The Pentode model is already implemented in the simulator. The problem is that I have no data to generate characteristiscs.
For tiodes I could create models based on manufactures characteristics. Such, unfortunately, is not for pentodes. To calculate the parameters of the pentode model I must have at least three/four families of anode characteristics for different screen voltages. I have the uTracer tester but at present I have no time to perform time consuming measurements of mentioned tubes. If someone has such data I willingly unlock simulator to support pentodes. I will try to make such measurements as soon as possible.
For tiodes I could create models based on manufactures characteristics. Such, unfortunately, is not for pentodes. To calculate the parameters of the pentode model I must have at least three/four families of anode characteristics for different screen voltages. I have the uTracer tester but at present I have no time to perform time consuming measurements of mentioned tubes. If someone has such data I willingly unlock simulator to support pentodes. I will try to make such measurements as soon as possible.
Last edited:
Greetings from Warsaw.
I think it is worth to measure the following tubes: EF86, EL84, EL34, KT88, 6L6. During the measurements, it is worth checking whether the second grid voltages of measured characteristics of a tube are similar (or rather the same) to shown by the manufacturer of the tube.
I have not heard of measurement data repository. There is however a lot of models developed on the basis of measurements using the uTracer tester.
Unfortunately, according to me, these tube models frequently do not work in real conditions. This is due to the fact that the measurements are often performed for the voltage ranges not used in practice. For example, I downloaded a EF86 model parameters and done simulations using the simulator and found that the characteristics in range shown on the uTracer website overlap (this means that the pentodes model in the simulator is implemented correctly), but for the voltages and currents in a more useful range, characteristics do not coincide with the factory data. I think the Spice model parameters were determined for improperly selected set of data. In my opinion, set of measurements should include the characteristics for grid voltages given in the factory datasheets and additionally for the voltages between factory values.
Hi,
Your developments work wonderfully and are the perfect marriage for the uTracer. I think the tube community owes you a great deal.
Find below my collection of setup files for the most popular tubes:
https://www.dropbox.com/sh/u6wrw1l9kak5aq3/AAAYStf99Un9BlO-vlKY-iPWa?dl=0
Just installed the version of the GUI (3.11.6) featuring also reference values for mu, gm and ri, so these have not yet been introduced. It will be progressively done over the following weeks.
Also, for some testing purposes of the Optimizer, you can use this data of a D3a wired as triode.
https://www.dropbox.com/s/n5ceuay7r4n6fka/D3A.uts?dl=0
I have lots of the tubes mentioned; will start doing some dataset dumps ASAP, although day work will keep me quite busy this and the following weeks.
Use everything at you own risk!
Your developments work wonderfully and are the perfect marriage for the uTracer. I think the tube community owes you a great deal.
Find below my collection of setup files for the most popular tubes:
https://www.dropbox.com/sh/u6wrw1l9kak5aq3/AAAYStf99Un9BlO-vlKY-iPWa?dl=0
Just installed the version of the GUI (3.11.6) featuring also reference values for mu, gm and ri, so these have not yet been introduced. It will be progressively done over the following weeks.
Also, for some testing purposes of the Optimizer, you can use this data of a D3a wired as triode.
https://www.dropbox.com/s/n5ceuay7r4n6fka/D3A.uts?dl=0
I have lots of the tubes mentioned; will start doing some dataset dumps ASAP, although day work will keep me quite busy this and the following weeks.
Use everything at you own risk!
Hi again,
I tested with my uTracer a good example of a NOS EF86 in perfect shape, at different Vs voltages (60, 100, 140, 180, 220, 260), plus the quick test report.
Find them here:
https://www.dropbox.com/sh/gq5bgbdi3i4t5k7/AACXxlunqDxdW5FFlmfYgwVJa?dl=0
Hope this can help you in fine tuning the pentode model. Let me know what tube types you would like me to test next.
Cheers
I tested with my uTracer a good example of a NOS EF86 in perfect shape, at different Vs voltages (60, 100, 140, 180, 220, 260), plus the quick test report.
Find them here:
https://www.dropbox.com/sh/gq5bgbdi3i4t5k7/AACXxlunqDxdW5FFlmfYgwVJa?dl=0
Hope this can help you in fine tuning the pentode model. Let me know what tube types you would like me to test next.
Cheers
Ale & I have discussed this particular issue recently, perhaps Dr. Reefman will update the Extract Model to allow for generating more useful pentode models. Or you can incorporate the scaling/fudge factor in your app, either way, it has to be done, otherwise, as you said, many pentode models generated with the uTracer/Extract Model would not work under real conditions.
