<div dir="ltr">Jon, <div><br></div><div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:small">>I maybe should have mentioned that my runs have been in 2D (cylindrical).  So have you done such runs?  </span><br></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline">No, I have done runs just in cartesian coordinates. </div></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline"><br></div></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline">>Also for the conduction, I'm using unsplit diffusion with electron thermal conduction and power law conductivity.  </div></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline">Exactly. </div></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline">You might </div><span style="font-family:arial,helvetica,sans-serif;font-size:small">also </span><span style="font-family:arial,helvetica,sans-serif;font-size:small">be interested in </span><span style="font-size:12.8px">saturated conduction, for this FLASH offered </span><span style="font-family:arial,helvetica,sans-serif;font-size:small">the runtime parameter </span><span style="font-size:12.8px">diff_eleFlMode. </span></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline"><br></div></div><div style="font-size:12.8px"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline">>You mention being careful with split vs. unsplit diffusion, but you don't say which one works - or >what that means to be careful.​</div> <div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small;display:inline">​ Or is it just that parameter setting addThermalFlux? I'll try >setting that to False to see if that helps.</div></div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:small">Sorry for not being clear. I wanted to say that I have set up</span><br></div><div style="font-size:12.8px"><span style="font-size:12.8px">unsplit/MHD + </span><span style="font-size:12.8px;font-family:arial,helvetica,sans-serif">Diffusion/</span><span style="font-family:arial,helvetica,sans-serif;font-size:12.8px">unsplit </span></div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:12.8px">or </span></div><div style="font-size:12.8px"><span style="font-size:12.8px">unsplit/MHD + </span><span style="font-size:12.8px;font-family:arial,helvetica,sans-serif">Diffusion/</span><span style="font-size:12.8px;font-family:arial,helvetica,sans-serif">split, </span></div><div style="font-size:12.8px"><span style="font-size:12.8px;font-family:arial,helvetica,sans-serif">obtaining good results with both but I'm not sure if the second option is completely compatible. In both cases you have to set </span><span style="font-family:arial,helvetica,sans-serif;font-size:small">addThermalFlux=false. </span></div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:small"><br></span></div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:small">In regards to what Klaus </span><span style="font-family:arial,helvetica,sans-serif;font-size:small">mentioned</span><span style="font-family:arial,helvetica,sans-serif;font-size:small">, </span></div><div style="font-size:12.8px"><div style="font-size:12.8px"><span style="font-size:12.8px">>This indicated that the variable you are trying to diffuse - temperature, </span><span style="font-size:12.8px">I assume - is extremely >high already on entry to the diffusion solver. </span><span style="font-size:12.8px">This could be either from previous calls of the >diffusion solver, or </span><span style="font-size:12.8px">from something else having gone wrong in previous time steps.</span><br></div><div><span style="font-family:arial,helvetica,sans-serif;font-size:small">at least in my case, the above units' setup have worked well for temperatures as high as 20MK and densities of the order of 1e-15 gr*cm(-3). </span><br></div></div><div style="font-size:12.8px"><br></div><div style="font-size:12.8px"><br></div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:small">>To answer your question, I'm not using super-time-stepping.  Didn't seem necessary since with Crank Nicholson diffusion you have unconditional stability.  So I haven't looked into how to use that.</span><br></div></div><div style="font-size:12.8px"><span style="font-family:arial,helvetica,sans-serif;font-size:small">Even in this situation I have turned on STS obtaining speed up and good results. </span></div><div class="gmail_extra"><br><div class="gmail_quote"><br></div>
</div></div>