Discharging A Charged 10,000 Volt Capacitor

[EJ Jurich]

While troubleshooting a Harris 5KW AM transmitter problem, it became necessary to discharge the 10,000 volt filter cap. The problem was that as soon as the plate was turned on, the transmitter had an immediate overload shutdown. This was a 4CX5000 final vacuum tube rig with a 10,000 volt power supply. To trace the problem, I started disconnecting the high voltage supply back towards the power supply, each time trying to get the plate voltage on. Finally, I opened a point and the high voltage supply came on. I knew what the problem was. However, walking around to the back side of the transmitter, I heard a sizzling sound and I could smell ozone. I realized that when I opened the last supply point, I also removed the bleeder resistor from the supply. With no load, the cap was probably charged considerably more than 10,000 volts. I looked at the station manager and said "there is only one way to do this". Using the shorting rod, I tapped the capacitor terminal. The bang was incredibly loud.
Although I was a bit lax in not following the diagram enough to realize the bleeder would be removed from the circuit, Each time I opened a point, I used the shorting rod before opening the next point. In other words, I made sure the cap was discharged each time before putting my hands in. My senses cluing me in that the cap was charged did save me a fright. I knew what to expect when discharging the cap.
I've worked with voltages few people doing electronic work will ever need to deal with. The highest was a 50KW AM transmitter with a 40KW, 4amp DC supply. Putting your hands in a tube amp chassis can be equally dangerous. My father told me that while working on a radio chassis at home, he picked up the chassis while powered on and his hand caught the high voltage. The only way he could let go was to fling the radio out of his hands.

 

[Nigel Goodwin]

While troubleshooting a Harris 5KW AM transmitter problem, it became necessary to discharge the 10,000 volt filter cap. The problem was that as soon as the plate was turned on, the transmitter had an immediate overload shutdown. This was a 4CX5000 final vacuum tube rig with a 10,000 volt power supply. To trace the problem, I started disconnecting the high voltage supply back towards the power supply, each time trying to get the plate voltage on. Finally, I opened a point and the high voltage supply came on. I knew what the problem was. However, walking around to the back side of the transmitter, I heard a sizzling sound and I could smell ozone. I realized that when I opened the last supply point, I also removed the bleeder resistor from the supply. With no load, the cap was probably charged considerably more than 10,000 volts. I looked at the station manager and said "there is only one way to do this". Using the shorting rod, I tapped the capacitor terminal. The bang was incredibly loud.
Although I was a bit lax in not following the diagram enough to realize the bleeder would be removed from the circuit, Each time I opened a point, I used the shorting rod before opening the next point. In other words, I made sure the cap was discharged each time before putting my hands in. My senses cluing me in that the cap was charged did save me a fright. I knew what to expect when discharging the cap.
I've worked with voltages few people doing electronic work will ever need to deal with. The highest was a 50KW AM transmitter with a 40KW, 4amp DC supply. Putting your hands in a tube amp chassis can be equally dangerous. My father told me that while working on a radio chassis at home, he picked up the chassis while powered on and his hand caught the high voltage. The only way he could let go was to fling the radio out of his hands.

Quite the opposite, many people doing electronic work will have been repairing TV's where the EHT for the CRT was 25,000V or more on a colour TV. The danger (and a great danger it is) of your situation was the high current - presumably the 10,000V would be capable of 1A or more, for a 5KW transmitter - EXTREMELY fatal. The EHT on a CRT TV was limited to around 1mA maximum, too small to be dangerous - apart from jumping back, falling over, and breaking your neck.

Tube (valve) amplifiers and radios are no where near as dangerous, it's only around 300V at relatively low current, it 'could' be fatal, if you were really unlucky - similar supply voltages were present in TV's of course as well.

The only really 'dangerous' domestic appliance is the microwave oven, with high voltage (about 2500V DC) and high current - not as bad as your 5KW transmitter, but you don't have one of those in every house

I'm surprised you didn't have a proper discharge probe?, which included a high resistance, to gently discharge the capacitor, without all the drama of shorting it out - and without potentially damaging the capacitor as well.

 

[schmitt trigger]

As Nigel says, transmitters are very, very dangerous.

I once saw the awful result of an accident with a transmitter tube. The clamp on the tube's anode was making an intermittent contact. A technician "attempting to fix" it, powered down the transmitter but didn't discharge it. He was kneeling when he touched the HV. The current came out from his left knee, which caused enough damage to require extensive reconstructive surgery. A little more energy, and it would have required an amputation.
A dark stain on the floor reminded everyone of the accident for several weeks. I am glad that I didn't personally witness the accident, people who did mentioned the awful smell of the burning flesh.

