Using a constant current source is very gentle on the heaters and is easy using an LM317. The heater supply AC is rectified to DC that gives the CCS the voltage compliance headroom it needs to work properly. Another technique is a negative temperature coefficient thermistor - it does require a bit of experimentation and does drop a bit of voltage even warmed up. With newish tubes with good "emission" running at just below the rated voltage doesn't seem to hurt performance. I was told that the "flashy" heater tubes were meant for series connection off the mains voltage using a ballast resistor that drops the excess voltage and limits the inrush current. This was popular with cheap radios and TV sets where a penny saved equaled more profit! I completely agree with Mr Carlson that gentle treatment of the heaters helps prolong service life. The one possible exception might be directly heated triodes where the flash helps blast off cathode poisoning (there were several tube "rejuvenators" that worked on this principle).
That was awesome Mr.Carlson. Wish I was twenty five years younger I'd go back to the electronics course I'd dropped of from and complete it successfully with so much inspiration that's coming from a person such as you. Your videos rekindle the passion for this art and science. Thank you so much and best wishes to all those aspirants out there. Good luck, cheers! 😊👍
@@MichiganPeatMoss I became a patron just out of curiosity because of that amazing capacitor tester - best $2 a month I've *ever* spent. What a great teacher.
My hat's off to Mr Carlson for spotting something I have surely missed for many decades. Seeing the 'flash' was something I associated with european-made versions of the twin triodes, or american-production parts marked with the european-standard "ECC83" type number. In all those decades I never once looked closely enough to observe just one (only) of the two heaters exhibit this trait. He's right, that can't be correct. But I am at a loss to explain it. Just goes to show that even after nearly 50 years of experience with a technology, there's always something more to learn. Probably what's kept my interest alive for so long. Thanks again for your very-observant work.
Love all your videos, so informative and helpful. It always amazes me how much you actually know and can remember. :) Sometimes I forget what day it is.
Outstanding video. This issue has always reminded me of the Phoebus cartel back in the 1920s . That was suppose to have gone away, but instead I think the practice of these sorts of shenanigans multiplied.
Ah so now I know what this little phenomenon is! Ultimately I was suspicious of 2 Eico-branded 12au7's (Mullard) and just replaced them both (VTVM and a homebrew tube amp). I kinda figured they were not quite made right, but I had no idea this was so widespread, or has a fix for that matter. Thanks for sharing. Love your videos. I've learned a ton from them.
Mr Carlson, Another flash of brilliance :). Really enjoy your tech tips. Great using that end cap as a base for the test fixture. Another useful PVC enclosure is an electrical entrance el. I have used them in photoelectric projects. Thanks again for all your great videos.
I've noticed that in series filament (rectified filament with 60 or more volts across 5 preamp tubes) preamps, that this problem is multiplied, and that with 12AX7A's the chinese rubiy tubes seem to be the best . Not only do they draw less current when on, but the cold resistance is not such a small fraction of the warm resistance. Also they seem to be very consistent from one tube to the next.
I'd guess this circuit would be appropriate for tubes that DON'T exhibit the flash problem as well. There IS an initial in-rush of current whether or not it flashes (16:06).
When I was in the NAVY the Destroyer (ship) I was on had hundreds of 12au7 tubes that were the shortest lived tubes. I replaced the with type 6189 type tube. They lasted much longer.
Just for info: I have always used a "soft start" set up for all of my vacuum tube circuits. Some have been manual, with a switch and resistor, and others have been automatic with a time delay relay and resistor. Both types have always been successful in eliminating any filament or heater flash and in saving all of my favorite (and rather expensive) tubes used in my audio circuits. I also use the same type of soft delay in the B+ power supply circuits for all of my pre-amps and power amps. Good luck and happy listening to everyone!
I have a Swan MK I H-F amplifier. How I save my expensive 3-400Z tubes is to use a Variac rotating the control on the Variac slowly upwards taking about two minutes to reach full line voltage. This way you do not shock the tube's filaments. I've had the amp for over 25 years, the tubes have not gone soft or lost power at 15 or 10 meters. I do have a matched pair of 3-400Z's stored in a cool, dry place, but they may never get used unless I get rowdy and drive the amp to 3.5KW with 100W drive from my Drake Twins. Your way for smaller tubes is great, I will have to remember this.
