to increase the current. But it is not a very linear tube to use as an amplifier.
Alain
Code:
* 6J6 Spice 3F4 model
* Modified Koren model (8 parameters): mean fit error 0.146816 mA
* Traced by Wayne Clay on 2/12/2014 using Curve Captor v0.9.1
* from General Electric data sheet
.subckt 6J6 P G K
Bp P K I=(0.02975503554m)*uramp(V(P,K)*ln(1.0+(-0.03870606743)+exp((2.609078816)+(2.609078816)*((44.4681031)+(-320.4695508m)*V(G,K))*V(G,K)/sqrt((45.24212568)^2+(V(P,K)-(7.633778567))^2)))/(2.609078816))^(1.325356539)
Cgp G P 1.8p ; 0.2p added
Cgk G K 2.4p ; 0.2p added
Cpk P K 0.9p ; 0.2p added
Rpk P K 1G ; to avoid floating nodes
d3 G K dx1
.model dx1 d(is=1n rs=2k cjo=1pf N=1.5 tt=1n)
.ends 6J6Wayne
Thanks for the PSpice 12AX7 model Wayne. I see Curvecaptor uses the LIMIT function in lieu of URAMP.
I also see that Curvecaptor does not model grid current I assume the diode was placed there as a mod to the model?
Fortunately, in the EXCEM models, grid current is included. Perhaps they can be combined.
Attached is an excerpt of the EXCEM/Rydel library converted to PSpice syntax and with Ip correction (rebounding Ip).
I also see that Curvecaptor does not model grid current
I assume the diode was placed there as a mod to the model?Fortunately, in the EXCEM models, grid current is included. Perhaps they can be combined.
Attached is an excerpt of the EXCEM/Rydel library converted to PSpice syntax and with Ip correction (rebounding Ip).
PSpice uses LIMIT, 3f4 and the like, use URAMP. Some simulators use/understand their own combinations i.e. Micro-Cap. Curve Captor outputs limit or uramp depending on what flavor of spice you've chosen.Thanks for the PSpice 12AX7 model Wayne. I see Curvecaptor uses the LIMIT function in lieu of URAMP.
I also see that Curvecaptor does not model grid current
The main intent (especially with the Koren 8 models) of the added diode is an aid in convergence. For example, as when simulating a CF that's directly coupled to the proceeding stage. It can be commented out if you so desire.
No
, Curve Captor does not model grid current. Nor does it model in the positive grid region.
Last edited:
After much frustration, I finally realized the 1626 model I found three months ago is not accurate. Plotting the triode curves in LTSpice confirmed this. So, I have created a new model (see below). It fits the negative grid curves very well, but is a little off with the positive grid curves. Definitely an improvement over the other 1626 model which is floating about...
Also, I noticed the 76 model ran very slowly in simulations. So, I updated that one as well:
Code:
.SUBCKT TRIODE_1626 1 2 3 ; Plate Grid Cathode
+ PARAMS: CCG=3.2P CGP=4.4P CCP=3.4P RGI=2000
+ MU=5.63 KG1=5010 KP=30 KVB=115.5 VCT=3.906E-5 EX=1.53
E1 7 0 VALUE={V(1,3)/KP*LOG(1+EXP(KP*(1/MU+(VCT+V(2,3))/SQRT(KVB+V(1,3)*V(1,3)))))}
RE1 7 0 1G ; TO AVOID FLOATING NODES
G1 1 3 VALUE={(PWR(V(7),EX)+PWRS(V(7),EX))/KG1}
RCP 1 3 1G ; TO AVOID FLOATING NODES
C1 2 3 {CCG} ; CATHODE-GRID
C2 2 1 {CGP} ; GRID=PLATE
C3 1 3 {CCP} ; CATHODE-PLATE
D3 5 3 DX ; POSITIVE GRID CURRENT
