Gapped DC Charging Reactor Performance

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "S & J Young by way of Terry Fritz <twftesla-at-qwest-dot-net>" <youngs-at-konnections-dot-net>

Coilers,

I am experimenting with a DC charging reactor power supply with my twin 4 x
23 TC setup.  Power is dual MOT with filtered doublers.  Charging reactor is
a MOT with welds sawed through to experiment with air gaps.  I removed the
primary and magnetic shunts.  My hope was that adding an air gap would allow
the RSG to run slower before it started power arcing with full power
applied.  The reactor is in series with a couple of MOT HV diodes to "de-Q"
the reactor resonance effects.

In summary, the air gap does show some improvement, but not as much as I
hoped it would.  Here are details.

Case 1:  MOT reactor with no air gap, L=16.5 Henry.  The RSG starts power
arcing for BPS of 350 or less.  By power arcing, I mean a thick drawn-out
arc several inches long and DC supply current going way up until the arc
extinguishes.  Nasty condition which must be avoided!  Power at onset of
power arcing was about 1260 DC watts.

Case 2:  Same MOT reactor with 15 mil (0.015 inch) air gaps between the "E"
and "I" core sections.  (The two sections are clamped together with hose
clamps with screw tensioners on each side to apply clamping force.)  L
dropped to 9.2 Henry.  Sadly, this setup power arced at 350 BPS also!  Power
was 1458 DC watts.  (Note - I don't have much faith in my kv and ma readings
at the onset of power arcing.  The RSG sounds rough and jittery and probably
the reactor is just starting to saturate.)

However, note the L dropped from 16.5 to 9.2 Henry, which is about 44% less.
L has a strong relationship to minimum BPS - the more Henries one has, the
longer the gap dwell can be before the reactor saturates.  So for the
comparison to be fair, one should compare ungapped and gapped reactors of
the same inductance.

Case 3:  Used an ungapped reactor (another MOT) with L=9.8.  This is as
close as I could get to the gapped L=9.2.  This required the BPS to be about
450 BPS or more to avoid power arcs.  So this shows the gapped reactor
definitely performs better (can store more energy without saturating).

Case 4:  Gapped reactor in series with ungapped reactor (another MOT) for a
total L=19 H.  This is the best I could do to approximate the original
ungapped 16.5 Henry reactor.  This time there was no power arc at full power
(1296 DC watts) at 350 BPS.  It starts to power arc at about 200 BPS.  Much
improved performance.

Case 5:  Forget the bother of gapped reactors and go with brute force.  I
put 3 ungapped MOTs in series which gave L to be 50 to 55 Henries.  This
worked the best of all.  BPS can go down to about 150 before power arcs
occur.  The downside is the extra space and weight of multiple MOTs, and the
need to insulate them from each other and whatever they sit on.  Note -
power loss in MOTs seems to be very low.  At 500 BPS the case 2 power was
1406 watts, and the case 5 watts was 1368 watts - the same considering my
measurements probably have about 5% error.

So I think one should select 2 MOTs with high inductance (they seem to range
from 10 to 20ish Henry - use the 20+ Henry ones), gap them both, and run
them in series.  Next best is to use 3 or 4 ungapped MOTs in series.  For
extra credit, construct your own 50 Henry gapped DC charging reactor!  The
alternative is to always use BPSs of more than 450 with small diameter gap
elements on a large diameter RSG (to minimize dwell times) and just use one
MOT.

Comments and your experiences welcomed please!

--Steve Young