• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Ultrasonic heaters

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I made a couple of these and they worked quite well during initial build and installation. When I finished the amps and hooked them up (monoblocks) they both developed a static noise. I have traced it to the ultrasonic heater. I cannot seem to remove this static noise. Does anyone have any ideas what may be causing this and how it can be remedied. Other than the static there is no hum and the amps sound very nice.javascript:smilie(':headbash:')
 
Hmmm, guess Pete knows what he´s talking about, but honestly, I also don´t get it. You would still need the iron... I used switching power supplys to heat up my latest monos with 4xEL34 each, switching freq is 135kHz - I had to pay even more attention on how to run the wires to not get artifacts of that freq coupled in... but other then that, it works really good and saved quite some pounds....
Cheers,
Marcus
 
Administrator
Joined 2004
Paid Member
I assume your ultrasonic heating scheme is generating emi which is getting into other parts of your amplifier circuitry. How well shielded is the generator circuitry and how clean is the waveform? You probably don't want a squarewave for this purpose. You could low pass filter the generator output with some inductors and caps just above the fundamental to clean things up.

The P. Millet article drew some interesting conclusions from the experience, in terms of noise and sonics the results were totally inconclusive in comparison to dc heating. Additionally getting the filaments to heat evenly was a problem due to resonance effects in the filament. (He was heating at a couple of MHz)

Another thought, is this noise audible with all sources, no sources or just digital. (Think beat notes here.) What is the output frequency of your generator design?

I frequently use constant current heating and usually add a couple of large chokes in series with the filaments in my designs. (Several mH minimum) I don't currently use common mode chokes although some do. I also add several thousand uF on both sides of the pair of chokes..

Good luck.
 
Hello Kevin,

but...

heating DHT tubes with DC is possible, but as I understand there will be a different DC potential across the cathode, which is not good, some parts will be worn out much faster than others.

I don't understand yet why RCA and other companies said in their datasheets that for example the 2A3 tube could be used in DC or AC heaters, without any significant performance modification: I could be wrong, but I think that this DC difference in parts of the cathode can strongly affect tube parameters...
 
Administrator
Joined 2004
Paid Member
Hi Giaime,
In practice there doesn't seem to be much discernable difference, the grid bias voltage is a little different for dc heating vs ac heating, but not much.

I have been running 300B's on dc for years, and have never had a filament problem with WE, SV, JJ, Vaic, Valve Arts or even Shuguang. I've built dozens of 300B amplifiers, both SE and PP.

I have a pair of 300B PP which have ac heating, as long as the tubes are highly symmetrical in construction and well matched they don't hum, otherwise they do.

I run 2.5V dht's (output tubes only) on ac primarily because I am lazy, cheap and I have not noted serious hum problems doing so.

It is pretty hard to get quiet operation with 5V, 7.5V, and 10V dhts though, so all of these designs have dc heating.

Some argue that dc heating sounds worse, but I have not noticed any significant difference except for lower hum. I do recommend constant current heating for 300B's although this must often be tailored to the specific brand of tube, and prevents ready interchange of tubes without modifications to the current sources. (Unless of course you make them adjustable.)
 
Giaime,

As far as uneven wear on the cathodes is concerned, the electrons are in a cloud around the cathode. They are held in place by the grid. As long as the tube is used within its limits... I would ignore any stories about uneven cathode wear.

;)
 
Dear friends,

thank you all for your answers. Please note that my objections aren't derived from reading what tube gurus say on the net, it's just a physical observation, am I an engineer or not? ;)

But...

electrons are in a cloud, so for normal operation there will be no difference between AC or DC heating.

But when the cloud is momentarily gone, as in a strong positive peak signal, maybe overdriving the tube, there will be a finite time constant for the electron cloud to be re-created, and no matter how little it will be this time, it surely (if I understand well) higher if I heat the cathode with DC: because heavy electron emitting areas are reduced.

Could this be a problem, let aside premature cathode wear (maybe an issue for old NOS pricey tubes)?
 
