I am currently working with this project and I am running out of ideas to solve two issue, here is a brief description of the project:
I am using a MOSFET Driver (IR2104), connected to two power MOSFETs connected in Half Bridge configuration (Infineon - 11N60CFD). Between the driver and the MOSFETs I have a small resistor (22 Ohm) connected with a parallel diode (1N4148) (For slow turn ON, fast Turn OFF).
The upper MOSFET is connected to an (aprox.) 150 Vdc source. And the bridge is switching at 50KHz.
The output of the Half Bridge is connected to a low pass filter (Inductor and capacitor) and I am using a big resistor as a load (> 75W (closer to 90W)).
At this moment the input to the MOSFET Driver is a simple 50% duty cycle signal.
At the output, I have a nice filtered sinus wave, just as I need it. The output of the half bridge (and input to the filter) looks nice, the signal looks square without ringing.
I have two issues that I am solving now and your comments will be greatly appreciated:
1) The Mosfets are heating to 90C (without heatsink). This seems a little high isn't it?
- I tried changing the values of the gate resistor to decrease the dV/dt and I couldn't see any difference
- I added more dead time (changing IR2104 to IR21844) and I couldn't see any difference, so it looks like this is not related to dead time.
2) The waveform of the current signal of the lower MOSFET (measured using a current sense resistor) has a lot of ringing (is hairy) during switching ON and Switching OFF, with peaks x3 times the order of magnitude of the current during steady state.
The two issues could be (or not) related.
Your comments are welcome.
I am using a MOSFET Driver (IR2104), connected to two power MOSFETs connected in Half Bridge configuration (Infineon - 11N60CFD). Between the driver and the MOSFETs I have a small resistor (22 Ohm) connected with a parallel diode (1N4148) (For slow turn ON, fast Turn OFF).
The upper MOSFET is connected to an (aprox.) 150 Vdc source. And the bridge is switching at 50KHz.
The output of the Half Bridge is connected to a low pass filter (Inductor and capacitor) and I am using a big resistor as a load (> 75W (closer to 90W)).
At this moment the input to the MOSFET Driver is a simple 50% duty cycle signal.
At the output, I have a nice filtered sinus wave, just as I need it. The output of the half bridge (and input to the filter) looks nice, the signal looks square without ringing.
I have two issues that I am solving now and your comments will be greatly appreciated:
1) The Mosfets are heating to 90C (without heatsink). This seems a little high isn't it?
- I tried changing the values of the gate resistor to decrease the dV/dt and I couldn't see any difference
- I added more dead time (changing IR2104 to IR21844) and I couldn't see any difference, so it looks like this is not related to dead time.
2) The waveform of the current signal of the lower MOSFET (measured using a current sense resistor) has a lot of ringing (is hairy) during switching ON and Switching OFF, with peaks x3 times the order of magnitude of the current during steady state.
The two issues could be (or not) related.
Your comments are welcome.
...and what is the capacitance?
If impedance of load (I mean input impedance of filter) is dominated by inductance, then you have 2A peak output current, which produces approx. 1...1.2 W conduction loss /device. This can make a TO220 case (what case do you use?) quite hot without heat sink. Maybe not 90 degrees, but if load is not only the inductance, the current can be higher.
Why do you use this FET at 150 V?
Do I understand correctly, supply is 150 V asymmetric? Then you could use IRF640N or something like that, with much lower conduction loss.
If impedance of load (I mean input impedance of filter) is dominated by inductance, then you have 2A peak output current, which produces approx. 1...1.2 W conduction loss /device. This can make a TO220 case (what case do you use?) quite hot without heat sink. Maybe not 90 degrees, but if load is not only the inductance, the current can be higher.
Why do you use this FET at 150 V?
Do I understand correctly, supply is 150 V asymmetric? Then you could use IRF640N or something like that, with much lower conduction loss.
Odnaizutra ->
If I disconnect the inductor, I disconnect the entire load and the entire filter and the MOSFET does not heat anymore (as expected).
Pafi ->
Capacitor (in serie with inductor): 3.7 uF
Case of power MOSFET: TO-220
I am using this FET because is the one I had in the lab, I started with IRF640 but I blew the ones I had so I switched to the clossest I had.
If I disconnect the inductor, I disconnect the entire load and the entire filter and the MOSFET does not heat anymore (as expected).
Pafi ->
Capacitor (in serie with inductor): 3.7 uF
Case of power MOSFET: TO-220
I am using this FET because is the one I had in the lab, I started with IRF640 but I blew the ones I had so I switched to the clossest I had.
Then leave the lab for two hours!
...or attach a heat-sink! ...or get used to the heat!The ringing can be a measuring problem and/or from the parasitic inductance of wiring and parasitic capacitance of choke. First you must be sure that probe cable (1:10 type!) is attached to current sense resistor without any loop (< 1cm^2), and current sense resistor has insignificant inductance (L/R=tau <20 ns)! If tau is too high despite of all efforts (it is in most cases), it can be cancelled by an RC LPF.
Alternatively you can use a current measuring transformer.
Pafi:
Thanks for your suggestions. It was a measurement problem. The ground loop was too big.
I solved it by placing the tip of my oscilloscope close to the shunt resistor, also, I placed a small wire (1 cm) connecting the ground portion of my probe to the grounds of my board. Thus minimizing the ground loop.
Now I can see a nice waveform.
The heating problem seems to be normal (now that I can measure the current). But I need further investigation on how to calculate the temperature change of the MOSFETs. Any suggestions?
Thanks for your suggestions. It was a measurement problem. The ground loop was too big.
I solved it by placing the tip of my oscilloscope close to the shunt resistor, also, I placed a small wire (1 cm) connecting the ground portion of my probe to the grounds of my board. Thus minimizing the ground loop.
Now I can see a nice waveform.
The heating problem seems to be normal (now that I can measure the current). But I need further investigation on how to calculate the temperature change of the MOSFETs. Any suggestions?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.