[Amps] TL-922 Revisited.

[Richard]

>> 1. This amp has had every known mod on the Rich
>>     Measures site done to it already. Don, KI6SZ, did
>>     the modifications. Thanks to Don and Rich for their
>>     help/contributions.
>
>Since parasitics don't cause tube problems, as a general rule, and 
>especially since the stock 922 is stable, that change should make no 
>difference. 
> 
€  Can parasitics cause tube problems?   In "Parasitics Revisited", 
*QST*, September/October, 1990, there is a photograph of a fil./grid 
shorted 3-500Z that apparently shorted when the amplifier sustained a 
"big-bang" anomaly.  There's also a photograph of a toasted TL-922 
bandswitch that burned up from what I believe was intermittent bursts of 
energy at c. 120MHz.  
-  When sending dits, even when a 922 is not oscillating, 120MHz 
damped-wave ringing can be detected near the anodes with a spectrum 
analyzer.  The reason is that there is a 120MHz resonant circuit at the 
anodes.  This resonance is formed by the 9.5pF of anode-C, the Tune-C, 
and c. 160nH between them.  If 3-500Zs have internal feedback-C and they 
can amplify at 120MHz, is is possible for them to occasionally oscillate? 

>> 2. After the initial mods were done, I suffered another
>>     grid short on one of the tubes. These tubes are the 
>>     3-500ZG's from RF parts. After getting and re-installing
>>     another set of tubes Every thing has been FB for about
>>     four months.
> 
>There are an abnormal amount of grid -filaments occurring in tubes. 
>It's getting more and more difficult to find good tubes.
>
Perhaps the problem is that amplifiers which don't use low enough vhf-Q 
parasitic suppressors allow enough vhf-gain to sustain vhf oscillation, 
which in turn puts stress on the cathode, thereby bending the hot 
filament toward the grid?  However, the drawback is that using low vhf-Q 
suppressors reduces power output by c. 2% at 29MHz. 

>In tubes I have examined, the filament has moved over into the grid 
>about half-way up the filament. 

Agreed.  This can be seen in the autopsy photo of a fil/grid shorted 
3-500Z in "Parasitics Revisited" in the Sept., 1990 *QST*.

>This happens even in tubes that never 
>have HV applied, and seems to be occurring in certain manufacturing 
>batches more than others.
>
I have not seen this in a new tube.  A typical, new 3-500Z or 3-400Z has 
a filament/grid withstanding potential of c. 8kV.  After a few big-bangs, 
the withstanding ability can be under 1kV.  When such a tube is hot, it 
shorts.   
- The amount of force need to produce a short in a 3-500Z appears to be 
considerable.  In a centrifuge, it takes 11G for 40-seconds to change the 
fil/grid withstanding voltage c. 5kV -- and that's with 5.7v on the 5v 
filament to make it more bendable.  
 
>> about 500 watts. Again, OK. But, when I upped the
>> power to full, about 1200 watts or so, I heard a POP!
>> Followed by the sound of my voice coming from the 
>> amp and accompanied by a bright white light.
>> Not good... Upon examination, I found that the two 1-ohm
>> resistors that are part of the new parasitic suppressor had burned.
>> This  suppressor is formed from nichrome wire and forms a
>> series-parallel network.
>> 
>> I replaced the resistors, and re-installed the suppressor.
>> Apart from an initial sput or zap sound from the amp, all
>> seems normal. After the initial zap sound, I did remove and
>> inspect the suppressor, suspecting that I might have burned
>> another resistor, but it looks normal. I also installed another
>> 3-500ZG to check against the first. This was done to see if
>> the initial tube had developed a fault. So far so good. I then
>> re-installed tube number one and tuned the amp on 10,15 and
>> 20 meters. Power out, plate and grid current, HV, input SWR,
>> all appear OK.
>> So now to  the 64 dollar question: Was this just some kind of glitch,
>> or was I visited by the dreaded VHF parasitic demon?
>
>Parasitic damage is greatly exaggerated. You are seeing the same 
>simple problems with tubes that everyone else sees. 

  I have never had a tube fil/grid short in my TL-922 or SB-220.  
>
>The first tubes had a grid-filament short and that is a very common 
>failure. The second was a momentary arc probably due to outgassing in 
>the tubes. This is more common with graphite anode tubes than metal 
>anode tubes, although many tubes have this problem. It, like the grid-
>cathode problem, is NOT oscillation related.
>
  the lifelong trip on the wide river in Egypt continues.

>All you did was finally reach the peak breakdown voltage of the tube, 
>and it flashed over.
>
A good 3-500Z does not flash over with 30kV between the anode and grid. 

>When the tube arcs, it often clears the fault. If the fault is a 
>small amount of gas the gas will sometimes getter from the arc and be 
>rendered harmless. If the fault is from a metallic whisker or debris 
>in the  anode/grid space the arc can clear the fault.
>
>The fault might return or might not. You just have to wait and see.
>
>One thing that helps is to always make sure you have the amp loaded 
>heavy enough to not create excessive and needless peak anode voltage. 
>Also be sure the antenna system has no problems like loose 
>connections or lightning arrestors or things that cause momentary SWR 
>problems.
>
Good points.  Tuner arcs tend to create tank arcs. 

>Another helpful check is to be sure the antenna relay in the amp 
>closes the output contact before the input contact physically closes.
>
>These are the common problems I have seen many times. If you are 
>worried about stability, put the original suppressors back in (but 
>leave the glitch resistors in, or better yet add the correct type of 
>HV glitch resistor). They have lower Q at VHF.73, 

In N7WS' tests of copper-wire vs. resistance-wire vhf suppressors [on my 
Web site], this was not the case.  Does it make sense that a better 
conductor produces a lower Q ? 

>Tom W8JI

-  R. L. Measures, a.k.a. Rich..., 805.386.3734,AG6K, 
www.vcnet.com/measures.  
end