Gassy 2a3?

[bepis69]

Hey, so I've been running the basic Sovtek 2a3s in my SII since I built it, and one of them seems to have developed a serious problem. I have confirmed that the problem is with the tube, not the amp, by swapping between channels. When switching on the amp, the tube emits a bright glow from inside the plate, while several pops can be heard through that channel. I would say the tube has developed gas but the getter is totally intact and visibly indistinguishable from the healthy tube. When peering down into the plate with a flashlight, the filament seems fine, it isn't bent off to one side contacting the grid or anything like that.

I recorded a video but it won't let me post the link.

When measuring voltage across the cathode resistor for the healthy tube, it gradually increases to 65.6V and stays there. With the problematic tube it shot up past 140V before I switched the amp off. The cathode bypass cap seems undamaged for now.

What do you think is going on here?
And what 2a3s would you guys recommend at a similar price point? Is there anything that's a meaningful upgrade from the Sovteks?

[hmbscott]

https://www.aliexpress.us/item/3256808416810940.html?spm=a2g0o.productlist.main.13.33587afdZEBydd&algo_pvid=467b3318-7a09-44ec-824a-55396c8a316b&algo_exp_id=467b3318-7a09-44ec-824a-55396c8a316b-6&pdp_ext_f=%7B%22order%22%3A%22-1%22%2C%22eval%22%3A%221%22%7D&pdp_npi=4%40dis%21USD%2195.98%2171.98%21%21%21692.00%21519.00%21%402101ef5e17427438541158525ec6b4%2112000045912706774%21sea%21US%211982106386%21X&curPageLogUid=T7LOj3d4BRJQ&utparam-url=scene%3Asearch%7Cquery_from%3A

The 300B version of that tube performed really well in a shootout I had a few years ago.

I just purchased a pair of PSVANE 2A3 2A3-TII a few days ago on this recommendation from PB, backups for my RCA black plate NOS tubes. Haven't received them yet so I can't comment on how they sound, but it's a higher spec tube than the Psvane 2A3C's my kit came with, so I presume it will be good. Price was pretty reasonable, $152 for the pair with tax and shipping.

[bepis69]

I just purchased a pair of PSVANE 2A3 2A3-TII a few days ago on this recommendation from PB, backups for my RCA black plate NOS tubes. Haven't received them yet so I can't comment on how they sound, but it's a higher spec tube than the Psvane 2A3C's my kit came with

When you get them, it would be great if you could post a comparison to the Psvane 2A3C and RCA black plates, as those were some of the other alternatives I was looking at.
For those who are curious, the video is at youtube dot com/watch?v=SgErqwa_A5Q

[Paul Birkeland]

That could be gas, or something physically coming apart inside the tube.  How many years do you have on this set? 

[bepis69]

That could be gas, or something physically coming apart inside the tube.  How many years do you have on this set?

About 2 years or so of frequent use. Probably 1000-2000 hours on the tubes.

[Paul Birkeland]

If this is an issue that happens constantly, it might be worthwhile to get a set of DC voltage measurements on the 4 pin socket while the issue is happening.  Typically I'd expect 5-10 years of moderate use for a new production 2A3 before they are exhausted.  Sometimes I have found that the solder joints in the tube base aren't always 100%, and that resoldering the tips of the tube pins with a little extra solder can help.  This kind of issue would be indicated by really odd voltages at the socket.

[bepis69]

This just got a lot weirder.
The issue is now occurring in both 2A3 tubes; I know for a fact that yesterday the second tube worked fine for listening in mono. I did a full resistance check of the amp as described in the manual and everything came out fine, so I'm really perplexed here. I also tested continuity between the 4 pins on each tube while they were standing vertically, and nothing was abnormal there. The odds of both of these failures being unrelated are extremely low, yet I can't find anything wrong with the amp.
Measuring the resistance to ground of the pins in the socket with all tubes removed, I get 1.25k for both heater pins, 248k for the grid, and an open circuit for the plate. Measuring voltages on the socket with the amp on but no tubes present, I get 448V on the plate, 0V on the grid, and a 2.5V sine centered at 0V on the heater pins. These measurements are identical for both sockets. I think it is normal for B+ to be higher with no tubes in.

I tested the filaments on both tubes and they both seem damaged in different ways. On tube A (the tube that failed first) it draws 2.5A at 2.5V but the filament does not visibly glow until I increase voltage to 3V, at which point it glows like it did when it was healthy. Tube B draws 3.1A at 2.5V and the filament doesn't visibly glow at all, even when increasing to 4V at 4.1A.

On close inspection, tube B's filament seems to be bent towards the grid in one spot but not quite making contact, and there is also what looks like some fine particles on the interior surface of the plate around that area. Tube A's filament and grid still look fine though.

I'm trying the tubes in the amp again and tube B is now not exhibiting any blue glow or anything at all, really.

Measuring voltages on socket pins & cathode resistor during startup, with 2 driver tubes present and 1 output tube:

Tube A:
plate: Voltage comes up to ~445V then gradually declines to ~430V then sharply drops to ~250-300V at the moment it starts sparking/glowing. The waveform then becomes very jagged, sort of like a bouncing switch, at this point I turn it off and the voltage smoothly falls to 0.
grid: Starts at 0V, slight ~50v transient when switching on, then after a few seconds, gradually climbs to 200V. The voltage here seems proportional to the brightness of the blue/red glow emanating from the tube. Declines to 0V over ~2 seconds when switching off. The plate and grid are an open circuit when the amp is off so somehow the tube is showing some conduction from plate to grid, or the grid is touching the filament but only when the tube is warm.
cathode resistor: Starts at 0V and is still only around 10-20V when the glow starts, but then climbs to ~210V when the glow is brightest.

Tube B:
plate: Voltage comes up to ~445V and just stays there indefinitely
grid: Starts at 0V, slight ~50V transient when switching on, then returns to 0V and stays there indefinitely.
cathode resistor: Starts at 0V and stays there indefinitely

I can only record these measurements for tube A for a few seconds before switching the amp off for safety.
I'm really at a loss here, I don't I want to get a new set of 2A3s before I can fully trust the amp, yet I can't find any issue with it.

[hmbscott]

When you get them, it would be great if you could post a comparison to the Psvane 2A3C and RCA black plates, as those were some of the other alternatives I was looking at.
For those who are curious, the video is at youtube dot com/watch?v=SgErqwa_A5Q

Watched your vid, the purple glow does look like gas and the flashing looks like arcing to me.

Here's my mini review of my kit's stock tube, Psvane 2A3S vs my RCA black plates NOS tubes back when I first got the RCA's:

I replaced the Psvane Classic 2A3-C's that came with my Stereomour II kit with matched NOS 2A3 RCA-construction black-plates (Sylvania branded), which I like a lot. Obtained from the very nice folks at Radio Electric Supply, $150 ea. The Psvane is a nice tube, but the RCA black-plates are definitely a cut above in refinement, most apparent in high-end smoothness without out any compromise in detail. This is in a new modified Stereomour II, (https://forum.bottlehead.com/index.php?topic=15167.0) that doesn't yet have the shunt regulator or DC filament supply upgrades added. I have very sensitive speakers so I hear significant the AC heater hum, and noted that the Psvains are definitely the noisier tube heater wise, but the RCA's aren't silent either. I have over a hundred hours on both tube types, and noted the Psvanes seemed to benefit much more from burn-in, especially early on, than the RCA's.

I haven't gone back and compared the stock Psvane 2A3S to RCA black plates after installing the shunt reg and the DC filament kits. I will say that the DC filament kit squelched the heater hum entirely. My speakers are DIY with estimated 100+ dB sensitivity. Also, not said in my original comment, the RCA sound is sweeter and more engaging that the Psvane 2A3S. If I had never heard the RCA's I would be totally satisfied with the 2A3S, but after comparing, the RCA is a lusher, sweeter sound and so they became my daily drivers. When I receive the Psvane 2A3-TII's and burn them in I will do a 3 way shootout and report back.

[Paul Birkeland]

Tube A:
plate: Voltage comes up to ~445V then gradually declines to ~430V then sharply drops to ~250-300V at the moment it starts sparking/glowing. The waveform then becomes very jagged, sort of like a bouncing switch, at this point I turn it off and the voltage smoothly falls to 0.
grid: Starts at 0V, slight ~50v transient when switching on, then after a few seconds, gradually climbs to 200V. The voltage here seems proportional to the brightness of the blue/red glow emanating from the tube. Declines to 0V over ~2 seconds when switching off. The plate and grid are an open circuit when the amp is off so somehow the tube is showing some conduction from plate to grid, or the grid is touching the filament but only when the tube is warm.
cathode resistor: Starts at 0V and is still only around 10-20V when the glow starts, but then climbs to ~210V when the glow is brightest.

200V on the grid does indeed indicate severe gas or a short.

Tube B:
plate: Voltage comes up to ~445V and just stays there indefinitely
grid: Starts at 0V, slight ~50V transient when switching on, then returns to 0V and stays there indefinitely.
cathode resistor: Starts at 0V and stays there indefinitely

This would be just a completely dead 2A3.

I too would be surprised to see a pair of tubes fail at right about the same time, but it has happened to me before.

[bepis69]

It's just bizarre that one tube fails, 24 hours later the second one fails in exactly the same way, then just dies completely a few hours after that, all with no indication of a fault in the amp. I'd never had any issue with these tubes and assumed they were still relatively early in their life.
The cathode cap and PSU caps seem unscathed by all this so I'll probably just get some cheap Chinese tubes and see what happens.
Thank you Paul B and hmbscott for your input