I built myself a 'spud' headphone amp a few years ago that's nothing more than a 500VCT power transformer with full wave solid-state diode rectification, CLCRC DC plate supply, 6.3VCT (AC) heater supply, two 12HL7 pentodes wired triode, a pair of Edcor 8k:50 SE OPTs, and the usual complement of jacks and a volume control.
The 12HL7s are being run pretty hot, at about 150V plate-cathode and 45mA Ip, so Pdiss is almost 7W. That's an attempt to keep the rp down low to compensate for the not-very-generous primary inductance of the very cheap OPTs. The max combined plate+screen dissipation for 12HL7 is 11W, so I figure 7W is low enough that the tubes should have a decently long service life.
Being that it's a headphone amp for bedtime listening, I sometimes fall asleep and leave it on overnight. Sometimes I forget to power it off when I get up, so it stays on all day.
I find that the 12HL7s go noisy after about a year and a half of use. One or the other of them will start to crackle, which is horribly uncomfortable in headphones! (Ouch.) My question is, should I expect these touchy high gm RF pentodes to go noisy after a year or two of use, or does it sound like something is wrong and I should start looking for a problem?
I have 1k ohm grid stoppers on the 12HL7s and 0.1uF 100V film caps bypassing the heater supply to chassis. Perhaps there are more oscillation suppression steps I should take? 10 ohm plate stoppers?
The 12HL7s are being run pretty hot, at about 150V plate-cathode and 45mA Ip, so Pdiss is almost 7W. That's an attempt to keep the rp down low to compensate for the not-very-generous primary inductance of the very cheap OPTs. The max combined plate+screen dissipation for 12HL7 is 11W, so I figure 7W is low enough that the tubes should have a decently long service life.
Being that it's a headphone amp for bedtime listening, I sometimes fall asleep and leave it on overnight. Sometimes I forget to power it off when I get up, so it stays on all day.
I find that the 12HL7s go noisy after about a year and a half of use. One or the other of them will start to crackle, which is horribly uncomfortable in headphones! (Ouch.) My question is, should I expect these touchy high gm RF pentodes to go noisy after a year or two of use, or does it sound like something is wrong and I should start looking for a problem?
I have 1k ohm grid stoppers on the 12HL7s and 0.1uF 100V film caps bypassing the heater supply to chassis. Perhaps there are more oscillation suppression steps I should take? 10 ohm plate stoppers?
I see two different styles of 12HL7s, one with a fairly bilaterally symmetrical anode structure that have a "Sylvania" look to them, but in Sylvania, Amperex, RCA, Zenith, GE, Westinghouse, Raytheon, and other boxes and labels, and another with different construction with a shiny anode face to one side, all in RCA boxes. What kind are you using?
All good fortune,
Chris
All good fortune,
Chris
My understanding is that these 'crackley' sounds develop because the heater-cathode limits have been exceeded in some way, and the insulation starts to fail over time.
I have a stash of 12HL7s.
Some are RCA, which have plate structure with a silver-color shiny flat side and a matte grey side.
Some are Sylvania with symmetrical looking, shiny black plate structure.
I was using a pair of RCA 'clear top' with the asymmetrical, shiny silver/matte gray plate, getter flash over the shiny silver side. Both eventually went noisy.
I have a pair of 12GN7A in the amp now, which are working fine, as I expected they would.
This is the schematic of the built amp. It's quite simple. Downright boring.
That's from a SPICE simulation, but the voltages in the made amp are very close (well within 10%, depending on the tubes being used).
The heater supply is plain old AC, from a center-tapped 3.15V-0-3.15V transformer. I have it lifted to +40V by a voltage divider. Hmmm... Perhaps the voltage divider lower resistor is too high of a resistance value for these tubes? Maybe I should just ground the heater supply center tap to 0V. The cathode bias is only about 2.5V, so that's not an issue.OldHector said:
I'll have to pull the amp out and open it up to remind myself what I did.
Well you won’t know for sure if anything is oscillating without an oscilloscope. But if the noise stops completely with new tubes then the tubes are the most likely culprit.
Yes, that's the mystery. Replacing with new tubes makes everything good again. The noise develops over time, much like how a tube with its heater-cathode voltage limit exceeded acts, as OldHector described. The mystery is that this has now happened twice, the exact same way, with two pairs of 12HL7s. After about a year of use, one 12HL7 goes crackly, really loud in one ear. Not pleasant. I'm wondering if I should expect a useful life of only 1 year from these tubes used this way.
It looks like I built this amp in 2018. How time flies...
I have a signal generator and a Rigol DL1052E oscilloscope. Hopefully those will be useful.
It looks like I built this amp in 2018. How time flies...
I have a signal generator and a Rigol DL1052E oscilloscope. Hopefully those will be useful.
These RCA types have smaller plates, and appear to be 11HM7 guts. Which are only rated at 7 Watts. They don't have the good triode curves of the bigger symmetrical design 12HL7 tubes (Sylvania ...) either. 50 mA appears to be the max DC rating for the real 12HL7s, 11HM7 types might be less.
( 12HL7 datasheet curves go up to a 100 mA knee, while the 11HM7 datasheet goes up to a 76 mA knee. So maybe 37.5 mA DC max. ) Frame grid types generally seem to have lower DC max current ratings than non frame grid types for the same Htr power rating.
Last edited:
How many hours would you say before the noise starts? 2 hours a day is like 700 in a year so if you are hard on them a year and a half might not be out of line.
"The heater supply is plain old AC, from a center-tapped 3.15V-0-3.15V transformer. I have it lifted to +40V by a voltage divider."
Why?
Why?
