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

Re: Cap safety & spark gaps



Original poster: "Gerry  Reynolds" <gerryreynolds@xxxxxxxxxxxxx>

To Mike and others,

After hearing many (seemingly) to not understand how you can get more voltage out of a transformer than its rated output, I thought it would be useful for the newcomers or those that maybe dont have a full grasp yet of electrical circuits to give a primer on some of the basics of the effect.

The transformer can be distilled down to a simple model that is a voltage source (open circuit secondary voltage) with a series resistance (R) and a series (L) feeding the TC primary. This simplifed model will suffice for purposes of understanding most of what happens in TC service. NSTs (current limited) and PIGs (not current limited) will be used to bracket the transformer type you use.

NST's have shunts in them to allow some of the flux (leakage flux) generated by the primary to bypass the secondary winding. The effect this has is to increase the leakage flux that adds more series inductance to the simplified model. This is how NST's current limit their output for short circuit conditions. The effective inductance can be determined by using either the nameplate data or the measured secondary open circuit voltage (Vs_oc) and the secondary short circuit current (Is_sc). The transformer impedance (mostly inductive - XL) is Vs_oc/Is_sc expressed in ohms. The equivalent inductance is XL/(2*pi*line_freq). Resistance has been neglected here as for most of these transformers, the resistance is small compared to the inductive impedance (reactance). It will suffice to remember that there is both primary and secondary winding resistance that shows up as the R in the model.

PIGs, by their very nature, are designed to keep the leakage inductance and winding resistance as small as possible. If one were to short circuit the PIGs output, the winding resistance and leakage inductance will not be sufficient to adequately current limit the transformer. Use of PIGs in TC service will require a "ballast" inductor (prefered) in series with the primary of the PIG to limit the current when the spark gap fires. This primary inductance can be transfered to the secondary for the model by increasing its value by the turns ratio squared.

When we add the TC primary load to the transformer's output and look at the resulting model for charging purposes only, we have a series RLC circuit (the R and L come from the transformer and the C comes from Cp of the primary). The TC primary inductance is very small compared to the transformer's inductance so it is neglected here.

This series RLC circuit is very underdamped (R is too small to prevent the RLC from having its own resonance). The consequence of this is the RLC will have a resonance of its own where its resonant frequency is dependent on the values of R, L, and C (yes - R effects this frequency to some extent).

The transformer is a frequency source of either 50 or 60 Hz depending on where you live. If the RLC circuit also has a resonant frequency of the same (worst case senerio). It will amplify the Vs_oc voltage out of the transformer by the factor of Q.

Q = sqrt(L/C) / R

If R is zero, (Q is infinite) you would theoretically get infinite voltage out or the RLC circuit (across the capacitor). Typically for NST's (I believe) the Q is about 5 or so. PIGs have much lower R resulting in higher Q's. This means that if you use a 10KV NST, you would have 50KV on the capacitor if operated at resonance with no safety gap and nothing else fails. With no capacitive load (open circuit), the output of the transformer would be it's Vs_oc. For Cp values between 0 and Cp=Cres (where RLC resonates with the line frequency), there is some voltage gain (we call resonant rise). With Cp=Cres you get the largest rise. Once Cp becomes larger than Cres (LTR - larger than resonance), the voltage gain starts decreasing. With a sufficiently large Cp, the voltage gain decreases back to unity. With Cp values even larger, the voltage gain becomes less than unity (the transformer can not charge the Cp to full voltage).

The above description is with the spark gap (SG) out of the picture. When the SG fires, a transcient is created that adds to the above steady state response. If the SG fires at say 10KV and the SG perfectly shorts the power source, the transcient response starts at -10KV (so the sum of steady state response and transcient responce is 0V). When the SG quenches and opens, another transcient response is started. Any lingering transcient response will "beat" with the line frequency and can also result in voltages that can go higher than steady state.

A third mechanism that is associated with SRSG's can also "pump" up the voltage. This mechanism is what allows SRSGs to effectively charge a larger LTR cap to full voltage (Vs_oc_peak) than for static gaps. The details of this mechanism is described in Richie Burnett's site:

http://www.richieburnett.co.uk

Gerry R.







Original poster: "MIKE HARDY" <MHARDY@xxxxxxxxxx>

Due to the recent discussions on safety gap, and air breakdown voltage, I need some clarification. Are you guys saying that if one has sufficient energy, and a resonant , or str cap, then charging voltage is dependent on spark gap length? This explains why output gets bigger as one opens the gap. So is this why a 35kV maxwell can blow with a 14.4 pig or PT? How does one keep there caps healthy in a ballast controled situation, where output is somewhat unlimited? I realize most ballasted coil configs use rotaries, but right now I'm using a sucker gap with my PT.