Electron Engine ™
Printed Circuit Boards by Emissionlabs
Tube matching with Pentode Board EE14
Though not primary intended for this, matching can be done also, with some care.
There is no analog read out for the plate current, but still the LED meters do a good job. If, for instance the red LED (for 60mA) is set such that it just begins to burn, it doesn't matter much for the matching if the current is indeed 60mA, or a certain percentage lower or higher. All we want is, all tubes must the SAME. Then, the quad finally is sweeped down from 60mA, all the way to 15mA, where only the green LED burns. During that sweep, we want to see no differences. At last, the cut off test is done, the bias is lowered, until the green LEDs are off, and what we want now, all four of them go off together. It is not allowed for one tube to stay "on" longer than the others. If the Quad passes all of this, it is well matched.
Lot size and costs.
When matching tubes, people often ask to pick a well matched pair. When I say, this costs (per tube) 12 Euro for double triodes, and 6 Euro for single tubes, they react surprized. They do not realize, I need to match 4...10 tubes to get a pair. For matching I need to burn them in first, and when I have bad luck, do all the work for 10 tubes, and get 2x 6 Euro for that. It becomes even more complicated with double triodes. If these don't match within one bulb (called unbalanced) these can not be used for matching at all. So here, having to do 10 tubes is nothing unusual. Moreover, unbalanced double triodes have to be sold as "unmatched", but cost and work for matching can not be charged on those, it is included in the 12 Euro per tube for the matching. So this is why good matching is impossible to be cheap.
Lot size and yield.
Yield is never 100%. Unless pentodes have a defect, in the end, most of them can be used for matching. There are good lots and bad lots. Do not think you can see that from internet pictures before buying. You can not see that from anything. With NOS, almost always, somebody has been picking from the tubes before, unless master cartons are sealed. Seals were not broken for no reason. Also those sellers who have 1000's of tubes per type number, and match them, it's normal they put back 100 good, but unmatchable tubes in an master carton and sell that as is. You will have a hard time with such a lot. So some lots are fine, others are just not like that. All of this, is included in the price of matching. I would say, it really pays off very much for the buyer, pay 6 Euro extra, and be sure to get the better ones.
From a GOOD NOS lot, of 100pcs, the numbers are approximately like this: One is defective. 50 years of storage does unveil some of the hidden defects. The other 99 test good, but initially unstable. So a burn in of appr. 30 minutes is needed. Now, generally you will see the following. If all tubes are more or less testing the same anyway, this is a perfect lot, and also you can say nobody has picked from the lot before.
Or, also not uncommon, they all test different. These are not bad tubes, but you will have more work with matching, and from every lot of 100, some 5..10 pcs are left, which are too far off to match them. Also such a lot needs longer burn in, which can bring some of the curious tubes back to more average values. Still some will always be left at the very high or very low end.
So, having read about the challenges, you will agree it makes really a lot of sense for the end user to have matched tubes and pay the charge for it.
With NEW made tubes, the situation is even less good. Generic quality of NOS is much higher.
Roughly it is like this:
- New made Chinese tubes have good quality, but they seem treated like old potatoes during the complete handling chain, and what they call matching is a joke. You need to remove the bad ones from the lot, and the good ones which are left, are very good ones. Be careful with unusual tubes, like 6SN7 in tennis ball glass, test data is terrible. But... the Chinese are improving, they are really getting better, these last few years.
- New made Russian tubes are handled a lot better, packed well, and branded with all those wonderful old brand names, for which they bought the rights. But electric quality is poor, and so are the data sheets. Optically they look a lot nicer as they really are. So you will remain with some very strange testing tubes, which look so nice.
How it's done.
You need a counter knob on the bias potentiometer, like pictured here. This allows to read the value of the 10-turns potentiometer from a scale of 0...100. This knob is just mounted like a normal knob. This knob is not included, but they are inexpensive. Check the axis diameter before you buy, some axis have 6mm, some have 6.3mm.
.
First number the tubes you have. Then for each tube, write the reading of the potentiometer at which each of the three LED begins to burn. Do this with only one tube inserted. So for a 60mA tube this will give three numbers. One reading at 60mA, one at 30mA, and one at 15mA. For a 30mA tube, there are only two numbers. Write those numbers in table. For a larger lot, Excel is a good idea. For a smaller lot, like 12 tubes or so, just sort it by hand
Such a table could look like this:
|
Sort by column by clicking in the arrows.Four tubes qualify for Quad#2, Five tubes qualify for Quad #1. So one can not be use for Quad#1. Number#3 seems an average kind of tube, so this is taken out, as it has best probability to be used later, for another round of matching. #6 and #7 are too far off. When you do enough matching perhaps you find matching tubes for those. Tube #12 has a bad cut off. So you see here, good matching requires many tubes, you understand now why this is not cheap. |
