Schematic with added notes included as attachment.
The problem is that the cathode bias changes over time. During tests, the voltage rises, from 6.5V to 10V and presumably beyond. I disconnected the .01 uF coupling cap between one of the gain stages, to confirm whether there was DC present across the cap. The cap is fine with no leakage. I also checked the treble control to make sure the grid was still referenced to ground, and it tested fine and with a path to ground.
Both V5b and V6b tubes have a positive grid voltage with respect to pin 4/gnd. When each grid (still talking about V5b and V6b) has been tied to gnd with an alligator clip and wire, the cathode bias returns to something normal, between 6V and 6.5V.
Could the tubes be bad? They test well in a tube tester, though I haven't left the tubes in the tester for 10 minutes to warm up and see what happens. The power tubes are ECL86 pentodes.
Thanks. Happy Holidays.
The problem is that the cathode bias changes over time. During tests, the voltage rises, from 6.5V to 10V and presumably beyond. I disconnected the .01 uF coupling cap between one of the gain stages, to confirm whether there was DC present across the cap. The cap is fine with no leakage. I also checked the treble control to make sure the grid was still referenced to ground, and it tested fine and with a path to ground.
Both V5b and V6b tubes have a positive grid voltage with respect to pin 4/gnd. When each grid (still talking about V5b and V6b) has been tied to gnd with an alligator clip and wire, the cathode bias returns to something normal, between 6V and 6.5V.
Could the tubes be bad? They test well in a tube tester, though I haven't left the tubes in the tester for 10 minutes to warm up and see what happens. The power tubes are ECL86 pentodes.
Thanks. Happy Holidays.
Attachments
The tubes could definitely be bad. Sounds like excessive grid leakage. The coupling caps in these German sets are known for becoming leaky over time and this accelerates tube wear. So regardless, these should be replaced. And if these are original tubes or not recently replaced, a new set is definitely in order.
Agreed,
The tube control grids may be leaky.
The 1 Meg may be too large, especially for leaky grids.
You already disconnected the coupling caps, and that eliminates them.
(unless they were leaky and caused the tubes to go bad).
With new tubes, and lower control grid resistor to ground, suppose that keeps the
grids from going positive (tests OK).
Then re-test the coupling caps in the circuit (real circuit, not just an Ohmmeter check).
Make sure the plates are at 260V and the screens are at 230V/260V.
If the B+ to those elements are too high, the tube may run hotter, and could
run away thermally (especially with the 'large' 1 meg Ohm to ground on the control grid).
The tube control grids may be leaky.
The 1 Meg may be too large, especially for leaky grids.
You already disconnected the coupling caps, and that eliminates them.
(unless they were leaky and caused the tubes to go bad).
With new tubes, and lower control grid resistor to ground, suppose that keeps the
grids from going positive (tests OK).
Then re-test the coupling caps in the circuit (real circuit, not just an Ohmmeter check).
Make sure the plates are at 260V and the screens are at 230V/260V.
If the B+ to those elements are too high, the tube may run hotter, and could
run away thermally (especially with the 'large' 1 meg Ohm to ground on the control grid).
SOP to fix this is: replace the coupling caps and replace the tubes. The upward current drift with time and the positive voltage at the grids are both classic signs of a gassy tube. Unfortunately, the ECL86 has gotten very expensive, so I can understand the OP's reluctance to replace them. I'm one of the few left in the Chicago area that actually repairs these sets (most of the old radio guys avoid them like the plague), and I've seen this behavior many times.
Yes, indeed. So, if there's a way to install an additional 14V 0.6A PSU into the device, I'd try to replace them by their series heated PCL86 TV counterparts, which are much cheaper, still.
Best regards!
I've done this for the eye tube substitutes in some of these radios, which are even more rare than the outputs. He could likely get away with just putting a 6.3 volt filament transformer in series between the transformer windings in the set and the output tube heaters. Voltages tend to run a bit high in these sets anyway. A PCL86 can be bought here in the states for about $6.00.
Unfortunately, series heated Magic Eyes, as the UM11, are as expensive as all the other round screen eyes, at least here in Germany. In this case I'd go for a Russian indicator tube whose type I don't have handy at the moment...
Edit: Found it. It's the Russian 6E5-S, an octal version of the 6E5. Not an exact replacement for two-range tubes as EM5, EM11, or EM35, but at least it keeps a similar look.
Best regards!
Edit: Found it. It's the Russian 6E5-S, an octal version of the 6E5. Not an exact replacement for two-range tubes as EM5, EM11, or EM35, but at least it keeps a similar look.
Best regards!
Last edited:
I have had the same experience with 6L6GCs - the grid resistor is too high.
I recall an experience H.J. Leak had with his Stereo 20 amplifiers in the day. He used 1Meg grid resistors on EL84s, and so many amplifiers were apparently returned that he had to recall the lot and advise decreasing the grid resistors to 470K.
