• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Lets settle the b+ on cold tubes issue!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
all the equipment I built over the years with soft start has not one failed or noisy tube....many have the same tubes
What I have noticed is that PLATE voltage has never been an issue when applied to cold tubes... It seemed that the SCREEN voltage was the culprit... My guess is that many Audio power Amps run the tubes at max voltages....with screen supply feed from the plate supply through a choke....ect..
At cold start-up the power supply may be at a very High Voltage due to no current draw.... these voltages may be insanely high and when fed to the SCREEN, thus a SCREEN to GRID or SCREEN to CATHODE arc....exceeding the MAX SCREEN voltage excessively...
When the tube is warmed up a bit, it can flow currents instantly upon hitting the STANDBY SWITCH, this avoiding exceddingly high screen voltages even briefly...
I believe most tube equipment such as TV's ..ect that have no stand-by switches have tubes operating in much more safer conservative ranges, thus can easily live through a cold start-up....
I do believe I agree. So does that AudioExpress article. In that article he claims a similar observation regarding the control structures being the most likely victims/culprits. He suggests that some of the excessive behavior comes from later stages trying to amplify surges from previous ones on top of the surge being suffered locally at the nearly same time.

So at this point I'm thinking a standby switch is probably a cheap and helpful idea, if used primarily to momentarily pre-start the heaters but, at this point, maybe not to maintain heat between active usage. (Although there are writings out there that say, basically, "leave 'em on, they'll last forever") Of course, electric rates do apply.

Soft starting the heaters is similarly likely to be a good idea.

These both cheap processes to implement, and there doesn't seem to be any significant downside. It seems there's no point to arguing about "need" when we're taking $5 in parts vs. a insanely expensive retube job.

Ordered startup of the amplification stages seems a bit more costly at higher B+ voltages, but still cheap compared to a retube. Again, can't think of a downside.

Baking tubes on arrival (new or NOS), seems to be easy enough as well. Seriously ticked off SO seems the only downside.

Running at lower than rated dissipation seems to be a major factor too. I'd love it if someone could scale that benefit for me. For example, would it be worth buying a set of KT120, or even KT150, an run them along the lines of a KT90 (linear load lines permitting, of course).

The only idea I've seen that I haven't really resolved in my head are...

One is the notion of starting the grid at cutoff and "slow starting" it towards the operating bias set point after full B+ is applied At, or near, the same time as we start ramping the heater current. I saw that somewhere, but I can't recall where. The idea being to more gently heat soak the assembly, and possibly, help retain the infant charge cloud.

The other is a question to if there is any likely benefit to positive pressure ventilation system, from either under the deck or above. The benefit question is open, but I expect if there is one an under deck system would cool the tubes much more evenly.
 
Tubes that are pre-biased do not have in-rush currents....
The empty capacitor can cause the in-rush....
Some reasons why inrush can be bad:
1. The capacitor can over heat and damage with excessive amps per second .... this is essentially ESR and current squared....integrated the area under that charge curve...

2. The BUS or mains that supplies the voltage may have regulations for surge current limits...
 
When not using standby (always on) on my guitar amp at turn on, I hear pops from the speaker till the tubes are fully warm.
What is the cause?
Really good question, and I don't think I have a good answer. But, some ideas...

1. Heating. Just like your glass tinkles when cooling, it suffers the same effects while heating. Metal does too (listen to a car engine cooling down). It could be that the supports/assembly are suffering those "earthquakes", shaking the assembly, and briefly disturbing the charge force balance. Then, what being an amplifier and all...

2. Highly charged particles are hitting the grids, causing a temporary "hole" in the field force, allowing more electrons to make it through. Again, what being an amplifier and all...

3. Ionization trails. If you look at a corona discharge you usually see little fingers of brightness concentration snaking around. Sort of tiny, low energy, sparks wanting to happen but they can't quite find a good path. A couple of gas molecules in the right position, and a small bunch of electrons hop scotch across them. Not enough to arc, but enough to overcome the grid's suppression. And, again, being an amplifier....

Btw, I think 2 is unlikely. It would take a pretty big bit of charge.

All three would quiet down with service. Ionization will be gettered and assembly stresses would equalize.
 
i built a lot of tube amps without B+ standby switch, no failure for far because of that, my oldest build is now pushing 5 years without any problems...

