Heathkit W5m delay timer

[trobbins]

Had you ever heard of the Stefan-Boltzmann Law applied to tube heaters before I wrote about it a few pages back?

Calm down - some of us were plowing through third year thermodynamics at uni nearly 40 years ago.

The contributions of thermal mass can't be dismissed in that video example - the delay in reaching incandescence illustrates how long it takes for bare filament to rise in temp.

The filament that is coated will dynamically transfer more heat to the coating than it would if it was just trying to radiate heat away from its wire surface - that conductive heat transfer would be a substantial path for heat flow in the temp rise time-frame [RCA 1962 p218-9] - indicating that the filament that is coated would not rise as quickly in temp as the uncoated wire. Radiation from the coating, although not very efficient due to low surface emittance, would not be insignificant due to the much larger radiating surface area compared to bare filament.

During that temp rise time the section of filament attached and extending from the terminal pin is also noticeably experiencing thermal conduction to the terminal pin (acting as a substantial heatsink). As such, if there was only a short length of exposed filament wire for the heater then the severity of filament flash would be much reduced and likely non-existent (as per the other heater in the video, and the other leg of the flashy heater).

 

[arts]

Actually,I was taking exception to the ''That's just really bad design'' part of your comment. As you seem to be very inclined towards research,you probably should have done some,and understood the context of my post, before you chose to slag my design skills.

This was,in fact,a classroom demonstration I once gave to show the correlation between the shift in vacuum tube characteristics as related to operation beyond published plate dissipation ratings,specifically vis a vis gas-related failures.This had nothing whatsoever to do with amplifier design,musical instrument or otherwise.

And while my post was,for all intents and purposes,intended to be funny,you will note that the term ''experiment'' was used.

 

[Retrovert]

Actually,I was taking exception to the ''That's just really bad design'' part of your comment. As you seem to be very inclined towards research,you probably should have done some,and understood the context of my post, before you chose to slag my design skills.
This was,in fact,a classroom demonstration I once gave to show the correlation between the shift in vacuum tube characteristics as related to operation beyond published plate dissipation ratings,specifically vis a vis gas-related failures.This had nothing whatsoever to do with amplifier design,musical instrument or otherwise.
And while my post was,for all intents and purposes,intended to be funny,you will note that the term ''experiment'' was used.

That wasn't a comment on your design skills in any way, shape, or form.

I remembered the paint-blistering comment (it is funny) without attribution or context, and in the process of noting the common red-plating issue from grid current (a big problem in many amplifiers, particularly older ones, and one I've been reading about in the context of prevent grid current) in the context of the routine practice of red-plating tubes, arcing, etc. I've also read about overheated metal tubes, and the cautions to ensure metal tubes are used only when one has verified proper circuit function. You can't see inside and it will readily melt down under adverse conditions.

My comment was not intended to reflect on you or your design skills in any way. I didn't even remember it was you or that it appeared here. My apologies for an offense which was never intended. Your comment was memorable as funny and accurate as to how hot a metal tube can get (glass ones can have the envelope sag), even if the context was long forgotten after I'd read it.

 

[JimmyEJames]

Thought for Retrovert with respect
Research by engineers' in material science goal to make better product during tube era and hoping to result in more durable machines
Tube computer from 50 years ago with hundreds and hundreds of tubes not same concerns as audio amplifier in scale of any problem you concern with
All a numbers game!
Worry over material science now losing game it is intended for other level and most so now today when hard to even find good tubes with proper welding inside
Install screen stopper, install inrush limiting on AC mains, install manual bias adjustment scheme and run tubes cool as possible for good sound and enjoy amp and not keep looking for things to worry about that not make any real world difference over long game
Well design audio circuit have sometimes original tubes lasting for 50 years as built and still test fine all day long even longer with few simple steps above
All your worry taken into account given consideration back when tubes were made well
Fight World War and go to Moon with them
Heathkit not supposed to keep you awake at night already built like tank with wide margins
Take care of power supply and last another lifetime!
Better ways to spend time on amp
JJ

