How to Test Electron Tubes
Contents
Working principles for tube testing.
Preface
I get many requests from people who want the "best" testing for their tubes. This is normal of course, and they have some ideas of what that means. This article can not quickly answer the many questions people have, but it does try to give some insight in tube testing, mostly for yourself. Do not focus on "nice to have" information, but require what is REALLY important. I will explain how to test in a professional, and meaningful way. I will also what kind of testing is useless. For those who own a tubes tester, please understand, there is a lot of nonsense information produced by amateurs and also by producers of new made tube testers. Some of which use a very low level concept, hiding this behind a nice digital display which at a first glance looks well done, and very technical. One of the most amazing things is also the misinterpretation of commercial data sheets by self claimed experts. So please don't be angry. if your tube tester does not get many positive points here, and if the test method it uses will be called unprofessional. Before we go into that, let's first look at what people are expecting.
Some requirements I was given
- Expectations by people who want 'the best'. They tell me: Please select for the best gain, highest transconductance, and lowest hum. Or re-select the selected and pre-packed tubes, to get them "extra selected". They seem to think this is little work, and costs almost nothing extra. What they do not accept, is when I must pick out the better tubes from a lot, they pay not only for the work of the picking, but also for getting the better tubes. Like when buying a used car, you may only be interested in one with shiny paint and new looking seats. For the seller it is little work to select one like that for you. However the price difference is not just the working time for this selection, you will agree. So why should that be different with tubes?
- Expectations by people, using a low level tube tester. These are difficult buyers, when they do not realize they have only a low level tube tester. They think the tester is great, because it matches their experience and needs. They will not understand why such a nice looking device, for what they payed a lot of money, is called pretty useless by me. They regard this an excuse, because even when the tester is not ultimate, at least it will produce useful information. Sometimes I have to say: No, not really.
What is a must and what is a want. The situation is, you "want" something, and to achieve this, you define a "must be" specification yourself. What most people want is of course good sound. Unless you are talking about electric guitar amplifiers, which are tube destructors. In many cases, the "want" is described directly as it is.
So what do we want? People say they want warm sound, because nobody wants cold sound. It should sound holographic instead of diffused. Bass must be tight instead of muddy. The highs must be clear but not sharp. And before we forget to mention it: The mids must be well defined, please no microphonics or hum. This are prepared to pay 10...15% more for this.. In the imagination of some buyers, all we needs to do is, use a tubes tester, and quickly pick such tubes from the lot, and sell the rest to others.
You already get it... things did not work this way. So why demand that? I don't know why buyers believe in that. Better is to divide your requirement, in must and wants. So do not want things which you made up yourself. I have sometimes tube lots which are to PERFECT quality, I can take two random tubes, and they are always a fine matched pair. Such as the Sylvania 7N7 (which can replace 6SN7 if you use an adapter). Costs for matching is a waste of money.
At the same time, with new production tubes, it does not work to buy 100 random tubes, and after putting them on a Hickok tester, we end up with 25 fine quads. Why that is impossible, read the next part for that....
Definition of terms
Quality means there is guaranteed data.
Reliability describes the chance that the quality is maintained during the Lifetime.
Lifetime is not easy to specify, sometimes a minimum lifetime is specified, but this is talk only, if the manufacturer has no instructions how to measure this. So lifetime in hours, you will hardly see. The most practical hint is to check the guarantee period for this. Really, what lifetime do you expect, if the manufacturer will not give more guarantee than 3 months, some give 6 months, some 1year, and from what I know, only Emissionlabs and Western Electric give 5years lifetime guarantee.
Avarage Data. This data is not guaranteed. Meaning if a tube has deviation, it is called a good tube, there is nothing wrong with it.
Maximum or minimum values. This may as well be a number which can eb used for incoming inspection, or a number which defines end of life. It should be clear, these numbers are only valid with the test circuit, belonging to it.
