Burned PVC Chimney
"the hotter it gets, the hotter it gets"
This is a PVC chimney which was used on a Russian GS-35B transmitting tube. The burned places appear at the lower edge of the bottom of the anode cooler and at the top of the solid copper ring that runs around the bottom of the cooler. These burned places appear on the chimney in a location which was near the back wall of the amplifier. There were no burned places anywhere else on this chimney or on the chimney of the adjacent tube. There are no streamers, holes or other traces of burning to the bottom of the chimney. There is no indication that there was an arc of any kind.
(click on the small pictures to see a larger picture in a new browser window)

1

2

3

4

5

6

7
.
Additional pictures of this amplifier can be seen at http:kc4bmx.gs35b.com .
       
This chimney failure occured in a BTI LK-2000 that had been retrofitted with a single GS-35B. Picture 10 shows a temporary fix after the PVC failure. Thanks to Brian, N9ADG for sharing these pictures.

8

9

10

.

 
Now we ask WHY?
The following explanation was received from my friend Phil Chambley Sr., K4DPK. His experience with plastics during his many years of working with them provides a significant understanding of this problem.
December 18, 2005
**
*Tony King recently replaced a PVC chimney in an amplifier that a friend had built.  It was burned by RF, and he has posted pictures on his web site.  Tony called me one morning a few days ago and told me about it.*
**
*Tony said that the spacing between the tube anode and the wall of the surrounding metal was on the order of an inch or less, and that the PVC chimney had burned on the wall side.  The tube had also failed.*
**
*Some years ago at RTP I was doing some work on one of the early forms of RF heating.  We were trying to use it to drive moisture off cakes of wet-extruded rayon tire cord.  We found then that, occasionally, a "hot-spot" would occur, and a catastrophic melt-down would follow.  I saw this again in the seventies, in my lab on North Avenue in Rome.*
**
*In RF heating, e.g. glue drying in furniture, drying of textiles after dyeing, etc., the material being dried is passed between capacitor plates in the anode circuit of a powerful oscillator.  Heat is generated in the material due to the dielectric loss of the water or solvents being driven away.  Residence time between the plates is adjusted, so that the field will not appreciably affect the material after is is dry.  This only works if the material being driven away (water, solvent, etc.) has a higher dielectric loss and a lower boiling point than the material to which it has been applied.  Absolute temperature control is, for the most part, assured.  Once the material being dried or driven off reaches its BP, it goes away as steam or vapor, and there is no subsequent heating.  In the case of water, the material never gets hotter than 212 F.  In the case of glycols, this figure is raised to above ~300, but in any case it is absolutely fixed.*
**
*Incidentally, in RF heating, the heat is not directly caused by electricity.  The heat generated in the material inside the field is actually due to friction, or work.  Because the material is dielectrically lossy, and it is confined in a field of oscillating polarity, the molecules try to align themselves with the instantaneous polarity of the field.  Consequently, they oscillate, or try to, to align themselves in series with the field.  This, and the friction with their neighbors, produces the heat.*
**
*An interesting thing we found is that almost all polymeric materials exhibit an increased dielectric loss with increased temperatures.  This means, simply, that the hotter it gets, the hotter it gets.  That is why uniformity of presentation to the field is important. *
**
*"Hot spots" absorb power at the expense of the surrounding area.  This is particularly true in frameworks where power is limited to just barely stand the process speed and workload.  What results is a catastrophic meltdown in a very localized area, when the surrounding material may emerge from the dryer still wet.*
**
*Now, to put this in the context of its effect on ham radio in general and RF amplifier design in particular, we have to look at what happened.  The PVC material was a (somewhat) lossy dielectric material between the capacitor, made up of the tube anode, and the shield wall at its closest point on the tube curvature.  Since the material wasn't moving, it simply sat there and cooked until it finally broke down.*
**
*Not only that, but the anode current that became localized in this particular spot on the chimney resulted in a hot spot through the internal structure of the tube as well.  I believe /that is what caused the tube to fail. *
**
*This brings up some interesting questions.  We, as builders, have all been asked at one time or another, "How close can I put the tube to the wall?"*
**
*Well, now we know that we can put glass tubes closer than external anode types.  **We also know that we have to be mindful of the material from which our chimneys are made.  Dielectric constant and loss are important.*
**
*Why is dielectric /constant/ a concern?  Because just about everything has a higher dielectric constant greater than air, and that constant is on top of the line in the capacitance equation.  The greater the K, the higher the C in direct proportion.  A capacitor straight to ground from the anode is an RF Bypass, and unless it is huge, it will not survive.  Even if it does, it is an un-necessary loss built into the amplifier.*
**
*Interesting how these things come to light.*
**
*Phil*
 
Links to problems using PVC in an RF environment:
 
 
Content of this page is copyright 2005 by W4ZT