[Prev][Next][Index][Thread]
Combined RSG Mails. Warning: Lengthy post
Hi Gary, Richie, Terry, John, all,
Let me see if I can answer all the RSG mails at once. Even
though I snipped quite a few passages, to keep it as short
as possible, it did get rather lengthy. But I think there is
valuable information in these posts. You guys are always
so fast in answering. I really have a hard time keeping up
(and a guilty conscience, too).
(Gary: I will go for the MMC stuff in a seperate mail)
Gary wrote:
">Did you mean across the cap or gap? I would favor the Gap, that way if
>it fires, the energy has a useful place to go, into the primary. If the
>safety is just across the Cap, it will see some frightful currents (no
>inductors to limit it) and do nothing useful. Clamping-wise, it makes no
>difference, the cap and gap see the same voltages."
RWB:
Actually (see diagram below), I plan on using a 3 piece SSG (safety
spark gap) across the NST, a second SSG directly across the total
spark gap setup (more on that in a second) and a safety gap with
limiting resistors directly across the MMC. That way, I have total
protection if anything goes wrong. Using a SSG across the cap will
have the benefit of routing the current (via the xformer SSG) directly
to the RF ground. I will have to play around with the SSG spacings,
but I think the cap SSG should be the narrowest, second narrowest
would be the xformer SSG and widest the SSG across the SRSG
(I love abbreviations.. ;o] ). I will be using two (two or three gap)
static gaps (blower cooled?) with large tungsten rods in series with
the SRSG. I hope this will further improve quenching at the high
primary amperage I will be running (see below for more details).
Ascii Schematic:
---------------------
|-RR-O O-RR-|
| |
| |
o---X||X----|----------|--------|---|----||-----|--P
X||X O | O P
X||X O | O P
X||X | | | P
X||X | O R R R P
X||X----|-GND R R P
X||X | O R R R P
X||X | | | P
X||X O | O P
X||X O | O P
o---X||X----|----------|--------|------------------P
NST XSSG RSSG SeSG CSSG Primary
SRSG CLRs
Cap
"O": Gaps
XSSG: xformer safety gap
RSSG: rotary safety gap
SeSG: series static gap
SRSG: sync rotary
CSSG: sap safety gap
CLRs: current limiting resistors
My original post read:
>>"However, this leads me back to my old question: Can I build a RSG
>>that is in sync with the mains, but has a higher break rate?"
Gary wrote:
">I have performed some PSpice simulations, looking for the best cap values
>to use with my 15kV/60mA NST. I simulated sync gaps, both at 120 and
>240BPS. For reference, using static gaps, best case was a 9nF cap, power
>throughput dependant on gap width, 534 Watts with a 25KV static gap. A
>mains-resonant cap would have been 10.8nF."
>For the 120BPS case, best case was with about 20nF, charging up to 23.17KV,
>or about 644 Watts throughput. Timing was adjusted to fire at the peak
>charging voltage.
>For the 240BPS case, best case was with about 8nF, charging up to 24.30KV,
>or about 567 Watts throughput. Timing was adjusted to make the two bangs
>per mains half-cycle equal in size.
>>So your reasoning about using a 200BPS sync RSG to allow the use of a
>smaller cap seems valid, although there will be some loss of throughput."
Gary wrote in another post:
>Did you notice in my picture how the copper tubing arced from the sides,
>rather that the ends? It seems that every other RSG I've seen has the
>arc coming from the end of a cylindrical electrode. I would think that
>the sharp edge of that cylinder end would promote premature coronal
>discharge as the gaps closed in on each other. Plus, it makes it
>essential that the motor shaft have zero longitudinal play.
Yeah. I have another (not production ready yet) idea. Tell me what you
think about it (courier font to view):
BBBBBB
||
------
__ DDDDDDDDDDDDDDDDDDDDDDDDD __
S| | | | | | | | |S
S| | | | | | | | |S
__ DDDDDDDDDDDDDDDDDDDDDDDDD __
------
||
To motor
D is the disk,
"S|" is a stationary
"|" is the rotating electrode
"----" is part of the mounting hub
"||" is a shaft
"BB" is a ball bearing and a mount
The two disks are made of G-10. The 8 electrodes are mounted
in hex-head bolts on each disk. It kind of looks like a squirrel
cage blower. I like this design because of a few reasons (in
addition to the ones you mentioned on the Pupman list:
1.) The stationary electrodes stay away from the motor and the hub.
2.) The squirrel cage setup has got to provide lots of air flow to all
electrodes. The air can only leave the "cage" through the sides.
2a.) I might need to drill a few holes in the first disk, so that the "fan"
can suck in the air here. This shouldnīt be a big deal tho.
3.) The electrodes are double secured. There is no way they can
leave the cage.
4.) If the stationary electrodes where to come lose, they canīt enter
the cage, as they too, are double secured.
