I can't find anything like this online, so I thought I would put this here as a subject of general interest. Anyone who has read the thread about my ill-fated Chinese tube amplifier kit build knows that it is now dead after only 2 days with a bad output transformer. Details can be found there for anyone interested.
Below is a Chinese 6P14 output tube that failed on me after 48 hours of use. I very carefully disassembled it, and here is what I found. I hope I have everything labeled correctly. This is my first time inside a tube. I suspect it may be the first time for others also, which is why I consider this a matter of general interest to tube enthusiasts.
I didn't make those bends that you see in the distorted rods. They are extremely stiff, and they were that way when I gently slipped the internals out of the sheet metal outer envelope, which I assume is the plate, set off to the side out of the picture. On the left side, there is a metallic brown discoloration where the material in the tube discolored the glass envelope. On the right side, parts actually are in physical contact with each other, and that's how I found it when I carefully slid it out.
The Chinese 6P14 tube is well known for having an extremely weak and fragile screen grid, as well as very poor quality control, but surely it can't be normal for failed tubes to look this bad inside? Does anyone have any knowledge on this subject? I'm not sure how many people actually take the time to dissect a tube.
Below is a Chinese 6P14 output tube that failed on me after 48 hours of use. I very carefully disassembled it, and here is what I found. I hope I have everything labeled correctly. This is my first time inside a tube. I suspect it may be the first time for others also, which is why I consider this a matter of general interest to tube enthusiasts.
I didn't make those bends that you see in the distorted rods. They are extremely stiff, and they were that way when I gently slipped the internals out of the sheet metal outer envelope, which I assume is the plate, set off to the side out of the picture. On the left side, there is a metallic brown discoloration where the material in the tube discolored the glass envelope. On the right side, parts actually are in physical contact with each other, and that's how I found it when I carefully slid it out.
The Chinese 6P14 tube is well known for having an extremely weak and fragile screen grid, as well as very poor quality control, but surely it can't be normal for failed tubes to look this bad inside? Does anyone have any knowledge on this subject? I'm not sure how many people actually take the time to dissect a tube.
Last edited:
If you saw the conditions under which they were made, you would not be surprised.
Last edited:
If you had a hot-running screen it could have melted into that shape. It also could have been that way from the get-go, which would have put screen voltage on g1. That would have resulted in very high plate current. What doesn’t make sense for the latter scenario is the lack of red plating. With a g1-g2 short it should have tried to conduct a good half amp or so and red plated quickly. The transformer would have tolerated it long enough to heat up and smell. Since none of that happened the hot g2 due to lack of plate current makes more sense. If the g2 failed over time because of a lack of plate current it would melt. Maybe the wires, maybe the supports, depending on the relative melting points.
I don't know if it did red plate or not. It had been running for the first 48 hours and I was asleep. Not a hint of electrical burn smell. Right from the start there was a problem on the left channel (with this tube) with ~22mA current showing on a meter with probes from plate-to-plate. Still no idea why. I didn't expect to find the tube in this bad shape internally, but then again, I haven't ever dissected a tube before. It's interesting (to me) to see what's inside and what might go wrong.
I suppose a possibility is that the screen grid and something else conducted enough current that the electromagnetic field bent the rods. That happens a lot with short circuits in building electrical systems and is why the wiring in your house needs to be in a conduit or stapled every so often. I'd like to think that it would require kA of current to cause this kind of distortion, though. That doesn't seem likely for a tube amp. But maybe if the OPT failed first and took the tube with it.
Tom
Tom
The screen rods with bowing in the middle suggest excess dissipation of the screen was the root or an early cause of fault. Later, the fault may have led to screen-grid contact but any increase in fault current would be limited by external resistance and sag and the plate current characteristic for the sagged screen voltage.
Did the amp locate the valve on its side, or was it in a typical vertical orientation?
Did the amp locate the valve on its side, or was it in a typical vertical orientation?
Correct diagnosis of the output transformer causing the problem, and thanks! 🏆
Well, it is apparent now. 😉
Nevertheless, I thought people might want to see the innards of a failed tube since I couldn't find anything on the Internet like my dissection of it. It's a surprisingly primitive device inside. Yet, they work so wonderfully to make music.
It's a long read from the beginning of this project:
https://www.diyaudio.com/community/...be-during-if-i-do.394962/page-10#post-7301336
Please post a complete and accurate schematic of your amplifier.
That is worth 1,000 words, and is also worth 1,000 analytical thoughts.
Prevent another failure (unless you will not repair the amp, and give up on it).
I an not searching through hundreds of posts in another thread, just to locate the schematic. My time is worth more than that.
Yes, for certain circuits, if there is no plate voltage, the screen will do its best to act like a plate.
Answer the question: Which can take more power dissipation before it melts, the screen or the plate?
Want a good example:
Hard wire a pentode tube/beam power tube in triode mode.
Use a wire from the screen to the output transformer primary (and no 100 ohm low wattage series resistor); and a separate wire from the plate to the same output transformer primary.
Then, a bad thing happens, the plate wire is not well soldered at both ends, and the connection opens at one end or the other.
Now, the screen "becomes a plate". Ouch!
That is worth 1,000 words, and is also worth 1,000 analytical thoughts.
Prevent another failure (unless you will not repair the amp, and give up on it).
I an not searching through hundreds of posts in another thread, just to locate the schematic. My time is worth more than that.
Yes, for certain circuits, if there is no plate voltage, the screen will do its best to act like a plate.
Answer the question: Which can take more power dissipation before it melts, the screen or the plate?
Want a good example:
Hard wire a pentode tube/beam power tube in triode mode.
Use a wire from the screen to the output transformer primary (and no 100 ohm low wattage series resistor); and a separate wire from the plate to the same output transformer primary.
Then, a bad thing happens, the plate wire is not well soldered at both ends, and the connection opens at one end or the other.
Now, the screen "becomes a plate". Ouch!
Last edited:
Schematic is in the first post of that other thread.
The fault (afaik) was identified near the end of that thread, and was a broken lead-out wire at the UL terminal/tap on one PP side (ie. the UL to plate portion of a primary half was not connected - post #195) of one channel, and apparently not easy to see, and likely happened after initial installation and resistance checks. Subsequent amp operation fooled the owner to thinking the stereo sound had not degraded - a not uncommon situation when listening to music even though one would think such a fault would be audibly discernible.
The amp orients the valve vertically (ie. nominal orientation). As such, the physical bending on both screen rods (and in the same direction) is not simple to explain. Its reasonable that the central section of each post would reach the highest temp as long as the screen wire remained intact, and the screen wire likely had a higher melting/deformation temp than the posts. Thermal expansion of the screen posts would be constrained by the screen wire, so some form of torsion or bending as shown seems plausible, if the posts could increase their length (eg. they may have expanded along their length and then hit a limit).
EIArte, afaik the fault did not lead to arcing, but red-plating and discoloration of the central anode section (post #153).
The fault (afaik) was identified near the end of that thread, and was a broken lead-out wire at the UL terminal/tap on one PP side (ie. the UL to plate portion of a primary half was not connected - post #195) of one channel, and apparently not easy to see, and likely happened after initial installation and resistance checks. Subsequent amp operation fooled the owner to thinking the stereo sound had not degraded - a not uncommon situation when listening to music even though one would think such a fault would be audibly discernible.
The amp orients the valve vertically (ie. nominal orientation). As such, the physical bending on both screen rods (and in the same direction) is not simple to explain. Its reasonable that the central section of each post would reach the highest temp as long as the screen wire remained intact, and the screen wire likely had a higher melting/deformation temp than the posts. Thermal expansion of the screen posts would be constrained by the screen wire, so some form of torsion or bending as shown seems plausible, if the posts could increase their length (eg. they may have expanded along their length and then hit a limit).
EIArte, afaik the fault did not lead to arcing, but red-plating and discoloration of the central anode section (post #153).
Correct. Upon very close examination yesterday, an unused UL tap on the left channel output transformer had wires that somehow became detached and were making intermittent connection. Just the slightest touch of the UL tap would make or break the circuit, I finally discovered.
Live and learn: I was all ready to blame the tube for the problem and went looking for the cause inside the tube. I was wrong. I never expected an unused UL tap making intermittent contact, eventually breaking altogether.
I have since attempted repair of not one but both UL taps on that transformer as well as the center tap, and the transformer is awaiting reinstallation and testing. The right output transformer did not have this issue because it has globs of glue protecting the wires where they exit the winding and attach to the tabs. Quality control somehow missed the glue globs on one transformer, which, although not very attractive, do prevent the issue. My repair work seems to be very secure, so we will see what happens.
In any case, now everyone knows what a small pentode power tube actually looks like inside, albeit a "little" distorted. 🙂 I always wondered what exactly was inside one because it's all hidden inside the sheet metal inside the tube. So simple, yet they sound so nice.
Correct.
I just realised I referred to 'red plating', but that wouldn't have occurred. Rather it was metal expulsion from the screen passing through the anode apertures - a not uncommon occurrence where the glass envelope shows metal coated sections (but not the getter) occurring over time and aligned to apertures in the anode and top mica, and sometimes coming from heater metal (ie. not just from a hot screen). The internal spray and deposition of metal over time can lead to increased leakage across mica discs between support posts and shows up as a form of input grid leakage current and its issues.
By flashing, I didn’t mean arcing, but I was incorrectly referring to the the barium gettering that is quickly consumed once a tube is leaking.
trobbins,
You said:
"Schematic is in the first post of that other thread."
What other thread? Post # 10? OK
https://www.diyaudio.com/community/...be-during-if-i-do.394962/page-10#post-7301336
That is 200 Posts.
The first post in that thread does have a schematic.
However that schematic is not Ultra Linear. But this thread mentions Ultra Linear.
And, if the screen became removed from the UL tap, that tube would have less current (almost zero plate and zero screen current).
But if there was a shared self bias resistor of the push and pull tubes, the current in the other tube would go up.
Only 17 posts, and I am so glad we have "solved" the original occurrence. Ha Ha.
Anatomy of a failed tube.
You said:
"Schematic is in the first post of that other thread."
What other thread? Post # 10? OK
https://www.diyaudio.com/community/...be-during-if-i-do.394962/page-10#post-7301336
That is 200 Posts.
The first post in that thread does have a schematic.
However that schematic is not Ultra Linear. But this thread mentions Ultra Linear.
And, if the screen became removed from the UL tap, that tube would have less current (almost zero plate and zero screen current).
But if there was a shared self bias resistor of the push and pull tubes, the current in the other tube would go up.
Only 17 posts, and I am so glad we have "solved" the original occurrence. Ha Ha.
Anatomy of a failed tube.
Last edited:
trobbins,
You are correct.
However, please go back and read my Post # 17 which I re-edited while you were typing your post # 18.
I stand by the fact that this thread mentions UL, but the schematic in question is not UL.
Just Saying . . .
You are correct.
However, please go back and read my Post # 17 which I re-edited while you were typing your post # 18.
I stand by the fact that this thread mentions UL, but the schematic in question is not UL.
Just Saying . . .
Yeh the devil is in the detail sadly. The OP used an OPT that included a UL tapping - but that UL tap was not used or connected to, but somehow one of the winding wires to that UL terminal had broken (and as such had disconnected the UL to P portion of the winding on one side of the OPT) - as per photo in post #195 in that other thread.