Leave on or turn off?

LeGrace wrote:I had a matched quad. Can I just order a replacement for Mr Burn Out? Or do I need to send one of the other ones back for matching? Or is not that critical?

Depends on who matches. I suspect that if you install wildly mismatched tubes, you may be repeating the process in short order.
I suspect, depending in your tube source that the "Matched Quad" you have was not really.  Less than 5,000 hours (to me) is an unconscionably short life. And if they were properly matched, something else is definitely going on.  If not, you now understand the consequences.

This is a McShane-knowledge-base issue. I have the means to match properly - few do. He would know as I am sure he does actual matching using means other than the phases of the moon and date codes.

EDIT:After a (very little) bit of research, it seems that the expected tube life for a new-production KT88 hovers between 4,000 and 5,000 hours. The majority seems to agree on the 4,000 hours citing actually achieved lifespans. The consensus above that seems to hover around 5,000 hours scattered equally above and below.

LeGrace wrote:Timely advice re "wearing out" LOL > just had my first KT88 power tube go south on me.  Major scorch marks evident on the glass. Estimate after 1000 - 1200 hours. Had hoped to get at least a year. Lesson learned!!!

Change brand !

My jj 6550 has been online 3000h now, still as new.

My tube source in this case is J. McShane. There are numbers hand inscribed on the top of each tube which closely match. I suspect this is just my normal bad luck.   Meanwhile I've switched to dual output tubes, so at least I'm back in business. If I cant get longer life I may have to stay this way.

LeGrace wrote: I suspect this is just my normal bad luck.

Probably not. You need to _REALLY_ check out that amp before slamming another set of tubes in it! Really!

Can you help with suggestions what I should check? I'm still a total newb, really have no idea. New dual tubes layout is working just fine, for now anyway.

On this, I will defer to others who have actually built and now use *your* amp. I will tell you that i take the test-point voltages and bias voltages very seriously in my very vintage 70s and 35. And I look for any sort of differences between channels as indicators of "where to start" as it were.

When you do this, be exceedingly careful given lethal voltages, and other issues. Shield your probes so you do not 'cross' anything while reaching inside. And make sure that *you* are not grounded. I run everything through an isovariac, so I can be dead-sure of input voltages and not dead myself....

Must be karma thinking about leaving the amp on 24/7.  One thing about a tube amp when they fail it is often with high drama.   

For starters check the 10 ohm cathode resistor on the tube socket (inside) where the tube failed.  Look for bulging, blackening, or excessive heating.  If so replace it with suitable resistor.  If it looks OK measure its resistance. 

If you can take good pictures then we can access and advise what to do next.

LeGrace wrote:My tube source in this case is J. McShane. There are numbers hand inscribed on the top of each tube which closely match. I suspect this is just my normal bad luck.  :evil: Meanwhile I've switched to dual output tubes, so at least I'm back in business. If I cant get longer life I may have to stay this way.

Picture of the tubes shows that the getter has been consumed, something causes air to
sneek in.
As two tubes shows similar leak it seems to be a common cause.
Are these tubes frequently removed from the sockets ? Minor movements in the
socket might damage the glass seal...

peterh wrote:

LeGrace wrote:My tube source in this case is J. McShane. There are numbers hand inscribed on the top of each tube which closely match. I suspect this is just my normal bad luck.   Meanwhile I've switched to dual output tubes, so at least I'm back in business. If I cant get longer life I may have to stay this way.

Picture of the tubes shows that the getter has been consumed, something causes air to
sneek in.
As two tubes shows similar leak it seems to be a common cause.
Are these tubes frequently removed from the sockets ? Minor movements in the
socket might damage the glass seal...

Your diagnosis looks spot on. Googling the term "flashed getter" has been educational. If air bleeds in normally will cause the getter coating to turn white along of course with inducing catastrophic tube failure. And indeed both getters on either side have turned white. Bottom tube in picture is a working one, coating is still silver. Just for comparison purpose.

Then this is looking a lot more like a tube QA issue rather then anything with the amp. And also meaning shouldn't be representative of life time I can normally expect from these tubes. Both of which make me feel a whole lot better. Thanks!!

Well I can see how the air bled in. On closer examination just discovered a crack in the glass!

Peter W. wrote:Not to be the least bit political - but do the most recent kerfluffles relating to Russia give one pause when considering tubes from that source?

