Question: Maximum voltage that may appear on plate for linear amp?

[K1JJ]

Hola,

Let's say you have a linear amplifier with 3KV on the plate. This is a standard single-ended class B stage, no plate modulation.

What is the maximum cumulative RF and DC voltage that could appear between the plate and ground?  Could parasitics or other operations (normal and not) cause this voltage to be higher at times?


A second question:  

I have found RF voltage can jump much farther than 1/2" per 10KV.  DC seems to adhere to this rule. Is it true that RF (like 3.8mhz) will arc across a bigger gap in air, or is it just a matter of RF breaking down insulation of some materials that are normally good insulators at DC, but not at RF?

Thanks.

Tom, K1JJ

[WA1GFZ]

I would think at least double in fly back mode when the tube is off

That is 1/4 inch in dry air any mositure will make it easier to jump.
Also any series L can have a tesla effect and raise the voltage at the launch point.
 a spark goes through the air in an interesting path like it is searching for home. We took some cool pictures of the LTI 2MV Marx making a 10 foot spark. All there little leaders searching for a path to ground. The picture showed a different path then the one we thought we saw real time.   

[K5UJ]

<<<That is 1/4 inch in dry air any mositure will make it easier to jump.>>>

I always thought dry air had easier ionization and lower breakdown potential.

Power line arcs always happen in cold dry air.  Do things change at RF?

[N4LTA]

The dielectric strength of air is approximately 25-30 KV per inch. The actual breakdown strength depends on the shape of the electrode. Pointed electrodes breakdown at lower voltages.

[K1JJ]

Yes, pointed electrodes do make a difference.

The other day I was running the 4X1 rig on high tap. Suddenly I saw a huge series of flashes around the RF final tube area. I thought the tube had arced inside and was toast. It took out my tiny wire fuse on stand-off insulators that protects the diode stack.

I found that everything was all right - the braided shield that connects the plate cap to the coupling cap had arced about 1.5" from the braid to the top of the cabinet. That's like equivalent to 40-50KV!  The rig had only 3KV on it X2 with modulation = about 6KV peak.  So tell me about RF and arcing in dry air.   I think the pointy wires of the braid encouraged the arc too. I replaced the braid and put in standard copper strap today.  It was quite the light show and glad it didn't do any damage other than depositing a small 1/4" long yellow/brown powder spot on the cabinet.
 
T

[K1JJ]

I would think at least double in fly back mode when the tube is off

6KV  peak when there's 3KV DC on the plate?   Could you explain what you mean by flyback mode and the reasoning for the 6KV number?

Now lets say there's no Tesla coil effect added in - what is the minimum distance away you would space a grounded surface from the plate cap and associated circuitry for 40M using 3KV on the amplifier?

Tnx.

T

[K5UJ]

I replaced the braid and put in standard copper strap today.  It was quite the light show and glad it didn't do any damage other than depositing a small 1/4" long yellow/brown powder spot on the cabinet.
 T

Hmmm.  Well I wondered about using braid (sorry Tom, couldn't resist) I know, I know my question had been about inductance not arcing.    anyway I'm sure glad no damage occurred. 

[Gito]

hi

Arcing can be caused by parasitic oscillation,You can find an article in the internet "    Improved Anode-Circuit Parasitic-Suppression For Modern..tube"

In the article it says that usually we used a high Q  L/R Parasitic suppressor That means using a low R cable like silver cable/copper cable,for the parasitic coil and also used the cable as the leads from the Plate RF tube to the RFC.

On the contrary with "modern tubes" the article  wrote to use a low Q L/R Parasitic suppressor. that means using a high R cable like stainless steel or nichrome Cable that has a 80 times the resistance of copper.
And also don't wound the parasitic coil using the R /,  around the R  but put the R and the L in parralel ,side by side.


Gito.

[VE3GZB]

Hola,

Let's say you have a linear amplifier with 3KV on the plate. This is a standard single-ended class B stage, no plate modulation.

What is the maximum cumulative RF and DC voltage that could appear between the plate and ground?  Could parasitics or other operations (normal and not) cause this voltage to be higher at times?

Assuming 100% modulation by a pure sine wave input, the voltage excursion on the plate should be 2X the DC voltage. In this case, 6kV.

On a non-sine wave input, expect higher peak voltages on the plates.

A second question:  

I have found RF voltage can jump much farther than 1/2" per 10KV.  DC seems to adhere to this rule. Is it true that RF (like 3.8mhz) will arc across a bigger gap in air, or is it just a matter of RF breaking down insulation of some materials that are normally good insulators at DC, but not at RF?

Thanks.

Tom, K1JJ

RF (especially if it's contaminated with parasitics) will generate transient peaks much greater than an average, an RMS value or a DC equivalent. Inductances and capacitances work to make a lovely breeding ground for parasitics and they'll pop their heads up for feeding any chance they get.

The breakdown voltage of any given dielectric doesn't materially change - all other things remaining constant - unless the dielectric is altered in some way, physically or chemically.

73s
geo

[WD5JKO]

Tom,

   A rule of thumb between for air insulation with flat conductive surfaces is 10 Kv / inch. This holds for DC on up. Switch over to two nails with pointy ends looking at each other, and the voltage standoff per inch drops a bunch. If the voltage is high, and one or both surfaces is insulated (like certain paints over metal), then the charge on the surfaces can get pretty unpredictable since the paint will not equalize the surface charge.

