ECC81 is available as Special Quality tube from Telefunken and is called ECC801S. For ECC81 there is only average data published. Reject limits have to be guessed by people who know how to do this. People who do not know how to do this, are confusing themself and others by this.

I have found sometimes ECC81 from really good German brands, which had 18mA anode current, vs. 10mA from the datasheet. Also NOS tubes which were well below 10mA, such as 7mA with fixed bias. In a way this is normal, because ECC81 is a high current, high gain tube.

Here is an important piece of electrical theory, you really MUST understand. The tolerance in fixed bias is with the Grid Voltage, and not with the Anode current. If you go into a tube factory, and make the anode distance of a tube smaller, the tube will appear to draw much higher current than typical. So something has changed with the tube. Does it have more emission? No, of course not! Emission comes from the cathode, and we changed nothing there. When investigating deeper, you will find what has changed is: The Grid Voltage, needed to get the typical anode current. For this altered tube, we could just use the curve chart of a normal tube, when we lower all the grid voltages a few percent. Like at the -5V curve we need to write -6V etc.

This may not seem much, just -6V instead of -5V, but since transconductance of that ECC81 is 5mA/V... an additional 5mA plate current will result!

Influence of mechanical tolerance on tube data. In fixed bias, ECC81 can sometimes produce up to 18mA anode current, and is just a normally good tube. The tolerance with ECC81 with very high anode current results (mainly) from two things. These are: Anode distance and grid wire distance. (each relative to the cathode). Tolerance, with unusual low anode current may in addition be caused by lower emission. It has to be clear, that any anode current deviation from average, with NEW tubes, is not so much caused by emission variation, but rather by mechanical tolerance. This tolerance has also an effect on the transconductance, but here it works in the opposite way. So when anode current is below average, transconductance is above average. If we would take an ECC81 during production process, and we would on purpose make the anode distance a little bit too wide, we would see anode current is below average, and transconductance is above average. From this you can already see, that neither anode current or transconductance as such, can be used as indicator of the emission. Rather BOTH these numbers must be judged together. (Please note, transconductance can only be compared with average tube data, when you set the tube for average current indeed)

To offer tighter higher tolerance, ECC801S was introduced.

ECC801S is a frame grid tube. Frame grids have several advantages, and one of them is, a much more constant distance of the grid wire to the cathode. So by nature, ECC801S will have lower tolerance. This is reflected by the ECC801S datasheet of course. In case it is not, this tube is no real ECC801S. Anyone can stamp this number on a tube. However, conformity with the original results not from gold anode pins, but from complying with the original Telefunken ECC801S datasheet in detail, and most specially at the points where ECC801S is different from ECC81. I think it is a real shame, today so called ECC801S tubes are sold, with datasheets of ECC81, and in fact that's what they are.

Tolerance once more. So tolerance of ECC81 is unspecified, unknown, just as you like. When looking at ECC801S, this tube is specified in detail, but only in auto bias. Now, in Auto bias, anode current has less variation than in fixed bias. Yet for ECC801S, we find there 7...14mA (in auto bias!). So a very wide variation, although this is in auto bias. This means, in fixed bias, values much wider than 7...14mA can be expected. This again justifies the existence of tubes like ECC801S, which deals with this problem of very wide variation.

Most of all ECC801S has a difference of max 15% between two triodes in one bulb in AUTO BIAS. This means, differences of ECC81 in Auto bias are higher than that, lets say maximum 25% , and in fixed bias a lot more. Due to the high gain, we must talk about 50% then. So an ECC81 with 8mA and 12 mA in one bulb, fixed bias, is not an average tube, but there is nothing wrong with it, and there is no quality rule, saying the more . Indeed such tubes can be found, also NOS by Telefunken.

BURN IN of this tube. This may be necessary for NOS ECC81, because this is a high gain tube and at the same time high current, which requires good a condition cathode, to perform up to the specs. For ECC81 which have been stored for decades, usually do not have this condition anymore, and a burn in is needed. This can take 10...20 hours or more, when the tube is really not nice condition. Or, it can take only one hour when the tube is almost fine to begin with. Also it can be observed one system tests 120% and the other 80%. Yet burn in will often bring both systems back to 100% or close. It is dangerous to burn in the tube with the adjustable bias tube, because any high gain tubes can l react suddenly to burn it, but the moment this happens can take hours to come. If you would not be checking the tube continuously, it can happen suddenly the anode current grows, and the tube becomes over heated. For this reason, UNOBSERVED burn in should only ne done with the auto bias card. Gain of this tube is too high for the adjustable bias card, to let it run unattended for hours. It seems ECC81 benefits a lot from burn in, also Telefunken. Wheras ECC801S generally comes out of the box with relatively good data, also without burn in.

Anode current at burn in. For burn in, both anode plugs must be inserted, but the meter must set to the "ideal" current in this condition. This done by reduce the anode voltage. For burn in, by my experiments, full dissipation seems not needed. So like this, current through each tube half is only half of maximum. So even If strong imbalance would occur during burn in, no tube half can get over heated.