Understanding Tube Testers (LONG)
David Forsyth
Maybe someone can shed some light on the whole topic of tube testing for me.
I finally got around to buying a reprint manual for my Eico 666 tube tester
the other day. Although the Eico had originally come with some scant
literature (basically some xerox copies) back when I first bought the tester
years ago, it wasn't nearly complete. I had since lost that documentation,
and so the tester has been sitting around unused for maybe a couple of years
or so. The reprint of the manual I just purchased for it has all kinds of
information on basic use, obsolete tube settings not on the roll chart,
(some) theory of operation, calibration procedure, etc.
This brings me round to my line of questions. I wanted to get the tester up
and running to test some old tubes I recently aquired. So, after reading
through the reprinted manual, I started by making sure the 666 was
"calibrated". Well, the only thing you really seem to be able to tweak is
the leakage and the line readings on the meter. There doesnt seem to be
anything to truly "calibrate" the meter, in the sense of making sure that
it's giving out meaningful numbers. So my question is, how does one truly
test a tube tester, to see if it's readings are accurate? I know different
tubes from different manufacturers will not be at the same "100%" point,
even when new. So what does 100% really mean? Perhaps I need a more
sophisticated tester? I see people who sell tubes listing a certain
transconductance, but how do they know for sure? I mean, wouldn't that
number really depend on the operating points of the test circuit, as
compared to the operating points used originally to derive the data shown in
tube manuals? In the ads you often see something like "tests strong at 1650
micromhos - rated minimum for this tube is 1200, new is 1850" How can you
know that the tester you have is putting that particular tube under even
vaguely similar conditions, so as to be able to compare to whats listed in
the manual? And even if it was *designed* to do that, how do you know it
still does, 40 or 50 years after it was manufactured? For all I know,
several if not all of the components in my Eico have drifted over the
decades, to the point where it's putting out meaningless information. So
what do I do - measure each and every component in the whole thing?
Well, yes I know you can still use the numbers in relation to one another,
like for testing balance between dual triode sections, etc. But even then,
those grid and plate knobs are so easy to rotate - and the markings are so
fine - how can you be sure you get consistant results even from one test to
another?
Occasionally I might like to sell a tube or two on ebay - and then I would
like to be able to rest assured that a tube I'm representing as "tests like
new" isnt in fact "marginal" - or worse! As an example, I tried testing
some old 4-pin antique tubes with the Eico and they tested very well - some
even slightly higher than 100%. Now from what I've seen out there in terms
of used antique 4-pin radio tubes, you dont usually see them testing above
maybe 40% to 70%. So thats what initially made me suspicious. Yet I then
went and tested some brand new ECC83S tubes from JJ and they tested at about
95% to 102%. So now Im thinking, ok the tester might be more accurate for
one type of tube than another type! How many tube types would I then have
to test in it to get an idea of how it's calibrated? Am I shooting for
something too idealized here? I guess I just imagined a tube tester to be
accurate for all tube types more or less equally, and to be able to give
meaningful and repeatable results. Is this possible to achieve with my Eico
or do I need to invest in something more sophisticated?
Well sorry for the long post - just had to get this out and see if anyone
has some insight into this whole tube tester business...
thanks in advance,
Dave
arizona cowboy
the Eico 666 leaves some things to be desired, such as the weak calibration
procedure you discovered and lack of a gas test.
However, the good news is that most of the lab grade testers out there (the
ones that typically bring $200+ on ebay) measure transconductance quite well
and have very thorough and accurate calibration procedures.
I sell many of the TV-7 series testers on ebay (more than anyone else) and
all of the ones I sell have all undergone an intensive factory calibration
routine by Dan Nelson of Phoenix, who is the world expert on this model.
The TV-7 series, especially TV-7A/U, TV-7B/U, and TV-7D/U are the best
selling high end testers of all time.
The are a military issue, using ceramic switches & sockets, mil-spec parts &
only a single capacitor and have an enviable reputation for durability and
accuracy when properly calibrated.
The TV-7 series was designed by Hickok, who make the lion's share of the
remaining high end testers out there.
The TV-7 series have data available for more tube types (over 4000) than any
other tester on earth.
There are also a handful of other really good ones (such as the Triplett
3444, Heathkit TT-1A, the Avo's and a few others)
but 95% of the tube testers for sale on ebay are emission types and barely
more than toys.
Kara Chaffee is a good one to calibrate the Hickok branded testers and their
Stark, RD, and WE clones.
post a brand & model you are considering here and I am sure you will get
many opinions on it (including mine!)
cheers!
wild bill
ebay: el34 AT comcast DOT net
PS - my favorite Hickok models are: 580, 539A, 539B, 539C, 533A, 600A,
800A, 752, 752A, 799, and 6000A -- with 533A, 539A, 752 and 600A probably
best bang for the buck.
"David Forsyth" <fod...@ncx.com> wrote in message
news:vm26pdm...@corp.supernews.com...
Nothing40
Thats my take on my Eico 667.The reading for different tubes are all
over the place. Used 6V6's test at 140+,and peg the needle,but a new
12AX7 measures ~60.
I can't make heads or tales of it.I just test a few different tubes,to
get a rough "ballpark" idea of what new ones test like,a few used
ones,etc and take it as it comes. (I just tested a batch of used
6V6's,and they were all over 120. and about 25% of them were off the
scale! (*tink!* goes the meter needle.)
The 667 also has the leakage trimmer,and one more (I forget what
for.),But it is also lacking any real "calibration" controls.I'm
eventually going to go through it and replace the caps and resistors.
Atleast get new,tighter tolerance parts in there,perhaps the reading
will become more meaningful? Thats about the only "calibration" for it
that *I* could think of?!