   A high voltage that is constant may challenge the insulation such that at the point of highest field strength, a corona spot may develop. Sometimes you can pick up corona on a portable broadcast AM radio as a screech across the band. Given time the corona produces ozone, and the 10 kv / inch rule is compromised gradually until an arc occurs.

   My job is the service and repair of high voltage ion beam accelerators. I am currently on a business trip that won't end because a machine has robotic encoder creep when running a Helium beam at 480,000 volts! To make it worse a Helium beam is modulated at about 70 Khz such that the A/D sampling rate and the noise is causing aliasing with unpredictable results. High voltage can be very tricky, and all kinds of little ticks and pops might never be heard, but boy the RF generated can sure nudge an incremental encoder. 

  So back to your 4 by 1 rf final, if you suspect a parasitic, make sure you have a big (beefy) current limiting resistance in series with your plate choke (> 10 ohms 100w?) so that the instantious peak plate current cannot try to approach infinity during the parasitic. Also hang a neon bulb near the plate circuit that will normally glow orange when you transmit. If it flashes purple on a modulation peak, then you have a VHF parasitic.

Jim
WD5JKO

[KM1H]

For HV, smooth round conductors are best to minimize arc paths.

A standard AC line fuse should not be used in a HV lead, what you saw was proof of that!  What you created was a plasma generator.  Find a real HV fuse.

Yes, a series HV surge resistor is always a good idea to limit fault current. Its purpose is to hold off the surge until a fuse or breaker can react and it must be sized to survive. This means a real wirewound, no cement junk. A 20-25 Ohm at 50W or larger should suffice for a reasonable size amp up to 2-3 kw out.

A 4x1 is notorious for parasitics, especially as a linear. If its running grid driven use a suppressor in the grid and plate leads. If GG put one in the run between the input network and filament coupling cap as well as the plate.

Always use a safety choke at the pi net output, its secondary purpose is to bleed off the DC charge across the plate blocking cap which can aggravate the Tesla effect across the bandswitch.

Carl
KM1H

[W2PFY]

when running a Helium beam at 480,000 volts!

May I ask what this machine is for??

[K1JJ]

Vely vely good information, thanks!

Carl, I like your idea of using a parasitic suppressor in the grid too, if needed. I have done that in the past.

I don't have any parasitic problems now, but I did have an arc-over in the 4X1 plate modulated rig cuz the plate cap lead was too close to the cabinet. By moving it all is FB.

Just to be sure... the opinions seem to say that for a simple linear amplifier with 3KV, (NO plate modulation) under NORMAL operation (no parasitics) there will appear X2 voltage on the plate = 6KV?  Is this correct?  I just need to confirm this for parts placement here in a new linear I'm finishing up.

** It seems to me that the RF tube is just acting as a switch here. When satuated, the 3KV goes to the load - when the tube is off the 3KV is across the tube. I don't see how 6KV is generated under normal operation. UNLESS -How does the tank Q and RF voltage across it figure into the total peak voltage on the tube's plate to ground?


Also, giving the same surface conditions, will an RF spark (clean RF at 3.8 mhz, no parasitics) jump farther than a DC spark in dry air?  Or does the RF ionize the air sooner and create a more likely event.  


Gito: Yes, I'm using nichrome wire and have the resistor mounted separately from the coil for the suppressor.

I'm also using 10 ohms at 200 w in the HV lead to the amp. I also have a fine strand of wire in the HV lead between two insulators. That ALWAYS pops when I have any kind of short. I also have one in the output of the diode stack before the filter cap - that has popped ocassionally and has probably saved the diodes. The diodes have been in there for at least 15 years now with lots of abuse.  I separate the ceramic spacers about 2"  apart and use the fine wire strands from RG-213 coax shield.

T

[WD5JKO]

when running a Helium beam at 480,000 volts!

May I ask what this machine is for??

  Yes this is an ion implanter machine used to implant ion beams into 150mm Galium Arsenide (GaAs) wafers. In this case this is an isolation implant at high energy with a very small ion. These machines strip one, two, or three electrons from a molecule (no more then two from He), and accelerate the tar out of it and ram them into a wafer.

   GaAs transistors and ic's are used a bunch in todays cell phones. The company has only two of these machines, and the one I'm working on is busted. That cuts their manufacturing ability 50%, so daily meetings, pressure, customer dictated troubleshooting methods, etc. are the norm. I've been on it 13 days straight now. It runs for 6-8 hours then begins to crap out exponetially as if something is heating up or charging up. I'm pretty baffled.

Anyway I rather not take any more space here on Tom's thread. But since you asked, I answered.

Jim
WD5JKO

[WA1GFZ]

Tom,
The RF plate choke does the same thing as the modulation transformer. When the tube turns off each 1/2 cycle the voltage takes off and loaded by the output circuit. The tank circuit gets a goose  twice per cycle. when the tube is on and when it turns off.
Carl,
In GG I like to AC load the heater choke with 1 to 2 Kohms to kill the Q of any VHF resonances. Then I hang a snubber across the input of 100 ohms in series with about a 100 pf doorknob. My 4-1000A rig is stable 160 to 10 meters. Actually the guy who has it just had it on last week after a year or so.