 

[Retrovert]

If I understand what you are saying, one could use MOSFETs as switches to pull the grids and screens to a voltage that would cause the tube(s) to be biased in cutoff? If so, do you connect the drain of the MOSFET directly to the grid? How does the capacitance of the MOSFET affect the circuit? Do you choose a MOSFET with low drain capacitance? Which MOSFET, part number, have you used in your amplifiers for this purpose? Is the control to turn on and off the MOSFET a manual control or do you use an automated or a semi-automated control circuit?

That's a good question.

I don't see a problem because of the minor capacitance. That is a primarily a switching time issue as it causes delay for turn-off. The MOSFET is being used a switch, not an amplifier, so it has at most unity gain. But in this case when it is in circuit it is driving tubes into cutoff, so there is no output so no frequency response. When off it won't affect anything because the capacitance is minor, and on the other side of the stopper, and probably no greater than the surrounding wiring. That capacitance is problem for HF circuits, not the LF audio. I don't think Miller Capacitance applies here.

I haven't yet added the grid/screen bias MOSFET. Still working that out. I started researching lifespan and turned up this as a good idea as part of the proper power on/off approach. I learned about that from RF power tube sequencing. As I said, there is a huge volume of material to process.

That's why I was leaning towards an Arduino. Easier to get all the coding right, including the ramping up, than build a sequencer with MOSFETs and RC constants.

Someone probably wants to build it with relay logic to avoid having sand in the amplifier. I don't take my tube gear to the beach, so I don't worry about sand...

 

[Retrovert]

Tube computer from 50 years ago with hundreds and hundreds of tubes not same concerns as audio amplifier in scale of any problem you concern with

Remember that the tubes we use are the same ones used in computers for these studies, so they have the same issues. Because they had thousands of tubes the problems are statistically significant. It's like have thousands of amplifiers being studied. And the results all show that tube failures dramatically dropped when the on/off cycling stopped. That to me says this is a problem worth studying.

Install screen stopper, install inrush limiting on AC mains, install manual bias adjustment scheme and run tubes cool as possible for good sound and enjoy amp and not keep looking for things to worry about that not make any real world difference over long game

Yes, basic protection as you outline is the baseline. But think about going one step further we can increase lifespan for tubes that aren't being made anymore and that are becoming unobtainable and thus have the price go sky high.

If ramping the heater and B+ up and down with power on/off increases the lifespan by 50% that is well worth doing for minor complexity. The cost for a matched set of NOS output tubes can now exceed $100 per tube. I wish I'd stocked up twenty years ago. Who knew?

Well design audio circuit have sometimes original tubes lasting for 50 years as built and still test fine all day long even longer with few simple steps above
All your worry taken into account given consideration back when tubes were made well

Yeah, well, I've never had one of those well-designed amplifiers. Mine are Dynaco, EICO, and Stromberg-Carlson. None have the full complement of protection features they should. Some omit basic provisions. So I'm going to re-engineer them for longevitiy, even if I'm not going to the moon with them.

 

[Retrovert]

The contributions of thermal mass can't be dismissed in that video example - the delay in reaching incandescence illustrates how long it takes for bare filament to rise in temp.

I disagree with that. The thermal mass may not play a part in heater flash. The flash appears to be near the region where the tungsten was welded to the pin wires, and it happens far more often in certain tubes from one manufacturer. The variables would be coating and physical construction. Now, the coatings all seem to be roughly the same materials. Few basic patents cover all of them. Not much difference. But the physical assembly is different. There may be a bend in the wire at that point as well. Tungsten wire, being brittle, does not respond well to bending. Filament lifespan depends upon diameter, roundness, and surface pitting, all of which can increase resistance at a point.