Do your builds include any features that may indirectly serve a similar function, or perhaps render it moot?

Like using a tube rectifier vs. SS type, for example? Maybe a well regulated HV supply that is very much within the tubes' specs? Even, perhaps, something like using thermistors on a dual transformer (heater separate from HV) might serve much the same function.

Does your 5 year baby use "big" output tubes? I've seen it said that signal type tubes generally live a great deal longer than than the big bottle power types.

Sorry for my obviously focused interest. I'm ok with re-tubing, but at between $500 and $1000 using current production tubes, I'm really in no hurry to do so. (The new toy is a 5 channel, 10 tube KT90, Ming Da). I'll mod it eight ways to Sunday, if it will help. I just need the information to do so in an both an informed and integrated way.

BTW, the thing came with a standby switch. Love to be sure if, or how, to use it.
 
Running at lower than rated dissipation seems to be a major factor too. I'd love it if someone could scale that benefit for me. For example, would it be worth buying a set of KT120, or even KT150, an run them along the lines of a KT90 (linear load lines permitting, of course).
+1
Conservative operation seems like the best recommendation to me. People routinely use resistors that are rated for at least twice the normal operating dissipation. It's standard practice. But those same people seem determined to bias their output tubes almost on the verge of redplating, and then complain when they don't last long.

The other common tube killers are:
1: Hot-switching of rectifier tubes. Don't do it unless the datasheet tells you to (e.g. gas rectifiers do need preheating).
2: Excessive screen grid dissipation, especially if the tube is forced into overdrive (mainly a guitar amp problem).
3: Excessive grid-cathode voltage in DC-coupled circuits when cold-starting. Traditionally, a pre-heating switch cured this, but these days we can use grid-cathode protection diodes instead.

Heater failure is a non-issue for indirectly heated valves, although a soft start won't hurt any more than the rock I put in my garden to keep tigers away.
 
wkirkpa said:
The idea being to more gently heat soak the assembly, and possibly, help retain the infant charge cloud.
What is an "infant charge cloud"? I have never seen that mentioned in all my valve theory books.

The other is a question to if there is any likely benefit to positive pressure ventilation system, from either under the deck or above.
For valves designed to require this, yes. For others, unnecessary.

Valves tend to be killed early by poor circuit design, including naive 'standby' switching.
 
What is an "infant charge cloud"? I have never seen that mentioned in all my valve theory books.
.
All things that go from 0 to N follow a function curve. So, at T=0 you have a hot (or cold) cathode with no free electrons (HV just hit and the assembly is just starting to charge, but the potential is still to low for them to emit). At T=N you have a fully formed cloud.At some point after T=0 you get your first electron jumping off, then the second, then more.

This, of course, happens very, very, quickly. But at different points in that function, different physics may apply.

By "infant", I'm referring to the zone of the curve where T is closer to 0. A charge cloud is starting to form, but hasn't achieved maximum emission.

For example, some observe that pre-heating the cathode reduces the tendency towards flash-over. By pre-heating we alter the function curve as related to building the charge cloud (it would presumably build much more quickly).

If pre-heating does alter the curve, and does reduce arcing, then it would seem there are, in fact, points along the curve (probably early) where less than satisfactory physics apply.

The other structures follow their own power functions, and these are designed to interact quite strongly. So, If the plate achieves any potential at all before the grid can suppress conduction, then the early cathode electrons will conduct away and extend the time the cathode spends in that lower T range of its function.

So, perhaps, by altering the power function of the grid (pre-charging in some fashion) we could also change the power function of the cathode. Perhaps preventing early conductions and shorting the time spent in the lower T range.

I have no idea, really. Thus my questions. I know power functions just are. But, I could not even begin to do the math.
 
Disabled Account
Joined 2010
Perhaps,

The answer is with the CRT the most critical tube and the most run tube that lasts for decades...how many TV's had a CRT fail..I have never had a CRT fail, magnetise yes fail no..not even in monitors in industry..I have seen them run in production environments for 40 years and even with burn lines on the phosphorous from lack of screensaver..Highest HT levels as well..

I believe that the cold start everything on is the way to go with perhaps a soft start with PTC or other. The problem is with hot turn on with heaters hot and tubes with possible conduction and discharged caps in cathodes or PSU..power surge and flash over are a real possibility.
The other problem is at power up with PSU reaching high levels of B+ with tubes cold and possibly exceeding operating limits of insulation etc.