I wonder what the throughput is - how long from factory-to-consumer?

Insofar as we're getting our new power tubes from Eastern Europe (Russia primarily) and from China, I wonder:
a:) what's the ratio of Asian consumer tubes to Russian tubes in service: and, b)
What are the failure rates of new tubes from those two sources.
Now that the Chinese seem to have cleaned up their act, I would take a SWAG and say the percentages are about the same, failure-wise.

deepee99 wrote:

Peter W. wrote:Not to be the least bit political - but do the most recent kerfluffles relating to Russia give one pause when considering tubes from that source?

I wonder what the throughput is - how long from factory-to-consumer? :|

Insofar as we're getting our new power tubes from Eastern Europe (Russia primarily) and from China, I wonder:
a:) what's the ratio of Asian consumer tubes to Russian tubes in service: and, b)
What are the failure rates of new tubes from those two sources.
Now that the Chinese seem to have cleaned up their act, I would take a SWAG and say the percentages are about the same, failure-wise.

Don't forget EU made tubes : JJ
+ several low volume mostly 300B and similar fron tjeckia and germany

LeGrace wrote:I had a matched quad. Can I just order a replacement for Mr Burn Out? Or do I need to send one of the other ones back for matching? Or is not that critical?

I believe that the ability to adjust the bias for each individual tube eliminates the need to require a 'matched' quad.

Dogstar wrote:

LeGrace wrote:I had a matched quad. Can I just order a replacement for Mr Burn Out? Or do I need to send one of the other ones back for matching? Or is not that critical?

I believe that the ability to adjust the bias for each individual tube eliminates the need to require a 'matched' quad.

You are wrong.

Matching, properly done, includes that the Gm ( amplification ability ) is also selected
as a parameter. If not matched it would be the equivalence of different size pistons
in the same engine.

You are wrong.

Matching, properly done, includes that the Gm ( amplification ability ) is also selected
as a parameter. If not matched it would be the equivalence of different size pistons
in the same engine.[/quote]

OK - let's cut to the chase. There are two (basic) reasons for matching:

Electrical - so that one tube of a pair does not become a current-sink and wear itself out very quickly at best, slag itself at worst.
Performance - so that the tubes are approximately even in amplification performance.

The reasons for the former are pretty obvious, and the consequences for ignoring the need equally obvious. But, individual bias capacity eliminates the penalty. Further to this - electrical matching is a momentary thing given the generally poor quality of modern tubes. Once upon a time electrical matching would remain tolerable over the life of the tube as QC was very nearly infinitely better than now, such that the aging process was reasonably predictable. A tube match to 5% today was likely within 10% or so withing its service life. That level of predictability is no longer possible. Individual biasing is the proper answer. Writing only for myself, were I to have to use modern production tubes in my 70, I would immediately move to a VTA board or-equal as individual biasing would be necessary for safe operation.

Now to the latter: In point of fact, have a look at the process behind the Atkinson Cycle engine (patented in 1887). There is no penalty for "different size pistons" (not quite, but very close to that) in the same engine if that is accommodated in the design. That one tube may do 70% of the work, another only 30% is not at issue as long as there is no electrical penalty.

As we have established that the tube-under-discussion suffered a mechanical failure, not an electrical one, I would revise my advice to LeGrace - yes, replacing one tube of a pair, while not ideal, is perfectly OK as long as the bias across the board is set-and-correct. Once more writing for myself, I prefer replacing in pairs. But if not electrically dead, my 'pulls' will be retained for individual re-matching as future short-term replacements. typically only one of any given pair is either compromised sufficiently for disposal or stone dead.

As compared to any given Revox solid-state amp, those of us driving tubes are back in the stone-age in terms of complexity. These devices are rugged, and with very little mystery attached.

Again, horses for courses.

The "piston" thing might be an broken allegory ( or what to say)
But if one tube has more "amplification" (Gm) then the other in a pair the up's and down's will be
unequal. NFB ( negative feedback) will reduce this, but not eliminate, providing that the amp
has NFB ( which a st120 has)

The thought of replacing 4 tubes when only one has gone bad is alarming. Especially in a case of premature failure as I've just experienced. But I appreciate the benefits of matched sets. Do any of you order matched quintets as a backup solution?

