A note from de Department of misstakes
Please understand what I write here is my personal opinion only. There is no special intend to write negative about companies products, but we do need to call things by their name.
Software Error #1
In both testers, the heater voltage is connected with one end to ground, and the other end is heated with DC. This is for practical reasons, and I would have done so the same way. You will say, ok so what? For indirectly heated tubes that is no problem, but this gives a real issue when testing directly heated tubes (DHT). The reason for this is, with DHT, the cathode's electrical reference is by definition the center of the filament, and really not one of the two ends, to your own free choice. There is little understanding about this, and it is amazing how even professionals are totally surprised when you bring up the subject, and even start to argue about it.
Now you can connect a tube like 300B with one end of the heater to ground, and the other to +5V. Sure that works, but you might just as well connect that other end to -5V. Still as you can understand +5V or -5V will result in another grid voltage needed for the same plate current. So perhaps if you thought this makes "no difference", well here you have already a difference since that needs another grid voltage. There is this rule saying the cathode of a DHT is the center of the heater, and there is no rule you can take either end to you free choice. There is no outside world connection to the center of the heater, but inside the glass bulb it is physically present indeed. Just not connected. So if you connect the heater with one end to +5V, the physical center of the heater is at +2.5V now. You have lifted up the cathode reference point by by half of the heater voltage. Not much of an issue when the heater center is at +2.5V and the grid at -60V, and you have a low gain tube such as a 300B. The error is 2.5 Volt only and you would not even notice at this low gain. However it is a major issue with a tubes like RE134 or the Emission Labs 30A, that work at low grid voltage and have high gain. The 30A has -3 Volt grid voltage, a 5Volt heater and a gain of 30x. You can not just allow an error of 2.5Volts on a grid of -3Volts. To put this in a numeric example: Suppose you tell the tester to apply -3V to the grid, that's what it does. Then to the heater is applies +5 Volts. So the center of the heater is now at +2.5 Volts and the Grid at -3Volts. The effective heater voltage is in reality -5.5 Volts and not -3Volts. All of this would be no problem, BUT..... the software still thinks it's -3Volts Grid voltage. (This is the ERROR we talk about) So when the test results are printed, the tube had EFFECTIVELY -5.5 Volts applied to it but it says -3Volts everywhere. The Amplitrex AT1000 (and the Sofia) make the same mistake. The fact that doesn't have to be that way, is proven by the Russian L3-3 (analog) tester which lets you even choose between AC and DC heating, and does NOT make this mistake. Test results are identical.
Now I heard someone call the "one end grounded grid" another method, that Jac doesn't agree with. This is an excuse. I see no problem with the one end grounded grid at all, but I see a problem with SOFTWARE ignoring the effects of this. Let me explain the error here, by using the 211 tube curves. The 211 can have some limited positive grid voltage, but let's look at the 0 Volts line first. You see it starts exactly where the two axis cross. This is important to notice. This how you recognize this (only) tube curve without grid influence, or in other words with 0 Volts grid. Here, for the 211 tube, you see the -20Volt line starts NOT at the graph corner. There is no Plate Current below +200 Volt. So if you need a certain plate voltage before something happens at all, this means a negative grid voltage was applied. Even so at positive grid voltage (which the AT1000 and Sofia can not do) the effect reverses. The curves of +40V and -40V grid look like mirrored. Only at 0Volts these "dead" zones fail.
Here comes the issue with the AT1000 and the Sofia, because with a 211 the cathode center t would be at +5V, the way this tester is constructed. Now, with a 211 the effect in reality will be small since +5V is not a big error on a swing of -200V as you can see in the curves. I just picked this tube to explain the curves and have a look at what positive grid voltage and negative grid voltage do. Also to show the 0Volt line is unique in character, and it can be recognized clearly.
However real problems come up when we test a tube like for instance RE134. This tube has 4Volts heater and works at much lower bias voltage, so suddenly it makes a great difference if the cathode is grounded by one end. As can be seen in the previous explanations, the cathode center is now lifted up to + 2Volts. Suppose now you apply -2Volts to the grid, the cathode-to-grid voltage is effectively -4Volts.
Lets bring this to the point: By applying -2Volts to the grid, the tester is going to plot the -4Volts line. (This is one of the curves in the chart here below). Now, a problem comes when you plot those curves with the Amplitrex AT100 (or the Sofia). Suppose you would be able to test the same tube, as you see here below, also in steps of 4V. Then you would get the same curves as you see here below, right? BUT..... where it says -4V, the Amplitrex (and Sofia) prints -2V. So the curves are the same shape, but the problem is, there is "-2V" printed at the -4V line, "-6V" printed at the -8V line, etc. The whole chart is given the wring voltages. And NO this is not just a matter how you see it, because when you print the 0V line, it is in reality the -2V line, which you will see because it starts at +15V plate, and not at 0V plate, as can be seen from those curves.