Portrait of a Tube

Description of 5963

5963 tube


Link to some data sheets:

12AU7 General Electric

Many tubes exist with similar data like 6SN7, the mother of many famous small signal double triodes. 6SN7 is a universal tube, meaning it is not made for something specific. So 6SN7 has good parameters for everything, but none is at trimmed at it's maximum on the cost of others. So gain is a fair 20x, and this allows the plate impedance to be reasonable low. So what ever parameter you look at, it is reasonable and useful.

Then, for miniaturization, 6SN7 was put in a noval socket, and that is the 6CG7. Due to the smaller size of 6CG7, I would say 6SN7 should be a bit more rugged.

At the same time, the series 12AX7, 12AU7 and 12AT7 was designed, of which 12AU7 is relatively close to 6SN7, l but not identical.

In the age of tube computers, there was an great need for tubes which can be used in the cut-off condition for a longer period. (So the heater is on, but there is no plate current flowing). We need to stand still a little bit with this. With normal tubes (not made for this requirement....) the cathode looses some of it's emissive, pure Barium surface when it is glowing, and this is regenerated from within by an Electro chemical process, called electrolysis. This converts the Barium Oxide from within, into Metallic Barium and Oxygen gas. (Yes, gas). Any openings in the cathode metallic layer get closed by this process. However electrolysis requires electrical current, so anode to cathode current , and here comes the problem when a tube is cut off for weeks or months. The metallic Barium will disappear, and when you would decide to switch the tube on suddenly after three months, the cathode is weakened. This will recover, but not immediately, when you want to use it. Computer tubes (we better call them relay replacement tubes) have a double layer cathode, and this layer is made such that it can maintain it's metallic Barium layer without electrolysis for a longer time. (not for ever, but much longer). 5963 is such a tube.

When talking previously about universal tubes, there is a thing with 5963 which is not universal, it is the very long cut off period which it can withstand. Also 5963 can go down no a possible low voltage if under current, which is simply convenient as relay replacement, and it will outdo 6SN7 or 12AU7 (ECC82) here. Logically this can only be achieved on the cost of other parameters. There is no lab report, saying how the engineers did that, any why, but we can look at the 5963 data sheet and find some details in there. So it appears better made for low voltage use, and consequently not so ideal for high voltage use

Here is a small lesson in reading data sheets. I used the RCA data sheet for this.

With data sheets, first things come always first.

It says all on top: 'For on-off control applications, involving long periods of operation under cutoff conditions'.

Second thing which draws my attention is the first line of data. Tolerance on heater voltage is 10% vs 5% for commercial tubes. This is another virtue of the double cathode layer.

Let's go just through the data sheet.

Microphonics: Untested. So what that means, microphonics doesn't have to be bad as an average, but there is no lower limit.

Next come specifications for capacitance, now that is important for fast switching, so logically that information comes right at the beginning.

Then comes characteristical use as a class A amplifier. Notice, the grid voltage of 0 Volt. This is not an audio application yet.

Then come a lot of mechanical data. So when you put them in a computer you need them by the 1000's and the wanted no disaster at RCA with wrong glass dimensions.


What do we read here:

How to use 5963 as frequency divider, or more on-off applications. It gives all the maximum ratings here.

What you see now, the 5963 MAY replace 12AU7 /ECC82 in probably the most applications, but maximum limits are less good with 5963. However as that table says clearly, that is for digital applications. If you would compromise on the speed, or extreme long cut off, you would be able to exceed those specifications. Only, logical RCA is not saying what happens if you do not stick by the intended use.

+ + + +

Can this tube replace ECC82?

Probably yes, but not guaranteed, as with some of the ECC82 Upgrade tubes (which is something totally different). Some of the upgrade tubes are:

  • 8514 = 12AU7 with better grid quality, and tested for lower grid leakage, etc-
  • 6189WA = American Version of ECC802S, just not as tight specified as ECC802S, but a much better tube than 12AU7.
  • ECC802S. This tube can do anything. So it's long life, can replace ECC82, long cut off, etc. Anything, you name it, and ECC802S has it.

Here are some of the requirements I know: Comparing the TUNG-SOL historical data sheet with the SYLVANIA 5963. The temperature inside such tube computers was very high, and even the ambient temperature was a LOT higher than for a household radio or TV. Such a computer room was heated by thousands of tubes. To have a number that you can verify always, the glass bulb temperature of the 5963 is specified (at 120°) vs. 'unspecified' for 12AU7 / ECC82. The defect rate of a lot of -say- 1000 tubes should be as low as can be, and not just initially but during 10.000 hours of service. It is really a pain to exchange single tubes, and control them with a tube tester, etc. So the idea is, such a tube runs extremely reliable during it's specified lifetime, and after that you replace the whole batch. So along with a digital tube computer comes a storage room, with a few thousand spare tubes. This makes us hopeful, new stocks will pop up every now and then :) Higher peak current of the 5963, 100mA vs 60mA for 12AU7 / ECC82. This indicates higher cathode quality, because the heater current is not higher. Note, a higher heater current, would also allow more peak current, but logically with all that heat inside a computer, that was no option. The real solution was: USE BETTER cathodes. 5963 sure is the winner here. Less grid leakage, which reflects in the possibility to use 500K grid resistor for the 5963 vs 250k for the 12AU7, in fixed bias that is, which is always harder. Specified grid dissipation for the 5963. vs. "your own problem" with 12AU7. Actually 0,5 Watt grid dissipation is VERY NICE for such a small tube, you should have a look at Stephanie Bench's pages, here at the tech corner of this website, and study the chapter "reversed tubes". That is a way to use the grid as an anode, and use the anode to control it. Maximum plate 300V for 12AU7, vs 250V for 5963. This is negative for the 5963, but when you give in on the grid dissipation which you do not need for hi-fi, the 100mA peak that you do not need, the Bulb temperature of 120°C that you probably don't have, etc, somewhere above 250V is possible, but how far above I can't say for sure here. Cut off of 5963 is better. Whereas that is not beneficial for HiFi, this is no disadvantage. G1 to G1 capacitance is specified for 5963, but that doesn't means this is better as 12AU7, but it is specified. Implied by the use for digital systems, the tube can be in stand by mode for a long time. Other than 12AU7 which can NOT be used in stand by, or you loose cathode quality sooner as you like! Last but not least, look inside... you see the long plate system similar to Telefunken ECC802S. That explains better what pearls you have here. Don't bother to compare it with Telefunken ECC802S, because THAT data sheet is unbeaten, but for the price of the 5963, this tube is sure a good alternative. Conclusion: is 5963 a better tube than 12AU7? I think yes. Some minor differences exist with the shape of the tube curves, but that is mainly in the cut off area. In the normal (Single Ended) 8.5mA area, the curves look the same. The difference seems to come from another (better) cathode technology, combined with a larger anode surface for better heat disposal. Same situation as with TELEFUNKEN ECC802S. Link to Sylvania 5863 data sheet Link to Tung-Sol 12AU7 data sheet (please send me the Sylvania 12AU7 data sheet if you can find it)

Replace ECC82 / 12AU7 by 5963

This will be possible most of the time, as long as the 12AU7 / ECC82 is not used at extremely high voltage, which is never a good idea anyway. The tube curves are a bit different in general, BUT they do match quite well in the typical bias point of 12AU7 / ECC82.

So yes, you will often be able to do this replacement. In cases where it will not work, you may see higher microphonics, or another bias. This does not have to be so. You just need to try it out, and many people are using 5963 successfully. Most of all this will be a very long life tube.