Conclusion. If a tube is specified as "low noise" but has more noise as average, it has a quality issue. If it is the noisiest of all, but it was not a low noise tube, this is not a quality issue. Quality means only, is the product "as specified", or not. If yes, quality is 100% . If a tube passed quality test at incoming inspection but doesn't pass any more later, though still in the life time, it has not a quality issue, but a reliability issue.
Working principles for tube testing
People have sometimes the idea of exactly measuring electric data from tubes, with highest possible precision, as if that precision is the key to good sound. Such an approach is not totally wrong, but it does belongs in the category "nice to have". This can be interesting, when you have the time and the money, and the engineering knowledge to understand what tester you need for this, how to obtain one that is in good condition, how to use it, and how judge the data it gives. Like this, it can be a great hobby. Lets put this approach at the one end.
At the other end, we see extreme basic approach. Meaning you have a tube of unknown condition, some equipment you want to use, and a waste basket. The decision is, to throw the used tubes away, or still use them. There is no interest in the "how and why" of tube testers. Like an Ebay seller, saying the tube tests "strong" on whatever tester he uses. This approach is all on the other end.
Then, everybody is somewhere positioned in between these two extreme positions, let's say in the middle. In this article I try to explain why it is important to set the expectation right. So choose a kind of tube tester that it matches your personal needs and capabilities.
Not all possible working principles are listed here, but these are the most occurring:
Gm testers, without control of the working point. This approach we find with most Hickok testers. Test conditions, you may or may not be able to extract this from the tester. How testing is done, is unclear, and you can't compare data with the data sheet. Just some faint stories about the great patent. Calibration procedure is vague, and many other things too. Yet the idea is: When the tester itself of good, it produces a reliabe result. A good Hickok is as reliable as the Funke W19, but they are useually in not so great condition as you would like. .
Gm testers, with user control of the working point. In this category, keep apart the AC testers from the DC testers. An AC tester simulates the DC test, though latest generation testers can do so really nice. It stays however a simulation, and precision is limited. Such testers are AVO Mk2, Mk3, M4. Later model Taylor tester, the polish Elpo, some of the Metrix, and the Central 752. AVO VCM 163 is the top product of those, but unfortunately hard to find now. Sure I forgot many here, and this list is not intended to be complete.
Emission Testers. These intend not to give you tube data, but only let you know if the tube is weak or strong. You are leaving this decision up to the tube tester, because you have no interest in data, and you need to get forward with repair and your work. Professional testers of this kind, will do this job probably better than you can do it yourself.
DC testers, without Gm test possibility. This approach we find with the Metrix 61 Series, Neuberger 370, 37 may other testers too. The advantage with those is, you can compare the results to the avarage data sheet values. At the same time, this is the weakness, as avarage data sheet values are no reference, and besides they do not use the recommended test circuit. So you end up here, with precise data, for which you have no use, leading to false expecatations, and self made quality criteria, which others may not accept.
DC testers, with Gm test possibility. There are ideal, and a LOT Better as the above. Now Gm can be tested at a specified DC current, for which you should take the avarage value. And no, it is dead wrong to set the tube for the avarage grid voltage, and then wonder or complain, the plate current is not avarage. Why should it?
Instead, Gm should be close to the datasheet plate current, because Gm directly corresponds to that value. At the same time, this provides you with the grid voltage deviation from avarage. Which number can be used to calculate the emission.
Digital Impulse testers. The makers of those don't like this designation, and the users are not always aware of the working principle. These test the tube with a very short impulse, in which current and voltage is measured. This allows a very tiny power supply, since power is supplied from a charged capacitor. By nature, such testers can not directly measure Gm, but must calculate it from the difference of some measurements. Issues with those are, the failing heart development in he tube. Defects or leakage, related to heat development can not be found. Also leaving the anode cold, will give an error with the result. This error is never zero.
Digital full power testers. These are very rare, but they do exist.
Patented Testers. These have all one thing in common. They replace classic, straight forward testing circuitry, by something 'patended', and less costly. These are tools, and not instruments.