Modern Tube computing

From: CRC <technobug_at_comcast.net>
Date: Wed Jan 21 21:02:03 2004

On 21 Jan 2004 15:55:40 +0000, Gordon JC Pearce <gordonjcp_at_gjcp.net>
keyed in:

> On Wed, 2004-01-21 at 15:43, William Donzelli wrote:
>>> The problem with NOS is that the glass is slightly porous and the
>>> valves
>>> will have gone a bit soft by now.
>>
>> This is basically false.
>
> Well, I call it as I see it. I have literally dozens of new old stock
> valves, in their original cardboard sleeves. Some are useable, some
> are
> absolutely shagged. It's not all just cheap brands, too.

Space heaters generally decay due to the escape of gasses into the
envelope. This can be caused by two mechanisms: outgassing and leakage.
When the tube is manufactured care is generally taken in choosing the
metals/glasses to ensure that they have a low gas content. Initially
the envelope is evacuated under fairly high vacuum and baked to allow
any adsorbed gasses to outgas. The longer the baking, the better the
vacuum. Consequently, the cheapies will use cheap metals/glasses and
shortcut the bake-out time. Finally, the tube is sealed and the getter
is fired (i.e. evaporated). The getter is an active metal that you find
plated around the base of the tube. The getter generates a high vacuum
in the tube by combining with active gas species and trapping the inert
ones. Since getter is expensive, cheap tubes go short on this item with
the subsequent outgassing of the guts causing the decrease in
conduction.

The second mod of failure is by leakage of helium. Although this gas
exists at about 5 ppm in air it abhors a vacuum and will over time leak
into the tube. Glass is permeable to helium. (HeNe lasers fail due to
the leakage of He. You can take a dead HeNe and resurrect it by placing
it in pure helium for several weeks...). The inclusion of gas into a
tube is evident by the blue glow one sees in a tube that is going over
the hill. Moral: don't own tube equipment if you're in the balloon
business.

As an aside, during the '60s is saw a control computer being developed
for the military that consisted of several hundred tubes integrated
into a volume of approximately 20x20x20 cm. The "tubes" were fabricated
by forming cavities in ceramic layers with beta emitters for cathodes
and plates and interconnection evaporated onto the ceramic. These
layers were then stacked to form the controller. The average tube was
on the order of a 1/2 cm^2 or less. This baby didn't glow, but I would
like to have the bucks that went into its manufacture...

        Claude Ceccon
Received on Wed Jan 21 2004 - 21:02:03 GMT