I searched for a bit but was unable to find anyone with exactly my problem.
I have a bunch of commercially made boards with IR2110 drivers and big chunky mosfets in a full bridge configuration. I'd like to use the IRS2092 to drive them but it requires a split supply and my boards are set up for single supply. I don't want to destroy the existing boards which have a proven layout and are good to well over 1000 watts in a very demanding application.
Assuming the output of each half of the bridge swings around half of supply voltage, could I simply use a resistive divider and cap to float the irs2092 at half of my single supply voltage, connecting the NFB as if the amp were the datasheet half bridge irs2092 application? I would provide the supply voltages for the irs2092 from resistors as indicated in the datasheet and possibly add external regulation if needed. Seeing as the input circuit draws little current and the irs2092 gate drivers won't be doing any real work other than driving the IR2110's I figured I could get away with such an arrangement.
I should mention this is a prototype and I'm ok with overcurrent and protection not working correctly- future designs will utilize better design practice with a microcontroller driving the gate drivers directly. I'm trying to quickly input an audio signal and determine if these readily available boards will have sufficient output for my application.
Can you recommend any alternative input chips or designs to utilize my existing mosfet/ir2110 boards with a single supply?
Jamie
I have a bunch of commercially made boards with IR2110 drivers and big chunky mosfets in a full bridge configuration. I'd like to use the IRS2092 to drive them but it requires a split supply and my boards are set up for single supply. I don't want to destroy the existing boards which have a proven layout and are good to well over 1000 watts in a very demanding application.
Assuming the output of each half of the bridge swings around half of supply voltage, could I simply use a resistive divider and cap to float the irs2092 at half of my single supply voltage, connecting the NFB as if the amp were the datasheet half bridge irs2092 application? I would provide the supply voltages for the irs2092 from resistors as indicated in the datasheet and possibly add external regulation if needed. Seeing as the input circuit draws little current and the irs2092 gate drivers won't be doing any real work other than driving the IR2110's I figured I could get away with such an arrangement.
I should mention this is a prototype and I'm ok with overcurrent and protection not working correctly- future designs will utilize better design practice with a microcontroller driving the gate drivers directly. I'm trying to quickly input an audio signal and determine if these readily available boards will have sufficient output for my application.
Can you recommend any alternative input chips or designs to utilize my existing mosfet/ir2110 boards with a single supply?
Jamie
I would use what you have got and add a centre voltage with a pair of constant current amplifiers. Say 10mA draw from the +ve to -ve giving 50% rail as a pseudo 0volt ground. There is no need for more current, the high current provided by the output stage is balanced and a pair of resistors centre tapping will probably do as long as there is decoupling between the rails. Regulation in not needed as long as the DC is fairly smooth.
When you say constant current amplifier do you mean a current source? - like a bipolar with a resistor at the emitter and some diodes to the base? Perhaps use two of them stacked with the midpoint (collector of bottom and emitter resistor of top) connected together? I suppose that would do a better job of actively resisting voltage change.
What about the circuit in the "single supply" application note? Could that be made to work with a pair of high voltage but low wattage bipolars and used only for the input ground supply?
Thanks for chiming in!
Jamie
What about the circuit in the "single supply" application note? Could that be made to work with a pair of high voltage but low wattage bipolars and used only for the input ground supply?
Thanks for chiming in!
Jamie
A zener at half supply voltage would do if the supply voltage is stable. The idea of a constant current source of equal magnitude would give a lower impedance reference voltage. If you had tqo 1N4148/914 etc, in series limiting the base voltage and a 100R* in series with the emiter of a complimentary pair of transistors, tie the bases together with a 10K resistor to bias them on. If the gain is equal, the centre voltage would be 1/2 the rail voltage. To compensate for gain differences, use a 100R on one and an 82R with a 47R* preset in series to the other. That will give a fairly predictable centre voltage. Make sure the two transistors and diodes are thermally connected, this will ensure good regulation and equal current regulation. As the temperature of the transistor rises, so does the hfe. Thermally connecting the diodes to the transistor will help to compensate.
Should I consider a different part that would be happy to drive the IR2110's referenced to ground of my single supply? It looks like that may be a better option. I was looking for an off-the shelf PWM controller but I suppose I could make one if I had to and let the IR2110's handle the level shifting.
Jamie
Jamie
colleagues,
I have followed the reasoning of all but my doubt is that the current on the MOSFET at the output stage will not burn the resistive divider and bipolar.
Another question is about the two GND, the main source and the splitter. How to join to have the same reference ??
I have followed the reasoning of all but my doubt is that the current on the MOSFET at the output stage will not burn the resistive divider and bipolar.
Another question is about the two GND, the main source and the splitter. How to join to have the same reference ??
- Status
- Not open for further replies.