Cycle your tubes once a year. You don't want them to go gassey! Gas is a real problem in large transmitting tubes like 3-500, 3-400 and 4-400 tubes. It a really RISKY buying NOS (new old stock) transmitting tubes from the likes of eBay. They can be gassey and no good even though they have never been used. 73, Jim W7RY
I agree. He looks very young compared to his level of knowlege. I am in my mid 60s, a Ham by hobby and somewhat experienced with older gear, yet his knowlege and logic surpasses my expectations exceedingly. Almost seems suspicious ... hybrid human perhaps, maybe a time traveler, or maybe just really smart, who can say, but he makes excellent videos full of useful information at any rate.
Great tips. I once had a 1950's Allen 3-manual church organ which used 12AUX7 tube oscillators, one for each note. They were built into 6-note chassis and the chassis were built on to a large floor rack. There were three racks so you can imagine the number of tubes involved. I never counted them but when I switched the organ on the lights in the house dimmed for a second. It took about as much current as a clothes dryer. But it sure shook the house when I "pulled all the stops".
I've ben arguing about this for years . I knew it was a build problem . So many people look for the flash as somthing cool , when its like having a tiny ark welder on the filiment .. Great idea , defintitly adding it to my tube testers.
Going back decades, I’d seen the flash many times, but wrongly, never attached any importance to it. The demonstration shows clearly that the bare, uncoated section of the filament, which is thermally-insulated by the hard vacuum, with nowhere for the heat to go except by radiation, gets the hottest. The demonstration reminded me of the use of a thermistor to protect the filament of a 35mm projector bulb, which could be costly to replace. I never had a single failure after inserting the thermistor in series; yet the bulb life was given as only 10 hours. The thermistor had a high resistance when cold, but a very low resistance when hot. The residual resistance when hot, was small compared with the filament resistance, and so caused negligible reduction in bulb brilliance. I am left wondering whether a thermistor could be used to protect a tube filament(s) in a similar way, with the advantages of reduced space requirement, less circuitry and simplicity.
You could also use a thermal time delay relay to short out the current limiting resistor. It has the simplicity of your switch solution and the autonomy of your mosfet solution.
+Eric Wasatonic LOL, You know, just outside the shot to your left at the test fixture scene, I had a 115NO20T on the bench.... But I figured, who is going to go get one of these old thermal delay relays. Maybe like you mentioned, I should have shown it. Thanks for your input Eric!
+Mr Carlson's Lab, I use thermal relay tubes all the time, lol. I have never used them to ramp up heaters, so that's next on the list. Brilliant idea - thank you!
I just bought some 115N010s, 115NO15s, 115N020s, 115N030s, and 115N060s. The prices on those have really skyrocketed recently. It's crazy. If you have opened up that Lambda PSU on your bench, you know it has a 115 volt thermal time delay inside it, too. I think mine has a 60 second inside it. I like the way the old octal thermal time delays will let me short cycle something off and back on without waiting a minute for the time delay to it's run full cycle again.
