Gradual HT ramp up

[Tubetec]

I was wondering if there was a simple way to do this .
Essentially mimic the way a tube rectifier slowly ramps up the Ht at turn on ,
Could we dump current at switch on for maybe 30-60 seconds , meanwhile the power supply C|R|C networks take time to charge up .
If we had a resistor and thermistor( lo-hi resistance )PTC? in series across the Ht supply then another thermistor(NTC) in series with the smoothing filters would that work?
Lets assume a passively smoothed HT with CRCRCR filter , 220uf caps and resistances in the order of a few kilo-ohms .
Could we incorporate NTC thermistors into the HT bridge rectifier circuit in such a way that initially they drop a lot of volts and only allow the reservoir and filter caps fill slowly.


I know theres HT delay systems involving relays they still spike the HT on the caps and tubes although the heaters will have had time to warm , or maybe some kind of mosfet that soft starts things but I'd much rather a way to do it simply, passively , two legged components only .

'It should also be noted that valves actually last longer if they have some cathode current flowing while they are heated, which again is contrary to the theory of the standby switch. Source: Radio Designer's Handbook, 4th edition'

 

[Tubetec]

Standby/relay delay is better than cold start with full HT applied but ramp up makes life much more gentle on the valve thermally and is idiot proof as far as user interaction is concerned as no front panel controls are needed .

If you have the chance use a laser thermometer to measure the cathode temp in a power tube after a suitable warming up period ,with heater voltage applied , when you apply HT the cathode temp rises sharply , meaning every other components within the glass bulb also has to get used to this 'new normal' , its obvious to me why both initial current limiting of the heaters and ramped application of Ht is a good thing and makes valves live longer ,standby puts an unquatifiable bump in the process of your tubes coming up to working temp ,so now they have to withstand cold(ambient), heaters on temp and with ht applied ,with simple ramp up strategies both in LT and HT your not forcing the structures inside the tube to bear undue temp rise/fall all of a sudden , a smooth curve , not hitting a brick wall like unloaded plate volts on cold cathodes or the thermal steping your tubes get with a standby switch . Effects on the tubes themselves aside your Ht caps wont complain about a gentle rise to volts , unlike the SMPS first filter cap which gets full mains ,rectified DC applied 'balls to the wall' ,without the loses of the transformer to buffer it every time its energised.

How long does it take a piece of tube gear like a mic or preamp to reach some kind of thermal stability where your happy to record with it ? whats a minute or two waiting for the HT volts to ramp up worth? could make the difference between 5 and 50 years or more in the life of the power supply components and save many trips to the repair man for new tubes because of thermal fatigue of one form or another .

Do you want your gear only to last till the next service interval or should we not be playing a longer game for the greater good?

 

[Winston OBoogie]

Easy to get ramping up H.T. - Use a MOSFET as a gyrator/cap multiplier. With a high value resistor into the cap on the gate.

I generally used a 10M R into a 10uF polyprop. C. * This ramped up at about the rate that felt right for a decent degree of the valve heaters warming up.

I had more stuff going on - current source into a zener or a resistor for voltage regulation. This was then fed to the 10M and 10uF. But you don't need this extra stuff.
You will need a protection diode for the MOSFET if there isn't one internal to the part. Put the end of the 10M right at the gate of the Fet, otherwise, put a small 100R right at the gate for parasitics etc.


*

= 100 seconds here of course so, adjust to taste

 

[rock soderstrom]

I am not sure if this is really necessary. Many old tube radios with sand rectifiers show that it works very well without.
The components needed for this are themselves at risk of becoming a problem, especially in long-term operation.

 

[Winston OBoogie]

A very good point Rock. Truth is, I don't really know if it's better for the tubes or not.

As far as I'm aware, none of the higher end units I built have had issues with a failed MOSFET.

My thinking at the time was more like: I want the filtering and low impedance output of a MOSFET gyrator on the H.T. line anyway, whether regulated or not. And since the part is already there, why not have a longer time constant too by putting in a much larger value resistor.

A few units of mine still have the same tubes I installed over 25 years ago. These were nice NOS European types. I'm OK with that as far as value for dollars spent. If I'm still around in another 25 years I'll see where it stands .

 

[MaxDM]

A rectifier tube tends to limit current in a similar way to a resistor

While it is true that it is not perfectly linear, like a resistor, it's close enough that you can just look at the charts and use voltage drop related to current and find a resistive value which is in the ballpark.

