[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: LTR cap BPS?

Original poster: "Barton B. Anderson" <tesla111@xxxxxxxxxxxxx>

Hi Gary, All,

Well, I put in your coils data (including sec/pri/topload). Javatc predicted 514kHz and autotuned at 7.54 turns for the 0.012uF cap. Note, I always start with the full primary turns, ID, and OD. I then run Javatc. Following the first run, I then set it for Auto-Tune and run again. Javatc reconfigures the primary for best tune. BTW, predicted sparklength would be somewhere between 12.9" and 16.4" (input power prediction, static gap prediction).

I looked down at the static gap section (first using the 4/20 model). As expected with the partially removed shunts, there was a lot of error.

4/20 NST, Rp=3.9, Rs=9k,

Cap        Meas    JavaTC    Gap
Size       BPS      BPS     Width    error
------     ----    -----    -----    -----
.012uF     297      251     0.05"    15.5%
.02uF      207      114     0.05"    44.9%
.025uF     157       73     0.05"    53.5%


I then tried out your idea for the 4/38 (short circuit measurement), but as you see, it isn't that simple.
4/38 NST, Rp=3.9, Rs=9k,

Cap        Meas    JavaTC    Gap
Size       BPS      BPS     Width    error
------     ----    -----    -----    -----
.012uF     297      499     0.05"    40.5%
.02uF      207      299     0.05"    30.8%
.025uF     157      221     0.05"    28.9%


I then took a stab at trying something between the nameplate and your short circuit measurement. The numbers came out better and shows that the NST unknowns are affecting outcome. Some of this error would also be in the gap itself.
4/29 NST, Rp=3.9, Rs=9k,

Cap        Meas    JavaTC    Gap
Size       BPS      BPS     Width   error
------     ----    -----    -----   -----
.012uF     297      381     0.05"   22%
.02uF      207      207     0.05"   0%
.025uF     157      147     0.05"   6.4%


Javatc just isn't setup for NST's with removed shunts. JavaTC's static gap bps for NST's is emperical around off the shelf units, and then there are still uncertainty's I have when applied over the range of units out there. It's just the programs best guess at the information it has available to it, which could certainly be improved upon.

As for the original posters question regarding the 60 bps number, it could certainly be near 60. Most likely, the cap size chosen in relation to the NST power is predicting this. The error has a large dependency on modifications, losses, etc.. Most likely, the NST isn't modified (considering the daunting depotting task involved) and the prediction will fair much better to the actual. The poster should also realize that the actual gap components and setup will play a significant role with this number.

All should be aware that for the lower sections of the program (power, rsg, static gap), the numbers predicted are close approximations, but even these approximations can vary with the various setups we use. I personally use these approximations to get an idea and feel for what I can expect in a setup. I am aware there are a myriad of modifications I can make to all parts of the circuit to change the outcome. This is just a fact of coiling. The original posters question is a perfect example and had the program not shown that the BPS could be something other than 120, the poster may have just assumed it. The answer to the coiler is yes, it will probably be near 60 bps if you build the gap as dimensioned and use the NST modeled.

But also, you can get this up to 120 or higher with a smaller cap size, more power, or gap changes, etc., also keep in mind the program is not determining missed firings or multiple firings beyond the charge to arc voltage (which are very real and do change with time over consecutive runs).

Take care,
Bart B.



Tesla list wrote:

Original poster: "Lau, Gary" <gary.lau@xxxxxx>

My 4/20's Rp is 3.9 Ohm, Rs is 9.0 KOhm.  Single gap today is .060", I
can't be certain if it's been opened since the experiment, but I'm
certain that it did not change between measurements, and that it was
between .050" and .060".

Regards, Gary Lau
MA, USA

> -----Original Message-----
> From: Tesla list [mailto:tesla@xxxxxxxxxx]
> Sent: Thursday, December 02, 2004 10:06 AM
> To: tesla@xxxxxxxxxx
> Subject: Re: LTR cap BPS?
>
> Original poster: "Barton B. Anderson" <tesla111@xxxxxxxxxxxxx>
>
> Hi All,
>
> I'm still here, I've just been in lurk mode since March. I've been too
busy
> with real work to do any coiling or spend much time with TCML. The
last day
> I've had off at work was back in 2003.
>
> The BPS stated at 60 must be a rather large cap size or small NST
power.
> JavaTC does take into effect running at higher input voltages,  but
does
> not account for inductive kick. Javatc will also account for resonant
rise
> if measured Rs and Rp are entered (there are inputs for measured NST
> values). The BPS can be above or below 120 depending on electrode size
and
> gap width, cap size, and nst specs. It is about as good as can be done
with
> calculation that I know of at this time, but of course, there are
> parameters within a gap setting which will effect your actual measured
> value. For example, heat!
>
> Because Gary showed his bps values for the various cap sizes, maybe he
> could run through Javatc and see how it faired, but seeing that it is
data
> from an old measurement, I know that gets difficult to do. Hard to do
anyway.
>
> For example, Gary's 4/20 NST run through Javatc (without actual
measured Rs
> and Rp values) showed the following "assuming 2 electrode static gap
at
> 1/4" electrode diameter" (of which I have no clue what it is). I also
have
> no clue of the effects of Gary's partially shunted 4/20. That really
throws
> things off. The only way I know about doing this is to input the NST
and
> known cap size and change gap widths until BPS is relatively close and
then
> wonder if gap width calc'd is near to the actual. From my vantage
point,
> this is impossible to do with any accuracy. It's in the hands of the
> builder to measure and test and show error. This has really not
happened as
> it should have with gaps. There are a few, but only a few (over many
years).
>
> Cap          Meas.     Calc.        Gap
> Size          BPS        BPS   @  Width
> ----          ------      -----         ------
> 0.012uF   297         299           0.045"
> 0.02uF     207         210           0.025"
> 0.025uF   157         154           0.021"
>
> It may be that the actual gap widths were all identical. But with the
> shunted and unmeasured Rs Rp values, who knows how it would really
play out.
>
> Although Javatc is fantastic with primary and secondary specs, I am
not as
> confident on the remaining power, rsg, or static gap sections of the
> program. Comparisons of measurements and calcs are still needed in
these
> remaining parts of the program to get that warm and fuzzy feeling.
>
> Take care all.
>
> Bart B. (in desperate need of a coiling vacation!)
>
> Tesla list wrote:
>
> >Original poster: FutureT@xxxxxxx
> >In a message dated 11/30/04 9:07:31 PM Eastern Standard Time,
> >tesla@xxxxxxxxxx writes:
> >
> >
> >>Well, my question comes from
> >>http://www.classictesla.com/java/javatc.html
> >>Seems like if I use its LTR suggestion, it states 60~ or so BPS at
any
> >>decently wide gap setting for the NST.
> >>Is this just the program screwing up and stateing 1/2 the BPS? or is
it
> >>really 60bps at those gap settings?
> >
> >
> >
> >Bart's JavaTC is very good, but I don't think I exactly agree with
> >his method of figuring the BPS, etc.  I don't think JavaTC accounts
> >for the extra power that can be drawn out an NST beyond its
> >name-plate rating due to resonant or inductive kick and
> >input overvolting.  This extra power permits the breakrate to
> >be higher than expected using a given cap value.  Perhaps
> >Bart has updated his program, I don't know.  A too-wide
> >gap will tend to reduce the breakrate of course.  I do agree
> >with Gary Lau's previous comments regarding the breakrate.
> >
> >John
> >
>
>
>
>