[Prev][Next][Index][Thread]
Re: Gap Losses
I am curious if anyone has tried a gap between the power supply and tank
to allow feeding of the tank only when it needs it? I've wondered about
having a low frequency current source in opposition to a high frequency
current at the spark gap. Theoretically the transformer protection circuit
impedance should decouple it during the ring but I wonder if part of the
tank Q could be affected by it in a practical circuit.
I have noticed that when I put in a big common mode choke to protect
my driver transformers the spark became wimpier (still same length)
but the power (voltage measured at transformer output and current
measured at transformer input) was the same?
Barry
----------
|From: "tesla-at-poodle.pupman-dot-com"-at-PMDF-at-PAXMB1
|To: Benson Barry; "Tesla-list-subscribers-at-poodle.pupman-dot-com"-at-PMDF-at-PAXMB1
|Subject: Re: Gap Losses
|Date: Friday, October 04, 1996 3:19AM
|
|<<File Attachment: 00000000.TXT>>
|From MALCOLM-at-directorate.wnp.ac.nzThu Oct 3 22:47:38 1996
|Date: Fri, 4 Oct 1996 08:55:07 +1200
|From: Malcolm Watts <MALCOLM-at-directorate.wnp.ac.nz>
|To: tesla-at-pupman-dot-com
|Subject: Re: Gap Losses
|
|Hi all,
| Last one from me before going on vacation. Just like expand
|on this excellent post from Richard...
|
|> I still can't see how presicely measuring the the input energy (even to
|> the nearest microwatt) Would give us even the remotest idea of gap
|> losses when there are a lot of circuit losses starting at the power cord
|> in the wall through the ballasts, transformers, variacs, capacitor and
|> primary circuit. The gap is of course the biggest loser of energy, but
|> the ballast system would be right on its heels.
|>
|> In my water arc explosion work, (uses a simple capacitve discharge
|> circuit with a switch or a gap), I have found that the gap or switch
|> losses are often between 30 and 40% of the input energy. and the
|> inductive tank circuit (wiring and capacitor) is good for another 10%. I
|> have changed the wiring by 1 foot plus or minus and get another gain or
|> loss of 5%. The stuff is just too touchey to pin something down relating
|> to the gap alone with generalizations made from one measurement way back
|> at the head end. System efficiency would demand a precise measure of the
|> output energy (spark) which can't be done by the use of common
|> instruments.
|>
|> Using the ammeter in the power transformer secondary circuit won't do
|> much for us either, as the gap is a dead short across the transformer
|> secondary when firing. It might tell us more about the percent impedance
|> of our transformer than gap losses.
|>
|> True gap losses can only be obtained with a wideband CT in series with
|> the tank circuit components and a simultaneous voltage measurement across
|> the gap. Integrating these with time will show the actual energy expended
|> in the gap which can then be subtracted from the input energy to yeild
|> still another inaccurate idea about how much energy reaches the Tesla
|> primary. (Circuit losses and magnetic losses in the Tesla primary circuit
|> are still ignored.)
|>
|> The gap losses are ever changing throughout its conduction period. The
|> drain on the power line or mains is increased during this time by
|> shorting of the power transformer. Resonant energy from the discharging
|> capacitor in the tank circuit is just one of the currents which flow
|> though the gap when it is firing.
|>
|> Richard Hull, TCBOR
|
|It has become abundantly clear to me during my experimentation this
|last week that while absolute losses in various parts of the circuit
|are difficult to quantify, there are several rules that come into
|play:
| Power losses in all normal resistive parts of the tank scale
|with Ip^2.
| Power loss in the gap scales with Ip.
| Outside of the tank circuit (transformers, power cord etc.) all
|losses are related to line current squared.
|
|For the tank alone: using the best possible components you can get
|causes the gap to be the dominant loss. Using poor capacitors will
|cause them to be a major loser (this has been seen when people switch
|from homemade caps to commercial pulse caps). Insufficient copper area
|in the primary coil and poor interconnects will cause them to be bad
|losers. In all cases, reducing peak tank currents as far as possible
|will reduce all losses and stress on components most likely to fail
|(caps) and reduce gap electrode wear. The best way to go about this :
|design a primary with the highest possible surge impedance. The best
|way to go about this is to run at a looow frequency with moderate
|tank capacitance and high tank inductance with a high cap voltage to
|maintain bang size. Running caps in parallel is a great idea to
|reduce current per capacitor (I^2*R loss). This might sound counter-
|intuitive when you consider that Xp = 2PIfLp but the big bonus in low
|frequency running is that cap reversal frequency is greatly reduced
|which reduces dielectric losses. You have to up power, you have to up
|input voltage.
| Also, I think it may be time to re-examine some gap
|arrangements. I know cylindrical series gaps are widely used and
|the use of blowers might well be improving their conductivity. BUT:
|for every series gap in the circuit you are suffering a voltage drop
|which equates to running a lower voltage with just a single gap. I
|think the quench characteristics of a jet of air through a single
|long gap vs the series arrangement should be compared with the
|series gaps. With a single wide electrode spacing it may be possible
|to effect an even better quench. And the use of a highly pressurized
|enclosed gap might be far better again, esp if a gas flow can be
|maintained through the space between the electrodes. Perhaps someone
|who has the gear can look into this. I will in time.
| If you are able to monitor the tank circuit _on its own_
|with an oscilloscope, look for the shape of the ring down. If it is
|triangular (linear), your gap is the major loser. If it is logarithmic
|(curved), you can do much better in the cap/wiring dept. I found out
|the hard way.
| Finally, raise the primary coil well above the ground to
|minimize shunting losses in the ground/building floor.
| Many people's experience have contributed to the above
|conclusions (to name some would be unfair to others) and I want to
|acknowledge them all and thank them for sharing their ideas and
|experience.
|
|Bye for a week or so,
|Malcolm
|
|PS - input on the above welcomed of course.
|
|PPS - many thanks to Bert Hickman for input on the modified Q equation
|for a primary where the gap is the major loss. Xp/Rp no longer applies
|in this situation as Rgap varies sinusoidally with tank current.
|
|MJW
|