Lesson: Although the manufacturers may have meant well (or did they??), one should not operate this spec at maximum. There's too little telling that all power tubes will be safe there, particularly with the very close G1-C spacing of high-gm miniature tubes. I experimented with 6L6GCs. With a slightly lower grid resistor than the maximum indicated, many tubes ran at +2V over the grid resistor after some 30 minutes on, increasing as time went on. (As most members would know, this is a runaway action.)
One must also allow for tolerances in mains voltage, tubes, resistors, etc. Thus I will favour looking at this aspect first. Hoping you have a 10Meg impedance multimeter, you can slowly increase the grid resistor from say 270K until run-away threatens. (The 1Meg used for the ECL86 is on maximum!)
That would be my starting point before resorting to expensive tube replacement as mentioned. (I also found an unacceptable tolerance generally in the characteristics of present-day tubes - but that is another story )
I recall an experience H.J. Leak had with his Stereo 20 amplifiers in the day. He used 1Meg grid resistors on EL84s, and so many amplifiers were apparently returned that he had to recall the lot and advise decreasing the grid resistors to 470K.
Lesson: Although the manufacturers may have meant well (or did they??), one should not operate this spec at maximum. There's too little telling that all power tubes will be safe there, particularly with the very close G1-C spacing of high-gm miniature tubes. I experimented with 6L6GCs. With a slightly lower grid resistor than the maximum indicated, many tubes ran at +2V over the grid resistor after some 30 minutes on, increasing as time went on. (As most members would know, this is a runaway action.)
One must also allow for tolerances in mains voltage, tubes, resistors, etc. Thus I will favour looking at this aspect first. Hoping you have a 10Meg impedance multimeter, you can slowly increase the grid resistor from say 270K until run-away threatens. (The 1Meg used for the ECL86 is on maximum!)
That would be my starting point before resorting to expensive tube replacement as mentioned. (I also found an unacceptable tolerance generally in the characteristics of present-day tubes - but that is another story
)Yes, that's a good substitute that works with minimal rewiring. But you can only get these from Ukraine or Russian Federation. To the States, it can often take as long as a month for them to arrive. I keep a supply of the common 1629, which is basically a 6E5 with an octal base and 12.6v heater. These were used heavily by the US military, so there's still a fair amount of surplus supply still available. But again, prices for all these tubes keep going up.
You already checked the leakage of the coupling caps (some failed to notice).
I do like the idea that one person mentioned earlier in this thread.
Use individual 220 Ohm resistors, one for each cathode (and separate bypass capacitors for each one. That allows you to see what each individual tube you have now is doing (and also anytime you want to check at a later date, or when replacing those expensive tubes).
And as already suggested, a grid leak resistor of 270k Ohms should be low enough for any good tube to work without any significant positive grid voltage.
Of course the gain of the input stage will drop with the extra load, and the other series values will have to be changed to get the same frequency response and negative feedback loop stability.
Perhaps a 470k Ohm will work, and not affect the above so much that you would have to adjust the other values.
I do like the idea that one person mentioned earlier in this thread.
Use individual 220 Ohm resistors, one for each cathode (and separate bypass capacitors for each one. That allows you to see what each individual tube you have now is doing (and also anytime you want to check at a later date, or when replacing those expensive tubes).
And as already suggested, a grid leak resistor of 270k Ohms should be low enough for any good tube to work without any significant positive grid voltage.
Of course the gain of the input stage will drop with the extra load, and the other series values will have to be changed to get the same frequency response and negative feedback loop stability.
Perhaps a 470k Ohm will work, and not affect the above so much that you would have to adjust the other values.
Last edited:
Reiterate what Wavebourn and DF96 said - I had this happen with a 1625 tube where it would start out just fine. But as the minutes passed on by, it would start to draw more and more current, eventually causing the bias resistor to get very hot.
The effect followed the tube.
Luckily the 1625 is - still - dirt cheap. But I now put current meters in my amps so I can keep better track of what the output tubes are doing.
The effect followed the tube.
Luckily the 1625 is - still - dirt cheap. But I now put current meters in my amps so I can keep better track of what the output tubes are doing.
Not wishing to belabour, but I repeat my post #15. Gassy tubes could certainly be responsible, but a resistor is rather cheaper than a replacement tube! As DF96 said, 1 meg is quite high for a power tube grid resistor (I would dare say excessive).
I have it that the intensity of glow inside a gassy tube can reveal gassiness. (By that I mean overly intense blue glow. Some blue glow is natural for most power tubes.)
Question to DF96 and others with the necessary experience: In a gassy tube, does the anode current keep rising with temperature/time as in run-away, or does it only give a higher than normal (constant) current? While I have limited experience of gassy tubes (perhaps I was fortunate!), I always assumed (experienced??) the latter regarding higher-that-spec readings on a tester.
I have it that the intensity of glow inside a gassy tube can reveal gassiness. (By that I mean overly intense blue glow. Some blue glow is natural for most power tubes.)
Question to DF96 and others with the necessary experience: In a gassy tube, does the anode current keep rising with temperature/time as in run-away, or does it only give a higher than normal (constant) current? While I have limited experience of gassy tubes (perhaps I was fortunate!), I always assumed (experienced??) the latter regarding higher-that-spec readings on a tester.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.