Good Advise .........people today like automatisation but the simple way is often more reliable , in this case a switch and resistor . One has to notice that while " preheating " vacuum tubes one has to take in consideration the design of the HV power supply in order not to stress components like capacitors and their designed max voltage ratings. ( max voltage and max voltage surge ) A good rule of thumb is to use a voltage at 70% of the voltage rating of an electrolytic capacitor. Without a load on the power supply one may exceed this max voltage , ( special attention with choke input filtering is needed ) I saw well renowned amplifiers with two 250 V capacitors in series ( sometimes without a resistor/voltage dividing network ) with 490 V HV supply under load ( and when tubes are fresh !!!! ) As well as 450 V capacitors ( low grade quality ) with 435 volt HV rail ( no room for line voltage fluctuations, voltage spike at start up or ageing of the capacitors , they tend to lower /loose their max voltage rating when they age ) Besides the value is often not the bette or best one , economic considerations or audiophile overkill, rarely the right amount and quality... To extend the life of your ( precious NOS ) tubes .....: 1) SOFT start of the filaments ( MC INTOSH used to install an NTC resistor on his AC input which performed a slow start up for the filaments as well as the HV .......) Was this to simple ? 2) PREHEAT the filaments for more than 30 seconds ( Powertubes like KT88 for more than a minute ) BEFORE applying the HV SHUT DOWN the HV before the filament supply to drain the HV. Install a bleeder resistor if your amplifier does not have one . 3) USE vacuum tubes always within the limits f their electric design .... 4) DO NOT OVERHEAT OR UNDERHEAT the filaments , apply the correct voltage for what the tube was designed.....6.3 V and not 5.9 or 6.9 !!!! 5) Correct line voltage !!!!!!!!!!!!!!!!!!!!! On the funny and flashy side of vacuum tube history : " The BBC did never shut down amplifiers .......think about it ." AT LAST.............Pull up the reference voltage of the heaters of small signal vacuum tubes ( dividing network on the HV /small +small smoothing cap ) in some designs , it will avoid noisy tubes and reduce the heater to cathode voltage difference not to exceed .....( this precaution is often forgotten by modern manufacturers and leads more than once to repeating problems for which there seems no evidence other the heater to cathode isolation related problems )
I remember the new TVs with is instant on the film tubes I believe ran at half voltage that way when you turn the TV set on it would come on right away delay circuit would keep your tubes lasting longer I believe it's a good idea to test your new tubes if you see a flash like that you need to send your new tube back and get another one! Remember you paid for it a good way to test your new parts you just bought 🤗
i made a circuit for my PCL-82 tubes using a mosfet power resistor and a few smaller resistors with a capacitor to make a delay circuit.. When power is first applied the mosfet is off... allowing the power resistor to send half the current into the 16V filaments to preheat them, after a short delay the capacitor charges and switches on the mosfet allowing now the full 16V to enter the filament , by now the tube was fully preheated and no heater flash occurs. Wanted an automatic function so I can simply turn on my amplifier and not have to worry about any other steps
Remember the instant on TV's in the 60's and 70's in which the filaments were run at a reduced voltage when the set was plugged in and turned off? Did that extend or shorten their lifespan? I occasionally pick one of these up at a garage sale etc. to tinker with. Thank you
Another elegant way is to use just one resistive component, an NTC Thermistor, in series with the filament line. At t=0, the thermistor has high resistance. Current drawn through the thermistor self-heats the component, and it's resistance drops (typically 5%/C). Within a few seconds, the thermistor is much lower in resistance than at t=0, and nearly full power is being dissipated in the filaments. True, you have to make a guess about which temperature coefficient you want, and how quickly (the thermal time constant) you want the thermistor to act (and yes, the thermistor body gets hot). In the 1960's, I supplied NTC thermistors to a large scale manufacturer. Their scales had an incandescent, 7-segment, multi-digit display (this was pre-LED days), and a small thermistor in series with each lamp greatly increased the life of the incandescent lamps.
Since everyone is chiming with other ways, why not simply a series choke? I'm not sure how large it would need to be but seems like that would provide the protection against large spikes we're seeing here.
If you leave the filament on always they will never burn out. A very good way to do that is with a center tapped 6 volt secondary transformer and a switch for standby. Thermionic emission devices rule.
Mr. Carlson has such a wonderful and clear didactic style. My Dad was a Seargant Tech at WWII end with a radio operator's license He built his own current-limited power supply, lights and all, just as Paul demonstrates elsewhere. He used to fix all kinds of fifties-sixties-era HiFi stuff for his friends while I played in the basement while he worked. Some of it 'rubbed off' on me. Pertinent to this video: When I was a teen I used to fix guitar tube and discrete transistor amplifiers, organs, etc. I recall that the VOX AC100(four EL34 push-pull output the original Beatle amp) had a open wirewound resistor in a frame below a bimetallic leaf switch with contacts that closed when the wirewound heated it up - the current limit resistor was in parallel with this bimetal, mounted insulated from the wirewound heater. Clearly, it was engineered with filament flash in mind - a clever automatic solution - and this video made me finally realize just why it was designed. I should have kept up interest in electronics and done a degree in it but life interrupted and I ended up sidetracked before learning the theory underpinning all the practical repairs I had learned to do. Sigh.