LeGrace wrote:The thought of replacing 4 tubes when only one has gone bad is alarming. Especially in a case of premature failure as I've just experienced. But I appreciate the benefits of matched sets. Do any of you order matched quintets as a backup solution?

This is one of the reasons I haven't gone for the M125s. That and I can't afford them.

peterh wrote:The "piston" thing might be an broken allegory ( or what to say)
But if one tube has more "amplification" (Gm) then the other in a pair the up's and down's will be
unequal.   NFB ( negative feedback) will reduce this, but not eliminate, providing that the amp
has NFB ( which a st120 has)

Peter:

Let me try to give you a better analogy.

You are aware that an ellipse is a 'circle with two foci". A matched pair of tubes would give a balanced (symmetrical) ellipse as a response. Weaker/stronger pair would give an egg-shaped response exaggerated to the extent of the difference. It is not as if one tube has all of the "ups" and one all of the "downs".

Does that help?

[quote="LeGrace"]The thought of replacing 4 tubes when only one has gone bad is alarming. Especially in a case of premature failure as I've just experienced. But I appreciate the benefits of matched sets. Do any of you order matched quintets as a backup solution?[/quote


OK - I can be a naturally snarky individual if given the opportunity - I will try to curb my natural response to the entire 'matching' debate.

a) The critical reasons for matching these days, especially in vintage equipment, is electrical. Greatly mismatched tubes will cause one tube to 'eat' all the current in a pair, greatly increasing the stress on it and leading to premature failure *IF* each tube in that pair is not biased separately. ANALOGY: Consider the voltage divider - a very simple concept of several resistors in series dropping voltage across each junction. Now consider this:  Two parallel resistors each rated at one (1) watt. Keyword here is "RATED". In point of fact, one is about 10% better than its rating, the other about 10% worse. Now, these are 10% resistors at 100 ohms. One is actually 90 ohms, the other 110 ohms. As it happens, the 110 ohm resistor is the 'better' one for rating. Put two watts of current through that system. Guess what happens to the other one at an actual 90 ohms. Think before reacting. Each resistor is visually identical to the other.

b) Whereas electrically matched tubes can give a better performance than unmatched tubes, if bias is addressed, the only actual penalty is in output rating. So, two perfectly matched KT88 tubes driven to the bleeding edge might give 75 watts at 8 ohms. Two mismatched tubes - again. bias addressed - might only give 50 watts.

There is NO REASON for a matched quad in a paired output stereo amp. Matching within a channel is sufficient. Matched quads are reserved for four-out devices. Keep in mind that matching, while painstaking, is not entirely exotic. What it means is that several electrical parameters in each tube are measured within 2% - 5% of each other, and all are within correct operating parameters. Any given pair can be high in this number, low in that as compared to another pair - but all measurements are within the limits.

You would have a better chance by ordering four random tubes to find the one that you need. But you would need a tester that allowed you to measure those things. Few do.

One reason ( the only ) to buy a matched quad is that if one tube is broken there is three with
identical parameters. After replacing the failed pair with a new, you still have three good tubes
and can copy with a second loss of the original quad and still have a working pair.

And buying a matched quad won't cost more then 2 matched pairs ( as far as i know).

Looks to me like LeGrace just had a gas leak in the one tube. For precisely the reasons Peter W. delineates, I prefer the ability to bias each tube individually, while still looking for closely-matched pairs although, after a few months those closely matched tubes may be fairly far out of whack. I wouldn't waste money on a whole new quad. It's like replacing all four tires when you've just got one lemon.

Adding to the confusion (which I love to do) is that the signal tubes (dual triodes) are also "matched" but it has a totally different meaning than "matched" pentode (output) tubes. Matching in the smaller signal tubes used in pre-amps (12A?7, 6SN7, etc) and the drivers is an internal measurement so that the paired elements in EACH tube are functioning at the same level, 50/50 being ideal, I think 60/40 being the outer limits of what's acceptable. Most tube testers can achieve this measurement.
Here's a good layman's overview:
http://www.thetubestore.com/Resources/Matching-and-other-tube-info/Matched-Sections-Service

Here ya go:  An explanation with charts and graphs:

http://www.augustica.com/Tutorials/Tubes/matching_triodes_and_pentodes.htm  

Imagine trying this with dual-triodes? Multiple dual-triodes?