Hello, totally new here. I have some experience with LTspice, but not with these tube models (2 of the "Ayumi_LTspice", contained in the zip).
Could someone with experience tell me why this simulates so poorly (takes ages), and the crazy behavior when I actually attach the "HV" output label to the left end of R15 (it's disconnected in the attached schematic) - i.e. instead of the +366V I'm seeing a large negative voltage, as if the tube circuits do some black magick there.
Notes apart from that: the outupt transformer model may be crap, it's after measurements with an ebay LCR meter on a transformer I have laying around, but the L2 winding seems awfully low in series R and inductance - then again I have no other output transformer to compare to. Also, the power source is not as in the orig schematic, it's the supposed output of some cheap power transformer from a tube radio I just "shot" on ebay, arrival pending.
Could someone with experience tell me why this simulates so poorly (takes ages), and the crazy behavior when I actually attach the "HV" output label to the left end of R15 (it's disconnected in the attached schematic) - i.e. instead of the +366V I'm seeing a large negative voltage, as if the tube circuits do some black magick there.
Notes apart from that: the outupt transformer model may be crap, it's after measurements with an ebay LCR meter on a transformer I have laying around, but the L2 winding seems awfully low in series R and inductance - then again I have no other output transformer to compare to. Also, the power source is not as in the orig schematic, it's the supposed output of some cheap power transformer from a tube radio I just "shot" on ebay, arrival pending.
I was told elsewhere that the tube models have "many voltage-mode B sources", posing problems for stability of the simulation.
But they all tend to look like that, if I understood right what this refers to...
But on this site, with a 180+ pages long thread, people seem to be happily(?) using such tube models... Then again, the warning sounds pretty much like what I'm experiencing?
But they all tend to look like that, if I understood right what this refers to...
But on this site, with a 180+ pages long thread, people seem to be happily(?) using such tube models... Then again, the warning sounds pretty much like what I'm experiencing?
I don't know, but I see several errors in the schematic of your second post:
1. The power supply isn't connected to node HV, like you already mentioned.
2. The feedback appears to be positive rather than negative.
3. The label SPKRb that's directly connected to ground could cause a node naming conflict.
4. The load resistor isn't connected to SPKRa and ground, but to two other node names.
5. I don't know if it is correct or not, but the coupling factor of the output transformer seems rather low to me.
By the way, sometimes choosing the alternate solver can speed up LTSpice considerably.
1. The power supply isn't connected to node HV, like you already mentioned.
2. The feedback appears to be positive rather than negative.
3. The label SPKRb that's directly connected to ground could cause a node naming conflict.
4. The load resistor isn't connected to SPKRa and ground, but to two other node names.
5. I don't know if it is correct or not, but the coupling factor of the output transformer seems rather low to me.
By the way, sometimes choosing the alternate solver can speed up LTSpice considerably.
Thanks for replying!
1. as soon as that label is connected, R15-left will measure ~ -64V, if it's not connected, it will have the extected 366V. That's the black magic I'm speaking of.
2. is odd as the text to the Mullard schematic says one should "try out" the correct way of connecting the feedback there, lol... no start/end of coil mentioned. I replicated the schematic there. But I tested without feedback, which does not make things better.
3. interesting... conflict, not thought of it, I'll try if this does anything
==> Nope, not changing anything, if I remove the GND-connected speaker label and just GND the lower load resistor's end, too.
4. yeah that's only in the screencap there, not in the zip. Some playing around with disconnecting things by renaming.
5. playing around with things... I have not measured my transformer that thoroughly. I still need to read the text on how to measure the coupling factor. Changing that to 0.99 or so doesn't help, reversing the feedback doesn't do anything either.
1. as soon as that label is connected, R15-left will measure ~ -64V, if it's not connected, it will have the extected 366V. That's the black magic I'm speaking of.
2. is odd as the text to the Mullard schematic says one should "try out" the correct way of connecting the feedback there, lol... no start/end of coil mentioned. I replicated the schematic there. But I tested without feedback, which does not make things better.
3. interesting... conflict, not thought of it, I'll try if this does anything
==> Nope, not changing anything, if I remove the GND-connected speaker label and just GND the lower load resistor's end, too.
4. yeah that's only in the screencap there, not in the zip. Some playing around with disconnecting things by renaming.
5. playing around with things... I have not measured my transformer that thoroughly. I still need to read the text on how to measure the coupling factor. Changing that to 0.99 or so doesn't help, reversing the feedback doesn't do anything either.
Is R9's right side supposed to be connected the way it is connected?
That -64 V is clearly unphysical. Simulators usually produce unphysical results when they don't converge, but you should also get an error message then, or when models are used far outside the range for which they were intended, or when the models are all wrong.
That -64 V is clearly unphysical. Simulators usually produce unphysical results when they don't converge, but you should also get an error message then, or when models are used far outside the range for which they were intended, or when the models are all wrong.
If my optics aren't twisted, it looks like it ;D But I'm not deeply enougn into what the circuit does how exactly, to understand the purpose.
Here is the original article - and I kept the part designators the same in my spice file, to not confuse myself unnecessarily:
http://www.r-type.org/articles/art-003h.htm
(the power supply is spec'd differently than my lower peak voltage - as I was trying to simulate what impact it would have to use a transformer I have available)
Ok, the 6V6 model is faulty. It uses ^instead of ** syntax... argh.
Changing that fixes it.The 6V6 is not equivalent to the EL84 anyway, but I have a supposedly close-ish one 6P1P, so I wanted to see what happens 😉 (its specs are lower)
I meanwhile learned that 6BQ5 is close to EL84, which is a model in the Ayumi pack, so if nothing else is wrong, I can make 2 circuits, a changed one for the 6V6, & compare.
Changing that fixes it.The 6V6 is not equivalent to the EL84 anyway, but I have a supposedly close-ish one 6P1P, so I wanted to see what happens 😉 (its specs are lower)
I meanwhile learned that 6BQ5 is close to EL84, which is a model in the Ayumi pack, so if nothing else is wrong, I can make 2 circuits, a changed one for the 6V6, & compare.