So as with all things it should be controlled with any operation being designed in or a calculated modification. The other problem is DC coupled tubes and taking precautions as discussed on DIY audio neon or diode protection.

Things like fixed bias being up at all times before tube conduction helps..even mixed bias or CCS can help. The problem arises with Bias powered down and not being up before tube heaters and tube conduction..so perhaps a stepped approach ensuring bias is up before heater and B+ are on.

Think about it, the amp is hot and powered down the bias is off and then you power up with all NTC thermistors still hot and no bias supply OP tubes conduct almost immediately and there is no limit to conduction..your asking for a flash over or damage.

Think about a bias supply fed from the B+ transformer and that's the supply switched at the mains with a B+ switch is a recipe for disaster.
While the heaters are hot with the power switch on. Its as bad if not worse than the bias winding going open circuit after power up.

So where do you stop? perhaps a B+ turn of if Bias fails..a soft start that does not require cool down period after turn off. A slow ramped B+.
All easily doable with nothing technical. Perhaps a pre-condition circuit<<normal procedure with drive circuits in industry.

Regards
M. Gregg
 
Last edited:
A charge cloud is starting to form, but hasn't achieved maximum emission.
The matter of how and why damage to cathodes might occur if they are allowed to saturate during warm up was settled on another thread, ad nauseum.

Parts of the cathode heat up and emit first, and, if the external field allows, all available charge carriers are conveyed to the anode - this is saturation, there is no 'cloud'. Locally within the cathode, this conduction produces heat which locally increases emission which creates heat which increases emission - thus local damagingly hot spots on the cathode might form during warm up, if the cathode is allowed to saturate during the warm up transition.

Mostly, this is a hazard for rectifiers, and these are the valves where advice on B+ sequencing and/or cathode current restriction during warm up is specified. Otherwise, for normal valves, saturation during warm up is unlikely and, according to ancient literature, generally not a hazard to valve life.

So, IMO it's a settled matter in a way that tallies well with common sense and experience. B+ sequencing is unnecessary, but harmless, for the vast majority of cases. It's not very exciting or worth thinking too much about really.
 
Last edited:
I like to use two ON switches for my tube gear, the first to turn on bias supplies and filaments, the second to apply B+. It's just how I do things, and it does help in initial debugging and troubleshooting, being able to selectively turn off portions of the units.
 
Do your builds include any features that may indirectly serve a similar function, or perhaps render it moot?

Like using a tube rectifier vs. SS type, for example? Maybe a well regulated HV supply that is very much within the tubes' specs? Even, perhaps, something like using thermistors on a dual transformer (heater separate from HV) might serve much the same function.

Does your 5 year baby use "big" output tubes? I've seen it said that signal type tubes generally live a great deal longer than than the big bottle power types.

Sorry for my obviously focused interest. I'm ok with re-tubing, but at between $500 and $1000 using current production tubes, I'm really in no hurry to do so. (The new toy is a 5 channel, 10 tube KT90, Ming Da). I'll mod it eight ways to Sunday, if it will help. I just need the information to do so in an both an informed and integrated way.

BTW, the thing came with a standby switch. Love to be sure if, or how, to use it.


not that big, mine was a 6LU8 sweep tube using both ss and tube rectification in a hybrid mode...
but have done a marantz 9 clone using only SS rects, i used NTC resistors in the primary winding of the mains traffo...

not yet 5 years old but....uses ss rects...
IMG_4924_zps3ccb7f81.jpg
[/URL][/IMG]
 
I like to use two ON switches for my tube gear, the first to turn on bias supplies and filaments, the second to apply B+. It's just how I do things, and it does help in initial debugging and troubleshooting, being able to selectively turn off portions of the units.
Does it seem to help any with tube life?

I'm seeing signal tubes rated for up to 10K operating hours, power tubes in the low/mid thousands. And, also, some power tubes lasting only into the hundreds of hours. I hear about tubes in old ATT switches lasting decades in operation, with only occasional and random failures.

The Audio Express article claims to have taken sets of tubes from hundreds (300ish) to thousands of hours (2500ish). His underlying circuit be what it may, his solution would seem to have circumvented something. Perhaps he could have rather engineered out a design fault, but the solution seems general enough to correct a host of possible faults that might happen to exist any any 'ol circuit.