I did this to replace a rectifier tube in a 3W guitar amp and it worked surprisingly well, IMO

 

[rmburrow]

I like indirectly heated rectifiers like the 5AR4, 5V4, 6X4, 6AX5, and some of the TV damper diodes. Directly heated rectifiers like the 5U4, 5R4, 5Y3, etc. bring up HT before the heater-cathode types conduct. There are some circuits using tubes like the 6AS7, 6L6, etc. as series pass voltage regulators that also bring up HT slowly. There are regulator circuits in the UTC transformer catalogs, etc.

 

[Tubetec]

I found a simple ht delay circuit in a Philips tape recorder ,
It connects the lower end of the HT supply(post rc filters) to ground through a 270 ohm resistor via relay at switch on , after a predetermined amount of time the other dc rail energises the relay and the dummy load is dissconnected ,

I found a suggestion on another page about using a bimetalic thermostatic switch as a ht delay , the same kind found in almost any home appliance,

I see thermostats come in both NO and NC contacts with switching values down to 20 degrees C ,

A combination of NC and NO Stats thermally coupled to dummy load or series dropper resistances or other heat sources could form a staged ramp up circuit , like an automated standby switch without the need for human input,
resistances could be suitably proportioned so that in the event a bimetalic strip malfucntions nothing bad happens ,

There has to be an idiot proof electrical solution rather than having to resort to electronic means , with all its potential variables , over time ........
Stats of a given value are likely to be available from a nearby electrical retailer for as long as we inhabit the planet .
There normally designed to switch 10A 240v AC , if its only switching 10 mA DC it should remain reliable for a very very long period of time .

 

[abbey road d enfer]

There has to be an idiot proof electrical solution rather than having to resort to electronic means , with all its potential variables , over time ........

A bi-metallic thermostat is not my idea of a reliable and stable element. They are known to fail.
I'd rather have a circuit based on MOSFET.

 

[Tubetec]

Thermostatic switches are often rated for 100,000 cycles at 10A 250V .Switching the HT rail on a preamp is very light duty compared to real world use.
Ive seen a few go wonky over the years , I would say its very rare though ,
Washing machines based on mechanical timers and thermostats are incredibly reliable , all too often its the logic PCB that lets you down in the modern machines .
The recycling yards are full up of nearly new white goods nowadays because you cat get someone to fault find and replace defective electronic parts .

In use you would expect to see some variation in the switching temp depending on ambient/opperating temp and over many years as the metal parts fatigue it may drift , but in any case failure of the thermostat to switch at a given temp or within a certain time frame doesnt mean a catastrophic fail , it only means your HT line wastes heat , if that happens its a really simple and cheap part to replace .


I dont know how Mosfets die but imagine the device shorts , that could blow up the crap out of your tube mic or preamp .

 

[abbey road d enfer]

Thermostatic switches are often rated for 100,000 cycles at 10A 250V .Switching the HT rail on a preamp is very light duty compared to real world use.

Yet many need to be replaced. Faults are nor electrical, they are mechanical. I had a vendor who said his first income came from replacement of thermostats in electric heaters. May be he exaggerated.

Ive seen a few go wonky over the years , I would say its very rare though ,
Washing machines based on mechanical timers and thermostats are incredibly reliable , all too often its the logic PCB that lets you down in the modern machines .

I personally had to dump two washing machines because their mechanical programmer had failed and was irreplaceable.

in any case failure of the thermostat to switch at a given temp or within a certain time frame doesnt mean a catastrophic fail

Indeed.

, it only means your HT line wastes heat , if that happens its a really simple and cheap part to replace .

For a DIY product it makes sense.

I dont know how Mosfets die but imagine the device shorts , that could blow up the crap out of your tube mic or preamp .

I don't think it would result in catastrophic failure either, whatever the fail, open or short.

 

[trobbins]

Tubetec, you started this thread a few years ago with assertions about shortened life of valves, and capacitors, and concerns about equipment making it to the next maintenance interval etc. And now you are rationalising away a solution using switch contacts, where much more failure data is ever going to be available to indicate the old adage of 'jumping out of the frying pan into the fire'. I'd suggest there is scant actual failure data related to what you originally propose as a problem needing a solution, and you don't provide that data or relate it to other failure related data of this type of equipment (ie. provide a context and balanced view of what other failure events happen), or relate it to failure rate data of any type of 'solution'.

The topic of fault events in tube equipment has lost its opportunity for meaningful statistical data to be amassed - that opportunity ended more than 50 years ago. Nowadays fault events are randomly described on forums, and increasingly biased by valve parts that were made in the last decade or two (with their own issues of quality control and batch issues and different manufacturing plants). There seems to be a continuous supply of solutions seeking problems without any meaningful context.

 

[Tubetec]

My frame of reference with tube equipment goes back a lot longer than 10-20 years ,

Its pretty clear to me why SMPS tend to blow their load , often in only a handfull of years ,
look at the inrush current spike when its energised , its enough to make a spark jump with an audible click when you plug it in .
You have a mains filter , which is only a few turns of wire , you have a rectifier and you have a filter cap . If your lucky the manufacturer has included a metal oxide varistor but the ramp up time is still more or less instantaneous , that stresses the filter cap .

Contrast that with an amp with a tube rectifier , there is no spike on the Ht line at start up , just a graduall ramping up over the course of around 1 minute before it reaches more or less full voltage .

Ive had a few customers over the years who will not use the standby switch properly , despite being told ,
they continue to flick both mains and standby on at the same time and this has a drastic effect on tube life .

There are tube amps with sillicon rectifiers and no standby , of course we have a mains transformer and its associated wire resistance to limit inrush ,but it still applies full unloaded HT volts long before the cathode has a chance to warm . Again without exception amps of this kind show very poor valve life compared to amps with either a tube rectifier or properly used standby switch .

I picked up an early 70's Marshall 20W amp around 30 years ago , for a few years I was constantly changing tubes early , in the end I got tired of the wastage and installed a standby , I havent needed to replace the tubes since .

A good example of how not to do a tube amp is the Peavey classic 30 ,
it has series connected heaters in the power tubes and no standby , it runs hot enough to fry an egg on .
And it absolultely eats and shits out tubes . Peavey of course had warehouses full of decent Russian made tubes back then and at a good price , but thats no longer the case .

I have all the evidence I need to know HT ramp up times do matter . good tubes are becoming more and more hard to find at a reasonable price , so theres every reason to make sure they live as long as possible .

By using the thermostat/resistor I save myself the need for a standby switch and the chance its misused ,
I also dont need a transistorised ramp up circuit in an otherwise passively filtered HT supply .
As I said previously the metal cased thermostats are made for 10A or16A switching current , they perform this job day in day out for years on end ,for instance so your emmersion water heater or kettle doesnt boil over , Its my guess that using a thermostat to switch a HT current of only a few 10's of mA it should last many years .

If you dont like my idea Trobbins , thats absolutely fine , but this suggestion Im basing my knowledge on some bullshit Ive read online is wrong and contributes nothing to the subject under discussion .

 

[trobbins]

Tubetec, my post is not in response to your suggestion/idea of preparing or developing or installing ramp-up mechanisms - no issue from me.

What I am concerned about is that others may read this thread and get the view that their valves/parts will last longer by introducing a ramp up. That is not a fact based on failure statistics, just a generalised assumption.

While no doubt well-intentioned, you seem to add a degree of snake oil with comments like "Do you want your gear only to last till the next service interval or should we not be playing a longer game for the greater good?" and "could make the difference between 5 and 50 years or more in the life of the power supply components and save many trips to the repair man for new tubes because of thermal fatigue of one form or another".

 

[Tubetec]

As I said and I standby it , from 35 years experience fiixng tube gear , any tube amps which apply full unloaded HT volts at the moment of switch on fail prematurely and every component in the chain is stressed needlessly ,

Im not relying on someone elses failure analysis , or Hartley Peaveys long winded explanation of why he was too cheap to include a standby in the Classic(30) series of amplifier .

If you have seen a study into the effects of prematurely applying HT to tubes you might let us know .

 

[Tubetec]

https://sound-au.com/valves/ht-delay.html

 

[trobbins]

There are special applications that require 'heater on first' operating conditions - I know of one early vintage rectifier tube, and as I understand it many transmitter tubes require it, as well as mercury arc rectifier tubes, and I think there are some specialist low leakage tubes for instrumentation. I recall the vintage rectifier had lower limits of HT voltage until the tube was sufficiently heated. Transmitter tubes are often referenced when people raise concerns of 'cathode stripping'. A lot goes on in a tube, and the level of engineering involved was huge when tubes were the only device available - I can recommend RCA's 1962 Electron Tube Design book (on-line) as an eye-opener summary of what was generally known in the industry.

