Gradual HT ramp up

[Tubetec]

Didnt I say I was attending to heater inrush current ?

If you have something meaningful to contribute to the thread Trobbins do , if you dont dont .


I'm incorporating pair of MOV's in series with each leg of the LT transformer secondary winding , they will show around 15 ohms each at power on , dwarfing the resistance of the winding(2-3 ohms ) initially then smoothly dropping back to a fraction of an ohm in a few tens of seconds .
So instead of the heater current vastly exceeding the nominal at start up so our tubes get a gentle current ramp , not a knee in the bollocks at switch-on .

 

[trobbins]

Perhaps be a little more precise then with statements like in #45. It would be sad if people got the wrong impression of 'stress' and thought their valves would die 'early' unless they changed their practices.

NTC, not MOV.

Which reminds me of 'heater flash', and how many people thought that was obvious and blantent overt stress that would surely lead to premature failure.

 

[Tubetec]

Im still waiting for anyone to tell me of a reliable tube guitar amp with silicon diodes and no standby ,

Heres another example of an amp that Ive owned , that has no standby ,thats hard on tubes ,
I sold it onto a friend and in turn it was sold to another guitar player I know , so I was able find out how it worked out , it was nowhere near as unreliable as the classic 30 ,but still after only around a year of home usage ,you'd need to change out the power tubes .

Seeing as your so vociferous on the subject Trobbins , maybe you know of one ?

 

[beatnik]

Preamps and all kinds of tube pro audio gear never have a standby. Do you think small signal tubes are not affected by this problem because the currents involved are smaller ? I guess the current capability is proportional to the physical size of the tube elements so I would imagine if this problem exists with power tubes then it's the same with preamp tubes ?

I am seeing a lot of people complaining modern production power tubes really not having the same maximum ratings as the vintage tubes the amplifiers were designed on (despite the datasheets claiming otherwise) so it might be the lower quality of modern production tubes is increasing the frequency of tube replacements, until the circuit is tweaked to be a bit less demanding for the power tubes ?

What about a power relay to delay the HT line ? would that be more reliable than the kind of switch you suggested ?

 

[5v333]

Varistors in series in order to constrain inrush current to mains capacitors and/or heaters seem like a brilliant idea.

Can you find a part that increases in impedance rather then decreases when voltage across it is high?

Can you find a part with low enough impedance at lower voltage across it?

Will the part be avaliable in a couple of decades when it fails and need replacement?

 

[Tubetec]

Its not to say small signal tubes arent effected but the fact that you always have large series resistances and capacitances in the filter network from the supply means it takes time before the HT appears . This certainly isnt the case with the power tubes which are subject to the full force of the spike at turn on .

My experience with these amps I mentioned goes back many years , when good Russian made Sovteks were cheap and plentiful , although I agree the quality of many modern tubes is suspect .

There are NTC and PTC devices , there very commonly used in SMPS ,
Im happy with the way the bimetal switch works for the HT ,
I will be trying NTC devices in series with the heater supply .


Maybe a PTC device like a Polyfuse could present a low resistance at switch on for the HT but these devices have curve with a sharp knee and are designed to opperate in a switch like manner in a matter of a few seconds .

I already showed the relay based HT delay from the Philips tape machine earlier in this thread , I prefer the bimetal switch shorting a dropper resistance approach as it doesnt require an auxilliary rail to energise the relay coil instead it relies on the heat wasted in the resistances to opperate .

 

[trobbins]

I've been able to use an NTC for heater soft-start as a way to avoid B+ surge - section 5 of link: https://www.dalmura.com.au/static/Power supply issues for tube amps.pdf

 

[Tubetec]

First I tried a pair of MF72 type NTC's in series after the rectifier and then in series with the winding , although I used a dummy load resistor in place of the fillament it very effectively supresses the inrush current spike


Im loosing around 0.8V dc across each NTC after they have a chance to warm ,
I started out with a 15v 15VA transformer, with around 300ma dc flowing I end up with 13v at the PSU output , but I need around 14.6v

The test so far is with the NTC's out side the PSU enclosure , they run moderately hot to the touch but not hot enough to burn , air currents have an influence on the voltage you get , for instance if you blow on the NTC it cools and reduces the voltage .
I calculated their resistance value under load to be around the three ohm mark , but I think they can go to around 1ohm at max current .
The other thing is the max reservoir capacitance is specifed at 110/220v , I see the value quadruples for 110v , I have around 6600uf with a 15V winding and the NTC's seem happy enough so far .

