[AMPS] filament life again
Richard W. Ehrhorn
w4eto@rmi.net
Tue, 30 Nov 1999 21:36:55 -0700
Hello John...
ETO had some life data on the 8877s used in our MRI amplifiers by GE Medical, Philips Medical and others. It was presented as a curve as number of tube failures due to all causes (per month or some such time unit) versus heater-on hours. After a smattering of infant mortalities in the first 100-300 hours or so, the failure rate stayed very close to zero until approaching 15,000 hours. As I recall there was a large "bell shaped" hump between 15 and 20k hours. And presumably not many left after that.
The hump was presumed to encompass normal end-of-life, usually due to low emission, although the failure curve included all catastrophic failures due to broken filament welds, inter-element shorts/leaks, and so forth. Not many of the latter group at any time however.
Notable exception to all of the above, of course, was the 12-18 month 8877 production catastrophe that resulted from the ill-considered cathode support design "improvement" in the 1986-88 time frame. For about 12 months' production, the failure "curve" would have shown the usual few infant mortalities, a large spike between about 200 and 400 hours, and not very many left after that.
I think data on the 3CX800A7, which is of similar design but smaller, is quite similar.
73, Dick W0ID
(John Lyles had written the following in response to previous discussion of thoriated tungsten vs. oxide cathode power grid tube lifetimes):
-----Original Message-----
From: John Lyles [SMTP:jtml@lanl.gov]
Sent: Monday, November 29, 1999 7:06 PM
To: amps@contesting.com
Subject: [AMPS] filament life again
Thoriated Tungsten vs oxide cathodes
At 11:46 PM -0500 11/24/99, Amps Digest wrote:
>Finally what factors influence cathode life? Is the number of Power Ups a
>major or minor factor? Does the overall burning time play a major role or
>>is the amount of tube conduction time the primary factor?
Both. Although the mechanical changes in TT filaments in larger tubes
with handles are more prone to failure than the lower power heaters
of oxide cathodes. For instance, we have found that the number of
power ups/downs of the filament has led to problems in X2159
tetrodes. These are also called the 8974 I think. What happened was
that the cutoff bias was changing. As tubes aged, there began a
trickle of DC plate current, which got higher and higher, as the tube
was cycled. It came to the point where the tubes were in bad shape,
with a lot of Xrays being generated along with probably internal
cutting of the anode due to the DC beam. It is believed that this was
due to the filament sag over time, changing the spacing of the grid
to filament. We have since gone to a program of leaving the filaments
lit as long as we can.
The high power filaments in TT tubes move around a lot when they
transition from cold to hot. This requires good mechanical
engineering to develop the supports to compensate, and not allow the
filaments to bow radially into the grid cage.
We see failures in Burle 4616 tetrodes usually after raising the
filament voltage. These are oxide cathode, 400 kW peak tubes at 200
MHz. We only run them at 120 kW peak for 10% duty. The filament is
low power, about 500 watts. Oxide cathodes are nice here as they do
offer high peak emission for very short pulses. A lot of it relates
to aging of the filament bars, which are painted with oxide in these
tubes. Failed ones show signs of metal fatigue and cracking.
Even though they are ramped up, when they get brittle enough, they
fail, again when turned on. So in this case, the failure is going to
happen, but the turnon is what causes it to finally occur.
>
>It's the last point that I want to address here. Both the The Care and Feed
>and Svetlana's How a tube works treat this matter in a cursory fashion.
>While it is clear that OXI's due have a shorter useful life How much
>shorter is yet to be proven conclusively. But of these write ups present
>their data in an annecdotal fashion. As this one particular Tungsten
>transmitting tube lasted 80,000 hours in commercial service. But is that
>typical? What can the average user expect? In CCS and in ICAS?
>
80000 hours is not typical in my experience. A properly designed
tube, when used within ratings, may reach this. However, it is my
experience that 20,000 to 30,000 hours for a very carefully handled
tetrode such as 4CW250,000, TH555, 4CW100,000 size tube. 15,000 for
the Burle 4616 oxide cathode tetrode.
As for the smaller tubes like 3-500Z, 4CX800, 1500B etc, I am not
sure what to expect. Broadcasters like to see about 20,000 hours or
more from their 1000 to 30000 watt tubes, which is just over 2 years
of continuous operation. Any more is wonderful, a bonus, but not a
guarantee.
An extreme example, we are getting over 100,000 hours on some of our
1.2 MW 805 MHz klystrons, so they have been in the sockets since
1980s.
John
K5PRO
--
FAQ on WWW: http://www.contesting.com/ampsfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
--
FAQ on WWW: http://www.contesting.com/ampsfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm