Re: ScanTesla Version 7.50

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>

Hi dest,

At 10:16 AM 6/1/2006, you wrote:
> Hallo Terry.
>
> > Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>
>
> > active streamer loading
>
> again. as i see from "ScanTeslaInstructions-740.pdf", you considering
> Xload connected to the secondary all the time - from the start of the
> bang to its end. how can this be possible? i have big toroid with a
> considerable roc, it just can`t break down at say 200kv and i doubt
> that high ambient ionization can greatly help here.
Version 7.60 will have FULL dynamic streamer loading with breakout 
voltage inputs and even ground strike distances and ground strike 
modeling :-)))))
There is always power in the "streamer load" even before breakout or 
there would not be energy heating the, yet to be created, streamer 
channel.  Not realizing that was always the problem with dynamic 
streamer load before.  I will probably have to write another paper on 
how my dynamic load model works...

> >>1. why it so slow?
>
> > When it is slow, it has found a new best match and it is recording
> > the data.
> > Most (80%) of the time is for file writing.
>
> then why haven`t you implemented some sort of buffer for this - in
> ram? memory is cheap today, but proggy generates modest volume of
> data, so imo such buffer would work great even on 32mb of ram : )
Good idea ;-))  I actually open and close the files all the time 
too...  I "good" programmer could probably fix a lot of my 
mistakes...  Just "getting it to work at all" is my goal ;-))  I 
usually print files "all the time" for debugging in case it crashes.

> > You can actually have another program going in a different directory
> > to get to about 100% CPU usage.
>
> so i can divide range of say Csec into 8 intervals, run 8 instances of
> the proggy on 4 of my piv3g, and still would have consistent results
> in the end?
Yes.  You could run Csec = 10 to 20 in one program and have another 
program running Csec = 21 to 30.

> i just got right answers from the proggy - correct lenght of the
> sparks, values of Lpri, Csec, and so on - to do this i had to use 12'
> of sparks loading instead of planned (and reported in the end of this
> run) 7', but i think i know why this is happening - see the beginning of
> this post.
Yes,  The dynamic streamer loading in version 7.50 is not very good 
yet.  You need a minimum Cload or the streamers don't start 
right.  The new version will fix that.  Version 7.50 had the program 
set up for dynamic streamer loading but it was waiting on those 
models to become available.  I guess mine got here first now.  I 
think there is also a problem where the model keeps running "forever" 
in V7.50 with dynamic streamer loading.  I have been working hard on 
these problems!!
I hope to have the new version out in a day.

Cheers,

        Terry