Tungsten wire is known to have hot spot issues because of swaging issues, contaminants, crystallization, elongation, etc. This was, again, well studied in light bulbs. I was reading some work from the 1930s about bright spots in light bulbs always caused by irregularities in the wire.

So based upon the issues for tungsten filaments in light bulbs, I would expect the same physical (mechanical) issues exist in vacuum tubes.

The filament that is coated will dynamically transfer more heat to the coating than it would if it was just trying to radiate heat away from its wire surface - that conductive heat transfer would be a substantial path for heat flow in the temp rise time-frame [RCA 1962 p218-9] - indicating that the filament that is coated would not rise as quickly in temp as the uncoated wire. Radiation from the coating, although not very efficient due to low surface emittance, would not be insignificant due to the much larger radiating surface area compared to bare filament.

This is where we differ. My reading of patents and research papers suggests that the coating doesn't provide enough thermal load to stop that thermal runaway, and I think the the simpler explanation is that the filament has hot spots, particularly at or near the welds or wherever the wire is bent.

The thing to keep in mind here is that only certain tubes have a major flash issue: those from some of the Phillips factories, but not all of them. It is not commonly seen in the US tubes or Russian tubes. The coatings cannot be that different, but the welding and bending to make the pin connects can be. This is a construction issue, i.e. welds or bends, not a materials (coatings) issue.

The problem is I can't find any of the Olde Ones who are still alive to ask what they knew about this issue; they all died to thirty to forty years ago. If only I had known I would need to ask them questions, I was even living in Jersey at the time and could have interviewed them and gotten a definitive answer on the subject. Missed opportunity.

 

[dcgillespie]

MOSFETs would in fact be ideal devices for ramping bias and screen grid voltages -- but you'd only need to ramp one of them to accomplish a soft landing at shut down (or soft start at turn-on). At shut down, you could ramp the Eg1 voltage up (more negative) to bias the tube into cut-off, but a more practical way would be to ramp the screen voltage down for pentodes, or the B+ down for triode and UL designs. By placing the ramping MOSFET in the power supply section, the capacitance of the MOSFET would not come into question with regards to the audio circuits.

Dave

 

[arts]

You've been using dramma, a lot

"High failure rates...." ; No numbers, nothing seems high about anything you've posited yet. What's high?

"Shortens lifespan..." ; This is the same type of imprecise, dramatic language that TV commercials and the nightly network news anchors use to support their stories.

"Overheating ....kills cathode... over time...." ; If none of my amps "overheat" anything, am I OK... over time? I expect my tubes to last 18.7 years according to my daily hourly average use rate over the last 37 months.

"It's a pile of tubes which failed a lot less when it was left on all the time. ..." ; You seem so scientifically biased at times, and then other times, not so much.

While you do have a point in the strictest sense of scientific methodology and research,I think that if every statement were given a footnote citing origin or author of material quoted,we'd be running out of space on the server

 

[tcdriver]

That's a good question.

I don't see a problem because of the minor capacitance. That is a primarily a switching time issue as it causes delay for turn-off. The MOSFET is being used a switch, not an amplifier, so it has at most unity gain. But in this case when it is in circuit it is driving tubes into cutoff, so there is no output so no frequency response. When off it won't affect anything because the capacitance is minor, and on the other side of the stopper, and probably no greater than the surrounding wiring. That capacitance is problem for HF circuits, not the LF audio. I don't think Miller Capacitance applies here.

I haven't yet added the grid/screen bias MOSFET. Still working that out. I started researching lifespan and turned up this as a good idea as part of the proper power on/off approach. I learned about that from RF power tube sequencing. As I said, there is a huge volume of material to process.

That's why I was leaning towards an Arduino. Easier to get all the coding right, including the ramping up, than build a sequencer with MOSFETs and RC constants.

Someone probably wants to build it with relay logic to avoid having sand in the amplifier. I don't take my tube gear to the beach, so I don't worry about sand...

