I am looking for some kind soul to advise on a suspect valve. To me it looks like heater is failing still working but I am not risking it. I dont want to have it fail and cause other probs. Right hand valve toasty on one filament.... On switch on there is a disconcerting spark (seemingly from valse but its so quick hard to tell). One filament is clearly heating but others not. I guess its time to replace (it is only 8 months old and light usage!). The valve circuit must be complete for it to light up at all so I guess there is a short of some sort. My instinct is just to replace (again). Am I right to just bin it?
If the tube's heater is broken, it is no loguer usable. Which part number it is?
Perhaps your transformer voltage is too high for the tube. Else add a soft start circuit in the heater transformer primary.
Perhaps your transformer voltage is too high for the tube. Else add a soft start circuit in the heater transformer primary.
Kinda looks like part of the filament has broken and fallen down, shorting out the remaining filament wire which now glows extra hot
Sometimes a knock on the amp while on can move almost molten heater wires.
If DC heaters they can get a shock when turned on.
I prefer AC heaters as the transformer is self limiting on power up.
As yet I havent had a noisy amp due to AC heaters, the problem was always elsewhere.
If DC heaters they can get a shock when turned on.
I prefer AC heaters as the transformer is self limiting on power up.
As yet I havent had a noisy amp due to AC heaters, the problem was always elsewhere.
Applying a mechanical shock to an operating filamentary valve (as the OP's in the pictures) risks damage and shorts.
Once the filament of a valve has been damaged like the one in the picture, it's best to stop using it: transformers in the anode may be subject to excess current, for example.
DC from a battery may give excessive startup current, but properly implemented DC heating has controlled safe current limiting.
Meanwhile, a typical AC heating arrangement will give 3x to 5x the normal running current to a cold filament - not effective self limiting.
Brute force DC heating for a 300B that works, and has a soft start:
300B filament cold is about 1 Ohm
300B filament warmed up is 4 Ohm
Start with a 6.3V filament transformer that is rated for at least 4 Amps; its voltage will not drop much, so may be about 6.7V or so, under medium load.
Use a Schottky bridge rectifier, a 20,000 uF cap, then a 2 Ohm resistor, and another 20,000uF cap.
The 2 Ohm series resistor between the two 20,000 Ohm capacitors provides a soft start supply:
2 Ohms resistor in series with 300B 1 Ohm resistance (cold), the 2 Ohm gets 2/3 of the voltage, the filament gets 1/3 of the voltage.
As the 300B filament slowly warms up, the filament voltage goes up, and becomes 5VDC.
The 2 Ohm resistor may need to be adjusted to get 5VDC across the 300B. Wait for at least 2 minutes until the 300B filament voltage stabilizes, before you can see if the voltage needs to be adjusted.
The ripple voltage is about 1 or 2 millivolts, it will not cause any appreciable hum at the amplifier output.
Worked in the USA. Is working in Japan.
Have fun!
300B filament cold is about 1 Ohm
300B filament warmed up is 4 Ohm
Start with a 6.3V filament transformer that is rated for at least 4 Amps; its voltage will not drop much, so may be about 6.7V or so, under medium load.
Use a Schottky bridge rectifier, a 20,000 uF cap, then a 2 Ohm resistor, and another 20,000uF cap.
The 2 Ohm series resistor between the two 20,000 Ohm capacitors provides a soft start supply:
2 Ohms resistor in series with 300B 1 Ohm resistance (cold), the 2 Ohm gets 2/3 of the voltage, the filament gets 1/3 of the voltage.
As the 300B filament slowly warms up, the filament voltage goes up, and becomes 5VDC.
The 2 Ohm resistor may need to be adjusted to get 5VDC across the 300B. Wait for at least 2 minutes until the 300B filament voltage stabilizes, before you can see if the voltage needs to be adjusted.
The ripple voltage is about 1 or 2 millivolts, it will not cause any appreciable hum at the amplifier output.
Worked in the USA. Is working in Japan.
Have fun!
Last edited:
Have you swapped the valve with the one in the other channel? Do this first, just as an easy sanity check. If the problem follows the valve, it's toast.
All good fortune,
Chris
All good fortune,
Chris
Before you bin it heat the heater pins up with your iron or maybe suck out the old solder, squirt a bit of flux up the pin and re-solder. It's worth a punt.
Andy.
Andy.
Many thanks to all your very kind replies. They are 5U4G valves. I swapped valves for some older ones (I tend to replace before they fail) and all works fine. On balance having read comments not sure I want to risk the suspect one, it will now go in my large box of working valves that needing testing. I don't currently have such kit but maybe will find someone who does.
Thank you all.
Thank you all.
Its up to the designer to specify a transformer that doesn't do that.
You can get bad AC and DC heater designs.
My DC heater design uses 12V SMPS and it's noiseless WRT to the audio. Indirectly heated tubes though.
And Nigel is correct to a certain extent, but I would say a trafo that limits the current on cold starts is overloaded/undersized in the first place.
But I notice that if I use a larger transformer, the tubes light up faster.
If I use AA carbon cells, the heater will light up but it takes far longer than if I use a lead acid battery.
If you're worried about that, and you use DC for the heaters, you can just use a depletion MOSFET like IXTH16N20D2 as a current limiter...
And Nigel is correct to a certain extent, but I would say a trafo that limits the current on cold starts is overloaded/undersized in the first place.
But I notice that if I use a larger transformer, the tubes light up faster.
If I use AA carbon cells, the heater will light up but it takes far longer than if I use a lead acid battery.
If you're worried about that, and you use DC for the heaters, you can just use a depletion MOSFET like IXTH16N20D2 as a current limiter...
Yes, I would say so too.
In any case, attempting to throttle the startup current with the transformer will cause other problems that we can do without.
For example, the 300B requires 5.0V across the filament, but the current drawn at 5.0V varies a great deal. WEs may take as low as 1.08A, while EML300Bs may reach up to 1.4A.
If we choose a 5Vrms 10VA PT in the hope of limiting the initial power, the regulation of such a PT is usually too high to give good results.
A 50VA part may have 12% regulation (the difference between unloaded and full-loaded voltage), and it gets worse as the VA Rating lowers. A 10VA PT might have 30% regulation, or even worse, so the open circuit voltage is 6.5V or higher, forcing us to choose some power resistor to get the voltage back down to 5.0V. Now we have a resistor whose value is only optimised for one DHT (on account of the variable current); swapping in a new one 300B may shift Uf by whole percentage points.
Adding this potential error to the slack tolerances of modern line voltage(+ - 10% here) is a recipe for a shortened lifetime, for sure..
Some 300Bs also shift performance with drooping filament voltage - the TJ Fullmusic is notorious for shifting gm here.
In any case, many constructors - including the commercial suppliers - use a single PT with HV and heating windings on the same core. In this case, the PT starts with zero load on the HV , and so the heater voltage rises, & most of the power is available to really give the filament a big startup kicking.
Either way, not a great solution for managing a (pretty delicate) handmade valve costing hundreds.