[Home][2014 Index] Re: [TCML] Findinding the facts of a 3 P unknown transformer 50:1 "black Box" and Charging inductor(s) design. [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] Findinding the facts of a 3 P unknown transformer 50:1 "black Box" and Charging inductor(s) design.



Hi Stephan,

Teslalabor wrote:
Hi Bert & all,

as you know I'm also actually constructing a similar DC resonant
charging system. At the moment the construction of the 400bps rotary
(400bps @ 4000 1/min with very low dwell time) is in progress.
Until it is finished, there are still 2 questions, I'm wondering about.

First question is about the design of a properly rated 6-pulse
rectifier. My system is powered by a 3-phase transformer, which delivers
9,44kV output between it's phases. So after 6-pulse rectifying I get
13,35kV DC for feeding the system, so far so good.
My first try were 6 diode strings with each only 16 diodes in series,
each 1kV / 1A. This 16kV strings died after short time during my initial
halogen-dummy experiments, when increasing the voltage more and more.
Then I build a new 6-pulse rectifier, constisting of 6 strings, this
time each using 39 diodes in series, each 1kV/3A. So I now have strings
each 39kV / 3A. Couldn't test it until now, because new rotary is still
in progress.
But I have the feeling 39kV is still to less, because I saw 6-pulse
rectifyers, using 60kV per string and these were only powered by MOTs.
So, what are the important design parameters, when building a 6-pulse
rectifier for a DC resonant charger? Are there special diodes, one
should use? Im actually using 1N5408 1kV/3A Diodes. Do you think, they
will survive?

The 6-pulse rectifiers live in a relatively benign environment. The total PIV rating of each stack must be at least 2xVdc (~26.7 kV), so using almost 3XVdc (39 kV) should certainly provide adequate voltage margin for your system. 60 kV strings on a MOT system is definitely overkill. I haven't analyzed your system, so I don't know what the RMS current is for your rectifier and dequeing diodes under various conditions. However, converting from 1A to 3A rectifiers will provide greater current margin.


Second question. By studying Greg Leyh's amazing DC resonant charging
systems, both the 40kW coil and also Electrum, I recognized, that he
uses an aditional air coil inductor and a capacitor, to protect the
charging reactor (and the DC supply?) from HF, going backwards. In the
case of the 40kW coil he uses a 9,5mH inductor combined with a 45nF cap.
Isn't the charging reactor itself blocking all of the HF, because of
it's high inductance?
Is this realy needed and if so, how critical are the values of L and C
of this protection circuit? Because I use a homemade reactor, I want to
minimize the risk of killing it, it was a lot of work :-)

That's a good question for Greg to address. :^)

Remember that the design for the 40 kW coil was done way back in 1990-91 - long before the army of TCML experimenters arrived. Today, Greg might swap the position of the tank cap and spark gap to prevent the RF that appears across the tank cap from backing up into the chokes. In his original design, a lower frequency filter (~8 kHz) was used to help block TC operating frequency RF from backing up into the charging chokes.

However, even if we swap the gap and tank cap, using a HF filter may still be a good idea. But use a modern Terry filter (that shunts and dissipates the transient's energy), since the TC's RF is now shunted by the gap. But, even with the gap shunting the HV charger output, it still generates tons of high-amplitude conducted (and radiated) VHF and UHF "hash" every time it fires or reignites. Unfortunately, the high dv/dt transients don't uniformly distribute themselves across the choke winding. Because of inter-turn and inter-layer capacitances and winding inductance of each turn, the transient voltage stresses become concentrated across a relatively small number of turns that are electrically closest to the gap. If not blocked, these have the potential (NPI) to generate corona and partial discharge damage between turns and even nearby winding layers in the choke. This can cause premature failure of the choke's insulation system. Introducing a lossy low-pass filter harmlessly bypasses (and dissipates) these transients before they can reach the choke. Consider it as cheap insurance... :^)

Bert


Regards,
Stefan


----- Original Message ----- From: "Bert Hickman"
<bert@xxxxxxxxxxxxxxxxxxxxx>
To: "Tesla Coil Mailing List" <tesla@xxxxxxxxxx>
Sent: Saturday, July 12, 2014 8:12 PM
Subject: Re: [TCML] Findinding the facts of a 3 P unknown transformer
50:1 "black Box" and Charging inductor(s) design.


Hi Jim,

Remember that, like distribution transformers, your large transformer
can easily deliver 2-3X its "nominal" rated power without breaking
into a sweat for typical TC run-times. It takes some time to heat up
the large thermal mass and surrounding oil.

Now, regarding charging chokes...
Although you can wind your own, or use a series of MOT's (with air
gaps added to the cores), it looks like you could reconfigure parts of
your existing charging choke to get the lower inductance you need.
Based on a back of the envelope calculation, you'll need something in
the range of 4-6 Henries. A charging choke should not saturate during
the charging cycle, it should not have excessive DC resistance, and it
must be capable of standing off the full DC supply voltage across the
winding(s). Unfortunately, the closed magnetic circuit of your
existing choke will cause it to saturate when used in a DC charging
system, and its high winding resistance will also cause unacceptable
losses.

Reviewing the images of your power supply, your existing choke
consists of two identical windings connected in series with fluxes
aiding. The series combination of the two windings is 150 Henries, and
the DC resistance is 563 ohms. Since the windings are identical, we
know that L1 = L2 = Lw (the inductance that a single winding would
have on the same core). The mutual inductance of the pair of windings
is M = k*Lw. Because of the closed magnetic circuit, k for your choke
is probably in the range of 0.95 or more. This allows us to estimate
Lw by using the combined inductance. The case of two identical coupled
inductors simplifies to:

Lseries = (1+k)*2*Lw

Solving for the individual (isolated) winding inductance (Lw) and
resistance:
Lw = 150/(2*(1+k)
Lw ~ 38.46 Henries
Rw = 563/2 = 281.5 Ohms

Suppose we now simply reconnect the individual windings so that they
are in parallel or anti-parallel. For the parallel (aiding flux) case,
the combined inductance equation simplifies to:
Lparallel = (1+k)*Lw/2
          = 36.54 Henries - better, but still too high.

And, for opposing fluxes, the equation becomes:
Lanti-parallel = (1-k)*Lw/2
          = 0.96 Henries - still too low

However, suppose you disassembled your existing choke. This provides
you with MUCH greater flexibility, allowing you to create a charging
reactor that meets your needs and doesn't saturate. After
disassembling, install each winding on its own independent I core to
create two identical chokes (using some of the surplus core material
you mentioned), and then connect them in parallel. The parallel
connection reduces winding resistance to 1/4 of its current value, or
about 140 ohms, and also doubles the choke's previous current-handling
capability - both important for resonant charging.

You now have some further options:
Case 1: Keep the two chokes magnetically isolated (k=0 in the above
equations)
Assuming similar core size and closed flux paths for each choke, each
new choke would have about 38.5 Henries, and the two in parallel would
have 19.3 Henries. This is getting closer, but is still too high.
However, if you insert small gaps into the magnetic paths of each
choke, you should be able to reduce the inductance down to the to the
range of 8 - 12 Henries for each so that the parallel combination
falls in the desired 4 - 6 Henry range. Core saturation is also now
avoided.

Case 2: create two magnetically-coupled windings (0 < k <= 1)
You will gain much greater flexibility by magnetically coupling the
cores through adjustable air gap(s) that link the two cores. You can
see by the equations above that reducing the gap(s) (i.e., increasing
k with flux aiding) INCREASES the combined inductance while reducing
the gap with fluxes in opposition REDUCES the combined inductance. As
in the above case, introducing the air gaps dramatically reduces the
effective inductance of each winding. It will take a few inductance
measurement and gap adjustment cycles, but you should be able to dial
in the combined inductance to the 4-6 Henry range that you need.

Resonant charging calculations:
Plugging in some numbers into a resonant charging system calculator:
Based on a 0.05 uF tank cap, a desired max break rate of >600 BPS,
maximum power output output of 10 - 12 kW, and a 14 kV HVDC supply, a
charging choke in the range of 4-6 Henries is suggested. Assuming 140
ohms of choke resistance, the above inductance range supports a
maximum break rate of 650 - 730 BPS, with peak choke and dequeing
diode current in the range of of 1.2 - 1.5 amps. Bang size will be
about 17 joules at a maximum tank voltage of about 26 kV. Charging
inductor power dissipation ranges from about 11 watts at 200 BPS to
about 100 watts at 600 BPS. YMMV based upon the actual 3-phase output
voltage from your transformer, but these numbers should be in the
ballpark.

Hope this helps and best wishes,

Bert
--
Bert Hickman
Stoneridge Engineering
http://www.capturedlightning.com
***********************************************************************
World's source for "Captured Lightning" Lichtenberg Figure sculptures,
magnetically "shrunken" coins, and scarce/out of print technical books
***********************************************************************

Jim Mora wrote:
Hello List,

I have been fascinated by 3 Phase, 6 pulse resonance changing since I
first
read Richie Burnett's pages 11 years ago. Bert Hickman has been very
helpful
in calculating the parameters of my repurposed charge transformer.

Important things to know are the cross-sectional area of the core of
coarse
(known).
The turns ratio (approx determined).
Another thread suggests measuring the inductance of a primary winding
(shorted in) and its DC resistance (Z) - good idea!

I have learned a lot from this exercise. My core cross section area
is bit
disappointing. But on the other hand, I don't have enough room for
very long
streamers but hot ones would be awesome. Bert had an excellent
suggestion to
rewire the secondary from wye to Delta which he points out has a 58%
drop in
voltage and a bonus substantial output charging current increase.

The likely output will be 10KVA based on the core. I have both a .125"
electrode x 12, and a .5" really beefy x 8 rotaries. Richie suggests
paying
attention to the dwell time to avoid streaming power arcs. I guess
that will
be tested "empirically" (I have lots of cut .125" electrodes for
distructive
testing) and 3/8" for dual series fixed. I haven't made the .5"
towers yet,
I'm thinking 9/16"

*** I now want to turn attention to the charging coil ***

My parameters are not too different as Richie's test case. I have a
new .1
uf (50KV)General Atomics that can be series to .05 (100KV). I think
having
(2) inductors is a good thing for voltage stand off and flexibility
and a
variable gap too.

*** I am going to study the many inductor design sites but any school of
hard knocks would be welcomed.  I have quite a bit of core material
around
both EI and C cores for low freq operation.

*** I fully intend to over build the diode strings but will start a new
thread on the Dequeing diode robustness. I have (500) UF5804 which are
milspec 3 amp 1000v pk inverse diodes, 500 RMS. I intend to over build
beyond that. 3 amps forward(1.5) "should" be plenty for three phase 6
pulse
current at 10KVA.

Thanks for all you do! I am having fun getting back in Tesla coiling!

Jim Mora

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla





_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla





--
Bert Hickman
Stoneridge Engineering
http://www.capturedlightning.com
***********************************************************************
World's source for "Captured Lightning" Lichtenberg Figure sculptures,
magnetically "shrunken" coins, and scarce/out of print technical books
***********************************************************************
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla