Hello friends!
I'm building an ultrasonic filament supply from cheap chinese "electronic transformers" sold for halogen lamps. These drive a small toroid trafo with a self-oscillating half-bridge topology at about 60 kHz :
The conversion requires inserting a big-ish electrolytic after the diode bridge so one has to limit the inrush current to avoid busting the drivers. The usual method is to replace the on-board fuse FS1 with an NTC thermistor.
Works well enough but I'm designing a replacement for the very cheap stock PCB that uses a self-oscillating half-bridge IC (IR2153). I want the new board to fit in the existing small enclosure so the extra needed parts create a severe board real-estate constraint even when using mostly SMD parts.
So I wonder if I could put the NTC off-board at the secondary output - specifically, thinking about an NTC on each leg to keep balance relative to a center tap to ground. Also note that if there were multiple tubes to be driven in parallel there wouldn't be a pair of NTCs per tube but rather a pair at the start of the heater wiring chain.
Do you see any issues with that? Can this create noise on directly heated tubes?
As always, thanks in advance for any insights,
Joris
I'm building an ultrasonic filament supply from cheap chinese "electronic transformers" sold for halogen lamps. These drive a small toroid trafo with a self-oscillating half-bridge topology at about 60 kHz :
The conversion requires inserting a big-ish electrolytic after the diode bridge so one has to limit the inrush current to avoid busting the drivers. The usual method is to replace the on-board fuse FS1 with an NTC thermistor.
Works well enough but I'm designing a replacement for the very cheap stock PCB that uses a self-oscillating half-bridge IC (IR2153). I want the new board to fit in the existing small enclosure so the extra needed parts create a severe board real-estate constraint even when using mostly SMD parts.
So I wonder if I could put the NTC off-board at the secondary output - specifically, thinking about an NTC on each leg to keep balance relative to a center tap to ground. Also note that if there were multiple tubes to be driven in parallel there wouldn't be a pair of NTCs per tube but rather a pair at the start of the heater wiring chain.
Do you see any issues with that? Can this create noise on directly heated tubes?
As always, thanks in advance for any insights,
Joris
Last edited:
"The conversion requires inserting a big-ish electrolytic after the diode bridge so one has to limit the inrush current to avoid busting the drivers."
What 'drivers' get busted ? How is that related to testing you have done with NTC in lieu of fuse that 'works well enough' ? Have you then tried the NTC placed on the 12Vac output ?
How does your steady state output current/voltage with heaters compare to existing halogen loading?
What 'drivers' get busted ? How is that related to testing you have done with NTC in lieu of fuse that 'works well enough' ? Have you then tried the NTC placed on the 12Vac output ?
How does your steady state output current/voltage with heaters compare to existing halogen loading?
In unmodified units the self-oscillating circuitry shuts off the drivers around the mains zero crossing. Below is the 120 Hz output to the halogen lamp and the HF waveform riding on it:
With a capacitor in place the unit always operates on the peak input voltage at all times. At power on with a cold filament as the output load, the inrush current is sometimes too much for the first transistor that turns on, which is T2 in the OP schematic. As these are very cheaply made, the original transistors don't have much operating current margin. My replacement boards will have better specs in that respect.
Since in modded units the switching transistors are always operating throughout the mains cycle, I derate the load down 50-60% (10-12 watts for a 20 watt unit). Below is my circuit driving an EL-34, 6.3V DC at 1.5A. As you can see there is some serious switching ringing to adress, although some of it might be due to sub-optimal breadboard wiring.
Last edited: