Heathkit W5m delay timer

[Retrovert]

Ouch, it appears I may have unintentionally poked a nerve. Looks like you've done a lot of research and given a lot of thought to this.

Well, not so much a nerve, as repetition of the same issues I previously addressed in this thread and elsewere, and about which others have extensively written. I've above explained heater flash, which is identical to stress in lightbulb filaments. (Forcing so much current through a filament that it glows like a light bulb is a bad idea.) RCA put out entire books about tube construction, which explain how the cathodes function, and others have explained the coalescing of emission islands to form an emitting surface.

This issue of heater damage has been extensively studied and the original sources are out there. In transmitting tubes power-on/off cycling destroys the filament. The discussion in the RF arena is about "black heat" (idle is 25-40% of operating temperature) vs "orange heat" (idle is about 80% of operating temperature). It turns out that moving from orange heat to fully hot is not traumatic, but moving from black heat to fully hot is very traumatic. Oh, people say, that's for RF tubes and our tubes are different! Really? They don't have cathodes and plates and filaments and grids? Oh, sorry, didn't realize that.

The Miller-Larson effect explains grain reorientation in the filament at about 600 to 700 degrees C which makes the wire brittle. Over a hundred years ago the lifespan of a 2,000 hour tube was dependent upon the thickness of the filament, and it was defined as a reduction in about 10% of the thickness. That indicator really hasn't much changed. We know that current inrush into a cold filament is many times the operating current, sometimes as much as a tenfold increase, and this causes damage to the tube and burnout.

We know (not disputed) that filament lightbulbs burn out because of on/off cycling and they generally die on inrush with a flash just like heater flash. Leaving a filament lightbulb on and running it at a lower voltage makes the bulb last a very long time. Decades, in fact. Sometimes longer. One Edison bulb is famous for having been continuously on for a hundred years because runs at low voltage.

Reducing thermal stress improves lifespan. This is particularly bad with RF tubes as their filaments are very high power. The VOA had its member stations not turn off their tubes, because this generated a 20:1 ROI. (Tubes are expensive, standby electricity for orange heat is inexpensive.)

But in HiFi land, when a tube dies the response is, oh, it's a consumable, just replace it. Guitar amps burn out tubes at a furious clip because the limits are exceeded. No surprise there. This is just an expected outcome.

Extending tube lifespan is very easy. Elevate the heaters to stop cathode-to-heater leakage, add stopper resistors to prevent arcs, don't exceed grid or screen dissipation limits, regulate the heater turn on current, don't bake the tube with excessive heat, add flyback diodes to prevent arcing in tubes from flux collapse, add over-voltage zeners to the grids and screens, etc. All basic stuff that we shouldn't be arguing about because the benefits ought to be obvious. Yet these positive improvements are honored more in the breach.

People ought to do their own research instead of blindly believing lore which is facially problematic, if not obviously wrong. As Josh Billings observed in 1874, "Wisdom don’t consist in knowing more that is new, but in knowing less that is false."

 

[drillher]

Tubes 4 Hifi sells a delay board kit for 35.00 on their website.

 

[Retrovert]

Tubes 4 Hifi sells a delay board kit for 35.00 on their website.

That is only for B+ and uses a relay. Over time the arcing will chew up the contacts and they'll spot-weld together. I don't know if it handles hot-switching. It only adds 17 seconds, according to the specifications, which is not enough to ensure the heaters are hot. Particularly if one is using delayed heaters.

You can build your own with a MOSFET for a fraction of the price and get better service.

You don't want the B+ to suddenly turn on full value, a ramp up is better.

Also this doesn't ramp up the heater voltage to gradually heat the tube.

 

[sKiZo]

The fact that manufactures of consumer electronics:

(a) did NOT understand this issue in the 1950s or 1960s
(b) could not add circuitry that competitors did not and still remain cost-competitive
(c) repeatedly defied specified limits for tubes and other components
(d) could not simulate designs
(e) used design topologies we would not​

I'd add ...
(f) And got away with it, keeping many tube lovers happy for many many years.
(g) Much of this equipment is still in service today, untouched.
(h) Restoration with modern day components, following the original design and spec, works wonders.
(i) Tweaking the design can certainly result in improved performance and sound quality, but remains optional.

 

[Retrovert]

I'd add ...
(f) And got away with it, keeping many tube lovers happy for many many years.
(g) Much of this equipment is still in service today, untouched.
(h) Restoration with modern day components, following the original design and spec, works wonders.
(i) Tweaking the design can certainly can result in improved performance and sound quality, but remains optional.

Which I'd modify as:
(f) ... and reduced tube lifespan which was never noticed by consumer because tubes were seen as consumables like lightbulbs, and who really tracks the lifespan on a lightbulb? (See: en.wikipedia.org/wiki/Phoebus_cartel )
(g) ... until one day those capacitors fail taking out an entire string of tubes which cost half a month's rent or half a mortgage payement.
(h) ... for both performance and lifespan.
(i) ... provided you have plenty of money to replace tubes with shorter lifespans because of preventable failures caused by heater and cathode damage, or exceeding limits set forth in the datasheet.

In my car (3 litre, 6-cylinder high-performance engine) I only use synthetic oil and change it every 6,000 miles. I could use ordinary oil, save myself 30% on the oil change, and change it every 10,000 miles like a lot of people and save even more. I have heard of people who lease cars and change it almost, not quite, nearly never. Not their problem. Car still runs until one day it doesn't, and the engine is shot. Some people expect a car to last 50,000 miles, others 100,000 and still others 200,000. A lot of this is design, but a lot of it is maintenance.

Tweaking a tube amp for lower stress on the tubes is like maintenance. You don't have to do it, but doing it makes everything last a lot longer.

 

[whoaru99]

That is only for B+ and uses a relay. ...It only adds 17 seconds, according to the specifications, which is not enough to ensure the heaters are hot.

17 seconds is in the ballpark.

Here's a measurement I made in my amp that uses SS rectifiers. B+ is the gold trace.

Note at power on the B+ begins to rise (relatively) slowly as the caps are charged through resistors. But, around 12 sec after power on the B+ plateaus until the time delay shorts out the resistors at roughly 15 seconds. The plateau starts where the tubes are beginning to heat up and conduct. Doesn't necessarily mean they are fully warmed up, but enough to keep the voltage from overshooting as it would without the time delay.

(btw, I made my TDR board have some range of adjustment via a trimmer).

 

[Retrovert]

Under normal circumstances it's probably close. It isn't clear that all heaters are fully hot at 17 seconds under normal circumstances. I've seen measurements of the current load which suggests it can take over a minute for some tubes to fully stabilize. Harper's work showed 14 seconds for a 6SN7, but there's a tail on that curve.

If the heaters are being delayed to 120 Sec for fully hot, however, then 17 Sec of B+ delay is insufficient.

Remember, heating in 14 seconds is 156 °C / Sec which is enormous. Using 120 Sec changes that to about 18 °C / Sec, a far gentler rise which gives the W time to equilibtrate and recrystallize without shock. There certainly will not be any heater flash with that gentle rise.

 

[trobbins]

Bob, can you link to the relay board? Does it have one or two contacts to use for power control, and are they 250VAC rated?

If so, then they should only really be used for mains AC side switching, or with a push for secondary HT CT switching (preferably with series contacts) - which looks to be your only option given that you want B+ delay.

With ss diodes (apart from higher B+ issue) or even valve diodes, you are likely to get multiple contributors of turn-on surge through the heated and ready to fully conduct KT66's due to the 100ms time constant of the cathode bias, plus the 70ms time constant of the grid coupling circuit with the grid conducting current, plus the CLC filter overshoot peaking around 50ms (if you set up PSUD2 to check). That is looking like a very bad surge scenario for your KT66's, compared to the original circuit without B+ delay (for either diode type option).

Ciao, Tim

 

[whoaru99]

Under normal circumstances it's probably close. It isn't clear that all heaters are fully hot at 17 seconds under normal circumstances. I've seen measurements of the current load which suggests it can take over a minute for some tubes to fully stabilize. Harper's work showed 14 seconds for a 6SN7, but there's a tail on that curve.

If the heaters are being delayed to 120 Sec for fully hot, however, then 17 Sec of B+ delay is insufficient.

Remember, heating in 14 seconds is 156 °C / Sec which is enormous. Using 120 Sec changes that to about 18 °C / Sec, a far gentler rise which gives the W time to equilibtrate and recrystallize without shock. There certainly will not be any heater flash with that gentle rise.

In my case, I'd say roughly 25 seconds to effectively full cathode heat given the cathode current flat lines (as measured by voltage drop across cathode resistors - blue trace) at about 25 seconds after power on. Tube compliment is 4) "super" 6BG6GA, 2) 12AU7, 1) 12BH7.

 

[Retrovert]

Ok, so they're not hot until 25 Sec and it's turning on at 17 Sec.

That's 30% premature, which, over time, damages the cathodes.

It's easy enough to build one of these circuits with a MOSFET instead of a relay, and use a programmable delay from a 555. Waiting 180 seconds isn't onerous and even 120 seconds would likely be sufficient.

 

[whoaru99]

Ok, so they're not hot until 25 Sec and it's turning on at 17 Sec.

That's 30% premature, which, over time, damages the cathodes.

It's easy enough to build one of these circuits with a MOSFET instead of a relay, and use a programmable delay from a 555. Waiting 180 seconds isn't onerous and even 120 seconds would likely be sufficient.

Right. Although I think it's a matter of hitting them closer to orange hot than black hot as you mentioned earlier.

Excellence is OK for me. Perfection is unnecessary, IMO, as it'll probably outlive me even at the point of excellence.

 

[Retrovert]

With tubes, orange hot is the new black hot.

For $30 I'd buy a 555, a MOSFET, a capacitor, two small resistors, a zener diode, a small trimmer pot, a power resistor, and lunch, and have a much better turn-on time delay. There's really no need to buy something like that gadget.

Good enough is the enemy of longer lifespan.

If a delay made your tubes last 20% longer and perform better at the end of their lifespan would you do it?

Given that a matched set of four nice NOS output tubes is $400, I sure would.

Now, I don't know what the actual extension in lifespan is. It's very hard to determine and I've extensively studied this. In the large-scale computer setups with heater delays the overall failure rate went from days to months. For the heaters the failures dropped to thousands of hours. Many added B+ delays as well. The papers describe how the failure rate dramatically dropped when power-on stress was eliminated. The ENIAC and Whirlwind stopped failing when they were left on, and that was because heater stress at turn-on was significantly reduced.

 

[whoaru99]

Can't really answer. I don't know what 20% longer would represent.

Certainly one can appreciate in the case where there are hundreds or thousands of tubes, the nature of complexity working against reliability where one failure results in critical downtime, every precaution is probably worth it. In other cases though, I think there are reasonable concessions to be made.

 

[trobbins]

In the large-scale computer setups with heater delays the overall failure rate went from days to months. For the heaters the failures dropped to thousands of hours. ... The papers describe how the failure rate dramatically dropped when power-on stress was eliminated. The ENIAC and Whirlwind stopped failing when they were left on, and that was because heater stress at turn-on was significantly reduced.

Could you link to the papers with that detail - I wasn't able to locate anything with those conclusions when sampling through the Whirlwind papers, or searching IEEE.

 

[Sam Cogley]

With tubes, orange hot is the new black hot.

For $30 I'd buy a 555, a MOSFET, a capacitor, two small resistors, a zener diode, a small trimmer pot, a power resistor, and lunch, and have a much better turn-on time delay. There's really no need to buy something like that gadget.

I'd love to see the circuit for this.

 

[drillher]

If relay contacts will fail with use switching the B+ voltage,how else could that B+ be switched with a delay circuit?I have a phono stage that seems to waste the 6922 tube pretty fast.It has a tube shunt regulated power supply.I was thinking that a delay or thermistor might help prolong tube life.

