why tube tester? Because it saves
time an money. There are two
types of tube auctions on ebay- expensive auctions for guaranteed
tested matched tubes and the "lemon" auctions, when we also pay a lot, or
sometimes very little - for unknown tubes. I like to hunt
bargains and special deals - buying lots, bags, boxes, died radio
repair uncle stocks etc.
Thanks to my
tube tester for example I discovered that ALL of my super brand
tubes are rather bad. If say there is an auction for a pair of Siemens
NOS but
WITHOUT the guarantee of measurements and matching - chances that the
seller of such specialized item does not know what he is selling are
ZERO! He is knowingly selling a pair, which is imperfect and he does
not
claim it is perfect. The chances of
getting a 30 or 50 % different halves in a tube are 95% if the auction
is not clearly declared as mint and measured.
I bought
recently a box of 100 Soviet military 6H6P's and they were slightly
used, not measured and not guaranteed. I paid relatively good price. I
was sure they are so so at best. Wrong. I measured that lot of 100 and
4 were 35% off the perfect measurement , but 96 were GREAT . That's
shocking good result. But again - the Soviet quality standards were
extremely high ( I mean for the goods they cared about - not for
consumer goods ;-)
If you use SRPP
like me, the good thing is that the two halves do not need to be
exactly equal - SRPP
takes care of these small differences, but it is important to have similar difference in both channels. So from my lot
I first picked the exact perfect pairs, say a 16/16 and 16/16,
and 17,5/17,5
with 17,5/17,5. That is my premium collection (about 20 pairs from my
lot. They will end up in my Fikus DAC)
Then the second
sort - also very good for SRPP - is like two tubes similarly imperfect:
15/17 and 15/17. or 18/16 and 18/16. These will be good for the
circuit, no
problemo.
I reject things like 12/20 etc.
Of course for
anode follower schemes - a perfect balance is more important.
I installed a
pot that sets the bias - so we can trace the whole tube curve - say
from 0 V to -4,5V bias - the whole practical range for the purpose like
lampizator.
I connected in
parallel 3 sockets: octal, noval 1 and noval 2, being heated by pins
4-5 or 4/5-9 respectively.
All 3 sockets
have triode one paralleled fully (Grid, cathode-grounded and anode
supply) and triode2 also paralleled all three sockets.
My V meters show
voltage drop on the anode current limiting resistors.
My
resistors are
2K and voltmeters are 20 V so the full scale means 10 mA. Of course we
could add more resistor choices and have a switch - selector. Like
being able to choose 0,5K, 1K, 2K, 5K to be able to use the
tester for many different tube types. The smallest current expected
from a certain tube is 1 mA, the largest is 40 mA in the practical
life.
It would be nice
to add a third meter - for bias voltage, but we can use our DMM meter
here - everybody has one, right ?
Tube tester in
detail
Oh I forgot to
draw the pot. It connects between the AAA cell stack - negative and the
ground. The viper goes to tube grid. We also need interrupting switch
because the pot (100K or 50K) will discharge the cells.
Of course
instead of cells we can build own electronic supply of variable voltage
from minus 5V DC to 0V.