5.) Even if you where to stick something into the rotating cage, I doubt
if it would "explode". The whole setup should be pretty safe from a
"flying tungsten bullet" standpoint.
RWB:
Richie wrote me a private mail. With his permission I will repost
most of it here.
Richie wrote:
>"And that is why I have designed and built a 200 BPS synchronous gap for
>50 Hz. I have carried out very many Microsim simulations, and the
>results do indeed show reduced cap voltages, during both smooth
>running and also in the event of incorrect phase or even a missed
>firing."
RWB:
I love this list and the info on it. I have two left hands when it comes
to MicroSim. It takes me about 3 or 4 times as long to draw just the
schematic in Spice, than it does via pen and paper. Itīs good to
know that some of you guys out there can Msim, thatīs for sure.
Richie:
>"From what I have seen there are three significant disadvantages of the
>100BPS synch rotary:-
>1. If you have a high current source, the cap size needs to be BIG.
> You want a big cap to make best use of the available power
> because the charging time is relatively long."
RWB:
Exactly. Like I said a MRC (mains reso cap) for my coil would need
a 254nF (at 4500VA) cap. Tests have shown I will be able to get ~
6000VA with my 6 NSTs with ease. This would need an even bigger
cap. Itīs not just the cash it would cost, but also the size (even as a
MMC) would simply be monsterous.
Richie:
">2. If the firing angle is set incorrectly the electrode presentations
> could take place when there is little voltage across the gap. This
> results in the gap not firing, or giving a low energy bang. The
>outcome of this is that little or no energy is removed from the
>charging circuit, and voltages can reach VERY high levels due to
> resonant rise. Goodbye transformer and cap.
>3. If the gap misfires, the voltage skyrockets even higher on the
> following half cycle if a safety gap does not catch it. Result, same
> as above."
RWB:
Correct. Of course you should (I will, see above) use a SSG anyway,
but everytime the SSG fires, you will be wasting energy and you
will have to power down. So the less it fires, the better.
Richie:
">I repeated the same scenarios using the simulator for a 200BPS synch gap,
>and noticed the following coresponding advantages:-
>1. Using the same cap with a 200BPS gap, reduces the voltage by 30%,
>this results in the same overall power throughput as at 100BPS, but
>lower stresses on components.
>2. When the firing angle is set correctly, TWO EQUAL BANGS occur on
> each half cycle. If the firing angle is not set correctly, the two
>bangs will be of differing amplitudes. The worst that can happen is
>that one presentation occurs at the zero crossing, and the gap does
>not fire here. The other presentation occurs at maximum voltage.
>This is exactly what a 100BPS synch gap does when it is set up properly.
>What I am saying is:- At worst the voltage is no higher than that
>found in a correctly setup 100BPS system. (Maximum voltage is 141%
>of normal 200BPS voltage.)
>3. If the gap misfires, the next presentation occurs in only 5ms instead
>of 10ms. Since the time is shorter the voltage has less time to rise
>before the next bang. (Maximum voltage is around 140% of normal.)"
RWB:
This was pretty much my gut feel. It makes me wonder why not
too many coilers have gone this route. I will accept a slightly shorter
spark length output for a longer component lifetime anytime. Input
voltage is of very little importance in spark output. One of the major
factors in spark length seems to the input Joules. How you arrive
at these (low voltage, big cap or high voltage, small cap) doesnīt
seem to matter too much. Even if the low voltage setup has slightly
higher losses, I can compensate for this during construction. Worst
case: I will have to add a few more caps (more joules) to my tank
circuit.
Richie:
">I initially considered the 200BPS idea because I was lucky enough to
>obtain a couple of power transformers, and I found that the cap would
>need to be massive and would see rediculous voltage at 100BPS.
>With 200BPS I found that about 75% of resonant size gives max power."
RWB:
Hmm. That seems to coincide with Garyīs results. This is kind of interesting,
because it doesnīt match John Freauīs idea of using LTR caps. Did you or
Gary try simulating the 200 bps idea with LTR caps? If not, could you try
this. What happens if you use a cap that is sized 1.75 x MRC size? I.E.
does anything "mysterious" happen at or near this value? It is just a wild
thought. If not (probably), then one has two choices: Either a 200/240 bps
SRSG with an 3/4MRC (our idea) or a 100/120 bps SRSG with LTR cap
(John Freauīs idea).
Richie:
">I think that I have made the right choice, and the proof is in the
>results. So far I have run the 200BPS synch rotary on my coil using a
>6KV 600VA neon without any component mortalities. Spark action was
>very smooth and the sound was a crisp whine."
RWB:
Have you compared spark lengths from a HBRSRSG (High Breakrate Sync
RSG) vs. a good (many gap) quenching static gap? If so, what differences
were there? My gut feel tells me you will see pretty much the same results
that John F. has seen in his experiments. One simple reason being the
throughput is much smoother and more efficient in a HBRSRSG, than with
a simple static gap (which fires "at [non constant] will").
Richie:
">As a side note, you will also notice a better "efficiency" since the
>power factor is better at 200BPS."
RWB:
Now that I DIDNīT expect. Very interesting. What was the difference
(roughly in percent) in power factors? And vs static gaps?
Richie:
">The only DOWNSIDE that I can see is that the same overall power at 200BPS
>results in a lower Toroid voltage that at 100BPS, because of the reduced
>capacitor voltage. There has been much debate on the mailing list recently
>about whether high break rates give the same spark performance or not.
RWB:
Of course you can compensate for this by using a bigger toroid (which you
will need anyway to match the higher current flowing in the primary) and
a smaller discharge toroid on top (instead of a simple bump). This should
get you great streamer/arc lengths for a certain input VA. My gut feel has
always been that very high breakrates (letīs say above 4-6 * Fmains) will
be very inefficient. Of course as soon as you start feeding a coil 10+kVA,
you wonīt have much of a choice. Cap, electrode, etc sizes and cost will
start to get out of hand. I did a quick calculation on my coil, xformer and
MMC. The biggest MMC I am willing (costwise) to build would be around
230 nF. This happens to be an almost perfect fit for a 75% HBRSRSG
designed MRC at 6kVA. As I can fine tune my NSTs to the cap, I can
actually make it a perfect fit. The primary current will be in excess of 3kA
(holy cow!!). However, even at this level, my MMC wonīt see more current
per string than at the level I running it at now (being ~60A per string).
If I donīt get at least 100" sparks from that setup, I WILL surely be
disappointed. Main toroid size will be 10" x 35", secondary toroid will
be 5.3" x 30" and the "discharge bump" will be 4.3" x 31.5 toroid. All
in all, this should drop my free resonance from 217kHz to about
100-130kHz.
Richie:
">I initially did some experiments at 100BPS and blew a neon transformer
>immediately at switch on due to incorrect firing angle. For that reason
>I took to simulating before trying so that I could find a suitable firing
>angle. After all, how else do you know what is good until you try it ?
>If you have a HV probe and scope you can tell which way to adjust the firing
>angle of a 200BPS synch gap from the cap voltage waveform. If the first peak
>is higher than the second, the gap is firing TOO LATE, and the motor
>needs to be rotated in the direction that its shaft turns in. Likewise,
>if the second bang is higher than the first, the gap is firing TOO EARLY
>and the motor needs to be rotated in the opposite direction. You will
>easily tell the second bang from the first. The first bang occurs at
>the end of a steep voltage ramp, whereas the second bang always occurs
>just after the voltage starts to decrease."
RWB:
Yeah, I am still trying to figure out a safe and easy way to find the
sync position before I wire it up to the coil. My idea was using a
modified timing light. But I like your HV probe/scope idea.
Richie:
">If you consider the 200BPS synch gap, let me know, and I can send
>you several simulation results, graphs and waveforms showing firing angle,
>power throughput, power factor, capacitor sizing. etc...
>It is far easier to explain these things with a couple of pictures, and
>they may help you decide which route to take.
>If you choose the 200BPS option I might be able do some simulations for
>you if I get time."
RWB:
Great. Tell me the parameters you need and I will send them to you
off-list.
(from the sync motors post):
I wrote:
>Are you using a capacitor start motor or a centrifugal
>switch start-up unit? Maybe cap start motors don=B4t
>apply enough juice to the starting windings on modified motors
>(this is the only difference I can think of) with the original cap.
Gary wrote:
">I'm not real familiar with motor terminology. My modified motor has an
>AC cap going to a 3rd lead, but there's no sort of starting relay or
>centrifugal switch. The motor that I will ultimately use is a 3600RPM
>sync unit removed from a Teletype machine. This unit has a kind of
>starting relay that apparently switched in (or out?) the cap at startup.
>Perhaps this is what was needed for my modified fussy-startup unit.
RWB:
That is interesting. My motor doesnīt use a capacitor at all. During
standstill to about 1200 rpms the centrifugal switch is closed. Above
these speeds it opens and the starter winding is disconnected. The
current draw actually rises, if the starting winding isnīt connected
during start up (from about 2A to over 7A). This was the case, when
I found the motor (see orginal post: one lead and the switch contacts
were shot). Your modified motor sounds more like a capacitor RUN
motor. Maybe these canīt be sucessfully modified. Your Teletype
sync motor most likely switches the cap OUT at or near 3600rpm
and in at lower speeds (i.e: startup).
Last but not least: Many thanks to all who have answered and helped
me out. (and for winding your way through a long, but hopefully,
informative post).
Coiler greets from germany,
Reinhard