R1 2 5 {RGI} ; POSITIVE GRID CURRENT
.MODEL DX D(IS=1N RS=1 CJO=10PF TT=1N)
.ENDSAlso, I noticed the 76 model ran very slowly in simulations. So, I updated that one as well:
Code:
.SUBCKT TRIODE_76 1 2 3 ; Plate Grid Cathode
+ PARAMS: CCG=3.5P CGP=2.8P CCP=2.5P RGI=2000
+ MU=14.5 KG1=3270 KP=144 KVB=4896 VCT=3.906E-5 EX=1.46
E1 7 0 VALUE={V(1,3)/KP*LOG(1+EXP(KP*(1/MU+(VCT+V(2,3))/SQRT(KVB+V(1,3)*V(1,3)))))}
RE1 7 0 1G ; TO AVOID FLOATING NODES
G1 1 3 VALUE={(PWR(V(7),EX)+PWRS(V(7),EX))/KG1}
RCP 1 3 1G ; TO AVOID FLOATING NODES
C1 2 3 {CCG} ; CATHODE-GRID
C2 2 1 {CGP} ; GRID=PLATE
C3 1 3 {CCP} ; CATHODE-PLATE
D3 5 3 DX ; POSITIVE GRID CURRENT
R1 2 5 {RGI} ; POSITIVE GRID CURRENT
.MODEL DX D(IS=1N RS=1 CJO=10PF TT=1N)
.ENDS
About the convergence problem, last week, I design a screen grid drive direct coupled amplifier schematic for simulation using a 6DQ5 and a 5965 medium-mu triode as direct coupled CF driver. It work very well until the voltage peak on the 6DQ5 screen grid get negative, then strange things happen ...PSpice uses LIMIT, 3f4 and the like, use URAMP. Some simulators use/understand their own combinations i.e. Micro-Cap. Curve Captor outputs limit or uramp depending on what flavor of spice you've chosen.
The main intent (especially with the Koren 8 models) of the added diode is an aid in convergence. For example, as when simulating a CF that's directly coupled to the proceeding stage. It can be commented out if you so desire.
No
I finally find out putting a diode between the anode and cathode of the driver tube did correct the problem. But applying a negative voltage on the 6DG5 model screen grid I have make no changes because the URAMP function keep it at 0V. But to tell you the truth, I don't know what happen when applying a negative voltage to a real pentode screen grid ? Nothing ? Tube explode ? I will have to try that one of these days ...
The 6DQ5 cost only 10$ at ESRC and the 5965 3$, much cheaper than a 12AU7 ... This twin triode was used in the first computers circuits and there is a lot of them for sale on the market.
By the way, thank you very much for your 6J6 models. Wayne Clay is a perfectionist and usually make very accurate models.
Thanks, this 12AX7 grid current model is very interesting.Thanks for the PSpice 12AX7 model Wayne. I see Curvecaptor uses the LIMIT function in lieu of URAMP.
I also see that Curvecaptor does not model grid current
Fortunately, in the EXCEM models, grid current is included. Perhaps they can be combined.
Attached is an excerpt of the EXCEM/Rydel library converted to PSpice syntax and with Ip correction (rebounding Ip).
The space in "u ramp" is a diyAudio editor problem ...
Thanks for letting me know. That's something I'll be on the lookout for from now on.
Thank you for this 6J6 model. After trying it and also the cogsncogs model, I will modify the best one to use with a new 6J6 symbol I just draw specially for this tube, with two grids, two anodes but just one cathode. I also draw a symbol for the 6HS8.6HS8 is indeed a bizarre tube, good luck finding or making a model for it...
Here is a model for the 6J6, I think made by Stephie Bench:
Code:.subckt 6J6 1 2 3 + params: mu=38.9 ex=1.484 kg1=780 kp=162 kvb=1176 rgi=2000 vct=.384 + ccg=2.6p cgp=1.5p ccp=1.6p e1 7 0 value= +{v(1,3)/kp*log(1+exp(kp*(1/mu+v(5,3)/sqrt(kvb+v(1,3)*v(1,3)))))} re1 7 0 1g g1 1 3 value= {(pwr(v(7),ex)+pwrs(v(7),ex))/kg1} rcp 1 3 1g c1 2 3 {ccg} c2 1 2 {cgp} c3 1 3 {ccp} r1 5 6 {rgi} v1 5 2 {vct} d3 6 3 dx .model dx d(is=1n rs=1 cjo=1pf tt=1n) .ends
Like you said, the 6HS8 is a very bizarre tube and I was sure nobody ever made a model for it ... But I will try to make one just for using in the particular conditions of this RCA graph :
An externally hosted image should be here but it was not working when we last tested it.
As you can see, it is possible to obtain a very high output voltage swing with this tube at low amplitude distortion. The load line I draw for a 300K dc load show a maximum voltage swing of 347V peak to peak for a 0 to -7V signal on grid #3, the so called "suppressor grid" usually tie to the cathode in most circuits (and tubes) ... This very linear 122.68V RMS can be useful in some circuits like a cathode follower output stage or a Futterman OTL amplifier. It can also be used at a very low supply voltage, as low as a 25V and 0.175mA operating point, for a really not dangerous circuit ...
The fact is the 6HS8 cost almost nothing, I buy 5 on eBay yesterday for 17$CAN shipping included and ESRC sale them for 1$ until their sale end this summer. Like it is written on the graph, there must be 0.1mA on grid #1, 67.5V on grid #2, the plate and grid #3 of the unused section of the tube must be tie to ground. Also, this tube can be used for a remote two channels volume control ...
As you can see, it is possible to obtain a very high output voltage swing with this tube at low amplitude distortion. The load line I draw for a 300K dc load show a maximum voltage swing of 347V peak to peak for a 0 to -7V signal on grid #3, the so called "suppressor grid" usually tie to the cathode in most circuits (and tubes) ... This very linear 122.68V RMS can be useful in some circuits like a cathode follower output stage or a Futterman OTL amplifier. It can also be used at a very low supply voltage, as low as a 25V and 0.175mA operating point, for a really not dangerous circuit ...
Very interesting tube, if you can trace the triode curves on the tubes you got, perhaps we can take crack at making a model for it using the Ayumi method. Then we still need to do a bit more work to add the supressor grid model to make it complete.
How to superpose two curves graphs
I notice Jazbo8 also do that, this is a complicated procedure and I will explain it once for all, this will be also useful for many other members of this forum.
To do this, you need four software, a simulator like SiMetrix with which you can adjust the graph size very accurately by changing the size of the window.
You also need a software to measure things you see on screen in pixels. I use the freeware "SMeasure.exe" but I don't know if it is still available. But there are many other similar online for free : Screen Measure Software Informer
The third software, everybody who are using Window have it somewhere in is computer, it is MS Paint ... You can also use other image editors.
The last one is a PDF reader, personally, I use Foxit reader : PDF Reader | PDF Viewer | Foxit Reader
To explain the procedure step by step, I will put the 6J6 model I get from cogsncogs on top of the 6J6A GE graph I get there : Ðàäèîëàìïà 6J6
It is different than the RCA one but in the 3F4 model cogsncogs give me, it is written the curves come from a GE datasheet ...
Here is the model :
I rename the tube 6J6_WC so I know who did it ... Wayne Clay ...
Step #1
The first thing to do is opening the PDF file and get the needed graph with a nice big zoom :
Then you copy the PDF viewer window image using the [ALT + Print Screen] shortkey, open MS Paint and paste the image on it [Ctrl + V] :
Using this image editor function, remove the unneeded part of the image, just keep the graph and write the name of the tube on it, or anything else useful for you :
The default bitmap is 24 bits and this make a huge file (1.26 Mo), save it first in a 16 color BMP format (211 Ko). Then save it again in a PNG format on the final name "6J6 GE.PNG", because it make a really smaller file than the BMP format (only 36 Ko), here's the result :
Step #2
After the tube model is in your model file and is associated to the correct symbol, you open the simulator and draw a little test circuit for the tube :
I name it "6J6_WC curves" for convenience. Looking at the 6J6 GE graph, you will notice there is 10 curves from 0 to -18V and the coordinate are 14mA by 500V. You must also measure the number of X,Y pixels of the GE graph using a measuring on screen tool or by looking at the coordinate in MS paint. I found the width of the graph is about 723 pixels and the height is about 508 pixels. I write about because the graph scan is not very rectangular ...
Now you have to choose the simulator analysis a DC sweep on device V2 (voltage source) from 0 to 500V with 1000 points for greater accuracy, this is for the anode voltage :
You can make the curves one by one by changing the voltage of the V3 device for the grid and run the simulator 10 times but it is much better to do it once using the "multi-step" analysis of SiMetrix :
After you define the step analysis like on this image, you can run the simulator and generate all the curves in on shot and you know then if the model is working ... You will first get a graph with a very high current but the correct anode voltage, because this is a triode :
Step #3
Using the cursor zoom area function, you can scale the graph exactly to 14mA by 500V :
Here is the hardest part now, you have to scale the SiMetrix graph to the same size as the GE graph using the on screen measuring tool, to do that, you have to change the size of the window using the mouse :
After I have adjust the size, I get exactly a 723 X 508 pixels SiMetrix graph, the tool provide a cursor magnifier, this help to see if the rule is properly align with the graph border, always use the inner positions for thick border ... Sometime, when the company graph is too high, you cannot adjust the SiMetrix graph the same size, then you can make the company graph smaller or cut the SiMetrix graph to a lower current, I do that very often.
After this is done, you copy the graph window the same way you do for the graph of the PDF file, using [ALT + Print Screen] and open a new image file to paste it on and do all the necessary clean-up :
I save this file under the name of "6J6_WC SIM.PNG" for future use like load line drawing, because this is the real curves of your spice model. But to put this graph on top of the other one, it is better to remove all what you don't need to see, like the voltages and current, also the thick left border :
This what will go on top ... In MS Paint, select all the image and copy it. Then open the GE graph image you first do. to be able to paste the SiMetrix graph on top, you must make the background of the GE graph transparent using the following button on the left panel :
Now you are ready to paste the graph on top of the other and save it under the name you prefer, personally, I name the file "6J6_WC superposition GE.PNG" :
Not bad ... The model fit pretty well the GE datasheet graph in the most useful area, the left bottom corner ...
This was a small question but a quite complicated answer ...
Doing all that can look as a big job but I think it worth it, after you make a dozen of superpositions like this, you will get use to and all will be done faster and easier. I hope many members of this forum will use my method, I don't think there is more convenient way do do that kind of manipulation.
Alain
This the question I was waiting for ...Thanks Alain.
That Rydel model looks pretty good
I have high hopes for the Rydel model in Curvecaptor.
How do you plot the curves over top the GE datasheet?
I notice Jazbo8 also do that, this is a complicated procedure and I will explain it once for all, this will be also useful for many other members of this forum.
To do this, you need four software, a simulator like SiMetrix with which you can adjust the graph size very accurately by changing the size of the window.
You also need a software to measure things you see on screen in pixels. I use the freeware "SMeasure.exe" but I don't know if it is still available. But there are many other similar online for free : Screen Measure Software Informer
The third software, everybody who are using Window have it somewhere in is computer, it is MS Paint ... You can also use other image editors.
The last one is a PDF reader, personally, I use Foxit reader : PDF Reader | PDF Viewer | Foxit Reader
To explain the procedure step by step, I will put the 6J6 model I get from cogsncogs on top of the 6J6A GE graph I get there : Ðàäèîëàìïà 6J6
It is different than the RCA one but in the 3F4 model cogsncogs give me, it is written the curves come from a GE datasheet ...