MF72 1209 is whats written on the devices , but I couldnt match it up in the catalogue .

 

[Matt Syson]

It is interesting seeing the 'dancing around' of myths and assumptions but only a few really 'get' the details of valve failure and how to prevent it. leaving heaters powered while there is no HT can cause cathode poisoning where the creation of the electron cloud is simply 'hanging around' the cathode so impairing it's initial performance. Usqing a NTC resistor in both the heater and HT supplies could be a practical solution to the perceived and probably real thermal shoch of switch on from cold after which a relay perhaps would sort the NTC so that normal supplies are resumed after heat up. Applyi,ng FULL HT with cold cathodes can also cause damage to screen or other electrodes so you have to consider a range of scenarios if you really want to 'protect' your valves AND it needs to be 'idiot proof' so deleberate or accidental switch on/off cycles consider the conditions of the valves. Silicon rectifiers 'should' have current limiting resistance to emulate the soft characteristics of valve rectifiers. Using multip^me transformerrs with a form of relay interlocking was used in some applications if high reliability was required. Ensurin g correct negative bias of the output valves when they are 'hot' and HT is on is also important. Decide what you are trying to 'fix' then work out a proper regeime using real physics not 'old wives tales'.

 

[beatnik]

it really seems like while a tube rectifier adds a bit more cost, in the end it potentially saves a lot of trouble.

why would you need to short the NTC with a relay ? I thought its resistance basically becomes uninfluential.

 

[Tubetec]

Thanks Trobbins , Matt and Beat

https://www.cantherm.com/wp-content/uploads/2018/08/MF72_AUG_2018.pdf

Type MF72 012D9 seems to be what I have ,
12 ohms cold , 1A current , 0,652 ohms at max current .

They get a fair bit hotter inside the psu enclosure and emit a peculiar smell of epoxy cement cooking .
They quote a max capacitance of 100uF at 240vac , 400uF at 120Vac so I think its safe to assume at a lower voltages even larger capacitances are permissible . If the permissible capacitance quadruples when you half the voltage , then it should follow that at 30Vac 6400uF is allowable ,

I had been thinking about the interlocking the elements , but in my mind it seemed to introduce another set of 'what ifs' into the equation , The Thermal switch and dummy load on the Ht works very nicely standlone ,
The two thermistors work great for supressing the heater inrush , down side is they run hot all the time and loose me heater voltage when the device is powered , in the current supply Im working on that means I need to move from a 15vac transformer to a 20v unit ,
I could try another two thermal switches to remove the thermistors , but theres no chance the switches will opperate at the same moment in time .

Maybe a single thermistor in series with the LT+ , between rectifier and first filter and a thermal switch to short the device out towards the end of the HT ramp up cycle , activated by the heat of a resistance in the HT line .
So two thermal switches and a thermistor get the job done with no worries about device mismatch .

 

[Matador]

leaving heaters powered while there is no HT can cause cathode poisoning

Applyi,ng FULL HT with cold cathodes can also cause damage to screen or other electrodes

So what are the specifications for these two criterion? How long can I leave filaments powered without HT voltage before poisoning occurs? What does "without HT voltage" mean? Like anything below 25V? 0V? Also, how do I definitively know the cathodes are no longer "cold" before I can apply HT to know I'm in compliance?

 

[Tubetec]

A laser thermometer could be used to measure cathode temp , but looking at the tube in darkness will give a good idea of how long it takes for the cathode to warm fully .

I did leave an amp on standby for a few days before , didnt appear to do any lasting damage .
Morgan Jones discusses standby in his books , he seems to favour a reduced heater voltage along with a small amount of HT present to draw off electrons from the cathode while the amp is in standby .

Tube rectifiers give all the components an easy start , but of course it will eventually fail , sometimes they can fail in a bad way that puts the transformer at risk . That being said Ive seen rectifier tubes that lasted many years .

 

[Tubetec]

Meant to ask ,
https://www.dalmura.com.au/static/Power supply issues for tube amps.pdf

Are these your own writings Trobbins ?
In any case its very nicely written article .