Random tube matching nugget

[km6xz]

Regarding my suspicion that the 'good old days' of tubes weren't that good, at least in regard of consumer grade tube variance, I just happened across RCA HB3 Tube Ratings http://www.tubebooks.org/tubedata/HB...be_Ratings.pdf
The 'range of characteristics' section p13 advises designers to allow for a variation of up to +/-30% of any of the average characteristic value shown for the type, though this includes ageing / wear.

A designer was being advised to consider 30% for circuit +tube aging, that is very different than tube incoming spec tolerance. If a product had a design intent of 5 years between overhaul, and what would they need as starting deviation? That advisory has little to do with this discussion, particularly when the standard resistors of the day for consumer products was 20%.
The tube aging charts were used to set maintenance intervals on industrial and military equipment PM and overhaul.
I never remember having unbalanced operation when plugging in a new tube in place of one of a quad when the majors were still making tubes. Have a bad tube...replace it and if the others are ok they will all match well enough, surely far better than required for guitar amps. TVs had much more critical circuits and they seldom needed shop alignment when replacing a tube. Communications gear was even more critical in tolerances yet realignment was seldom needed except in final tanks adjustments or neutralization touch-up when changing tubes.

[km6xz]

Tube data sheets from the USSR period are of no use in spec'ing tubes that are packed with. The 3:1 figure is a design max versus rated spec. The highest voltage tolerance of current 6L6's is the rebranded Soviet era surplus 6П3с-е and it is packed with a paper spec sheet that rates it as max anode potential of 250 volts. So designers knew the spec and what they could use them for were far over official levels. There were many factories churning out the same design tube and that all have to be plug replacement. The factories did not design the tubes, that was done at a research center called a design bureau. A few factories were also design bureaus such as Svetlana since they were the largest and oldest of any tube company. The competition between factories was in reliability tolerance and production efficiency. By setting the specs incredibly low, any factory could meet them in their evaluations but since there was lots of staff for QC, (employing people was more important than making a "profit"), every tube was tested and retested so what came out of the factories was usually very good indeed. There was constant process development so they improved as time went on.

There is no question that the variations in tubes now is bases on crude assembly methods that rely more on hand labor and less on precision jigs. IF someone was interested in building tubes with much better performance, longer life and resistance to abuse, it would be easy with the materials advances, precisions robots, and photo etching processes that are in common, routine service in every other area of electronics and metallurgy now. No question. But there is no market for modern tubes. All uses of tubes are less than 4 million annually so there is no incentive to invest in the production robots for just 1 month of output for one production line. Maybe when 3d Printing develops more people could make their own metal tubes using deposit type printing. modern lithography as used in IC manufacturing could make tubes that were the size of ICs and precision of ICs but used thermionic emission, possibly as not much above room temperature.

[Randall]

To chime back in here... one supplier not mentioned here is Valve Queen out of London, Ohio. She matches at what voltage you need and burns in for 72 hours. I bought a quad of Ruby KT66s matched at 500v for a problematic amp that came with a 6 month warranty. Can't beat that. So far so good.

[pdf64]

A designer was being advised to consider 30% for circuit +tube aging, that is very different than tube incoming spec tolerance...

I never remember having unbalanced operation when plugging in a new tube in place of one of a quad when the majors were still making tubes. Have a bad tube...replace it and if the others are ok they will all match well enough, surely far better than required for guitar amps.

Here's the KT66 info http://www.mif.pg.gda.pl/homepages/f...k/KT66_GEC.pdf
P3 notes that it's considered to have reached the end of its useful life when the power output has halved / gm fallen to <5.5mA/V
As the initial bogey gm was 7, that's only down 22%!

BTW, is CP1005, the British Standards Code of Practice for Use of Electronic Valves, cited in the above KT66 info, available on the internet?

When I started tinkering with guitar amps back in the mid 80s, GE 6L6GC and Siemens EL34 were readily available at reasonable cost from Raedek in the UK.
But I remember having terrible trouble with trying to set a suitable bias if I only bought enough for the particular amp, eg some would run at ~10mA, others at 50mA cathode / plate current, and I'd have to bodge in individual bias trimmers for each power tube to get it to work ok.
I found that it worked out a lot better to try to buy a few extra tubes each time, in order to get a reasonable balance, eg buy 6 to get a reasonably matched quad.
But even though the prices then look cheap now, musos have always been skint and want the re-tube as cheap as possible, and no one wanted to eat the cost for the outlier tubes that won't match.

Apologies to anyone faced with my bias trimmer mods when they open up a classic amp

[martin manning]

Y

Interesting discussion. I also recall reading that tubes testing at 70% of spec are to be considered at end-of-life. Pete, perhaps you were getting seconds or returns back in the day?

[pdf64]

I guess that's possible though suspect not; once biased ok, the strug tubes generally gave the expected power output and worked reliably.
I'm still keep a couple of those GE 6L6 that I bought back then (which were probably strugs from different orders) 'for best', as they sound better and put out more power than any current production that I've tried.

If, given identical electrode voltages, the plate current of different individual power tubes of the same type in fixed bias didn't vary unacceptibly, why would MO / GEC advise for each KTXX to have its own bias trimmer with a +/-25% range?
And the popularity of tube matching services have come to being?

[Old Tele man]

Sometimes (often?) one has to *do* what is necessary rather than what is desirable (wink,wink)!