 

[Tony Stewart]

When smelling or hearing Ozone , this is always due to the surface contamination of partial discharge (PD). While clean air can withstand 10k to 30 kV/mm from sharp to smooth surfaces. Any dust accumulation can reduce that to 300V/mm with moisture. So compressed air with isoprop is essential maintenance to restore some of the insulation.

Grid service technicians recently repaired some of the insulation on our 40 yr old street pad-mount distribution transformers that use 18kVac. They use a 2 m fiberglass rod to handle each wire plug for safety. The old ceramic insulators were protected by a neoprene glove and looked to be around 60 mm long which is equivalent to 18k/60 mm or 300 V/mm. Later this summer they will dig to the buried XLPE cable and inject fresh silicone to rejuvenate the cable's insulation.

In the 70's my 1st colour TV had similar tripler corona and I attempted to repair with silicone RTV without the critical and necessary task of deionized cleansing, which I now understand better. It worked but was not completely free of corona.

I estimate your supply has at least 10uF of storage at 15 kV and over 2 kJ of energy. When an arc rise time < a microsecond, this is equivalent to a megawatt radar pulse, so next time get a safety rod > 10 Meg and discharge for 5 minutes for < 20 watts of heat. Then short out with a jumper while in maintenance, as the memory C in electrolytics will slowly restore some of the voltage just like a battery from the double electric field effect. Although I suspect these will be plastic (PU or PE) which do not have such as strong an effect yet a very low ESR resulting in ESR*C = Tau = xx nanosecond range rise times in the microwave band.

The explosive sound is from the particles travelling faster than the speed of sound in a small gap.

Lightning does the same thing but travelling over a large gap of pre-ionized streamers and can be heard a thousand km away on any AM radio quiet channel.

 

[ronsimpson]

I use a plastic rod to push the blead resistor wire back into place. (acrylic, Plexi) I have used wood, but it must be very dry.
It is possible to add a new bleader resistor, but I think reconnecting the old one is best, not everyone has an extra 10kv resistor. The resistor does not have to make good contact for now.

I worked on several 100kw fm radio stations, (erp) and one small a.m. station.

Maybe you should get some HV resistors and make a bleader on a stick.

 

[EJ Jurich]

I'm surprised you didn't have a proper discharge probe?, which included a high resistance, to gently discharge the capacitor, without all the drama of shorting it out - and without potentially damaging the capacitor as well.

Broadcast transmitters only come with a very large shorting stick and a nice long handle. The high voltage caps are usually 4uF to 8uF oil filled.
That 50KW transmitter was at WAPI-AM in Birmingham, Alabama. One morning I got called because it would not come up one morning. I opened all the cabinet back doors and could not see anything obvious. I tried to bring it up and got another really loud bang and immediate overload. I always hate those, not enough time to get meter readings. Opening the back doors up again, upon closer inspection, it seemed the three-foot high insulators the high voltage buss ran across looked a little dusty. I took a clean cloth and wiped them down (after first touching the shorting rod to the HV line). That was it, the transmitter came right up. I forget what tube was used for the final, but it was big. Replacing the tube required turning a big screw under the tube socket that pushed it out of the socket. The new tube is pulled back down into the socket with the screw. I remember how much that tube cost, $10,000. I have installed new transmitter sites, both AM and FM. My last install was a digital instal in indianapolis.

 

[ronsimpson]

Broadcast transmitters only come with a very large shorting stick and a nice long handle.

I replaced with a bleader resistor on a stick. Big bangs scare me.

I had the big cap explode once. They are hard to get.

 

[Tony Stewart]

I found out BIL 100 bushings that are about 18" long are only rated for 100kV lightning transients but easily have creepage disharges at 15 kVdc. In order to test over 100kV adding another bushing in series was not enough. I had to put 4x8 teflon sheets to cover the top around the center vertical bushings as well as wipe down with isoprop. Then I put an AM radio around the transformer to pickup the PD static discharges which are the canary in the mineshaft to corona and then full discharges. There must have been a high humidity situation with the dust. Carbon trails can be another clue. I would suggest a PD detector to alert for cleaning for both the oil caps and the air bushing . DGA samples can be done for oil for H2 content.