The topic of ramped/slow application of heater voltage is sometimes raised in modern forums as a benefit to heater life, especially when heater flash is observed. There is little to no statistical failure assessment on that, with the only rigorous assessment I've seen based on the early computers where thousands of the same type of valve needed to have very low failure rates. They used special varistors to slow the heater turn-on, which was linked to using special rotating machine supplies for the high power demand from the heaters.

Few people go out of their way to assess why a tube has failed. That's reasonable for general servicing, as time is money, and there is no incentive. Some manufacturers collated warranty returns, and seem to have made initial assessments of those returns, with good reason as that affected profits and sometimes product prestige. Heathkit were known to have power transformer failures, and there was some comments from retired service staff, and some company product service sheets that went to adding surgisters to mitigate transformer failures, but examples like that are rarely in the public domain, and difficult to statistically assess. Even nowadays manufacturers keep warranty and fault information very much 'in-house'. A recent forum thread touched on output transformer failure due to internal arcing, and the amp manufacturer said it was a rare event and so didn't have much incentive to add additional protective devices - that seems like a common situation.

Yes I'd seen Rod's webpage, and imho he adds context to the topic and doesn't go overboard.

 

[Tubetec]

Theres a few different varieties of thermostatic switch available ,
The old style button types , KSD301 , usually 10 or 16A current rating
A smaller rectangular type , KSD9700 , 5-10A
and a TO-220 type , KSD01F , <5A
Switching temp can be anywhere from 15-250c depending on type .

I was able to use PSUDII to sim the ramp up circuit ,
At switch on a 20k resistor is in series with the rectifier output ,thats followed by a bleed resistor of 50kohms ,
The heat generated in the 50k resistor opperates the thermal switch , that shorts out the 20k series resistor after a predetermined amount of time .
With the component values I had it took around 60 seconds for the HT to ramp upto 70% ,after the 20k resistor is shorted its takes another 30 seconds to reach full voltage .

Im not even incorporating an on/off switch , you connect it too the device , plug it in , heaters come on , HT ramps up then switches out the series resistor automatically ,

I think the doubts about reliabillity are totally unfounded , the thermal switch is much more robust than any mosfet based ramp up circuit and it only requires two components , it also wont self destruct due to an inadvertant short of a probe tip .

Theres no electronics involved here , its thermal/electrical considerations , backround or ambient temp will play a part in the switching time but its all good compared to full unloaded HT at start up ,

 

[merlin]

There are various components that probably do live longer/easier lives with a soft start; the transformer primary, SS diodes, maybe the reservoir capacitor. The tubes are not one of them.*

*With the exception of DC-coupled cathode followers

 

[Tubetec]

Well my direct experience tells me otherwise ,
The standby switch is far from perfect , but its proven to increase tube life when opperated correctly , otherwise why else would it exist ?
Ive seen many examples of tube amps with solid state rectification and no standby, tube life is definately reduced , I seen this on both my own amps and amps Ive serviced for other people over the years .
Ive modified a few , like my Peavey Nanovalve and my Marshall 20W , adding standby ,
Both those amps were very hard on tubes , since the mod I never had to change the tubes again .

The ramping up of the HT ensures the structures inside the valve come up to opperating temp gradually , not in a stepped manner like a standby ,

Ive also seen evidence of the breaking down of the cathode material , the contamination likely causes the intermittant noises , also worsening microphonic performance in the tubes that were misstreated in this way .

I have some astonishingly beautiful balloon valves from the early 1920's ,
I not going to be waking them up for the first time in maybe 100 years with a gigantic kick in the ball sack

What happens the moment of switch on matters , both on HT and LT rails ,
PSUDII allows you enter values for both primary and secondary ohms of the transformer your dealing with , making for good accurate estimates of inrush currents .

I had a few thermal switches from household appliances I scrapped over the years to play around with , they make a surprisingly quiet tick sound when they opperate , much less obtrusive than flicking a toggle switch .

I have a tube pre-amp psu with a traditional standby arrangement , when the HT is engaged it creates a sound in the audio path very much like someone sucking in air through their teeth , theres a sudden rush of electrons ,
with the ramp up circuit audio fades up out of the induced backround noise seamlessly , without any bumps bangs or clicks or sharp intake of breath sounds in the audio . Likewise if the power cuts for any reason , the sound just fades to fuzzy grey , no disturbance .