Other simple, generalized, solutions (like grid+ heater) may also cover a host of other evils in any particular circuit. The key is does the solution suggest it may either insure, or exceed, expected operation? I don't think we'll ever see statistical certainty - but performance vs. published specs (which are also statistically determined) is a good second.

The way I figure it is, if someone is doing something that suggests they can get a "power" output tube *consistently* up to, and better yet beyond, its rated operational life... Then I'm game for trying.

Perhaps our standby switches, for those of us with them, should be wired for both heater - and grid. Is there any downside/risk to pre-powering the grid?
 
not that big, mine was a 6LU8 sweep tube using both ss and tube rectification in a hybrid mode...
but have done a marantz 9 clone using only SS rects, i used NTC resistors in the primary winding of the mains traffo...

not yet 5 years old but....uses ss rects...
Very nice. I was going to DIY the whole thing. Lurked around here for quite awhile. Eventually joined just to look at the pictures, actually.

My trouble was, by the time I priced the transformers, tubes, and chassis - I was already around the price of this unit from cattylink. What with sockets, connectors, I was well over. And I still had to do, well, something under the deck.

So, I laz'ed out. Bought the thing and saved an immense amount of time on procurement and purely physical construction issues. And it supposedly "works", if in only a relative way.

So, if I have to, I'll rip out the deck and do whatever. Still save time and money. On arrival the P2P wiring was, well, immaculate. Many above average and reusable components, nothing stellar tho.

As for the topology, I'm not sure yet. Still RE'ing the schematic. Could have got one from the factory, but they asked fully twice the price. Heck, if I needed a "warranty" I'd have bought two and had all the spare parts in the world, zero fuss, and a free compliment of tubes - and still saved about $20US over a factory unit. Alas, I decided to pay myself about $2000 to RE the circuit myself.

I haven't even played it yet. Maybe it's glorious. Maybe it doesn't need a complete refactoring, just some"fixing".

It does have a "Standby/Operate" switch, and I'd like to play it while I'm figuring things. But I don't want to prematurely kill a set of tubes. So for now, as my particular rubber hits the road, and like the OP, my inquiring mind wants a clue about how to "best" (most safely, tube life wise) use the thing.
 
tubes have sockets because they are meant to be replaced and does not last a lifetime...
life of tubes will vary depending on how hot you run them and the class of service you put them in..
I'm totally comfortable with that. The trick seems to be in "the class of service you put them in."

I read of similar tubes living hundreds of hours, I read of them living into the low thousands, and the high thousands too. It seems, sometimes, a seemingly small detail can have an amplified effect (yes, pun intended). Infinite would be just fine with me, but double might do nicely.

I've seen commercial claims of numbers like 7000 hours for tubes rated around 2000. Can't recall where tho, been scanning just too much stuff lately. Maybe Its proprietary "magic", just wishful thinking, or maybe just a bit ethically challenged. Still, an enticing pursuit.

If it's real, then we're always allowed to figure a proprietary bit out for ourselves.

For many, differences like that are very meaningful. A good NOS tube ain't cheap. A good set of current production tubes ain't always cheap either. But, maybe, its all just a Grail.

Anyway, for now, I think I have plan. Good, or bad, I guess.
 
We keep telling you that you don't need a plan, and a poor plan might be worse than no plan.

For long valve life, run it continuously but well within its voltage, current and power limits. If this is not possible, then use a single mains on/off switch - unless it uses high power valves in which case use some automatic sequencing. Do not use manual sequencing, as users (including us) are not reliable.
 
We keep telling you that you don't need a plan, and a poor plan might be worse than no plan.

For long valve life, run it continuously but well within its voltage, current and power limits. If this is not possible, then use a single mains on/off switch - unless it uses high power valves in which case use some automatic sequencing. Do not use manual sequencing, as users (including us) are not reliable.
Yes, that will do it. And don't do anything that slows down cathode warm-up ramp time, or intentionally cools the cathode from design spec ratings as to heater arrangement.
 
I've seen commercial claims of numbers like 7000 hours for tubes rated around 2000. Can't recall where tho, been scanning just too much stuff lately. Maybe Its proprietary "magic", just wishful thinking, or maybe just a bit ethically challenged. Still, an enticing pursuit.

i never really paid attention to this, why? a tube can get broken in minutes if you drop them on the floor from a height....;)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.