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

Re: Using HV COAX without stripping the shield

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

Hi Kurt,

Thankyou for the link. I've been googling on the blumlein effect recently and the results seem to indicate a transmission line effect used for pulse shaping. Assuming there isn't a broader definition, I believe termination will kill the effect whether its load or source termination. In the googled results, the effect was initiated by a spark gap firing (like our TC's) with a fully charged transmission line.

With a 50 ohm coax connecting the pig to the sparkgap, the sparkgap end (during conduction) will have an impedance much lower than 50 ohms and the pig end loading will be much higher then 50 ohms. Since the reflection coefficient is given by (ZL - Zo) / (ZL + Zo) where the ZL can be complex but I will keep things simple by assuming ZL = 0 at the sparkgap end and ZL very high at the pig end. Given this, the reflection coefficient (Kr) at the SG end is -1 and at the pig end is +1. If the intial charged voltage at the time of SG breakdown is 15KV , then a pulse (change of, or delta waveform) of -15KV will start propagating toward the pig end. When the pulse arrives at the pig end (incident voltage Vi), Vi will get totally reflected (Vr = Kr*Vi), and will result in Vr of -15KV. The voltage at the pig end is the sum of Vi and Vr or -30KV. Now the reflected wave will arrive back at the SG end as Vi = -15KV and will reflect as +15KV (K= -1). When the 15KV reaches the pig end again. it will double to +30KV and so on. The voltage at the pig end will oscillate between -30KV and +30KV (note this waveform is superimposed on the initial charge of 15KV).

Transmission lines with +1 Kr at one end and a -1 Kr at the other end are inheriently underdamped and have a resonance with a frequency corresponding to 4 prop delays (down and back twice). If there is a source frequency corresponding to this resonant frequency coupling into the coax, then energy can build and the +or- 30KV will grow in magnitude. The fundamental frequency of the primary is way too low of frequency to support growth (except perhaps for very long coaxes). Absorbing the energy at either end of the coax can dampen or prevent the oscillation. A 50 ohm series resistance at the SG end will result in a reflection coefficent of 0 and nothing lingers beyond the initial bounce off the pig end (the pig will still see a 30KV waveform). If the pig end is terminated in 50 ohms, then that end will have a reflection coeffient of 0 and only the initial Vi will be seen. Vr = 0, so Vi+Vr = 15KV in this case.

Someone also suggested, I think, that growth in the voltage at the pig end is due to the capacitance of the cable resonating with the secondary (I presume they meant the TC secondary and resonating because of the mutual inductance between the secondary and primary and at the fundamental frequency (sorta like resonate charging of the TC primary from the power source - but in reverse). If this is the mechanism, then reducing the coax capacitance and/or adding series resistance will be helpful. Comments welcomed.

Gerry R

Original poster: "Kurt Schraner" <k.schraner@xxxxxxxxxxx>

Hi Gerry,

please see http://www.pupman.com/listarchives/2005/Oct/msg00610.html

BTW: the braid of the coax should be shown grounded. Please view in fixed font.

Cheers,

Kurt


----- Original Message ----- From: "Tesla list" <tesla@xxxxxxxxxx>
To: <tesla@xxxxxxxxxx>
Sent: Thursday, November 03, 2005 4:40 PM
Subject: Using HV COAX without stripping the shield


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

Hi All,

It has been mentioned in the group in the past that there can be a high frequency resonance when using a coax cable such as RG8U to connect the HV from a PIG to the TC where the PIG is remoted from the TC. This resonance can cause a higher voltage standing wave on the PIG end of the coax. Dr R has suggested stripping the outer braid of the coax to minimize the transmission line effect, It seems like to me that doing so changes the characteristic impedance of the cable and doesn't really eliminate the problem.

I was wondering if anyone has experimented with adding a series RC termination to the pig end of the coax to effectively terminate the coax. The C could be a string of C's to get adequate HV capabilities and the R would be 50ohms if the shield wasn't removed or say 300 ohms if the shield was removed. The RC time constant would be set to be higher than the resonant frequency of the TC and yet low enough to effectively terminate the coax for the high frequencies where the coax length would resonant.

Gerry R.