Here is the model :
I rename the tube 6J6_WC so I know who did it ... Wayne Clay ...
Step #1
The first thing to do is opening the PDF file and get the needed graph with a nice big zoom :
An externally hosted image should be here but it was not working when we last tested it.
Then you copy the PDF viewer window image using the [ALT + Print Screen] shortkey, open MS Paint and paste the image on it [Ctrl + V] :
An externally hosted image should be here but it was not working when we last tested it.
Using this image editor function, remove the unneeded part of the image, just keep the graph and write the name of the tube on it, or anything else useful for you :
An externally hosted image should be here but it was not working when we last tested it.
The default bitmap is 24 bits and this make a huge file (1.26 Mo), save it first in a 16 color BMP format (211 Ko). Then save it again in a PNG format on the final name "6J6 GE.PNG", because it make a really smaller file than the BMP format (only 36 Ko), here's the result :
An externally hosted image should be here but it was not working when we last tested it.
Step #2
After the tube model is in your model file and is associated to the correct symbol, you open the simulator and draw a little test circuit for the tube :
An externally hosted image should be here but it was not working when we last tested it.
I name it "6J6_WC curves" for convenience. Looking at the 6J6 GE graph, you will notice there is 10 curves from 0 to -18V and the coordinate are 14mA by 500V. You must also measure the number of X,Y pixels of the GE graph using a measuring on screen tool or by looking at the coordinate in MS paint. I found the width of the graph is about 723 pixels and the height is about 508 pixels. I write about because the graph scan is not very rectangular ...
Now you have to choose the simulator analysis a DC sweep on device V2 (voltage source) from 0 to 500V with 1000 points for greater accuracy, this is for the anode voltage :
An externally hosted image should be here but it was not working when we last tested it.
You can make the curves one by one by changing the voltage of the V3 device for the grid and run the simulator 10 times but it is much better to do it once using the "multi-step" analysis of SiMetrix :
An externally hosted image should be here but it was not working when we last tested it.
After you define the step analysis like on this image, you can run the simulator and generate all the curves in on shot and you know then if the model is working ... You will first get a graph with a very high current but the correct anode voltage, because this is a triode :
An externally hosted image should be here but it was not working when we last tested it.
Step #3
Using the cursor zoom area function, you can scale the graph exactly to 14mA by 500V :
An externally hosted image should be here but it was not working when we last tested it.
Here is the hardest part now, you have to scale the SiMetrix graph to the same size as the GE graph using the on screen measuring tool, to do that, you have to change the size of the window using the mouse :
An externally hosted image should be here but it was not working when we last tested it.
After I have adjust the size, I get exactly a 723 X 508 pixels SiMetrix graph, the tool provide a cursor magnifier, this help to see if the rule is properly align with the graph border, always use the inner positions for thick border ... Sometime, when the company graph is too high, you cannot adjust the SiMetrix graph the same size, then you can make the company graph smaller or cut the SiMetrix graph to a lower current, I do that very often.
After this is done, you copy the graph window the same way you do for the graph of the PDF file, using [ALT + Print Screen] and open a new image file to paste it on and do all the necessary clean-up :
An externally hosted image should be here but it was not working when we last tested it.
I save this file under the name of "6J6_WC SIM.PNG" for future use like load line drawing, because this is the real curves of your spice model. But to put this graph on top of the other one, it is better to remove all what you don't need to see, like the voltages and current, also the thick left border :
An externally hosted image should be here but it was not working when we last tested it.
This what will go on top ... In MS Paint, select all the image and copy it. Then open the GE graph image you first do. to be able to paste the SiMetrix graph on top, you must make the background of the GE graph transparent using the following button on the left panel :
An externally hosted image should be here but it was not working when we last tested it.
Now you are ready to paste the graph on top of the other and save it under the name you prefer, personally, I name the file "6J6_WC superposition GE.PNG" :
An externally hosted image should be here but it was not working when we last tested it.
