12AX7 heater flare on power on

[jjasniew]

When I was a kid, my dad bought an AC/DC stereo amp with a 35W4, two 50C5s and a 12AX7. When you first turned it on, the 12AX7 heater filaments would light up like a light bulb and then calm down as the other tubes heaters came up. It never failed to amuse me - and also never failed operationally.

In my Jolida LM1875 based JD 1301, I'm running a pair of RCA smooth plate 12AX7s that I've owned for 30 years. Upon power on, their heaters do something similar. While I can understand the "mechanics" of the series connected filaments in that old AC/DC amp, I fail to grasp why this is happening in the Jolida. A quick measurement of the steady state filament voltage shows nothing amiss; 12 point something... Should I be concerned? Or will these tubes also never fail despite my observation? Thanks!

 

[jean-paul]

Cold and hot resistance differs. Although many disagree (regarding longer lifetime of the tubes/valves) it helps to have DC heater PSU's with slow start. I look at it like old light bulbs worked. They almost always failed at power on. Mains voltage at max. of sinus and lowish cold resistance was a good recipe for filament failure.

BTW that old amplifier probably had some of the heaters in series which is not good design when different tubes are used. Nevertheless the tubes were quite rugged as they rarely failed.

 

[Osvaldo de Banfield]

This kind of tubes do this effect. You can limit it using a NTC in series with the heater or any other soft starter mechanism for it.

 

[jean-paul]

NTC's will do the job but they are not the most reliable parts. At least they are less reliable than a DC reg slowly ramping up. Also no hum with DC.

What value and power rating would you recommend when using an NTC, Osvaldo?

 

[petertub]

When I was a kid, my dad bought an AC/DC stereo amp with a 35W4, two 50C5s and a 12AX7. When you first turned it on, the 12AX7 heater filaments would light up like a light bulb and then calm down as the other tubes heaters came up. It never failed to amuse me - and also never failed operationally.

In my Jolida LM1875 based JD 1301, I'm running a pair of RCA smooth plate 12AX7s that I've owned for 30 years. Upon power on, their heaters do something similar. While I can understand the "mechanics" of the series connected filaments in that old AC/DC amp, I fail to grasp why this is happening in the Jolida. A quick measurement of the steady state filament voltage shows nothing amiss; 12 point something... Should I be concerned? Or will these tubes also never fail despite my observation? Thanks!

Some philips does. Look the other way !Or place the amp out of sight.

 

[Galu]

One hypothesis suggests that the part of the filament wiring not 'heat sinked' by close proximity to the cathode heats up quicker.

My Brimar ECC807s have been flashing for 50 years, so go figure!

 

[jean-paul]

Some philips does. Look the other way !Or place the amp out of sight.

Best technical advice I ever read

 

[jjasniew]

This was the amp I "grew up with"... Fond memories of playing the Beatles White Album through this thing - into whatever speakers we happened to have lying around. I remember one has this metal frame that used to resonate...

Yes, these were series connected heaters, straight across the AC line. Those big fat 50C5 heaters against the little 12AX7, can understand the turn on transient.

Cant understand what's happening in the Jolida.

 

[jean-paul]

Can't understand what's happening in the Jolida.

Again, it has to do with cold/hot resistance of the filament. When switched on the resistance of the filament is low. Current will be higher (=more light!) and during/after heating up resistance will go up and less current will be drawn. You mention measuring in steady state shows nothing special. You should of course measure current at power on when the phenomenon occurs

So if you think that this can be improved, yes it can in various ways. If only to protect your NOS tubes. In 1966 now one cared as tubes were common and found on every corner. Also the technique to do this better was non existent and even if it was available it would have been too bulky and too costly for consumers.

If you think this is a problem many of the old tube geezers will say this was not a problem in the past decades. That is partly true as it is not really a problem when you can buy a new 12AX7 at the radio shop for a dollar. But why were filaments in oscilloscopes and TV sets (= more expensive devices) protected against this phenomenon? There must be at least some truth in it one would think. Now it is 2019 and one can do better and simply designs slow ramp up PSU's by using just one (1) cap extra. It won't hurt so no reason not to do it. I recently noticed that using a recent model and excellent semiconductor regulator for the filament is also seen as something unacceptable by some. Just LM317 seems to be accepted. I would think using all possible techniques may only lead to better performance.

Or just look the other way as suggested

 

[jjasniew]

>So if you think that this can be improved, yes it can in various ways.

In appx the same era, my grandmother had a GE TV set with the "instant on" feature. AFAIK, that was just a diode with a switch across it, which kept (I assume series strung) tube heaters warm in a stand-by mode.

I assume a carefully implemented scheme could keep this little amp "warm" in a standby mode as well, i.e. some heater current without any B+. I've no idea what a warm but dark cathode temperature would do to the longevity of the cathode emission, compared to shelf-life (OFF) of the tube.

For this particular amp with its LM1875 output, I also believe keeping the chip warm speeds up its time to thermal equilibrium and hence the amp's ability to sound better, sooner than a start from dead cold. Perhaps the two could be combined to render a beneficial standby scheme.

I have a number of kapton stick-on foil heaters, so "warming the chip" could be done without dissipating power in it. Perhaps that could be integrated with the tube heater warming scheme - (I wonder how many poles you can get in a 1/4" BAT handle ST switch? I think I happen to have one with 4)

My memory was jogged and the stereo SE amp my father bought used 35C5s, not the 50C5s. Another picture of the model I had as a kid; I think my dad got it at Lafayette vs Radio Shack.

 

[jean-paul]

Please no standby heating. Not OK with regards to energy consumption, possible fire/breakdown when unattended and the gains are too small IMO. Starting the tubes slowly is easier and safer. It won't take more than 15 to 30 minutes to have the chip at its optimal temperature I guess!?

 

[Osvaldo de Banfield]

NTC's will do the job but they are not the most reliable parts. At least they are less reliable than a DC reg slowly ramping up. Also no hum with DC.

What value and power rating would you recommend when using an NTC, Osvaldo?

A barreter, for example.

 

[jean-paul]

Had to look that one up but I read it has positive temp coefficient?!

 

[jjasniew]

jean-paul >Not OK with regards to energy consumption...

I actually have one of those "green plugs" to eliminate energy waste in my system. It's set for an hour. Typically I listen for about that long and when it goes off, I go off and do something different. As a retiree, it's easy to get lost in the sound (of some of the better jazz streams) with a couple-few hours easily getting past me before I realize what happened.

Which means I'm always listening with the equipment in a thermal transient condition. Unsure what that is compared to 1 hour, but I suspect it clicks off - just when things are getting good. I could put the heating part on a global timer to mitigate energy waste 24/7, as I have for my internet modem and router.

 

[jean-paul]

Well you already have a standby system then

 

[baudouin0]

At switch on if a small section of heater warms up slightly quicker than the rest then its resistance increases. This causes this section to dissipate even more power as w=i*i*r, and i is still dominated by the cold sections. Its like a thermal runaway of that section of heater. Once the rest of the heater warms up the current returns to its correct value and that section cools down. This is also true of series connected valve heaters too.

 

[Brice]

A lot of European tubes do that. It's rare in American tubes.
As previously said, it does not shorten the tube life or has other hill effects.
I also remember this on old TV set.
I am not a kid anymore, but still amuses me to see this

 

[6A3sUMMER]

My Telefunken ECC81/12AT7 has a flare at turn on.
They are just like that. It never failed yet.

If you are going to loose sleep over it, find something more important to worry about . . .

1. Did you fuse your amp with Both a Fast-Blow, And a Slow-Blow fuse in-Series?
Example: 2 Amp fast blow in case inrush current has a fault, and 1 Amp slow blow
in case the warmed-up current is rising due to a fault.

2. Do you have Bleeder Resistor(s) to keep the amp safe when you open it up?
Will they discharge the B+ fast enough before you remove the screws and bottom plate?
Safety first!

3. World Politics, etc.

 

[MarcelvdG]

A Haltron ECC81 in my phono amplifier also does that. I found it rather scary until someone from a Dutch valve radio forum explained the mechanism to me. As was mentioned before in this thread, the part of the heater that sticks out of the cathode tube gets hot quicker than the rest, hence its resistance increases before the resistance of the rest of the heater increases and it temporarily gets a disproportional part of the total heater supply voltage. The phono amplifier has worked fine with this for many years, but I included an NTC in my next valve project just to be sure.

 

[jjasniew]

Okay - appreciate the input and experience - I'll find something better to worry about and leave this observation back down in the noise-floor of life.