Got it. Thank you for the explanation. I can understand why you are leaning towards an Arduino for timing the sequence of the turn-on and turn-off events. I am very interested in seeing what the implementation of an optimum sequence will look like.

 

[JimmyEJames]

Remember that the tubes we use are the same ones used in computers for these studies, so they have the same issues. Because they had thousands of tubes the problems are statistically significant. It's like have thousands of amplifiers being studied. And the results all show that tube failures dramatically dropped when the on/off cycling stopped. That to me says this is a problem worth studying.

I do know tubes the same but not always same time though just same time in history
You not understand me the problem of filament being studied so closely when computer use hundreds maybe thousands of tubes was reliability getting real big problem fast if not addressed!
Might technician long time to service these
You not running ENIAC in your house banks and banks of heaters all coming on one time so all you need is inrush limiters and problem solved
Design elegant solution be fun but answer already at mouser for $2 why make parts count keep going up on audio amp and increase part count with transistors?
Final solution for ENIAC make tubes better same thing need today! ; - )
I know arc issue real but not problem like some people make out to be problem is bad quality tube even guy in video post 88 agree with me
I have more than one amp with 50 year tubes and they fine no arc ever I put limiter on ALL AC lines first day!
Put this on amplifier and you cannot have tube die from arc because you cannot have arc!
http://www.mouser.com/ds/2/18/AAS-920-325D-Thermometrics-NTC-Inrush-031814-web-850596.pdf

Understand what you doing and great! From internet writing article and know that ENIAC had best turn on system available during time still had failure every day
Too much for me to baby bad tube poor made with project when thermistors and other simple device handle problem on tube amp just fine
Solid state where problem is most not like thermistor! Fix that for me! : - )
I keep thinking speaker protection circuit for solid state and relay feature burn up something to check periodically never gets done and that is where high failure rate is many times BUT does serve useful purpose and save speaker's life!
Good luck with project!
JJ

Several tubes burned out almost every day, leaving it nonfunctional about half the time. Special high-reliability tubes were not available until 1948. Most of these failures, however, occurred during the warm-up and cool-down periods, when the tube heaters and cathodes were under the most thermal stress. Engineers reduced ENIAC's tube failures to the more acceptable rate of one tube every two days. According to a 1989 interview with Eckert, "We had a tube fail about every two days and we could locate the problem within 15 minutes."[21] In 1954, the longest continuous period of operation without a failure was 116 hours—close to five days.

 

[whoaru99]

I disagree with that. The thermal mass may not play a part in heater flash. The flash appears to be near the region where the tungsten was welded to the pin wires, and it happens far more often in certain tubes from one manufacturer. The variables would be coating and physical construction. Now, the coatings all seem to be roughly the same materials. Few basic patents cover all of them. Not much difference. But the physical assembly is different. There may be a bend in the wire at that point as well. Tungsten wire, being brittle, does not respond well to bending. Filament lifespan depends upon diameter, roundness, and surface pitting, all of which can increase resistance at a point.

Tungsten wire is known to have hot spot issues because of swaging issues, contaminants, crystallization, elongation, etc. This was, again, well studied in light bulbs. I was reading some work from the 1930s about bright spots in light bulbs always caused by irregularities in the wire.

So based upon the issues for tungsten filaments in light bulbs, I would expect the same physical (mechanical) issues exist in vacuum tubes.



This is where we differ. My reading of patents and research papers suggests that the coating doesn't provide enough thermal load to stop that thermal runaway, and I think the the simpler explanation is that the filament has hot spots, particularly at or near the welds or wherever the wire is bent.

The thing to keep in mind here is that only certain tubes have a major flash issue: those from some of the Phillips factories, but not all of them. It is not commonly seen in the US tubes or Russian tubes. The coatings cannot be that different, but the welding and bending to make the pin connects can be. This is a construction issue, i.e. welds or bends, not a materials (coatings) issue.

The problem is I can't find any of the Olde Ones who are still alive to ask what they knew about this issue; they all died to thirty to forty years ago. If only I had known I would need to ask them questions, I was even living in Jersey at the time and could have interviewed them and gotten a definitive answer on the subject. Missed opportunity.

I fully agree the systemic root cause is in the manufacturing process, in fact said I think it's a defect in the other thread about this flash phenomena. However, the technical root cause of why that section of filament flashes is consistent with thermal mass.

Here is the closest still frame I could get from the initiation. As you see, the heating is well away (relatively speaking) from the weld and nearly equidistant, but not quite, from the coating. Why? Because the thermal mass of the bare exposed filament is much less than the weld area and materially less than the coated area. The picture of the heating is quite consistent with greater thermal mass towards either end of the exposed filament.

It is unfortunate the frame rate doesn't allow a look a few steps back in time because I believe it would be even more clear.

 

[Retrovert]

You've been using dramma, a lot
"High failure rates...." ; No numbers, nothing seems high about anything you've posited yet. What's high?
"Shortens lifespan..." ; This is the same type of imprecise, dramatic language that TV commercials and the nightly network news anchors use to support their stories.
"Overheating ....kills cathode... over time...." ; If none of my amps "overheat" anything, am I OK... over time? I expect my tubes to last 18.7 years according to my daily hourly average use rate over the last 37 months.
"It's a pile of tubes which failed a lot less when it was left on all the time. ..." ; You seem so scientifically biased at times, and then other times, not so much.

These are not "dramatic" terms. They are ordinary, rather pedestrian and everyday language necessary used to describe the problem that is being solved. If I said the B+ voltage in an amplifier was "very high" we would understand what that meant in the vernacular as opposed to "very low". If I say it is very low, someone will say, check for a bad filter cap or something dragging it down to ground. If I said very high, someone would say check if you've lost a load resistor. Yes, it is helpful to define those as, B+ was 527 and it should be 300. But sometimes that's not necessary.

Similarly if I say a tube "dies" or the cathode "is killed" we know that means the tube has ceased to function or that the cathode's emissions have stopped. We do not always write as if for a lawsuit, with the most precise, flat terminology. Should I say "the tube has ceased to amplify, and by that I mean that its Gm has declined to the point that electrons from the space cloud no longer make transit to the plate, as controlled by the grid and screen" or "the tube died". If I need the mode of failure, it might be helpful to say, "ions poisoned the cathode".

High failure rate as opposed to low failure rate is not "drama". These are accepted terms to explain that a problem is better or worse.

I'm not dumping literally hundreds of hours of research here until I've had a chance to write it up and publish it. In the meantime I've shared conclusions that can be verified by others. You don't like it, that's on you. You are free to do your own research. Dig through hundreds of papers, patents, and books. Get footnotes, follow up on them. Request copies from librarians. Find out they've been shredded. Ponder. Look more.

I think I can say the ENIAC is a big pile of tubes. I'm adding big, to my previous statement, by the way. The ENIAC had 17,468 tubes. If that isn't a big pile, I don't know what is. The MIT Whirlwind, whose tech reports I have similarly read, had 5,000 tubes. That, too, is a big pile o' tubes.

 

[Retrovert]

I fully agree the systemic root cause is in the manufacturing process. However, the technical root cause of why that section of filament flashes is due to the lower thermal mass, despite what the research papers may suggest. I content the research papers are not talking about this specific phenomena, rather more general nature of things.

Here is the closest still frame I could get from the initiation. As you see, the heating is well away (relatively speaking) from the weld and nearly equidistant from the coating.

It is unfortunate the frame rate doesn't allow a look a few steps back in time because I believe it would be even more clear.

View attachment 785321

And you just proved my point! That bend has altered the tungsten's properties.

Yes! Altered it.

Tungsten wire is not like copper wire. The reason for this is that tungsten is a very brittle metal and it is difficult to make tungsten into wire. Here's the first writeup I found on a search: www.tungsten.com/tips/how-tungsten-wire-is-made/

Here's how it was made back in the day, with photos: www.r-type.org/addtext/add044.htm

The problem is that current carrying depends on the minimum diameter of the wire, its grain size, any voids, any surface pitting, and other properties. Back in the early 1950s, Larson and Miller worked out the lifespan of metal wires based upon these characteristics. Their work is still today considered to be the explanation for the physical behavior.

So that bend is very bad for the properties. (Oh, wait, I should say, "that bend negatively alters the properties in a significant way, specifically by increasing resistance by 18.24 percent at 500 degrees K" to avoid dramatic speech.) I specifically addressed this:

The variables would be coating and physical construction. Now, the coatings all seem to be roughly the same materials. Few basic patents cover all of them. Not much difference. But the physical assembly is different. There may be a bend in the wire at that point as well. Tungsten wire, being brittle, does not respond well to bending. Filament lifespan depends upon diameter, roundness, and surface pitting, all of which can increase resistance at a point.​

My explanation is simple and covers all of the observed phenomenon.

The problem with the Philips tubes is they bend the wire in a bad way which causes it to have higher resistance and, as per my previous writeup on the thermal runaway process, that causes the incandescence.

Quad erat demonstrandum.

 

[whoaru99]

Sorry, now you're just backpedaling.

 

[Retrovert]

It's mystery, imagination and hype that sells lottery tickets. Against the odds you can still win $1,387,264,312.03. Lots of takers. If you buy 568,765,892 tickets you get some really good odds of winning.... compared to everyone else... but there are no takers for those odds.

Don't worry about storage, the cloud will be infinite.

You are being silly.

 

[Retrovert]

Sorry, now you're just backpedaling.

No, I just explained the phenomenon and I'm not backpedaling. I believe these issues reduce tube life by a measurable and significant factor. By "significant" I mean more than 50%.

When you start posting well-reasoned facts maybe I'll believe you're in a position to comment on my work. All I see is a volume of one-liner postings to increase your posting count.

You've never responded to my detailed analysis of why your YouTube video did not contradict my physics.

My analysis, in fact, contains the clear hypothesis that some alteration of the wire (weld or bend) at the point of incandescence is true and the proposed phenomenon of runaway heating is correct. Your theory is coating, but if that were the wire has an entire uncoated region. We would see heater flash in every single tube at the uncoated region. Yet we do not! Why is the problem confined to Philips tubes? Why is it the bend that flares up? Because that bend creates higher resistance. Tungsten wire is not, as I noted, like copper. So that's where the thermal runaway starts.

 

[whoaru99]

When you start posting well-reasoned facts maybe I'll believe you're in a position to comment on my work. All I see is a volume of one-liner postings to increase your posting count.

I posted the best reasoned facts of all, a picture in vivid color.

Why would I post an artist's rendition, so to speak, when there is an actual photo?

 

[Retrovert]

Got it. Thank you for the explanation. I can understand why you are leaning towards an Arduino for timing the sequence of the turn-on and turn-off events. I am very interested in seeing what the implementation of an optimum sequence will look like.

Me, too.

I don't think the problem is very complicated, since the order is well-defined, but it does require automation. One could, of course, build this with a simple sequencer using a 3-bit counter, a 3-to-8 decoder used to drive the MOSFETs, four timers to control the duration and ramp for the on/off, and some glue. Too many chips which are more work to use than an Arduino. Amazing how we can just throw a computer at the problem and be done with it.

 

[Retrovert]

I posted the best reasoned facts of all, a picture in vivid color.

Why would I post an artist's rendition, so to speak, when there is an actual photo?

The photo just proved my point of thermal runaway.

My explanation of the physics and phenomenon was accurate and the application to your specific problem is left as an exercise for the reader.