What I have been doing for years is using an older style time delay relay like was used in commercial/industrial equipment back before PLC's became the norm. The kind with a octal tube socket base and a timing adjustment knob on top. Most are rated with 10 amp relay contacts so they handle a pretty good load. I ended up with hundreds of them years ago at a auction of a local relay manufacturing plant that was moving operations over seas. Went there to buy a Bridgeport milling machine and ended up coming home with the Bridgeport milling machine and hundreds of these little time delay relays along with a bunch of other goodies. Took me a while to find a use for them but I found this filament flash problem a great place to use them. The ones that have a time adjustment range of 1-60 seconds is perfect for this application. Just mount a octal socket, break the filament feed line and run it through the relay and install the dropping resistor across the relay contacts.
If your going to use mosfet might just as well put in current limiting circuit. 1ohm shunt resistor and some bias resistors and a transistor or two. Done.
Hello Mr Carlson. I enjoy and learn a lot watching our videos. I would like if you can make a video comparing the gain, quality and microphonics between similar vacuum tubes like 12AX7 ECC82 12AU7 12AT7 7025, 6l6 EL34. and how to calculate an output transformer SE or PP based on the data sheet of the output tubes. greetings from Argentina
I wished I would have known that back in the day I used to have an amplifier that would go through tubes like that get a bright flash when you turn it on it worked for a long time after about a year it's shorted it out I'd have to buy another one I was using cheap Japanese tubes at the time.. "hitchagotcha or something
You can solve the problem by placing an inrush current limiting thermister in series with the power cord, which actually drops in resistance as it heats. This also protects solid state diode circuits, though only when turned on cold. It will always protect the tubes, though, because they only flash when cold. My 11 tube RCA Color Television dot/color bar/crosshatch generator uses one of these, and always turns on gently. Amazingly this piece of equipment has ALL of its original components, which test and work perfectly.
I'd love to have some "suggestions" about using a current limited supply in place of individual circuits for each tube... time delay might change some uses quite radically.
...if the vacuum tube heaters were in SERIES, a single current limiting circuit for the entire heater string would suffice- but if the heaters were in PARALLEL, then EACH vacuum tube heater would probably need its OWN current limiting circuit.
Good stuff, and yes, even back 30 odd years ago, (when tubes were supposedly cheap ....), I did this. On my Hallicrafters Rx and Tx, I installed a big wire-wound resistor across the power switch (each rig is supplied by an individually switched power board back in my rack). So when the main switch for that particular piece of equipment is turned on, there is a soft start for all the heaters/filaments, before the power switch on the actual equipment is turned on to bridge the resistor and bring full power to the tubes. Good also for re-forming the capacitors - in that case, the HT switch is also turned on some minutes later with the power switch still 'off', so there is half LV and something like half HT in that mode. Not as good as running through a Variac, but a quick precautionary tactic.
...since practically ALL vacuum tube circuits (except battery tube radios?) use AC for the heater voltage, I wonder what effect a SERIES DIODE WITH A SHORTING SWITCH, would have had on the heater flash(?)
Very informative video. I've watched a bunch of your videos, and learned a lot, but this is my first comment. How would a hobbyist go about finding out the component specifications in that mosfet circuit? Let's say for a 60's vintage tube radio with 5 tubes with 6,3v heated filaments. I get a lot of the physics and the math, I'm just not sure what I should be referring to or where to start. Thanks a lot for your videos! I also really appreciate your thoroughness in your electronic work.
...I know I'm late- but first off- the circuits in this video, used DC for the heater voltage!! A "5 tube radio" uses AC for the heater voltage, and the FET circuit would NOT work with AC- but the "resistor and switch" circuit might work.
I don't see these modern junk tubes flare up like the old western European tungsten filament tubes did. Nickel chromium heaters generally will never flash or flare up, but just like their tungsten counterparts, before the NiChrome wire gets hot it's still almost like a dead short, for a few microseconds, until the heater filaments get hot. When the heaters get hot the resistance goes up. Old incandescent light bulbs do the same exact thing. Test a light bulb for continuity with a Fluke or even an old Simpson 260 and it measures almost like a dead short.
I did something similar with my old tube gear. I used a relay (with a simple timing circuit) to short out a current limiting resistor that I installed in series with the filament supply.
For automated operation an off delay timer set to 15 seconds with a NO and NC set of contacts, the NO set running through a resistor of whatever value and the NC side running no resistor., The timer itself running off of main power. There no nasty transistor in my tube circuit, lol.
Very interesting trick. Seems like the incandescent lamps also always burn out on ignition. I'm building a 5e3 tube guitar amplifier, and I wanted to replace the classic 'standby' switch (which I'm not convinced is doing any good) with your 'pre-heat' switch. Which resistor value would you recommend for a double 6v6, 12ax7 and 12ay7 setup (it also has a pilot 6.3v lamp)?
I've got a tube flasher buried up in my junk and had a bunch of adapters with it I mainly used it on picture tubes but it had about all of the tube adapters and settings..and running DC on those filaments would keep them a lot quitter but draw a lot more current on startup...but the European tubes used a different substance to Dope the bottom of the plate and grids and as the tube ages the material would crack and fall off and short across the filaments and cause problems everywhere in the tube...and then people would wonder why this happened...how did this substance get into the filament which is mainly shielded from the rest of the tube..running DC thru the filaments causes magnification opposite of what it happening inside of the tube so when the substance would start to fall it would be attracted by the filament and cause a short...AC across the filament doesn't cause so much of a problem..so if the filament was shorted and not open there was a pretty good chance to make the tube work again by applying AC across the filament at about 200HZ at twice the normal voltage and it would shake the snot out of the filament and the substance would fall to the bottom of the tube and no more short and give ya a little more life out of the tubes if you didn't smoke the filament in the process...Those European tubes stank ...Their theory stank...and on..and on..anybody want to get into an argument about tubes...LOL
I know this is 4 years late ... Could you make a simple automated circuit for AC filament using a triac and a small rectified voltage with a RC timer to trigger the triac ? Same principle as the FET solution you showed.
Some of the RCA CTC series tube chassis used a thermistor in the filament chain to provide a soft start. When we got a call for a dead TV, the first check was the thermistor if it was open, very common, we'd remove the body of the resistor and twist the leads together. That way, we didn't have to remove the chassis to solder in a new part. There was another thermistor in the degaussing circuit that could fail, but it rarely happened.
Amazing video Mr.Carlson, I know it is all simple to you, for me not so much ;) ... but a question, do any manufacturers use this trick in an amp, do you know of any major brands? It looks like a cheap trick that saves you big bucks.
ive been meaning to ask where you learned all that you know. vacuum tube technology has been out of favor for a lot of years, so i thought, "hes not old enough to get it on the job, so he must be a serious tinkerer. but just now you mentioned making a test tool to deal w pallets full of vacuum tubes that had heater flash so. that either means im way older than i wanna be or you absobed a lot of good info in your formative years!
Great info, I was not aware of this issue. I am getting back into vacuum tubes after 25 years in modern solid state electronics. For that fix with the MOSFET a Triac could be used for an AC system. The current will jump when the triac fires but should still eliminate the heater flash.
walking thru your tips and learning so much, i had no interest in electronics back in UnderGrad then got a job in Medical R&D and was really handicapped by the knowledge gap. i was stuck working ergonomic, man machine interface, beam guidance and materials science for Data clarity. this basic stuff should get its way into the STEM program here in the US or Scouts program for Technology. Great basics that can be used to understand circuits and know what and why of so much stuff in everyday use. Modern Day Mr. Wizard stuff that I used to watch as a kid. Thank You for providing this ground level 101 material, extremely well presented and most understandable you little test kits are great too. Now as a Patreon I want to explore all your material and get to where i can solve some on my hobby issues in electronics as well as setup all my test gear for these old radios. very encouraging and energizing for this old brain.
