[Phlogiston]

Or any 2 - 2.5 volt common cathode power tube.

Have you ever put a 45 or a 2A3 into your 533, 600A, TV7, checked for shorts (ALWAYS), pushed the test button, only to watch the needle go up and then start to drop?

This can be a bad sign on small signal tubs on a tester like a I-177. It MAY be putting too much signal (,5Volts AC) onto the grid, BUT that is not what I am talking about here.

I have heard several people say that they had discarded tubes like the 45 (245, 345) for such poor test results. Then wished that they had not.

First lets start with the tester cal. When you are setting the filament voltage on a lot of testers, you have to aim at 6.3V and up to 12V as accurately as possible since these are the most common tubes that you will test. However, many testers will have a significant error towards the upper and lower ends of the filament voltage spectrum. This is usually ok on the top end since the error is less significant in heating the filament to a valid test point and those 50V tubes tend to draw a lot less current.

On the low end, 1V, 1.5V, 2V, 2.5V the tester may be a bit low BUT the relatively high current consumption is going to cause the filament voltage to sag even more. On my 533, testing at the pins of a 45 gives me about 2.2V. AND when you push the test button the filament voltage sags even further -  this is assuming that you 0ed the line to start with. SO that 2.5V tube sees the filament voltage drop further to around 2.1V.

Now, I have heard, and it appears to be somewhat true, that tube testers were built to accommodate this issue and I see that even with a reduced plate voltage the bias is lower than would be expected in circuit.

If you test a brand-new 45 tube under these conditions there is no problem. If you test a well used tube there are issues. The used tube may have more problem with the reduced plate voltage in the tester and specifically with the reduced filament voltage.

How can I easily validate this? Well, a tester like the 533 has a "life test" that duplicates these conditions. Problem is that you are doing the life test without engaging that function. AND as the filament voltage drops further under test the output of the tube drops further which is observed by the dropping meter.

So what to do? I have marked the line setting in my 533 for 2.5V test value which is about 2 marks higher than 0(ed). This is not alone the solution. Instead what I am doing is the "Life test". A new tube is going to perform well at the reduced filament voltage and there will be little gain (possibly a loss) in the mho reading as the filament is adjusted to 2.5 V (which also raises the other test voltages). A well used tube will instead spring to life with the increase in filament voltage. A used-up tube will stay low with no or little improvement - it is still bad.

So what I am suggesting is creating a valid "Life Test" for these tubes before you decide to throw one away. A tube that tests well at 2.1V probably does not need to be tested at increased filament voltage. But you may want to save an expensive tube by retesting a tube that gives you 1000/1850 at 2.1V. This tube may give you 1850 at 2.5V and will work just fine in an old radio. The "bad" one will still test bad. You may also find that a filament has gone to sleep at a lower filament voltage and after testing at 2.5V the test results may improve at 2.1V. Do not confuse this with thermal inertia - let the tube cool a bit down to the 2.1V temperature before claiming improvement.

Do not test the tubes above 2.5V (filament).

[morzh]

I have this with many tubes on my 600A, when I push the P4 and see the needle going up and then slowly dropping a bit. Not too much but it will drop. I consider this "good".

[Phlogiston]

Well I was working with the HV issues on both Stratospheres, testing 16 #45 tubes and it occurred to me to write this down for the benefit of people here.

[klondike98]

Thanks Russ!

[TV MAN]

Thanks Russ! I have observed this myself, but being a 'radio packrat' I don't throw tubes away. Some pretty weak ones still sound decent enough, and since there will be "no more" of most of our old radio tubes, I sure dont want to see any usable tubes trashed. I don't notice this as much on the Precise 111 since it uses separate filament and power transformers but note it more on the Hickok 600a and my old Eico. I really appreciate your input 

[Phlogiston]

Yes, and I think you could throw the TV-7 in with those testers. I have an A/U on the bench that I almost never use - only when I am curious about the test values given on a tube I have purchased with test numbers from the TV-7.

This test also applies to #26 tubes  but to a lesser extent - they don't load the PS down quite as much.

[Arran]

Russ;
 I remember bringing this issue up before when Jayce in Ohio mentioned running into many bad testing #26s and #45s on his tester. I can't speak for the early to mid 1930s tube testers, as I do not have one, but with just about any tester from the 1940s onward the heater/filament current drawing down the filament voltage can be an issue. I think this is because the designers of the testers assumed that when one was going to be testing a tube with a 1.5 volt filament that it was going to be something along the lines of a 1A7 or a 1C5, not a #26 tube, and the same thinking seems to have gone on when it came to the 2-2.5 volt tubes, that someone would be testing a #33 or a #34 and not a #45 or a #47.
  One my Stark/Marconi tester what I have to do on #45s, and other 2.5 volt AC tubes, is set the tester to the 3 volt range and then adjust the rheostat with a voltmeter using the pins in another tube socket, for some reason this does not work on a Triplet I have with the tube in place. On #26s I have to set the tester on 2 or 2.5 volts and do the same, as a #26 draws about 1.2 amps of current verses milliamps of current like a battery tube would.
   One other thing to watch for on early tubes, whether it be an AC or battery type, is bad solder joints on the heater pins, grid pins, or grid caps, all of those can cause the tube to test poorly if at all, this includes '01As.
Regards
Arran

[Phlogiston]

Also, yes. I just got a bunch of 26s. I had to resolder about half of the pins.

[Phlogiston]

I have been inventorying some more 45 tubes and have come up with a more easily followed test using my Hickok 533, but, this should apply to most Hickoks of that vintage, 534, 600 and so on.

First, again, the 45 draws over an amp at 2.5 volts causing a sag in the filament voltage which is exacerbated by the additional draw in the "test" condition. The filament voltage drops below 2V. The more "used" a tube is, the worse this is for testing (which makes sense).

I was using the "line" adjustment to raise the filament to 2.5V. Problem is that it also raises the plate and bias voltage. They do seem to track so the tube would not be damaged but the test results are not going to track with the chart.

I found that setting the filament at 3V raised the applied voltage (AC) to 2.5V when the line was set (to the center of the meter). A prolonged "test" will still cause the meter to drop since the load on the PS is still increased. SO, keep your eye on the meter and record the measurement in the first second or so. Results testing 52 used 45 tubes were from 1200gm to 1750gm, probably still a little conservative but inline with expected results.

Setting the filament at 3V (centered meter) gives a plate voltage of 120V and a bias of -20V which is about right (21V would be better but I used the spec. setting of 61. If you really want to be accurate, increase the bias to 63 or what ever gives you the proper bias in relation to plate voltage).