[Old Tele man]

Here's the KT66 info http://www.mif.pg.gda.pl/homepages/f...k/KT66_GEC.pdf

BTW, is CP1005, the British Standards Code of Practice for Use of Electronic Valves, cited in the above KT66 info, available on the internet?

I believe these are 1967 & 1969 India versions of that British 'Code of Practice' CP 1005:

• 1967: https://ia601005.us.archive.org/21/i...597.2.1967.pdf

• 1969: https://ia801003.us.archive.org/17/i...597.3.1969.pdf

or possibly try this 1962 German source: http://m.freestd.us/soft1/559294.htm

[R.G.]

Interesting concept - components matched to 30% at end of life.

That's an incredible degree of **low** spread for un-selected parts.

A good semiconductor designer will expect 200-500% variation in parameters for bipolars, larger for FETs unless there were pre-selection and matching done. The change over from tubes to transistors essentially obsoleted the practicing *generation* of EEs at the time, and little things like this are why. A new generation of EEs had to find out how to and be trained to create designs with predictable results with such wide variations in parts.

[Rob's Radio-Active]

It consistently recurs to me that it is now possible to build much better and more consistent tubes than were ever built in the Golden Age of Vacuum, and most likely far cheaper than they are built in today's parasitic manufacturing market.

As an example, NASA and contractors are reviving the F1 rocket engine from the Saturn V. The F1 is the most powerful liquid fueled rocket engine ever built. Ever. The plans were not lost, as was thought, just stored so deeply that they could not be retrieved easily. But it's being re-engineered to modern manufacturing practices to eliminate the incredible amount of hand welding that was in the first ones.

Today, we have incredibly accurate and fast machines for making widgets quickly and reliably, to much finer tolerances that were ever possible with fixed tooling. We know more about materials. The functionality of a tube is almost purely a function of three things: the mechanical assembly accuracy of the internal parts placement, the vacuum chemistry of the materials, and the degree of vacuum. We can do all of these to a much higher degree than was ever possible before.

What we don't have is the economic will to go do it. The folks who could do this can make far more money making higher-profit-margin and higher-volume widgets for other purposes.

It's a pity.

As a (former) Manufacturing Engineer with a B.S. in Industrial Technology from a respected school of engineering, I agree.
The business model today, at least the American business model and those overseas firms that copy it, is only for the short term perception of gain--even if the company "saves" itself right out of business in a few years. Just look how much money is wasted on not investing money in employee training or incentives. It seems few believe in their own business enough to invest in it...calling it "spending" instead. They take their short term profits and invest in other people's companies, apparently thinking they are run better than their own. I see no logic in it.

Cheers,
Rob

[R.G.]

I think I know the business model - fear of change. Businesses even more than individuals live on guessing the future, as they are required to invest - that is, bet big chunks of money that future sales will actually happen - to survive at all.

So businesses are fundamentally risk averse. Start-up businesses often are not, as they have no clue what will work, and simply make their bets blind. The roaring startup that eats up the market is the distracting exception, where a blind bet happened to be a really good one. In general, the vast majority of new businesses are out of business in a year. We only notice the survivors.

Established businesses can either hunker down and live only on their existing business and products, so they stop growing except by accretion and luck, or set up small betting sand boxes to try out new avenues of wagering.

But it's at the interaction of the high level employees and the business that you get the fear of change. The high level employees in big businesses are generally the ones who have noted that it's easier for them to get promoted and make more money by looking good for a moment than by consistent performance reflected in the business. So they sabotage their fellows to look good. And as they near the top, they can't do that any more, so they sacrifice anything for the appearance of short term "wins". The lowest risk of short term wins is parasitism of the future - not investing in employees, not investing in infrastructure, MBA cream-skimming, and other things that make the stock look attractive in the market.

The winnowing process that gets upper management ahead has the side effect of eliminating most of the people who actually have the talent to grow businesses.

The lure of being able to dip into and out of other companies in the stock market is almost unavoidable. The hope is to grab a hot new startup, or get get into and out of another business in the days, hours, or minutes that a stock uptick will take. It's far less of a risk than investing in the actual company, and easier to get out of.

As a sanity check, though, it is very likely that we have just lived through what will be called something like the Second Industrial Revolution. The last 50-75 years saw the rise of computers in everything. That's an incredibly profound change.

[Rob's Radio-Active]

Well for what it's worth, I do buy matched power tubes and, for an amp restoration or repair anyway, usually retest them and write the actual micromohs and conductance reading on a small tag on the base.
For parts sales online, I do not do that anymore, because I have found the matches to be pretty good and feel good about what I'm selling. But for a local job I like to have the micromhos and conductance reading for if and when I see the amp again. I like to see how they are doing after a period of time.

Cheers,
Rob

[ThermionicScott]

Well for what it's worth, I do buy matched power tubes and, for an amp restoration or repair anyway, usually retest them and write the actual micromohs and conductance reading on a small tag on the base.
For parts sales online, I do not do that anymore, because I have found the matches to be pretty good and feel good about what I'm selling. But for a local job I like to have the micromhos and conductance reading for if and when I see the amp again. I like to see how they are doing after a period of time.

Cheers,
Rob

The thought of tagging birds and other critters before releasing them back into the wild comes to mind. Do you keep these measurements and install dates recorded somewhere?

[Rob's Radio-Active]

Just the tags I put on the tube bases.

[Old Tele man]

A "Tag & Release" policy for guitar amps!