[Amps] Ameritron AL-811H-- Again

[R.Measures]

On Sep 30, 2005, at 2:33 PM, Jim Reid wrote:

> Here is the reply I rcv'd this morning from Tom Raush, W8JI,
> who designed the AL-811H for Ameritron.
>
> Some personal stuff at the opening, is deleted, and he did give
> permision for me to pass this info along to those of interest::
>
> "Hi Jim,
>
> ...snip...
>
> 1.) The common failure in 811 and 572 amps is tubes. The
> common mode of failure is high vaccuum or gas arcs.

Note -- the attribution marks in the original were incorrect.

  The kaput AL-811 I examined had no loss of vacuum in the three tubes 
that failed.  The cause of tube failure was shattered filaments.  This 
malady occurred all at once when the amplifier was putting out RF into 
a dummy-load.  A pop was heard and the filaments went out.  My guess is 
that it was unlikely all three tubes had gas arcs at the same instant.

> Although
> most likely under periods of high peak anode voltage
> (operation), flashovers can occur under ANY condition when
> HV is present.

Even in tubes with good vacuums?

> Sometimes the arc itself breaks down the gas
> and getters the tube,

chortle.

>  sometimes a tube is not recoverable.

shattered filaments are undoubtedly that.
>
> 2.) The resistor in the grid are to improve balance between
> tubes. The capacitors keep the grid at chassis potential for
> RF, and this is to SHIELD the cathode from the anode. The
> grid is the ONLY isolation for feedthrough capacitance.

Feedthrough C is what gets through the grounded-grid.  A grounded-grid 
is not a "SHIELD", it's a semi-shield below it's self-resonance point, 
and at its self-resonance frequency, its not even that.
-- Murphy was right about things not being as simple as they first 
appear.
>
> The values were chosen to minimize unwanted coupling through the
> tube

It is possible to neutralize what gets from the output circuitry to the 
input circuitry at the operating frequency, however, it is not possible 
to accomplish this at the frequency of the HF anode circuitry's 
unavoidable VHF resonance. (In the AL-811 I inspected, the anode 
circuitry's resonance was c. 95MHz)

> and to equalize the drive to tubes over a normal range
> of replacement tubes. I wouldn't let a backyard mechanic
> without a network analyzer suggest new values based on
> emotional opinion.
>
> 3.) You can add as much nichrome as you like and you will
> have the same tube problems. The problem is not parasitics.

Explaining why not would have been helpful.

> The original suppressors are perfectly [adequate]..

The original suppressors have a Q of c. 4.5 at 100MHz.  Ni-Cr-Fe 
suppressors have a Q of <2 at 100MHz.  More Q in a parasite suppressor 
means more Rp* on the anode and more Rp means more VHF amplification.
/*  Rp = parallel equiv. R /
>
> 4.) The meters are already protected. 100% of [such] meter
> failures are caused by the negative rail of the
> electrolytics rising above ground when the HV dumps to
> chassis ground through the tube anode to grid flashover.

I have seen evidence that such flashovers commonly occur outside the 
tube -- typically across the Tune-C and/or from the anode cap to the 
output-compartment cover.
>
> The current flows back through the grid shunt to the negative
> rail, and that is what blows out the grid shunts and meters.
> If anyone took the time to trace the current path they would
> see a single negative rail clamp diode protects the metering
> system.
>
> 5.) The "bang" occurs because things outside the tubes move
> and arc when you dump over a thousand volts to ground
> through five or ten ohms of power supply ESR.

The ESR of currently produced electrolytics is less than this.

> If you double
> that resistance by adding a 10 ohm series resistor, you
> reduce the surge current to half the value. It will still
> damage things. 75 amps of HV current isn't much better than
> the 150 amperes.

Even for a 100a-pk rated diode?

> If you really want to do glitch current
> suppression you should add at least 25 ohms using a HV surge
> rated resistor.

Depends on the tube-type.  Small oxide-cathode tubes need more ½s.  
3-500Zs can use less.
>
> Anyone is welcome to do whatever they want, but the fact is
> if the tubes are good there isn't a problem.

The tubes in the aforementioned AL-811 were making normal P  until the 
instant that a pop was heard and output dropped like a rock.  The idea 
that an arc in a tube can getter a tube would be a good candidate for 
an April fool article in *QST*.

> If one or more
> is bad you'll have a problem even with dozens of circuit
> mods. If you have problems with an 811 or 572 amp it will
> almost always just be a tube or tubes.

It's always the other guy's fault --- a.k.a.: Not Invented Here 
Syndrome.

> That's because tubes
> are cheaply manufactured and inadequately tested and
> conditioned before being sold. 3-500Z's are getting that way
> also.

10-yrs ago, Chinese tubes had mfg problems.  The two 3-500ZGs we tested 
yesterday did not.
>
> Tubes are costly to build correctly, that's why Eimac dumped
> glass tubes years ago,

In the early 1990s, Eimac was building glass tubes in SLC that had 
inadequate anode welds which caused the anode to eventually drop down 
on the grid.

> and why the people who bought the
> glass tube line weren't sucessful.

They used the same inadequate amount of spotwelding-I that Eimac-SLC 
used.  For some reason, the dudes and dudettes at Eimac-San Bruno knew 
how much current was needed to produce strong spotwelds.
- - - - In college, in sheet-metal shop, we Freshman EE majors learned 
that a nice, clean spotweld was inherently weak and that a 
rough-looking spotweld was strong since the rough look was an 
indication that melting had occurred throughout the weld.

> We have to live with what
> is being manufactured and sold.
>
> For example, the Chinese manufacturers I've seen don't even
> highpot the tubes for voltage breakdown.

Undiluted hogwash.

> In the mid 1990's
> they were testing production 572B's at 200 watts output with
> 1700 volts on the anode!! This is fodder for people with an
> agenda because it means there will be significant numbers of
> field failures related to flashovers.


>
> If you want to increase tube life in the 811 series, plug in
> 572B's. They are tested about like they should be testing
> 811A tubes. They have high reserve dissipation, so you are
> not likely to "cook" the anodes by overdissipation. The
> stated intention when designing the 811 amp was to build a
> very cheap amplifier. The cheapest amplifier possible for a
> given output power.
>
> There could have been $100 dollars of additional circuitry
> like a 25 ohm 25 watt HV fault resistor, a series 50 ohm 50
> watt cathode feedback resistor, and even a bigger tank coil
> and more expensive meters. Even with all that the tubes
> would by far remain the weakest link in the system. There is
> little point in installing rear wings on the back of Hondas
> with weak 4 cylinder engines that go 100 mph maximum, but
> some people want to do it. Myself, I prefer to get the BIG
> problems first. Change the tubes. Chinese 572B's are about
> as good as old RCA 811A's were at the same power.
>
> 73 Tom"
>
> Very interesting information.
>
> 73,  Jim W6KPI/KH6
>
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
>

Richard L. Measures, AG6K, 805.386.3734.  www.somis.org