Last update broke things. about circuitjs1 HOT 8 CLOSED

Hi, could you be more specific? Could you send us the circuit, using File->Export as Link or File->Export as Text, and explain what's wrong with it? Are you talking about the version at falstad.com or lushprojects.com ?

from circuitjs1.

I didn't know there are versions other than at falstad.com.
Well, I think the main problem is convergence.
I had it running with "convergence failed" message on screen at one point — can't recreate that now.
Let me try mashing something up and I'll send you the link if I can before going out(6:30AM here).

from circuitjs1.

I don't remember GitHub's Markdown implementation parsing EOLs like that...

Here's a stupid little circuit. Take note that the simulation speed is one step up from 0.
Look at the ridiculous numbers.
Now if you put simulation speed to 0 and try disconnecting the circuit and resetting the simulation — conductors remain powered.

As an unrelated note: can you please make European resistor notation setting saved to browser cookies?

from circuitjs1.

Hi, I tried it on the original java version and got a similar result. There were ridiculous numbers, which seems right. So I moved the transistor out of the circuit and it still had huge current numbers. Definitely a bug, but the java simulator had the same problem. I have not tried it on an old version of the javascript version.

I tried resetting after disconnecting the transistor and everything seemed fine.

Yes, I plan to make resistor notation saved by cookies. Right now it looks at your browser language setting.

from circuitjs1.

Caught it. This little modification made the circuit run with "convergence failed" message on screen if you press "Reset" and then "Run/Stop" button and the current flow is pretty wanky.
Here's a demonstration.

Thank you for planned cookies.

from circuitjs1.

The unresettable bug only occurs with simulation speed set to 0.
(The simulation lagged a bit when I was removing the ground connection.)

from circuitjs1.

Ok I tried the "caught it" circuit in the java version and also got a "convergence failed" there.

If the simulation speed is set to 0, the simulation stops. Lots of weird things happen if you move the circuit elements around. You can even try this with the LRC circuit. The simulation never gets run to update currents. If you press reset, it resets the voltages but not currents.

Anyway, I definitely see some issues that need fixing here. But I don't think any of them are new problems.

If you try to put a voltage source across a bare transistor like that, it should probably just give an error rather than report an unrealistically large current.

from circuitjs1.

I believe Paul's revisions have fixed this in the JS version now, so closing this issue. thanks for reporting.

from circuitjs1.