Not bad ... The model fit pretty well the GE datasheet graph in the most useful area, the left bottom corner ...
This was a small question but a quite complicated answer ...
Doing all that can look as a big job but I think it worth it, after you make a dozen of superpositions like this, you will get use to and all will be done faster and easier. I hope many members of this forum will use my method, I don't think there is more convenient way do do that kind of manipulation.
Alain
This was a small question but a quite complicated answer ...
Doing all that can look as a big job but I think it worth it, after you make a dozen of superpositions like this, you will get use to and all will be done faster and easier. I hope many members of this forum will use my method, I don't think there is more convenient way do do that kind of manipulation.
Alain
As usual, a great tutorial. This is a great way to verify that the model matches up with the datasheet it was built from. I think for the triodes, it is hard to beat CurveCaptor, it is more accurate than Ayumi's model. If you do not care about the control grid (say for hi-fi applications), then that's the best one to use. But for guitar & PA amps, where the grid is often over-driven, then either use the Ayumi or the Excem models, where the positive grid currents are included, i.e., instead of relying on the silicon diode model.
6HS8 model
I look at the PDF datasheet and the curves are very complicated, I am afraid it will be almost impossible to make a complete model for. I look at many Ayumi models, some are very good but some are less good, very often is grid current models are wrong and I notice is pentode often have "square knees" and his triodes models are not fantastic, look at those four examples :
The curves at low screen grid voltage of is 6DQ5 don't fit, the current is much lower. But it is the only one I have ...
The 0V curve of the Ayumi 5965 model is horrible but the other curves are perfect. But this is unusable for my simulation, I use the Wayne Clay model instead.
His 12AY7 model, the 0V curve is not perfect and the curve under -3 volts, no comments ... But it can be use from -1 to -3 grid voltage.
This Ayumi 6SQ7 model is usable in most area, but he seem to have problems with the 0V curve ...
Accurate pentode model are very hard to do, I never see any with a suppressor grid pin, they are always tie to the cathode and not many datasheet said something about this grid voltage contribution. I believe every pentode with a separate suppressor grid from cathode can be drive using that grid but I never see a amplifier schematic working this way except the two I found using the 6HS8 for remote volume control purpose.
I like to walk off the beaten track in life, this is where the most interesting discovery can be done.
Yes, it is a very interesting tube, it is in the old RCA databook I have since over 40 years but I never paid attention to it until now ... I look at the 6HS8 page hundred of times but the only thing I notice about it is the basing diagram is upside down, this is the only tube in this book with a such printing error ...
I look at the PDF datasheet and the curves are very complicated, I am afraid it will be almost impossible to make a complete model for. I look at many Ayumi models, some are very good but some are less good, very often is grid current models are wrong and I notice is pentode often have "square knees" and his triodes models are not fantastic, look at those four examples :
An externally hosted image should be here but it was not working when we last tested it.
The curves at low screen grid voltage of is 6DQ5 don't fit, the current is much lower. But it is the only one I have ...
An externally hosted image should be here but it was not working when we last tested it.
The 0V curve of the Ayumi 5965 model is horrible but the other curves are perfect. But this is unusable for my simulation, I use the Wayne Clay model instead.
An externally hosted image should be here but it was not working when we last tested it.
His 12AY7 model, the 0V curve is not perfect and the curve under -3 volts, no comments ... But it can be use from -1 to -3 grid voltage.
An externally hosted image should be here but it was not working when we last tested it.
This Ayumi 6SQ7 model is usable in most area, but he seem to have problems with the 0V curve ...
Accurate pentode model are very hard to do, I never see any with a suppressor grid pin, they are always tie to the cathode and not many datasheet said something about this grid voltage contribution. I believe every pentode with a separate suppressor grid from cathode can be drive using that grid but I never see a amplifier schematic working this way except the two I found using the 6HS8 for remote volume control purpose.