Back in the 1960s, before solid-state TVs, some manufacturers had an "Instant-On" feature. They would keep a few volts on the tube filaments, including the CRT, all the time. Of course, this increased your electric bill. Some TVs had a switch on the back that allowed you to turn off the instant feature. It was probably around 3-4 Volts for the 5-Volt tubes and 7-8 Volts for the 12-Volt tubes.
The approach used as a filament saver in the past were “special” heater resistors in series with the heater strings. Now we know those resistors were just NTC thermistors. Isn’t that a simpler solution? NTC thermistors are cheap, work on AC or DC, and are a simpler solution, don’t you think?
I recall from my teenage years in the 60's dismembering old radios and TVs (In those days, TV repair shops would happily give away scrap sets to hobbyists). I would frequently see an arrangement of series connected heaters with a thermistor to "soft-start" them.
Thanks Paul for touching on the main problem I encounter when buying old radios. I cringe when I hear a seller saying he plugged in his Grandfather's old radio and it didn't do anything. Cold filaments & discharged leaky capacitors put a tremendous surge load on an already marginal power supply. Better if they would sell that radio completely untested and let me bring it up on a variac. I have gotten some with exploded electrolytics that popped metal foil ribbon under the chassis in the high voltage area.
A Temperature stable Vacuum Tube. A paradigm shifter in light of many preconceived ideas dating back to the age of drifting oscillators. Realy enjoied this little demonstration.
Excellent video! You are very knowledgeable on electronics. It seems just about anything that uses electricity has issues with high current draw when powered up. Capacitors, power transformers, filament light bulbs, motors, and the list goes on. Some devices NEED full current to kick em into gear, but most electronics would be happier if that "inrush" current could be suppressed a bit, like your flashing tube example. Items that use modern cheap switching power supplies should be left alone though. They arn't too fond of a slow voltage rise when powered up. But, old skool iron transformer equipment driving tubes would be a good candidate for some kind of ramp up to full voltage instead of slamming it w/full mains power IMO. An example: I have, yes have an old Heathkit solid state stereo amp that has no speaker protection circuit or relay like most modern amps to. When I powered it up, the speakers would thump something terrible. Ya, I learned if I switched the speakers off, waited about a minute, then switched them on again, no thump. But I had to remember to do that everytime. So, I built up a simple delay timer circuit and using a hi wattage lower ohm resister in series with the main lead from the transformer to the rectifier driving the B+ to the output transistors. When the amp was turned on, the relay would be open, forcing power through the resistor slowly charging the B+ capacitors for a few seconds. After the time out, the relay would close thus providing full power from the transformer! NO more thump! Amp is pushing 50 years old now! I made the mod around 20 years ago and amp has never had a failure!
Some manufacturers purposely design it to fail - so they can sell more and prosper. If it is too reliable, they will not sell more and be bankrupted in no time.
that's what happens when shit no plan manufacture policy is applied, because existential crisis is so good for innovation XD Nature works, cause there's a plan, but no personified authority to convince anyone, which naturally led to "Church" for ex. inventing One = Themselves as mediums
This idea would have saved at least half of the tetrodes which I was forced to replace in many of the CCA and Continental AM (broadcast) transmitters over the years! (mostly Raytheon were affected by premature filament failure) Now, Svetlana is the only available vendor for many of them! Many thanks, am sending this link to many of my friends!
Found an interesting comment on a blog from Eli Duttman entitled NTC THERMISTOR as SOFT START for TUBES..."BTW, the bright turn on flash associated with Philips, Mullard, etc., made tubes is intentional. Philips and the companies they owned (including Mullard) used such construction so small signal tubes would not cathode strip, when used in combination with fast starting, directly heated, vacuum rectifiers." Not sure what Cathode stripping is and if it is applicable to this discussion?
