>Still searching for an explanation why an anode-grid arc might blow the
>cathode zener, something might have come up from an off-list exchange
>with K5GW (reproduced here with permission).
>>
>>I have lost many cathode zener bias diodes in my 2m 8877 amp over the years
>>(none since installing the 50ohm plate resistor). I always assumed that when
>>the plate arced to ground and by way of grounded grid, perhaps plate to grid,
>>that the HV power supply b plus was almost instantly placed at ground
>>potential. If this is the case, the B minus lead must also nearly instantly
>>rise to a negative value close to the full HV dc voltage less whatever amount
>>has been lost during the nearly instant period of time. This being the case,
>>the cathode is now negative several thousand volts compared to the still
>>grounded grid. Looks to me like a large current would flow in this case,
>>possibly damaging either/both cathode and grid. I know enough current flows
to
>>blow the smithereens out of the grid meter!
However, most builders protect the grid current meter shunt resistor with
=>200a glitch diodes -- which also protects the grid current meter. With
glitch diodes, the peak potential across the grid current shunt R during
an anode/grid arc would be roughly 1.5V -- which is not enough to blow
the zener bias diode.
>>Am I overlooking something obvious here? 73 de Gerald, K5GW
>
>No - and maybe you've just provided the final clue!!
>
>As you say (and EIMAC confirm) the arc pulls the B+ line very close to
>chassis potential and tries to push the B- line a long way negative. If
>you sketch out a simplified circuit, it goes something like this:
>
>1. From B+ rail
>2. Through surge limiting resistor (if used)
>3. Arc to grid/chassis (about 50V according to EIMAC)
>4. From chassis "upwards" through grid current meter
through the grid current meter's external shunt resistor is normally the
case.
>5. Through anode current meter
>6. To B- rail.
semi-agreed
>The bias zener is connected to the junction of the two meters. The surge
>current is trying to drive this point a long way negative (with respect
>to chassis). If it succeeds, it will also drive the tube cathode
>negative, and then the cathode-grid part of the tube will conduct like a
>diode. Then part of the surge current will be flowing through the grid,
>cathode and zener - maybe enough to blow the zener and cause the much
>publicized damage to the grid.
- However, there seem to be some problems with the grid/anode arc
theorum:
1. I have never seen a kaput 8877 grid bearing an arc mark.
2. No amplifier builder has reported finding one.
3. The reported arcs/'big bangs' sound like an arc in the room. Arcs
in a near vacuum do not make a 'bang' sound. For instance, an arc in a
vacuum variable capacitor produces a faint 'tink' sound. Bangs need an
atmosphere.
4. Roughly 2/3 of the kaput 8877 grids I see have fairly uniform gold
blistering around the circumference of the grid.
It would seemingly not be possible to damage the cathode bias zener -
from an anode to ground arc - D.a.02.GIF if the builder had the forsight
to put glitch diodes across the grid current meter shunt R. .
>[Note to the reflector: this is a good case for a small GIF file to make
>the circuit more clear. If you want, and with the lsit-owner's
>permission, I could post one.]
>
I have used illustrations/photographs before on [AMPS]. I post the JPG
or GIF illustration on my Web site, provide the URL, and proceed with the
discussion. [AMPS] archive readers see the URL as a Link on their Web
browser. If they want to see the illustration, one click does it. For
example, to see how glitch diodes are connected to meter shunts:
http://www.vcnet/measures/D.a.02.GIF
>Part of the solution is the surge limiting resistor in the B+ lead. It's
>very significant that you haven't lost a cathode zener since you added
>this resistor.
>
If the typical 0.5 ohm grid meter current shunt resistor does not open,
the negative discharge path is not likely to be through the zener..
- There are two methods of protecting the shunt resistor: by limiting
the peak fault current with a glitch resistor in the HV positive lead --
and by protecting the shunt resistor with glitch diodes. I do both.
>The other part of the solution is a pair of enormous
1N5400 series diodes are not large, and they will handle 200a pk.
Rich...
>diodes across the
>meters, capable of handling the full surge current. There is a circuit
>on AG6K's web pages.
>
>I use a packaged 25A bridge rectifier, connected like this:
>
>"+" terminal to chassis.
>
>Both "AC" teminals connected together, and connected to negative of G1
>meter, positive of anode meter. This is also where the bias zener
>connects in.
>
>"-" terminal of rectifier to negative of anode current meter.
>
>This puts two big diodes in parallel with each meter. If the normal
>voltage drop across each meter is less than 0.5V, the diodes will have
>no effect on the calibration.
>
>When there is an arc, the surge current stretches the voltage drop
>across the diodes to maybe a few volts each, but the diodes hold it
>there until a fuse blows. The meters can usually handle that kind of
>mild overload, and come up smiling.
>
>In most published circuits you see wimpy 1N400x 1A diodes "protecting"
>the meters. Well, maybe their surge current rating is high enough, and
>maybe they would fail short-circuit - but nobody will guarantee that.
>It's gambling with several $$$ worth of meters and other parts, and
>possibly the tube as well. For a few pennies more it's worth using
>really big diodes that you KNOW will handle the whole surge.
>
>You can also use a 1N400x connected the reverse way for good luck - but
>I can't see how a surge could ever go "backwards" in the normal GG
>circuit.
>
>
>
>73 from Ian G3SEK Editor, 'The VHF/UHF DX Book'
> 'In Practice' columnist for RadCom (RSGB)
> http://www.ifwtech.demon.co.uk/g3sek
>
>--
>FAQ on WWW: http://www.contesting.com/ampfaq.html
>Submissions: amps@contesting.com
>Administrative requests: amps-REQUEST@contesting.com
>Problems: owner-amps@contesting.com
>Search: http://www.contesting.com/km9p/search.htm
R. L. Measures, 805-386-3734, AG6K
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|