My Triplett,ermm 3814 (??I think?) is much the same way, all 12AX7's
test about 51%,new or used..The *really* dead ones will test lower.
most other tubes seem to test somewhat normally,for the most part..a
few tube types won't ever test over,say 55%...I dunno why? If it says
they are bad,or marginal,I get a second opinion ;-)
Chris Haedt
This type of question comes up all the time, and requires some
clarification. Virtually all tube testers (except for the lab types)
were designed to find BAD tubes--tubes that were likely to cause
trouble in an electronic circuit. Tube testers are service
instruments, designed to help troubleshoot and repair electronics.
This is why the better testers (like the TV7, the 539, the Triplett
3444) provide you with a minimum value or reject point for each tube.
Tube testers were NOT designed to provide determine merit of good
tubes. So--when selling stuff on Ebay--it is best to specify the
tester used, the result obtained, and the reject point--FOR THAT
PARTICULAR TESTER. For instance, if you say a 6L6 tested "44" this has
no meaning unless you also say that it was tested on a TV7. Even when
a micromho figure is given--you need to know the tester used and the
reject point for that tester. Different testers subject the tube to
different operating conditions, and this results in different
readings. These readings vary a great deal. The concept of mutual
conductance is such that it varies by tube operating condition--so
different testers can yield different numbers and still all be
correct. We are not talking about an easy to measure and understand
concept like DC voltage. You might want to purchase Alan Douglas'
book TUBE TESTERS AND CLASSIC ELECTRONIC TEST GEAR. This book will
give you good background on the different testers and how they
operate.
The Eico 666 tester is one of the better emission testers, and has an
excellent interelement leakage test. This is not really a very useful
tester for audio work. Emission testers do not tend to agree much
with each other because the specs for tube emission were not really
defined by manufacturers (except for rectifiers and a few tubes like
the 6L6 used by the military). This means that every manufacturer had
their own ideas, and a tube that might test good on your Eico--might
test bad on a Jackson. The emission tester concept is based upon the
observation that emission of electrons from the cathode of a tube
tends to fall off as the tube wears out. This does work for finding
really bad tubes, but you certainly cannot compare Eico reading with
anything in the tube manuals. The 100% that an Eico reads was just a
number determined by them, and does not relate to much else.
Most tubes are used to amplify a signal, and the emission tester does
not tell you much about this ability. That is where the concept of
mutual conductance comes in. A mutual conductance test provides an
indication of a tube's ability to amplify a signal. This is really a
better measure of a tube's merit, and measures of mutual conductance
do not relate well to measures of tube emission. When you talk about
matching sections of a triode tube--the Eico is not useful--this
should be done with a mutual conductance tester. Small signal tubes
are voltage amplifiers--so mutual conductance is very good measure of
how they will perform their function. Mutual conductance testers (when
operating to spec) will tend to agree on what are Good and Bad tubes,
while emission testers will not. I base this upon my own experience
of restoring and calibrating hundreds of tube testers.
Hagstar
>
> - with 533A, 539A, 752 and 600A probably
> best bang for the buck.
Or the 540- these are often quite reasonable. If it doesn't test a tube,
I don't want it anyway.
John H.
Hagstar
>
> When you talk about
> matching sections of a triode tube--the Eico is not useful--this
> should be done with a mutual conductance tester.
You mean like the Eico 666 ?
John H.
Chris Haedt
Yes--The Eico model 666 is NOT a mutual conductance tester--in fact
Eico did not even make a mutual conductance tester. Certainly there
is great confusion about the 666/667 models. The Eico 666/667 are
"dynamic conductance testers"--which is a fancy way to measure
emission. It is the same system used by Jackson and Precision in
their better testers. In this test, proportional AC voltages are
applied to the tube elements, and the rectified DC is measured at the
plate of the tube. The relative merits of this test are discussed in
Alan Douglas' book. He does not think much of the system--but the
better system did tend to go along with better leakage tests and
transformers. So--testers like the Eico 666 and the Jackson 648 ARE
certainly better than some of the simpler testers like the Eico 625.
Chris Haedt
We had an excellent discussion of tube testers just last week on the
Antique Radio Forum--Test Equipment Forum. The thread was titled
"Heathkit TT-1A" and the last post was on 09-04-2003. Much of what is
being asked about here, was addressed at that time.
Chris
J. Vincent Collins
What is your opinion of the Hickok 799?
What would be a realistic price?
Vin
"arizona cowboy" <el34ISMYUSERNAMEATcomcast.net> wrote in message
news:akadnbee_q6...@comcast.com...
Chris Haedt
> Wild Bill,
>
> What is your opinion of the Hickok 799?
> What would be a realistic price?
> Vin
>
Hickok 799 something like a B&K 707 or 747--it is really unlike
anything else Hickok ever made. The 799 has lots of sockets and a
unique switch to convert each socket from I believe a red, blue, or
green coded tube type. Amp tubes are all tested as triodes. The
sample I had worked only fairly well and there was trouble with that
socket switch mentioned above. Of all the many Hickoks out
there--that would be way down on the list (last?). There are so many
better testers out there. If you like the multi-socket fast set-up
concept you would probably be better off with a B&K 747. The Hickok
799 had three or four versions, and one wonders if they ever got it
right. I was certainly not impressed with the one I had and gave it
to Wendell Hall--he wanted some parts from it. I can't give you a
fair price--as I would not buy one. Alot of the Hickok stuff that was
made towards the end was not always the best--well meaning designs
that never really got perfected. Some people might think that the
later Hickok solid state testers were better--but components such as
diodes were not so reliable in the early days. The 580A sort of
falls in this camp also. It takes some fancy footwork to get these to
work properly.
> >
J. Vincent Collins
Vin
"Chris Haedt" <ch...@airpump.com> wrote in message
news:9123781c.03091...@posting.google.com...