Powered on at least a few nights a week, and the amp gets left on for about 8 hours, sometimes longer. I often fall asleep forgetting to turn off the amp. Sometimes I forget to turn it off in the morning and it's left on all day. And according to what smoking-amp just wrote, it does look like I'm being hard on them.
Lower likelihood of hum from the heater supply, according to RDH4, Blencowe, Jones, et al. This is a headphone amp, so I was worried about any and all sources of hum, but I didn't want to add a separate transformer for a DC heater supply.pl802 said:
Great info, thanks.smoking-amp said:
I wonder if 12GN7A is a better choice for this low plate voltage/high plate current type of use. It has even lower rp than 12HL7-triode. I think the 12HL7 sounds a tiny bit better, but not enough so that I should be burning them up.
Or perhaps I should back off on the plate current (install higher value Rk). Cheap and easy to do. I did try it that way but noticed a reduction of bass response and a somewhat thinner sound. It still sounded fine, though. I was 'tuning' the circuit by ear and found that I liked the sound of it once the Ip got up past 40mA. Maybe that's just too much for 12HL7.
By contrast 12GN7A is rated for 11.5W plate + 1.5W screen dissipation, so I figure that type can take more plate current. I find the sound from it is a little different, but still quite good to my ears. I could save the 12HL7s for a different use.
I wonder if these frame grid pentodes are more sensitive to max heater cathode resistance than other types? I have 27k ohms heater to cathode.
PS - Have you seen the prices of 12GN7A these days?
Last edited:
A countdown timer makes a lot of sense for a nightstand Amp.
https://www.amazon.com/Century-Digital-Countdown-Repeat-Function/dp/B01D3QEK4E/ref=sr_1_52_sspa?crid=RHHY756HJ5W3&keywords=auto+turn+off+timer&qid=1668358075&sprefix=auto+turn+off+timer,aps,307&sr=8-52-spons&sp_csd=d2lkZ2V0TmFtZT1zcF9hdGZfbmV4dA&psc=1
12HL7 has very little distance between the cathode and grid unlike other tubes, maybe elevating it is working against it life-wise? The spec sheet does say not to use a tube tester where the difference is as low as 50V even. Maybe disconnect the elevator wire, put in new tubes, and see if those last longer? You could add 10 turn hum pots to each tube to precisely null each tube, since it is for headphones and any hum will be really noticeable, unlike speaker hum.
Attachments
Last edited:
I'm not understanding how the max grid-cathode voltage of 50V would influence the cathode to heater voltage limit. Are you thinking there's some kind of cathode to heater leakage current that creates a voltage differential between the grid and heater that becomes important?
You know, if that's the case, connecting the heater supply CT to 0V might make those noisy 12HL7s work again. Maybe.
I was out grocery shopping, so I picked up one of those old fashioned light timers. I'll set it to turn the amp on at 10:00pm and off at 2:00am. That should do the trick.
You know, if that's the case, connecting the heater supply CT to 0V might make those noisy 12HL7s work again. Maybe.
I was out grocery shopping, so I picked up one of those old fashioned light timers. I'll set it to turn the amp on at 10:00pm and off at 2:00am. That should do the trick.
smaller envelopes or bulb sizes don't do well with life expectancy
far as heat dissipation
running small tubes at high dissipation will reduce life dramatically.
has to do with pin spacing and extra heat that goes into components.
numerous causes of " Crackles" , usually carbon or insulation failure.
Or makes plate contaminants easier to build up with the tighter spacing
of a miniature tube
tighter pin spacing of 7 or 9 pin miniature tubes, tends not to do well
with high heat or high overall dissipation.
they made a " mini" version of the 6V6
rather amazing for the time. But the heat killed those little guys.
grid is way closer to the cathode in a miniature tube.
heat dissipation and failure from contamination.
spread out pin spacing and use a larger glass envelope.
tube life increases.
far as heat dissipation
running small tubes at high dissipation will reduce life dramatically.
has to do with pin spacing and extra heat that goes into components.
numerous causes of " Crackles" , usually carbon or insulation failure.
Or makes plate contaminants easier to build up with the tighter spacing
of a miniature tube
tighter pin spacing of 7 or 9 pin miniature tubes, tends not to do well
with high heat or high overall dissipation.
they made a " mini" version of the 6V6
rather amazing for the time. But the heat killed those little guys.
grid is way closer to the cathode in a miniature tube.
heat dissipation and failure from contamination.
spread out pin spacing and use a larger glass envelope.
tube life increases.
Another thought. These tubes weren't typically used for audio. If they did go a bit noisy in a tv would you notice?
Noisy or glitchy tube in a TV would be visual.
Depends on the circuit.
The more robust miniature 12 volt tubes in a generalized manner
more common with color TV.
heater with a hotspot, or starting to short to cathode.
will sometimes " fix" itself after warm up.
TV could have strange or intermittent behavior from cold start.
red green or blue randomly go in and out.
or not hold signal well if in AGC circuit
or nothing like the old days when banging on the TV might fix it. LOL
till the tube finally just failed.
Depends on the circuit.
The more robust miniature 12 volt tubes in a generalized manner
more common with color TV.
heater with a hotspot, or starting to short to cathode.
will sometimes " fix" itself after warm up.
TV could have strange or intermittent behavior from cold start.
red green or blue randomly go in and out.
or not hold signal well if in AGC circuit
or nothing like the old days when banging on the TV might fix it. LOL
till the tube finally just failed.
Last edited:
Yes. But the 12HL7 is a Video Amplifier. A small POWER tube. A Volt of signal at the grid. And video Signal-to-noise of 50dB is excellent.
I've been watching. I had similar trouble 50 years back on a TV Tuner tube: some months in audio (headphone) service and it sounded like crap. But it was days when you could almost name your price for TV tubes, because they were going out of style real fast. Never did sort that.