Home
AudioAsylum Trader
Tubes Asylum: Power Tube Reliability (Long) by Jim McShane

Questions about tubes and gear that glows. FAQ

For Sale Ads

FAQ / News / Events

 

Power Tube Reliability (Long)

71.194.241.37


[ Follow Ups ] [ Post Followup ] Thread:  [ Display   All   Email ] [ Tubes Asylum ]
[ Alert Moderator ]

If you all had the ability to read my Inbox in the last few weeks you would have found that:

1. SED/winged C tubes are bad quality

2. Mullard EL-34 reissues are bad quality

3. KT-88 EH are bad quality

4. KT-88 Genalex are bad quality

5. KT-77 Genalex are bad quality

6. Tung-Sol reissues are bad quality

7. JJs are bad quality

Why are they bad quality? Well, there have been instances of all of the tubes above failing in service DESPITE the most careful testing I can do.

I want to make something perfectly clear here. I don't stock tubes I find have unacceptable quality. I am in a position where I can make that determination based on a large sample of tubes, and that is indeed how I base it.

But yes, tubes I sell still fail on occasion, no doubt.

And what comes though in nearly every case is the mistaken idea that old stock tubes didn't fail. Hogwash...

Why do you suppose that those big tube caddies were invented? The fact is that repairmen needed a large tube supply because tube failures were the number one cause of a TV, radio, or hi-fi unit performing poorly.
In the last few years I've had failures of:

1. EL34 xf2 Mullards

2. RCA 6L6GC black plates

3. Mullard GZ34

4. Original KT-88 Genalex

5. GE JAN 6550A

6. Westinghouse 7591A

Now, you can say that "well, the number of those that failed is small". Agreed. But the number of those tubes still around is miniscule, so the number of failures is proportionately nearly the same as you find with many current production tubes.

The fact of the matter is that tubes are the least reliable component in a typical piece of gear. While tubes possess a certain ruggedness in that they can shrug off momentary issues that would detonate solid state active devices, they simply have a number of strikes against them that negatively impact reliability:

1. They have multiple relatively unsophisticated critical connections. At least a plate, screen, grid, cathode, and two heater pins all must maintain flawless connection.

The connection point is subject to some vibration and significant heating. Sockets endure large numbers of heating-cooling cycles. In many pieces of gear the number of insertions and removal of the pins has helped the socket to deteriorate.

2. The tube is essentially a hand-made device. While there is some machinery involved, the machines are dependent on their human operator for consistent and accurate assembly quality. It needs to be said that the electrical characteristics of a typical tube have much less variability than the typical discrete solid state device. But the number of mechanical and electrical connections that have to be made 100% right is of concern. Making 'em 100% right ain't easy!

3. Tubes are often installed in equipment that is working at far less than optimum condition. Vintage audio gear that is 40-ish years old is hardly the ideal place for any electronic component to thrive.

I want to share a story - I will not name my customer, as I have no desire to embarrass him or ridicule him. He's a good soul! But the story illustrates a point...

This customer has a couple Marantz 8B, a 40 year old classic vintage audio amp. He purchased eight Mullard reissue EL-34 tubes to put in his amp. After a few days one of the Mullards failed on him. CRAP! I HATE that! It seems no matter how carefully I test and retest a tube some inevitably fail after the customer puts them in service. So I sent him a replacement tube (for free of course) and I now felt nearly 100% sure that he's all taken care of. On RARE occasion a second tube will fail, but it's quite uncommon.

So when I got another email about a second tube failing it was extremely discouraging. It was not just a notification of the second tube failing, but it was accompanied by statements condemning Mullard reissue quality - the jist was "I want different tubes to replace this Mullard stuff - I want SED/winged C". What was ironic was I had just read a note from another fellow who told me how SEDs were no good, that's why they failed in his amp. So he'd like some other tubes, maybe some reissue Mullards??

Okay, you could chalk it up to coincidence, and believe me, I felt terrible about it. All I can do is replace the tubes and all should be good. A THIRD failure is almost unheard of. So I write back to the original person with the 8B and tell him I'll replace the second tube that failed, could you send me the matching info from the tube and I'll send one right out. He replied he'd have to track down the tubes, AS HIS AMP WAS IN THE SHOP with one channel out and the amp was getting some mods. What really happened here? Could the malfunctioning amp have damaged the tube? The first tube I replaced for him wasn't shorted or such, there was no way it did any damage. I doubt it was the tubes at fault at all. But this person is now convinced that the Mullards are no good - and here's yet another reason why he thinks this. His service repair guy informed him that the best tubes that he could get were the Winged C because of the Mullards being "re-badge" tubes.

Whew...

I am going to close my comments with three last observations:

1. Other than the JJ octal tubes and possibly the late KT88-98 there is no significant difference in reliability between any of the major brand new production large power tubes (6L6, EL34, 6550 etc.) when installed in a healthy, properly biased amp.

2. The reason you hear about more KT-88 Genalex failing than other tubes? It's because they sell at 4 to 6 times (or more) the rate of any other large power tube in my (and I'm sure other vendors') stock. There are WAY more in service than any other current production power tube, so the number of tubes that fail will be the largest. As well, it's rare that people post how they haven't had any trouble - most never mention it.

Yes, if you have a quad of SEDs/EHs/Genalex/etc. and one fails it may seem the failure rate is 25%. But mathematicians and statisticians agree that to make a determination you need an adequate sample size. Four isn't large enough of a sample. Could new production improve? Sure, but it's far from bad already. Anyone around in the late 80s/early 90s that tried to use Chinese power tubes of that era can vouch for that.

3. Old stock tubes used to fail too - but you don't hear much about it anymore since the number of large old stock power tubes is tiny as compared to the number of new production in service. Were they better than new production? I believe the best tubes of the era were - but not by a huge amount. They had plenty of failures too.

I'd like to hear other people's comments.


This post is made possible by the generous support of people like you and our sponsors:
  Kimber Kable  



Topic - Power Tube Reliability (Long) - Jim McShane 08:52:31 12/14/09 ( 41)