Last edited:
I did notice that the calculator doesn't work on my tablet. I don't know that you even care. You may not. It's not a huge deal to me, but I sometimes do this stuff when I'm away from home and only have my tablet, so it'd be convenient for me, but I imagine you have bigger fish to fry at the moment.
Nevertheless, if you feel like tackling it at some point, I don't think the HTML5 drag and drop stuff is compatible. You might check out this question on StackOverflow. I suspect the EaselJS solution is the one you'd want as, at least at first glance, it didn't seem to rely on jQuery and it looks like you've managed to avoid jQuery so far.
Nevertheless, if you feel like tackling it at some point, I don't think the HTML5 drag and drop stuff is compatible. You might check out this question on StackOverflow. I suspect the EaselJS solution is the one you'd want as, at least at first glance, it didn't seem to rely on jQuery and it looks like you've managed to avoid jQuery so far.
izozaya -> thank you for your time, to develop data and making them available. I already downloaded them and I will analyze data on Monday.
jazbo8 -> That would be really good news, if models were "adjusted". Next week I'll try to show the problem with Simulator using specific example of EF86.
pdavis68-> thank you for the tips. I'll check the contents of a thread on StackOverflow.
jazbo8 -> That would be really good news, if models were "adjusted". Next week I'll try to show the problem with Simulator using specific example of EF86.
pdavis68-> thank you for the tips. I'll check the contents of a thread on StackOverflow.
Hi Fantastic work well done we have been needing something like this for years
In the spice modelling could I request the modelling of positive load lines and grid current?
As per Dmitry paint curves for us A2 Nuts
Model Paint Tools: Trace Tube Parameters over Plate Curves, Interactively
Could I also Request models for PX4, PX25, 211
In the spice modelling could I request the modelling of positive load lines and grid current?
As per Dmitry paint curves for us A2 Nuts
Model Paint Tools: Trace Tube Parameters over Plate Curves, Interactively
Could I also Request models for PX4, PX25, 211
Not the uts file, but the utd file, i.e., the file containing the numerical data. In any event, here is the scaling issue gsmok was referring to (I think...) wrt to the uTracer generated models. Below is a comparison of the EF86 SPICE model from Dr. Reefman's library and the Philips datasheet, where it is quite clear that the two are no where close to each other.
An externally hosted image should be here but it was not working when we last tested it.
I'll check this tomorrow
. Today I have no access to my personal computer (I use public one).And it is precisely the problem with pentodes I mentioned before (that's why pentode module of the Tube Simulator is still locked). Probably parameters of available pentode model have been calculated on the basis of measurements for second grid voltages ABOVE 200V (on the uTracer Site there is the description of measurements of PF86 for screen voltages 200V/250V and 300V
). These values do not match the real tube circuits and the characteristics delivered by tube manufacturers like Philips (typical screen voltages 100V and 140V). The ExtractModel had simply no complete data set to determine the correct parameter values for lower screen voltages.
Last edited:
Those are for the pentode, could you please run one in triode mode, the result should look something like this (this one from the PF86 example given by Dr. Reefman):
Code:
Point Curve Ia (mA) Is (mA) Vg (V) Va (V) Vs (V) Vf (V)
1 1 0 0 -5.5 48.02 48.7 4.5
2 1 0 0 -5.5 56.2 56.88 4.5
3 1 0 0 -5.5 64.39 65.07 4.5
4 1 0 0 -5.5 72.57 73.59 4.5
5 1 0 0 -5.5 81.1 82.12 4.5
6 1 0 0 -5.5 89.96 90.3 4.5
7 1 0 0 -5.5 97.81 98.49 4.5
8 1 0 0 -5.5 106.33 107.01 4.5
9 1 0 0 -5.5 114.51 115.2 4.5
10 1 0 0 -5.5 122.7 123.38 4.5
etc.
...I appended these tubes to the list. As for the voltage ranges I'm planning that in the new version of the simulator, these values can be defined by the user.
Could you add the contents of a file you are using for data processing. That would help me in finding the cause of the error.
Last edited:
Thank you for adding the 845.
However, how can I change the Y-axis? I wish to create a much lower load resistance slope than normal, and need to increase the height to which the Y-axis can rise to.
Very Best, Synchro27
PS: sorry to be a pest, but I just remembered another feature that would be wonderful to add - the ability to double, triple or even x 4 the same tube. As in, in this case, parallel dual 845's.
However, how can I change the Y-axis? I wish to create a much lower load resistance slope than normal, and need to increase the height to which the Y-axis can rise to.
Very Best, Synchro27
PS: sorry to be a pest, but I just remembered another feature that would be wonderful to add - the ability to double, triple or even x 4 the same tube. As in, in this case, parallel dual 845's.