 

[peterh]

If relay contacts will fail with use switching the B+ voltage,how else could that B+ be switched with a delay circuit?I have a phono stage that seems to waste the 6922 tube pretty fast.It has a tube shunt regulated power supply.I was thinking that a delay or thermistor might help prolong tube life.

A tube that is used within it's design limits will normally last several 1000 of hours. An amp that "consumes tubes" has some inherent problem. Playing around with delays and stabilized voltages within that broken circuit will not save tubes.
Problems might be DC related, or AC(oscillation especially in power tubes), problems might be continuous or only at
certain events ( violent power on / off that causes overcurrent ).

 

[tcdriver]

...For $30 I'd buy a 555, a MOSFET, a capacitor, two small resistors, a zener diode, a small trimmer pot, a power resistor, and lunch, and have a much better turn-on time delay.

I second the request to see the schematic for the turn on delay you describe. Thank you.

 

[Retrovert]

Using a MOSFET for a heater delay was much simpler than the relay circuit since there's no relay and no flyback. Here's the rough outline.

The circuit I like is a MOSFET (as a switch) in parallel with a resistor. An RC constant charges to the point it is sufficient to turn the gate on and allow current to pass through the MOSFET. During the time the MOSFET is off, the heater is powered by a parallel resistor sized, as per Ohm's Law, to limit current to 300 mA. When the MOSFET turns on (again, as a switch) the resistor is effectively out of circuit. Otherwise it runs through the current-limiter resistor. A resistor bleeds the capacitor to ensure that if the unit was switched it wouldn't be instant on, and would go through the same delay cycle. A diode ensures the gate voltage is always within range.

This was the equivalent of the transistor-relay circuit, but without the relay contacts. It is easy enough to build.

For my next amp rebuild I was going to add replace the RC circuit so that I could better control the curve for initial startup. (Exponential instead of linear.) I'd also considered an Arduino for that curve, as it can perfectly shape it to match measurements of heater startup current. The initial part of the heating curve has much greater current demands and likely needs a slower ramp up to allow it to equilibrate. Remember, the heater is dramatically expanding during this portion and that the initial thermal stress is very high.

But that circuit didn't handle B+ delay, which I learned was required to reduce tube stress, so it was back to the drawing board to do both. I have not yet built my B+ delay, as I'm still working it out, but have sketched out the same circuit except using the MOSFET as a voltage-controlled resistor (VCR), instead of as a pure switch.

One could use the same technique, run B+ at half voltage and then turn it fully on, but I don't like that as it suddenly jumps. Which is why I'm not using it. A slower rampup is superior because it allows the islands of charge to properly coalesce over tens of seconds. Half power to full power after heater delay is better than nothing, of course. But a rampup using an RC constant for a VCR is easily done.

The same Arduino can be used to drive both circuits. I was leaning in that direction and wanted to get my Arduino expertise up to the point I could build that.

I will draw up something for the basic MOSFET delay when I have some time and post it, but it is a very simple circuit.

 

[Retrovert]

A tube that is used within it's design limits will normally last several 1000 of hours. An amp that "consumes tubes" has some inherent problem. Playing around with delays and stabilized voltages within that broken circuit will not save tubes.
Problems might be DC related, or AC(oscillation especially in power tubes), problems might be continuous or only at
certain events ( violent power on / off that causes overcurrent ).

All true and this is excellent advice.

Many times the design limits for grids and screens are exceeded, grid and screen stoppers are omitted (and arcs blow the tubes), chokes do not have flybacks, the heater-to-cathode limits are exceeded (gradually degrading the heater insulation), etc. Many amplifiers cut corners to save costs in the days when components were much more expensive. Yes, tubes were very expensive, but that was a consumer consumable, not a manufacturing cost issue. Many others have elsewhere outlined some of those cost-savings measures, and I've harped on them as well.

It amazes me that some people think arcs in tubes and heater flash are "normal" and nothing to worry about, when they are indicative of design failure. How can an arc be considered normal function?

peterh is absolutely correct that if the amplifier isn't properly working now, adding delays won't do much. But if the amplifier is working, adding delays will prolong the life of the tube and stress them less.