The chemical coating on the cathode will strip off with the electrons and stick to the plates, when the tube's filaments are not heated up all the way and a signal is applied to the tube. They will also strip some at idle too, but not quite as badly.
Once again something new you have bestowed upon me. Never had heard of filament flash, thanks for the education. I could listen to you for hours, you are very knowledgable.
Wow, that 12AU7 has the worst filament flash I have ever seen being fed from a stiff 6.3 volt source. I suspect the manufacturer did something to reduce the filament wire diameter where the hot spot is occurring. The RCA color TV's in the late 1960's used a negative temperature coefficient thermistor in series with the power transformer primary to solve this problem plus, limit inrush current in the power supply. It was very effective, took about 5 seconds for the power to come up fully when switched on. I don't recommend running DC through tube filaments, it causes DC notching (metal migration problems), which shortens filament life,
Are you saying the old US tube Mfg's made better tubes than the ones coming from Europe now that everyone seems to want? I was in the tube business with Tung-Sol and later with RCA and it didn't seem to be a concern back in the 50's to 70's. Wonder if it was due to poor welding of the filaments. Blasting away the tungsten with too much heat. I enjoy your videos on the old Ham gear W2HER
Hi Mr. Carlson I have seen this often in valves - firstly, when I was little, I noticed the flash in one of the valves in our B&W telly. This was a transformerless design with all the filaments tapped across the mains via heavy duty resistor to absorb the difference. The telly ran for approx nine years like this. When I git a bit older, I read in the 1950's hobby magazine that it was known about & "didn't seem to hurt the valve". In 2014 I built the Mullard 10+10 amplifier using two EF 86's & four ECL86's. Three out of the four ECL86's flash on startup so it will be interesting to note how long these filaments last. I have always assumed that the flash was normal - partly 'cos of our telly & partly because of the article. Thank you very much for posting this - it's the first time I've ever heard of it being a problem:)
Great demonstration. Within amateur radio RF power amplifiers, many have a Soft Start circuit, which eliminates the in-rush current to the electrolytic capacitors and power tube filaments.
so some tubes like new old stock Mullard or Amperex 12AX7s "flash" brightly when you turn on the power -why ? The answer is rather technical so bear with me. The two filaments have unequal cold resistances. When power is first applied, the filament with the lower resistance passes more voltage over to the other filament, which makes it glow brightly (Ohm's Law at work). As the lower voltage filament warms up, its resistance rises to the hot resistance of the other filament, allowing both to get equal voltages. This effect is pronounced in amplifiers with AC filament supplies because larger amounts of current are available for the tubes, DC filament supplies like those found in better hi-fi equipment tend to be current limited, reducing the opportunity to "flash.
But he tested for the flash condition with just one heater at a time, so how could what you stated be true? I think the reason the flash occurs might be because the alumina coating that should run the length of the filament is missing at the end.
Very interesting. I have seen this in the past and always wondered what was going on, it seems the tubes would eventually fail or have a shorter life but never knowing the actual reason ... just buy another tube I guess. Thanks for the insight, very informative and helpful !
Did you consider a PTC thermistor in series with each filament? Same idea as the PTC thermistors used in series with TV / monitor degaussing coils. One component, no user actions.
I've been working with tubes (TV and Hi-fi, later T&M) since the early '70s, and before that watching my father and uncle (hobbyists) as a child. I'd seen cases of heater flash but never gave it much thought or knew what to call it. In my years working in repair shops I think I saw some American brands that had flash. Of course there were many brands of tubes back then that are now gone and forgotten (Philco anyone?) Thanks for teaching me something new about tubes! It's been a long time.
Yes indeed, thanks for the thorough and well-paced explanations! I'm a tube/amp newbie, and when I was rebuilding some preamps about a week ago (the paint-by-numbers way :P ) I came across a tube that did this. I think it was a 12AU7, NEC branded, and original to the amp. A quick swap took care of my concern, and now I know what happened! Cheers!
