Also, in case of a shutdown (fex overcurrent detected=>stop drive) the magnetising inductance combined with any capacitance in the GDT circuit will resonance away, causing accidental turn ons of both the upper and lower FETs. Exactly how depends on the FETside circuit of the drive , ie DC restore or how it is implemented.
These turnons last until the energy stored is wasted in any resistance , and thus several milliseconds.
If both upper and lower FET turns on simulataneously in a halv bridge things might go very bad . And there is nothing to do about it since the drive signals are already stopped.
Lorylaci, well it was just my guess. The stacked secondaries made me think so but I could have misread the schematic. Makes sense that it is a purpose built Xformer, but with a overdone circuit...
I know the difference of the transformer designs. However most of the ATX supplies are 2 or 1 switch forward converters today, atleast of the ones I have opened up to repair or study.
Older AT supplies was typically a halfbridge converter with BJT high voltage transformers , driven by a GDT and a TL494 controller chip or any of the similar types. These are not so sensitive for above problem with resonance. Cant be becasue the GDT transformers typically is wound on an EI core thus having leakage inductance ranging into uH, might not be a problem when operating at about 40 kHz or so.
These turnons last until the energy stored is wasted in any resistance , and thus several milliseconds.
If both upper and lower FET turns on simulataneously in a halv bridge things might go very bad . And there is nothing to do about it since the drive signals are already stopped.
Lorylaci, well it was just my guess. The stacked secondaries made me think so but I could have misread the schematic. Makes sense that it is a purpose built Xformer, but with a overdone circuit...
I know the difference of the transformer designs. However most of the ATX supplies are 2 or 1 switch forward converters today, atleast of the ones I have opened up to repair or study.
Older AT supplies was typically a halfbridge converter with BJT high voltage transformers , driven by a GDT and a TL494 controller chip or any of the similar types. These are not so sensitive for above problem with resonance. Cant be becasue the GDT transformers typically is wound on an EI core thus having leakage inductance ranging into uH, might not be a problem when operating at about 40 kHz or so.
Last edited:
I have read replies on TI's forum where GDT's caused problems in LLC designs using the 25600,I would use silicon in HB design's and GDT' in Full Bridge and Two Switch forward converter's. The addition of a pnp and a couple of resistor's added to the gate drive results in a robust design in IMHO.I have seen design's up to 300 KHz using GDT's.A good source for GDT's is Pulse P0585 or 584.
Best GDT are when you put two secondaries inside a screen tube, which serves as primary. An ordinary audiocable , stereo for example. Wound on a toroid, high Al numbers (typically a CM filter core)
I did achieve < 100n leakage on a 1 mH primary-secondary, about 12-14 turns on a small ferrite core . Couldnt measure better due to stray inductances in test cables (short)
I did achieve < 100n leakage on a 1 mH primary-secondary, about 12-14 turns on a small ferrite core . Couldnt measure better due to stray inductances in test cables (short)
interleaved PFC is one of the ways to go for high power active PFC, as it effectivly splits the powerloss on severeal parallell switchers (spreading loss and enable use of less expensive components ) and by using smart phase control also reduces ripple stress on the output capacitors. I think it can also reduce EMI.
transformer design
My Litz wire arrived and I am going to wind my transformer in the following form:
1/2 primary,secondary over primary,1/2 primary for good coupling,will gap core and the leakage inductance will be supplied by external inductor since this will be for high power output.I will be using ETD49 core set and bobbin,the rest of the design is complete and after testing transformer should have supply about1.5Kw,the converter is designed to work below 150kHz.
My Litz wire arrived and I am going to wind my transformer in the following form:
1/2 primary,secondary over primary,1/2 primary for good coupling,will gap core and the leakage inductance will be supplied by external inductor since this will be for high power output.I will be using ETD49 core set and bobbin,the rest of the design is complete and after testing transformer should have supply about1.5Kw,the converter is designed to work below 150kHz.
I'm awakaning the thread because of IR1150 has been obsolete and a little hard to find
I was looking the IR1155 as a replacement
The only difference I've noted at their typical application diagram is the way they set their switching frequency. 1150 uses a resistor and 1155 uses a capacitor, both from pin2 to gnd
1150 VS 1155 datasheets
So here is the question
For a given frequency of 150Khz 1150 needs a 50K Ohm resistor
In case of 1155:Can the 150Khz be achieved with a 600pF capacitor instead of the resistor?
Has anyone tried this before?
I was looking the IR1155 as a replacement
The only difference I've noted at their typical application diagram is the way they set their switching frequency. 1150 uses a resistor and 1155 uses a capacitor, both from pin2 to gnd
1150 VS 1155 datasheets
So here is the question
For a given frequency of 150Khz 1150 needs a 50K Ohm resistor
In case of 1155:Can the 150Khz be achieved with a 600pF capacitor instead of the resistor?
Has anyone tried this before?
Old thread but I have a question about the LLC type converter.
Ik need multiply voltages, and one is 350 volts (100mA) and 2 x 120 volts (250mA) and 15 Volts 10 amps or 1 x 6,3 and 1 x 12.8 regulated.
so I be told flyback or forward is better for multiply voltages who are high.
But these converters are not resonant and have quit much EMI I do not want in the amp.
regards
Ik need multiply voltages, and one is 350 volts (100mA) and 2 x 120 volts (250mA) and 15 Volts 10 amps or 1 x 6,3 and 1 x 12.8 regulated.
so I be told flyback or forward is better for multiply voltages who are high.
But these converters are not resonant and have quit much EMI I do not want in the amp.
regards
Never tried high secondary voltages with LLC, but would not expects special problems there. The voltage tracking between different secondary windings maybe mediocre with LLC compared to flyback. But this can be greatly improved by using synchroneous rectification. For 1kW I would prefer the LLC due to its superb low losses and low component stress.
Hi there
It is for a hybrid, and also for experiment with some stuff.
Needed is for a hybrid who mosfets are driven in dc is.
2 x 120 volts 250 mA (driver mosfets)
1 x 350 a 400 volt 100 mA
1 x 6,3 volts 5 amp
1 x 12,3 volts 5 amps.
The amp output itselfs is separeate supply.
Because the filament voltages are needed very stable, I cna use feedback or maybe better I use a fixed frequency LLC and
use regulators for all supply, so I can get very clean voltages.
If I use no feedback smps resonance type, I have to set the frequentie on Lr and Cp resoance, I did test that and
voltage keeps quite stable.
quasy resonant half bridge or flyback is maybe a other option, more simple but do not now how clean it is.
I have a transformer from a big screen tv who has resonance type supply, the bobbin has space in middle and airgap. 172 mm2
thanks,
It is for a hybrid, and also for experiment with some stuff.
Needed is for a hybrid who mosfets are driven in dc is.
2 x 120 volts 250 mA (driver mosfets)
1 x 350 a 400 volt 100 mA
1 x 6,3 volts 5 amp
1 x 12,3 volts 5 amps.
The amp output itselfs is separeate supply.
Because the filament voltages are needed very stable, I cna use feedback or maybe better I use a fixed frequency LLC and
use regulators for all supply, so I can get very clean voltages.
If I use no feedback smps resonance type, I have to set the frequentie on Lr and Cp resoance, I did test that and
voltage keeps quite stable.
quasy resonant half bridge or flyback is maybe a other option, more simple but do not now how clean it is.
I have a transformer from a big screen tv who has resonance type supply, the bobbin has space in middle and airgap. 172 mm2
thanks,