[Alex Kitic]

Zitat:Original geschrieben von Rumgucker

Zitat:Original geschrieben von Alex Kitic
I am glad that you do not consider the HF AC as some anathema....

Das Problem bei Halogentrafos ist die 100Hz-Modulation.

Indeed it is: look at the following illustration:

[Bild: tube_scope1.gif]

Zitat:Original geschrieben von Alex Kitic
My solution includes two steps:
1) Added cap to improve rectification (much lower ripple) in the initial stage, at the input to the half-wave oscillator circuit.

Das ist gefährlich! Damit umgehst Du die Kurzschlussfestigkeit des Trafos.

Warum denkst du das es ist gefährlich?

Here is a nice schematics of a generic electronic transformer:

[Bild: 1.gif]

And now with short-circuit protection (Kurzschlussfestigkeit...):

[Bild: 2.gif]

Therefore adding the capacitor in parallel with the rectifier bridge will not compromise the short-circuit protection. What it will do is:

a) Improve ripple, therefore increase the DC voltage at the input of the half-wave oscillator circuit.
b) By increasing the DC voltage the diac is cut out of the circuitry, it is not needed to restart oscillation - this does not change the operation of the short-circuit protection (if present in the unit at all!!!).
c) As a result of the improved ripple, and thus increased DC voltage, the output AC voltage will also increase proportionally: approximately 1.4x, depending on the increase of DC voltage as a result of improved ripple filtering.

The output AC voltage is 40-60kHz on most units, and after the adding the cap the "dips" at 100Hz (what you call 100Hz modulation) will disappear and the result will look basically as the second half of the above illustration.

Zitat:Original geschrieben von Rumgucker
[quote]Original geschrieben von Alex Kitic
2) Adjusted number of turns on the secondary (and the primary if need be) of the ferrite core output transformer of the unit.

Ich habe am Original nur den Mindestlaststrom reduziert, damit der Trafo auch ohne Anodenstrom schwingt.

Actually, the minimum load is much more important on most electronic transformers. If the load is not high enough, and there is no "no-load" protection circuitry (basically, a fly-wheel over the output transformers, sometimes transformers are used that include the fly-wheel diode) the low load condition could damage the half-wave oscillator circuitry.

Zitat:Original geschrieben von Rumgucker
[quote]Original geschrieben von Alex Kitic
A full fledged solution would include a regulator in front of the half-wave oscillator, to get a regulated DC voltage (thus a regulated AC output) and get rid of all relevant ripple (a regulator will have additional -60dB ripple rejection).

Nein! Der Halogentrafo wird 100 mal pro Sekunde gezündet. Das ist der Trick, der einen Halogentrafo kurzschlussfest macht.

As explained above, the 100Hz ripple is what one needs to get rid of, and it has nothing to do with the short-circuit protection. With 100Hz ripple the whole thing is just an electronics engineering exercise, but without it it becomes some sort of innovation.

Adding a regulator circuit (a plain LM317 or 7805 placed on top of 250-300V of zener diodes will do!) should decrease the ripple another -60dB. After that, the electronic transformer gives out pure 40-60kHz AC, depending on the model...

Zitat:Original geschrieben von Rumgucker
[quote]
Ich habe große Siebkapazitäten verwendet, um die 100Hz Modulation zu sieben. Der elektronische Halogentrafo ersetzt lediglich einen konventionellen Trafo. Die Spannungen wurden gemessen.

Of course, high capacity is a good way to deal with it - but an even better way is to clean the ripple from the DC that will go into the half-wave oscillator. This HF AC for HF ripple DC (i.e. DC where the ripple is somewhere high above 20kHz where we should not hear it...) I have not yet tried, but intend to try on some preamplifier: could be a cheap and innovative solution.

[quote]Original geschrieben von Rumgucker
[quote]Original geschrieben von Alex Kitic
But my main interest here at the moment is creating interest in the use of HF AC for direct heated tubes, as a promising solution that might one day become "the" best sounding alternative.

Ja. Das seh ich auch so.

And that part I have tried and know very well: you can hear the 100Hz modulation as a 2nd harmonic 200Hz ripple when heating DH tubes, and thus you gain little or nothing: but adding a mere cap (like, 330uF/400V) makes all the audible hum disappear as if there was some magic involved! The direct heated tubes will have less residual hum and buzz than on DC (believe it or not). That is what I have tried, what works perfectly well (is safe, depending on the type of electronic transformer used: I have tried on Vossloh-Schwabe units that allegedly have short circuit and no-load protection, and after adding the cap they can still survive both being used without load, and having the output shorted - although that is something that is not going to happen with electron tubes: when a filament is broken, it does not make a short but rather a no-load condition).

Another thing I would like to add to safety is that it is a good measure to add a thermistor (NTC) to slow down the unit (the diodes) once the cap is added. Since a 330uF cap will require almost 100A of current burst, and very few diodes can survive that (1N4508 can, but most other cannot, and it is lottery whether you will have to change the diodes and the fuse...) - adding a thermistor will reduce the current at start-up perfectly, increasing both safety and the lifetime of the modified unit.

[Bild: Inrush%20Current%20Limiting%20Power%20NT...0curve.jpg]