A good year ago, I made an experimental unit to heat my 300B's at 100 kHz. It uses a square wave, which actually has some advantages over sinewave heating. Simplicity of the electronics is just one of them. Despite its simplicity it works very well, making an amp that has always had some residual hum dead quiet.

Switching artefacts such as ringing, uncontrolled current and voltage spikes can be real showstoppers in any switch mode power supply, because not only they radiate into the outside world but (even worse) they have the potential to upset every control loop in the power supply and associated circuitry. A filament supply is of course especially critical. If you use some kind of (half) bridge or push-pull circuitry, careful tuning of the dead time is very important. As is proper decoupling at the primary side of the HF power transformer (assuming you use one).

But some circuit details are very helpful in identifying the source of the problem. Don't give up!

Jurgen
 
Unbalanced current

Hi,

I have also given the "unequal emission" aspect some thought, and it might or might not be significant. In my amp, the 300B's are working quite hard, so I figured that they could use every bit of help to live a little longer.

A healthy 300B has a transconductance of 5.5 mA/V and at my setting (75 mA standing current) perhaps a tad more. Let's consider DC heating, then the "lower" half of the filament is biased 2.5 V below average, and the "upper" half of course 2.5 V above. As a result, lower half works 100*(2.5*5.5)/75=18.3 % harder than average, and the upper half likewise 18.3% less hard. Or, the difference in emission between both filament extremes is about 40%. I can imagine this can lead to a situation where one extreme begins to wear out, while the other extreme still has life left. This is of course less of an issue with a 2A3.

The significance of this is of course debateable, but working out the ultrasonic power supply has provided me with a very interesting (yet affordable) experiment. So the deal was good anyway...

Jurgen
 
SY said:
Guys, think about scale. Yes, there's a potential difference from one end of the filament to the other. But what's the potential between filament and plate even at max drive? 100-200V? The couple of volts from one end to the other is pretty insignificant in comparison.

Not really. In order to determine what the electrons are doing, you need to know the effective plate voltage seen at the cathode. Due to distance and negative grid bias, that voltage isn't very much, and 2.5V is significant. This leads to increased distortion at signal minimums since parts of a directly heated cathode can "disappear" when the voltage gradient due to heating current is less than the effective plate current. This, in turn, causes the noramlly 1.5 law characteristic to become a 2.5 or even a 3.5 law device, with the resulting increase in harmonic products and the slower rolloff of harmonic amplitude with increasing frequency. That's why 2A3s and 45s, and 300Bs run with reduced heater voltage sometimes sound better.

Not sure this has any relevance to the question of DC vs. AC heating, as AC doesn't help prevent the "dissappearing cathode" behaviour.
 
Due to the construction, the grid has a factor "mu" more influence on the electric field seen by the cathode than the plate. This factor makes the triode a useful amplifier. At typical operation points, the change in cathode-plate voltage due to filament voltage is indeed negligible, but the change in cathode-grid voltage isn't. The voltage gradient of a heater doesn't influence the emission of the cathode itself, but is does influence the electric field strength just outside the cathode, and as a result the fraction of the total emission being bled from the space-charge surrounding the cathode.

Due to the potential difference between both extremes of the filament, there is indeed a degree of extra nonlinearity near cutoff, due to the fact that the entire triode does not cut of at the same grid voltage. But a DHT filament approaces the ideal of a planar cathode far better than an indirectly heated cathode sleave, thus making up multiple times for the linearity lost.
 
Well, again, let's look at the numbers. The mu is about 3.5. Typical grid bias is -60-70. Heater voltage is 5. So as Kevin mentioned, there has to be a slight change in biasing for DC, but we're still talking about a 2.5V difference compared to 60 volts of bias and 300V on the plate. With a transconductance of perhaps 4mA/V, we're not talking about a major field difference from one end to the other.
 
Miles Prower said:
Not sure this has any relevance to the question of DC vs. AC heating, as AC doesn't help prevent the "dissappearing cathode" behaviour.

Thank you Miles, that was what I intended.
But why AC shouldn't resolve the problem? The filament has a thermal mass that cannot follow the variations of the filament voltage, so it will heat up uniformly. Or am I wrong? :xeye:
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.