I like to walk off the beaten track in life, this is where the most interesting discovery can be done.
The curve that you used was the Ep vs Eg2 for Eg=0V. The correct one is:
An externally hosted image should be here but it was not working when we last tested it.
Ayumi's 5965 model is based on RCA not Sylvania here is the correct comparison:
An externally hosted image should be here but it was not working when we last tested it.
Again, Ayumi's 12AY7 model is based on RCA not GE, here is the correct comparison:
An externally hosted image should be here but it was not working when we last tested it.
Too bad Ayumi's SPICE models do not indicate which datasheet they are based on, to see that you need to use R or open the pctube.r as a text file, where he does provide references to each of the models.
I also have a question on the lower left corner accuracy for the Eg=0V curve, since you were not the only one that complained about it. Of course, it is important but certainly not as important as the typically used linear region, no? For the best linearity and to avoid grid current, you would purposely avoid getting anywhere close to Eg=0V, in fact, Eg<-0.6V is often used as a "cut-off" point (RDH4). Just curious, that's all...
I can't see your images actually, maybe they will appear later ?The curve that you used was the Ep vs Eg2 for Eg=0V. The correct one is:
An externally hosted image should be here but it was not working when we last tested it.
Ayumi's 5965 model is based on RCA not Sylvania here is the correct comparison:
An externally hosted image should be here but it was not working when we last tested it.
Again, Ayumi's 12AY7 model is based on RCA not GE, here is the correct comparison:
An externally hosted image should be here but it was not working when we last tested it.
Too bad Ayumi's SPICE models do not indicate which datasheet they are based on, to see that you need to use R or open the pctube.r as a text file, where he does provide references to each of the models.
I also have a question on the lower left corner accuracy for the Eg=0V curve, since you were not the only one that complained about it. Of course, it is important but certainly not as important as the typically used linear region, no? For the best linearity and to avoid grid current, you would purposely avoid getting anywhere close to Eg=0V, in fact, Eg<-0.6V is often used as a "cut-off" point (RDH4). Just curious, that's all...
Of course, there are differences between companies graphs, also between their tubes, also between two tubes of the same brand and the same batch, also between different life stage, but transistors are worst. Spice models are really approximate but like you said it help when the source datasheet is indicated ... This is the only way to verify the mathematical accuracy of a model.
About using the bottom left area, it depend on the needs of the circuit, the 12AX7 is made to operate in that area, for small signal or operation at very low supply voltage. This circuit I design last week for example :
An externally hosted image should be here but it was not working when we last tested it.
It is a screen grid drive direct coupling amplifier. To do it this way, I have to raise the 6DQ5 cathode enough to give sufficient room for the 12AX7 output swing, and the voltage at the plate is only 50.63V, this is very low. If I use a higher operating point, the supply voltage already high will have to be higher ...
Also, the operating point of the 12AX7 produce a very low distortion. The signal needed to produce the 7.615W maximum power is only 212mV RMS, this is very low. The THD distortion is really low for a SE pentode with no feedback at all, but I simulate the circuit only 100mS to get those results, usually, I simulate the circuit 10 seconds and the distortion is always lower ... But the Ayumi 6DQ5 model I use is not very accurate at low screen grid voltage, perhaps, a real 6DQ5 will give worst results, maybe better results, who knows ?
I still have a lot of work to do on this schematic, I like to try the 6HS8 to replace the 12AX7 because it cost about nothing and can work at a very low operating point. I will probably build a prototype for this amplifier, when I got enough time, I got many small OPT sleeping in my junkbox I can use for that.
You can't see the images? That's weird, may be your ISP/PC is blocking Dropbox? Anyway, I only posted those comparisons to point out that Ayumi models were not so bad when the correct datasheets were used. In any case, I think most of these models should give you good enough indication to see if the circuit will work or not, perhaps just not in the extreme low V/I area that your particular design requires. Direct-coupled screen drive power amp, that you don't see everyday...